Embed Size (px)
Citation preview
The CECOM Center for Night Vision and Electro-Optics
C GTOEL_'.CTRONIC WMORKSHOPS
SUPERLATTICE DISORDERING K~J 'CAfonr For
December 7, 1988 /~U;'-'! i - 0 -, ,
SP C S
sponsor- djointly by
ARO-URI Cen~ter for 0pto-Electronic Systems ResearchThe Institute of Optics, University of Rochester
I UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAG3E
REPORT DOCUMENTATION PAGEIs. REPORT SECURITY CLASSIRICATION 1b. RESTRICTIVE MARKINGS
!Inr annif i d__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
2&. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION I AVAILITY OF REPORT
2b. DECLASSIFICATION IDOWNGRADING SCHEDULE Approved for public release;distribution unlimited.
4. PERFORMING ORIGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION IMPORT NUMIBER(S)
_________________________Ato ./ '-/O/E6NAME OF PERFORMIN4G ORGANIZATION 6b. OFFICE SYMBOL 7&. NAME OF MONITORING ORGANIZATION
University of Rochester ________ U. S. Army Research OfficetADES(City Stat, and ZIP Code) 7b. ADDRESS (City State, and ZIP Code)Te Insitute of OpticsP.0Bo 121
Rochester, NY 14627 Research TrianglePakNC27921
Ba. NAME OF FUNDING /SPONSORING 6ib. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION Of aplc"U. S. Amy Research Office I_____ ZA10-A-k-61073
Bc. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERSP. 0. Box 12211 PROGRAM PROJECT ITASK WOK UNIT
Research Triangle Park, NC 27709-2211 ELEMENT NO. I NO. N. rAcESSION NO.
11. TITLE (Include Security Claniflcation)
Optoelectronic Workshop XI: Superlattice Disordering
12. PERSONAL AUTHOR(S)Susan Houde-Walter
13&. TYPE OF REPORT 113b. TIME COVERED 14. DATE OF REPORT (YearllolonthDay) IS. PAGE COUNTTechnical IFROM _____TO ____ December 7, 1988 r
16.SUPLEENAR NOATONThe view, opinions and/or findings contained in this report are thoseofVhe authgr(g) and sh uld not be conatpid as an 2fficial Duartment of the Army position,
17. COSATI CODES 18. SUBJECT TERMS (Continue on revee Nf rwceawy and identify by block number)FIELD GROUP SUB-GROUP Workshop; superlattice disordering )
!9. ABSTRACT (Continue on 'fv's Nf necena'y an ident by block number)
This workshop on "Superlattice Disordering" represents the eleventh of a series ofintensive academic/ government interactions in the field of advanced electro-optics, aspart of the Army sponsored University Research Initiative. By documenting the associatedtechnology status and dialogue it is hoped that this baseline will serve all interestedparties towards providing a solut1 1 n to high priority Army requirements. Responsible forprogram and program execution are Dr. Nicholas George, University of Rochester (ARO-URI) and Dr. Rudy B user, NVEOC./
20. DISTRIBUTION/ AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATIONOUNCLASSIFIEDIJNUIMITED E0 SAME AS ftPT. 03 DTic USERS Unclassified
22a. NAME OF RESPONSIBLE INDIVIDUAL 22b. TELEPHONE (Includs Area Code) I22c. OFFICE SYMBOLNicholas George 716-275-2417
DD FORM 1473, s4 MAR 63 APR edition may be usduntil xhausted. SECURITY CIASSIFICATION OF THIS PAGEAll otiw editions are obsolete. UNCLASSIFIED
The CECOM Center for Night Vision and Electro-Optics
OPTOELECTRONIC WORKSHOPS
XI
SUPERLATTICE DISORDERING
'or
December 7, 1988I ,I -
1i K
sponsored jointly by I IARO-URI Center for Opto-Electronic Systems Research
The Institute of Optics, University of Rochester
OPTOELECTRONIC WORKSHOP
ON
SUPERLATICE DISORDERING
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 Systems ResearchOrganizer -- Susan Houde-Walter
Overview and Status of Work at Univeristy of RochesterSusan Houde-Walter
B. CECOM Center for Night Vision and Electro-OpticsOrganizer - John Pollard
The Status of MBE at C2NVEOJack Dinan
Electronic Structure Calculations: Impurities, SuperlatticesRonald Graft
Photoreflectance Studies of Group I-VI SystemsPaul Amirtharaj
Properties of CdTe/InSb InterlayersMichael Martinka
Non-Linear Optical PropertiesGary Wood
Laser Diode ArraysDavid Caffey
4. LIST OF ATTENDEES
1. INTRODUCTION
This workshop on "Superlattice Disordering" represents the eleventh of aseries of intensive academic/ government interactions in the field ofadvanced electro-optics, as part of the Army sponsored University ResearchInitiative. By documenting the associated technology status and dialogue it ishoped that this baseline will serve all interested parties towards providing asolution to high priority Army requirements. Responsible for program andprogram execution are Dr. Nicholas George, University of Rochester (ARO-URI)and Dr. Rudy Buser, CCNVEO.
2. SUMMARY AND FOLLOW-UP ACTIONS
Dr. Buser opened the session with remarks on the NVEOC mission in opticalmaterials research. Dr. Buser stressed the directions in which epitaxial growthand processing efforts must go to achieve sophisticated new device typesneeded by the military, and the Army in particular.
The first talk was a tutorial in response to a request from NVEOC personnelon the implications of the quantum size effect in III-V semiconductors forlasers, modulators, and detectors. The tutorial, given by Prof. Houde-Walter,emphasized the differences between performance and characteristics ofquantum well and double heterostructure lasers, bulk and quantum wellmodulators utilizing the quantum-confined Stark effect, and some newlyproposed GalnSb/InAs strained layer superlattice structures which may mimicthe absorption spectra of HgCdTe.
Prof. Houde-Walter next gave a presentation on the current research at theInstitute of Optics in superlattice disordering and its application to monolithicintegration in Ill-V semiconductor optoelectronics. Methods for improvingstate-of-the-art lateral resolution in AIGaAs/GaAs were described and resultsfrom preliminary calculations were presented.
Current efforts in molecular beam epitaxy at NVEOC were next given by Dr.John Dinan. Fabrication of CdTe/InSb buffer layers for HgCdTe structures andCdTe/CdZnTe superlattice structures was discussed, as was MBE growth ofInSb on CdTe and CdTe on InSb. The studies of InSb/CdTe interfaces revealeda chemical reaction between InSb and CdTe which resulted in non-abruptheterostructure definition. Improvement in interface quality was achieved bycareful control of component flux ratios during growth.
Dr. Ronald Graft, also from NVEOC, followed with a discussion ofcomputationally efficient numerical modeling techniques for simulating thepresence of impurities in bulk semiconductors and for modeling superlattices.A recursion method, used to describe the effect of impurities and pointdefects, and a renormalization method, which was used to modelsuperlattices, were presented.
Dr. Paul Armirtharaj gave a review of the photoreflectance technique anddiscussed his application of the method to Il-VI semiconductors. Ademonstration of the sensitivity of the technique to processing (implant andanneal) and characterization of CdZnTe alloys and CdTe/ZnTe superlatticeswere presented.
The properties of CdTe/InSb interfaces and molecular beam epitaxial growthof In2Te3 layers on InSb (100 substrates were described by Dr. MichaelMartinka of NVEOC. In an effort to achieve InSb/ln 2Te3/InSb andInSb/ln 2Te3/CdTe multilayers, InSb and CdTe were grown on epitaxial in2Te3layers. Characterization revealed pronounced mixing of the alloys, consistentwith thermodynamics of the InSb/In2Te3 interface.
Dr. Gary Wood of NVEOC gave a presentation of results from a jointNVEOC./Rockwell International/University of Illinois collaboration on opticalnonlinearities in semiconductor multi-quantum wells. The physical origins ofnonlinearities in bulk and quantum well semiconductors were discussed in
detail. HgCdTe multi-quantum wells with a sawtooth profile were designed,grown, and their electrical properties were characterized. Afterwards Prof.Houde-Walter compared NVEOC's data using a well-known figure of meritfor nonlinear materials (X( 3)/au).
The last presentation by NVEOC was given by Dr. David Caffey on laser diodearrays. Blocks of commercial diode arrays were used to pump zig-zag YAGoscillators. Adequate heat sinking in a dense array of lasers was a concern.Despite some thermal detuning of the optical emission of the pump, goodpumping efficiencies were achieved.
The seminar concluded with a lively discussion of the process of developingnew materials into state of the art devices. Dr. John Pollard expressedinterest in improvement of the noise characteristics of avalanchephotodiodes and Prof. Houde-Walter suggested that quantum wellavalanche structures might have improved signal-to-noise ratios.
Follow-up has included informal swapping of technical papers and continueddiscussions between Prof. Houde-Walter and NVEOC personnel on heatsinking laser diodes and the unique ability of impurity-induced disordering togenerate window transparency and reduce facet damage in high-power laserdiode arrays.
ARO-URI UNIVERSITY OF ROCHESTER AND C2NVEO OPTOELECTRONIC WORKSHOP
SUBJECT: Application of Superlattices and Quantum Well Structures
Small Conference Room, Bldg 305
Wednesday, 7 Dec 88 at 1000
AGENDA
1000 Introduction Dr. Rudy Buser1005 Overview and Status of Work at Susan Houde-Walter
University of Rochester
1100 The Status of NBE at C2NVEO Jack Dinan
1120 Electronic Structure Calculations Ron GraftImpurities, Superlattices
1440 Photoreflectance Studies of Paul AuirtharajGroup Il-VI Systems
1200 Properties of CdTe/InSb Interlayers Mike Nartinka
1220 Non-linear Optical Properties Gary Wood
1240 Laser Diode Arrays Dave Caffey
1300 Discussion
POC: John Pollard, X45780
INTRODUCTION.
RUDY BUSER.
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS
Dr. Buser indicated the importance of controlled layergrowth for the next generation of device structures and theimpact th.is has already had on the laser technology. Thedisordering of superlattices to obtain controlled electronicproperties was particularly interesting and provides thematerial scientists with new tools to fabricate specialdevices. fie also discussed in the opening remarks the futurepossibility of providing a new sensor-processor which wasintegrated on one focal plane. The conceptual structure isgiven in the following viewgraph which conbines a detectorplane with an electronic/optical "neural net" type processorcoupled with an associative memory and output display forautonomous recognition systems. This ill necessitate thedevelopment of the growth of superlattice type strutures insemiconductor materials and their appropriate modificationunder controlled conditions. The future investment at C2NVEOwill provide for the controlled growth of materials using
the methods of molecular beam and chemical beam epitaxy.
C2 P-4to =n Cy CCU to 4-0 to =
a = CV? CJ 4
cn Lon -r- "W a CU w~ r-C2 L L^~ 1= 01o C w C3 c - 'C
W-4 4J LL 3 F4 1 -
11 C3 C-3 %*A. =4 0 0.cz C3 0 I- Els- Ca4- Li C
oCa e- n -.A -* vi viLLon ci %A L. - W "I LA to
V- C3U C= 0- i 40 J LJ-LJ wl 4-a 16- C3L
L~L.
t7% Ca/LozLi 4' 40..
e5 ....
LIJ A - _ _ _ _ _ _ _
- _ cr0
-L 0U0
La.J U.cca
F-
a.- FL 4- L U
O= L t^ W-
___ an L. 4-0 C3 cdE 4'go . 9. Ca CL.t-
_c to = *p. #A. LL Z a
C.3a Ca -m40=~ En Li. a. 0 C2 < K
a.c cc
Lo Li mn =i cto to _
ca+f MM Aa
LI
%-A-
0i-
tA 0
-U.
IU o
wALz
wVA
THE QUANTUM SIZE EFFECTAND ITS IMPLICATIONS
FOR LASERS, MODULATORS AND DETECTORSIN OPTO-ELECTRONICS
andSUPERLATTICE DISORDERING
SUSAN HOUDE-WALTERURI Center for Opto-Electronics
The Institute of OpticsUniversity of Rochester
(716) 275-7629
In Collaboration withGARY WICKS
The Institute of OpticsUniversity of Rochester
SEMICONDUCTOR CRYSTALS
V(x) a
V(x+a) = V(x) I) x1 -D crystal with N atoms and period "a"
Solutions to Shrbidinger equation with periodicpotential are Bloch functions:
']'k(X) = eikx Uk(X)
where uk(X) = Uk(x+a)
k is crystal momentum:k=2
NaResult: certain energies are not allowed; "band gap"
1h2k2E=
Nearly free electron: E 2m*
E
parabolic ON -ON SAD " CONDUCT IO "N-o = -re, HEAVY-HOLE[ 11"DEgap L4L IHT1OES
k( -- .. .,[mo\-+
QUANTUM SIZE EFFECTS
Nearly free electron model (bulk):
density of states22
Particle in an infinitely deep 1 -D well of width Lz
Schrodinger eqn.:2 2
2m dz2
En t 2 n= 1,2,3,...
>> = A sin staes
>> discrete number of bound states.
3-D Semiconductor Crystal, 1-D Confinement
I
kz3 2 3 4
2dg.
0 Y
like 2-D nearly free electron result: constantmodel for each n density of states
for each n.
Heteroj unction Energy Bands
CondctonEC Band
E_____ II~ s BaI GapA_ AEv"" Val ence
Band
Ec
______ _____Ev
AIGaAs GaAs
Quantum Well - infinite potential barrier
- n=2
Eeff(l)
- na hhn=l lh
n=2 hhn=2 lh
Lz< X elec (-200 A in GaAs)
Eeff (1)
ILz)2
Eg (bulk)Lz
Particle in a 1 -D well of finite height, Vo, and width Lz:
In the well:
d_2_ 2 ___m
dZ2 + kT
In the barrier region:
d2T 2 k 2m(Vo - E)dZ 2 +1 ki1 T = 0 ki =h
dz 1 1z
A ek1 z <-- -
= B sin (kz) L < Z <
C e z >
Quantum Well - finite potential barrier
Eeff ()
/1I(Lz)2 z
Eg (bulk)
60-80 A(nGa s)
Form lattice-matched heterostructure, e.g.:(AlxGai .. AsfAtyGai .yAs)
AlAs GaAs AlAs
Ec Ec
Ec 6.-AAs Owwfwiiet Stmre.
Eg AlAs NetCA
TEgpGaAs got ieto
Ev FJ L Ev
x=1 y=O x=1
Lg .4000A
LL 3.4401
L,~ L .1401040
e,5
1100 15500 1600 1650170
ENERGYY IOVI
ffm: R7PiNd.LE., Fes K69.?MLMELEME XV (04S~)
OTHER FEATURES OF ABSORPTION SPECTRA
1. Room temperature excitonsexciton: hydrogen-like e-h pair bound together by
, Coulomb forceconfined in QW, e-h pair has smaller orbitradius==>stronger binding energy
CONFINED "BULK" EXCITONEXCITON
] Lz Z: 3004L z xIO0A
___FL
excitons absorb/emit light at hv = Eg - Eex where Eex isthe energy required to thermally ionize exciton
IIMil
2:o-
o. 00
.ENERGY liVi0 & -- ---,-1.42 1 30 I ll L" 1.74
P140T0#N E[ RGY (OV)
From: J. Ralston, Ph.D. thesis, Cornell Univ. 1988
Consequence: absorption spectra is further sharpened
2. Polarization Selectivity
Angular momentum selection rules yield differentoscillator strengths for light and heavy hole transitions:
Heavy hole: a = 3 n = 0Light hole: a = 1 n = 4
Light polarized parallel to layers: both lh & hh excitonsperpendicular -only lh excitons
01 1b0I
.wI00
- 1.0w0).I
] - --
01 0 . .
paee Eopt perpendicular toEopt parallel to layerslayers
13g 14d IS 44 1.
PHOTON ENd[c (.evi PHOTON [NEQGC I)
Modified from: Weiner, et al APL 50, p843 (3/30/87)
Consequences:birefringenceextra polarization selectivity in waveguides
Two Coupled Quantum Wells
(n coupled quantum wells = superlattice)
(n _uncoupled quantum wells = multi-quantum well)
Ih--
N .
1.0
~ .j
.. 4
01.55 160 1.65 1.70 1.75
1.5 L1h
-Le
o 0.5
1.50 t.55 1.60 165 1.70
1.5
L.-12.6 L£SO
3' L
S0.5
011.501.5.0 .510
ENERGY (IV I
How do dimensional constraints affect refractiveindex?
Dielectric constant:
C = C 1 + ic2 = (n+ik)2
where n = refractive indexk = extinction coefficient:
k=- h c a
Kramers-Kronig relations:
-a 10 (0 C ') dd=1+ C~J,2 do'
e2(co) - f Pj, 2 d0
Calculated refractive indices:
bulk semiconductor quantum well
12 ,A(hv-Egf"' £, {Ae(hv-E 1)*BJ(hv-E, )}
n n
I I
I I
I EEg hi' E hv
Experimental:ref: Y. Suzuki and H. Okamoto, J. Elec. Mats. 12, 397 (1983)
Compare refractive index (dispersion) of bulkAlo.3Gao.vAs to GaAs/AlAs SL's with various barrierthicknesses:38 L B.,
3.8 (LzI B)in E g3.7,," "'-- -35
cr 33 (34 40) '.. . ,~3.2 4'
1.3 1.4 1.5 1.6 1.7 1.
PHOTON ENERGY We)Note: for a fixed average alloy composition, one canmake large changes in refractive index by tailoring
heterostructure dimensions.
>>useful degrees of freedom in device designlaserswaveguides, lenses, modulatorsdetectors
QSE IMPLICATIONS FOR LASERS(based on invited talk by A. Yariv, LEOS '88 (THE-3))
gana f E r() f,(NE).f,(NE) h/T 2
(E-hu) h y-2
galinOnly these electronscontribute to gain
)oE
gain QW
First few electronscontribute to
Loss Line. maximum gain
(E
- ~ LA- t
1. Threshold Currentref.: Arakawa, et al, IEEE JQE, OE-2., p666 (10/85)
Semiconductor is transparent when conduction bandFermi level penetrates the density of states
2D
=E
NTr m* dEc2D = th2 j (E'C)/kT
e F-1+ 1
=ln2 m*kT,th2
.5 x 1012elec 1.x1018eleccm 2 cm 3
Transparency is independent of confinement geometry
eNTr
Tr - e T p V active
Quantum well lasers typically have 20-40x smalleractive volume >> proportionately smaller threshold
0 100 ---- -- -- --- -------
E saw
<1 mA >1 Os mA Itrs
Modal gain =confinement factor x gain
jl Eo~I dz
f Eopt dz
waveguide mode
\gain medium
Typically L
2000 A
2. Maximum Gaingain
max -
gain
0
F N
*Since maximum gain is approx. 100 cm-1, require lowlosses to achieve lasing in SQW laser.
*Increase modal gain by increasing r7: use MQW's.
>>Therefore, use MQW's (typically N=3)
gm~x(N QW's) =- N gm~x(SQW)MUNDoC ~ of
where N = number of QW's100
60 4
22
C,
20-
0 goo 200 300 400 500CURRENT DENSITY (A/Cm)lt
From: Arakawa, et al IEEE JOE OE.219 p666 (10/85)
3. More on Current Thresholdref.: Yariv, APL fil p1033 (9/19/88)
-r -- - 'w - L---
I S
I~~s 3(~L n3D Tr W l WLdI 3D aN Tr(WLL+-In'-+ a atg3D R g3D
mirror & scattering80% 20%
2D Nr(WL) g' In' WLd sThresh N2D(L +2D R '2D sca_Y
20% 80%*The dominant contribution to threshold approacheszero as end mirror reflectivities approach unity.
e.g. for R-->I and a=0 the minimum threshold currentdensity is
IThresh min/Cm 2 = 60Ncm2
>>> JThresh = 0.1 mA for typical device dimensions(W=1.5gm, L=120 gim)
4. Laser Linewidthref.: Derry, et al, APL 53, p271 (7/25/88)
QW lasers have narrower linewidths
A&j) = (1 + t2 ) A-0ShlwTow
where a = dXR/dN
and X = XR + iX , complex susceptibility
32~28
S20 Ra- 1
4
00 02 04 06 0 0 t 12 14 16 16 20 2? k4
IN4VERSE OUTPUIT POWLA (n.W 1)
FIG I Spectral inewidi h at, a futis ioni of~ 111C I cipreodI ,. 1 iii4 pill .
SCII SQW labe and fir t h,, ~ aintla waiall 1id Lc rdkci. ii.
R, W' 8Ksld AI - 103'/;
Slope: DH: 74.7 MHz mWQW (R=31%) 28.3QW (R=90%): 5.9
A-o-=6 MHz @ 1.3 mW
5. Modulation Bandwidthref.: Arakawa (as above)
1 2
~~4 -d0!-- -
100 MHz 500 MHz I GHz 2 GHz 5GHz 100G1iFrequency
fra( gr Orp
where g'= dg/dn
P o= power inside cavity
and rp =- photon lifetime in cavity
QW lasers have about twice modulation bandwidth ofDH lasers.
In sum:QW lasers can be designed to have:
lower current thresholds than DH lasers; these lowthresholds can be further reduced by reflectivecoatings,smaller (2-3x) modal gain than DH lasers (forsingle QW lasers); same modal gain for MQWlasers (typically 2-3 QW's),
narrower linewidths (about 5x narrower thanequivalent DH lasers),twice the modulation bandwidth of DH lasers.
QSE IMPLICATIONS FOR MODULATORS
1. Quantum Confined Stark Effectref.: D.A.B. Miller, et al, Phys. Rev. B, 32, 1043 (85)
i~o ioo
CON0IJTI09l ..
- Eappi perpendicular to,., _" , . iplane of layers
"ANo E=10 4 V/cm (typicallyacross 10-4 cm)
Eg decreases as Eappi becomes stronger
Eopt parallel to layers Eopt perpendicular tolayers
10)i10.000 0001
-004 004
I I I ;O..
t e
00020.
004 CC? .00
I002 -002
I 3 144 Is 44 15
PHOTON ENERGY (Ov) P 'O ON ENER'r WV1
From: Weiner, et al, APL 50, p843 (3/30/87)
A. Electro-absorption: waveguide intensity modulatorref.: Wood, JLWT §9, p743 (6/88)
AaMQW =50x &acGaAs bulk (is comparable to LINbO3)
residual aborption for typical device = 2 dB
typical device speeds = 3-5 GHz
B. Electro-refraction: Phase modulatorsref.: Weiner, et al, APL 50, p842 (3/30/87)
Zucker, IGWO '88, paper MA-2
Not useable near band edge due to unacceptable aborptionlosses
Far from band edge, 10-4 -An-c10)-3 with minimum chirp
Dominant electro-optic effect is quadratic- with Eappi
13 2
1200
300-
6001
00 s0 160 240 320
APPLIED VOLTAGE IV)
Bias to achieve small modulation voltages with min. chirp.
2. Other Modulators
A.Exciton Field Ionization
Eapp parallel to layer interface: rip exciton apartwith large fields (105 V/cm)2-5x larger An than QCSE (penalty is higher fieldstrength)B. coupled quantum wellsC. direct to doubly indirect switching ofsuperlattices
QSE Implications for Long Wavelength Detectorsref.: C. Mailhiot and D. Smith J. Vac. Sci. Technol. B A p1 268
(Jul/Aug 87)
Propose Type-Il strained layer GaInSb/InAs SLto replace HgCdTe:
*III-V's are relatively easy to process
*Relaxed tolerances (e.g. bandgap less sensitiveto QW thickness than on HgCdTe alloycomposition)*Reduced dark current
Type I Superlattice:
Bottom of conduction band and top of valence bandcoincide in space.
UL L
Type II Superlattice:
Bottom of conduction band and top of valence banddon't coincide in space.
GaSb InAsNegative bandgap in bulk;
- negligible overlap ofwavefunctions
GaSb InAsTuneable bandgap byQSE; negigible overlapof wavefunctions
GalnSb InAs Introduce strain by barrieralloy; suppressesbandgap so can furthersqueeze QW; result issignificant overlap of e-hwavefunctions
WAVELENGTH (Ism)o 12 10 8 12 10 8
InAs-Goi .xI nxSb Hg1i-xCdxTe5 X:0.4 X=O.21 5
[001]
22LAJ
,oa a b=3210 3 l 3 0Z0
S5 EgallOmeV E 'llOmeV 5
0 1' 92 .- 2
100 120 140 160 100 120 140 160PHOTON ENERGY (meV)
from D. Smith and C. Maihiot J. Appl. Phys. 62 p2545(9/15/87)
QSE materials is a new class of materials with extra
useful degrees of freedom
Many parameters:
QW thickness, barrier thicknessNumber of QW'sWell, barrier compositions, dopingExtent of optical confinementEtc.
>> well suited to design ana engineering of opto-
electronic circuity
>> good candidate for monolithic integration
Heterostructures are grown epitaxiallymolecular beam epitaxymetallorganic chemical vapor deposition
epilayers(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
ii 0
04
E 0
0
as
0
Co of
0.0
0 0EW £4:
Cu1IN3NOvynv
Partial Mixing of heterostructure interfaces:
Ec
(as-grown)
Eg Eefleceffetiv
(as-grown (Gsmingwell)
AlAs Ga AlAs
z
Use to tune emission wavelength of lasersref: M. Camras, et al., App!. Phys. Lett., 54, p5637 (1983)
Mixing is enhanced by impurity transport
ref: W.D. Laidig, et al, Appi. Phys. Lett. 38, 776 (1981)
MASKED Zn DIFFUSION INTO
SL
Zn2As3
Si3N4
____ IJAs/GaAs SL
small bandgap large bandgaphigh index low index
- *ow -ndex
Waveguides:ref.: F. Julien, et al., Appi. Phys. Lett., AQ, 866 (1987)
disorder boundary
AlAs
GaAIAs SL GaAs OWSL
AlAs
GaAs substrate
Grating Couplers:ref.: J. Ralston, et al., GaAs & Related Compounds (1985), 367.
350A
Masked Se Implantation, grating period A = 3500A, X = 1.15 g~m
Lasers:
Reduce facet damage by increasing Eg at outputwindows,
ref.: R. Thornton, et all Appt. Phys. Lett, 429,1572 (1986).
WL'SOOM
'.5-.A
WL 0 Offi
CW CUAEW (A)
Monolithic Integration:
Laser, waveguide and modulatorret: R. Thornton, et al., I. Ughtwave Tech., .,786 (June 1988).
Ita moeulasel SectSO"
ggp.Uuled Waweouls
110-90 Amplie Secisafi
General Phenomenon:
Alloy Species Site
P, Ga, Al, Vili
AIxGai.xAs/AlyGal.yAs Zn Ili
Si, Ge
S. Se
GaP/GaAsl.xP_ Zn V
lnP/Gai.y A 1. Px Zn Ill V
Electrically inactive species:Ga, Al implant and anneal- collisional mixingP implant and anneal - collisional mixing, strain(?)Vill - Ga gettering into cap layer
Electrically active impurities:Impurity diffusion from ext. or grown-in sourcesIon implantation - two stages:
collisionalimpurity mixing
Can pattern by beam writing or photolithographywaveguides and gratingslow threshold lasershigh power lasers
THIS WORK: Brian Olmsted
(in collaboration with Gary Wicks)
GOAL: Monolithic Optoelectronic Integration
>>improved lateral resolution for component definition
Zn diffusion
gjm
Si diffusion _ _-
1 n
Control diffusion of mixing species by applied electricfield
1.0
0.8
0.6
required depth of miig>p
"I .
0.2
0
0.0 0.5 1.0 1.5 2.0 2.5 xlO 4
depth into heterostructure (arb. units)
drive in mixing species erpendicular toheterostructure layers; little time for sideways diffusion.
mask edgediffusion dominated
drive in mixing profile;little time for sidewaysdiffusion.
field-dominated
Use evaporated source or gettering layer:
Zn2As3
Si3N4
AlAs/GaAs SL
Use to study liD mechanism, fabricate components.
*Improve lateral resolution
currently sideways mixing is greater thanpenetration ==> improve by order of magnitude
*Design low loss structures
currently a=3-4 cm-1 ==> a<1 cm-1
*Build integrated optoelectronic systems including
basic optical component types:
sources
lasers
modulators
electroabsorption
electrorefraction
passive components
waveguides
couplers
detectors
06
i Uw
wLU
-o.z0
4A
0iLUm
I--m
LU
0wUL
THE STATUS OF MBEAT C2NVEO
JACK DINAN
DECEMBER 7, 1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
MBE at CCNVEO
CdTe / InSb
11 VI SUPERLAT17CES
HgCdTe
IN SITU ARRAY PROCESSING
CdTe / InSb
substrate/ buffer for HgCdTe epitaxy
2D gas
resonant tunneling structure
quantum well detector
quantum well laser
vU
FAIUA 0 0
a, I I I I l i
CD a
*A .
0 0 0
,. *** * * **/r
0,, 0 0 0 0 U *b
,D I 0 0 0 0 00
llt II IAIIin
oo
*Od OM 1100 AIn 99:91 81L1
af)
Ic'J
'0L
Iq
11.4 I,,I
rz ~ : :TTis :;! ;;+ , *d :
I~ Is
13 I.2~ * .'~ ...... h1 .1 *.. .
.... '.....
-m-6 .....fl .....10 it
'1~" .j~ ~. V~7~T'3F -7Tjj
I
y
.11
'I-
*%
1.- a* '~,*.. ~
-- *a..
I *
OVI a%) um vL " um '
t- atU . . ..
0 •
I I
q wI
III I l
I
U-
cii
ri
0
Os:
009 0 F 06COi' 0 0(V) jua0>
HI VI SUPERIXETCES
Cdre / Cd&Te superlattices
dislocation-filter for HgCdTe epita-cy
00O
0,Nw
LL.
VTto
bc
0
ClY
(24!un KJ04!qJID) ApsuSjUJ
IN -SITU PROCESSING OF ARRAYS
growth
mesa etching
metalization
psmivation
0n.
-t
.
40
VZ
1D
ELECTRONIC STRUCTURECALCULATIONS:
IMPURITIES, SUPERLATTICES.
RONALD GRAFT.
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
ELECTRONIC STRUCTURECALCULATIONS
OUTLINE
Recursion MethodGeneral Remarks.Application to point impurities and defects.
Renormalization MethodGeneral Remarks.Application to Superlattices.
COLLABORATORS: Prof L. Resca and D. LohrmannThe Catholic University of AmericaWashington, DC
Prof G. Pastore-ParraviciniUniversity of PisaPisa, Italy
16 O *
0 I
Ugi
w cc1
_ I
00
CCC
CL .
IM I=
m, M
0 000CC E
0
>--
*0
z w 0 0 0 "
0< 0wg 0 x 0 c": "a. <
< w0 0 0_wa 0 V65
0 0 0 -a Ac
o 4)W L.2
.0*0 N 0
0I. L) 00
E 0*
LU
00
wMI
z U
wi A
0-
w -
A I-
> '-
w * A
*1E
0
EE
00Do -0
E o*0 0>
0 0
0000 ~ 0- 0 0 .0- C os 01 0 00i a- 0. q..(aMdEo 0 > .. 0E
L)~ ~ 0 L
0 00. 0.C.0 0 0 0cc 'lo = 0. CL 0
OE .0 m0E 0 0to %4w EE.... 0 - V 0 0
0C 0
0 -0 r -
w3 o-~ 0.0 C r0 0 0CE
WL M0 . 0 0 0.0 :2~' 0Eo 0 -- C0.0 0 E0= 0L 0
.E 0 0 0
0L - L- -- ~ - Eco> L0 . 0 ~ ~ 0 0 0
0 w. 0 L 0 0 'I
0 0 0 0 _
0 00 CL .04- > o00 E E E 0
CL 0 0- 00L CO 00 5,0 L e
CU0aC lw 00 . OE
0 o
00
o 0
0
w LU
z U-0
0 ULP
0 )LU
E %
w b-U.>00
N caLU
0
LLU4
Q LL o 0
z 0 0
Z
F- I.- _o
-oLUUIL co w
2 LU 0L) 0 i
m x cc
wi i
(:D
0
ww
0 4)
< E m1 >
NE w .0 w0
_
%..0 0 11 0--
0
Rt 0
z0 co 7E 0
o .0z 0
E c cu 000 0
o0o
A E0 b0 co
-Z EccGow00
000
0 E x 2
co 0.iCL
00
CL
00
4) E C
%- EN 0 E 4
III.- 0 0L - C
0 E0
0L 0. C O0 0-0
0.- - 00a 0' 0
- -$-- E_ 0L 0 0
co 0 a0 ~ 0. C.L
C 0m 0. o
0 4
00
o 0
-0 0
0- 0 0 E t
0w E WE h.0 . -- 0 "a ~
4' 0...4 0. ~
~0 Zm 0 o
E -4..0 0 .0
c 04 0 0 &_
00 x cC0CL
= 00ca 0
wL 0 L 0 z02 E ~ 0 v.2
E 0~ E. Qum
00I0 c 00, 0-== < tM~ 0 0 0
* ~ E4' L 2 4 0. N
CO0 ~ ~0 0~ >~EI**E 0 E 0~.
0 W6000 4 C 0C.
CL
0P2-
20I-
L>
0z
U'
-UzWLL
0
00u00vs
PHOTOREFLECTANCE STUDIESOF GROUP II-VI
SYSTEMS.
PAUL AMIRTHARAJ.
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
PALLt. Ae,,fov
-4 PL'r A/t. r e VA
GaAs
0.4
I dR
ER (0K)
0-1
EEl E0 0AEb E2 &
0 2 4 6Og~ E (ev)
O.C. ~ N 4AC, righ1 V44 ,
&mottE~ LeIACT ksFa r ee
F&L.Ib "tbL LA OVE-
C EL6U-Q 1" &&F EC~Corr
Loi - %. 4 W7
IJJ E- L-'T SL..ytQb L6C 004i c6
>
P1.16 ~ ~ w LEC - E
US: 0kLkok-TP ~ OrV44A LA
0.2-
0.1-
L&J.
- 0.1
-F
-0.3-
2 0-
xFig. IZ Te Franz-Kefdysh lineshapes F and G. plotted with x-axis reversed to simulate electric
fieI&_ rodulian, of the imaginary and real parts of e of an Mo critical point plotted against increasingenergy.
2a
IGe - lKE0 TRANSITIONS
0-.(A (eV-
Fig. 13. Electroreflectance lineshapes for Ge for two fields taken under conditions where the
/ Franz-Keldysh theory is applicable (Aspnes 1975).4 At~~hM4 O~_~~?vI~eNG~.'d4,VOL.
ELECTRIC FIELD MODULATION
VPLATINUM
WIRE v
50A Ni- SAMPLE
SEMICONDUCTOR
E SAMPLE200 A Af1o 3 AQUEOUS
ELECTROLYTE
(a) LONGITUDINAL
V
SEMI- INSULAT INGSAMPLEt F
eiE
(b) TRANSVERSE
A~'i~*e-sa Xb-4 O~o~ 0 - 9 49 AJ J cCA 'r0&,,
100
amCDI
~.i, 'uI - Cgs
- - .v96
22
PHUTOREFLECTANCE - 770KClxZn xTe - E
2-
Cdle
-2-
2-
64 CduZn1 .2Te
-2
x140 CduiZnaTe
0
3.4 3.5 3.6 3.7 3.8 3.947W MgVIA".AI ENERGY (eV)
61 A.1, v~12.~ em) FIGURE 3
t I+ C16 -7\-- -- -,
• .... C tS A7
- . -.>;., = o-.
10 PHOTOREFLECTANCE - 770( - E
LPE Hg., Cd Te (x- 22)
5
2.1 2.3 2.5 2.7
ENERGY (eV)
2~~~0 42:7 cu C . 2~
LIi
+
LJL
0
00
wL
44*/
4 0t
44/* 2C
/U C/L
IU 11 CF.
= Ch
IL
Lu4.4 44
Cu J
1 Wm C I I
I I "l I I
4- PhotoreflectanceCdTe < 100 > 77K-E 1
2 "
UNIMPLANTEDIANNEALED X0.1
A E
z
IMPLANTEDIUUANNEALEO X2-- ------------ -----------
-22
IMPLANTEDIANEEALED Ioc C
-2
! I I I
3.0 3.2 3.4 3.6 3.8 4.0
Energy (eV)
' 41't~A3 P- 8t Cv-44 / 4& k/
S\& £Gv 8
4 PhotoreflectanceCdTe <i100> 77K -Eo
2
XG.004o - -- -- -- - - -- j
9 A
-2 UNIMPLANTEDIANNEALED
-
cc 8
-2 - IMPLAUTEIMUIERD
4
2
IMPLANTEDIANNEALED-2
-4tti
1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Energy (eV)
/TM II-.rm fAj &r Ai: / U Tf~ ' / 4AY
Photoreflectance IMPLANTED AND2 CdTe <100> - 770K - Eo UNANNEALED
6328 APump
S---------- -- - -------
2-
x 5& 1- - -- -- -- -
0-O- --
3-
2-1 [32o0- 400A
1 - I Pump _!f x 1.5
-1I I I I I
1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9Energy (eV)
4 ?I4 I & I IO 3+4
AGaAs
GaAs BUFFER ai AI sA
PHOTON
FAI G As 5 UANTUM AIGaA$
WELL CONDUCTION
fl:1 {lh 1 1 E0CAIGaAs)
(3) (1)
hh-Ih
VALENCE
BAND
' N i
j4 4x
000 CRU
wow Z
0 1 .
(9-0L X) H/iUv
r
B
S.
S
* -
_________ S
CdTe/ZnTe 50/50 SL
----------- - ----
Transition Square Parabolic SL-1 SL-2
---- --------- m ------------- ------------
z,, 1.61 1.67 1.7 1.65
92h 1.88 1.82 1.88 1.83
E3h 1.98 1.97 1,92
E,, 2.13 2.09 2.08
Ell 1.65 1.71 1.74 1.74
-- --- m----m- ---- -------- m--- mm------------5
-75'A .. w' .
%A
06
%w
<2zZ .a
0 tA
>0
ZLJ20~
owU
L
~0
0uwAu
PROPERTIES OF CdTe/InSbINTERLAYERS.
MICHAEL MARTINKA.
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
Molecular Beam Epitaxial Growth andCb xacterization of In2 Te3
T.D.Goldint. P.R.Boyd, KA Maftnka, P.M. Airtharaj and J.H. DinanU.S. Army Center for Night Visin and Electr-OpticsAMSEL-R D V-IT, Fort Belvoir, Virinia 2200-567, U.S.A
S.B. OadrlU.S. Naval Researd Laboratories. Washington D.C. 20375, U.S.A.
D.RT. ZahnDepartment of Physics. University College, P.O. Box 78, Cardiff CFI IXL U.K.
C.R. WhitehouseRoyal Signal. and Radar Establishment, Malvem, Worcestershire. WR14 3PS U.K.
tPermanent address: Cavendish Uaoatory, University of Cambridge, Cambridge, U.K.
Abstract
We report studies of the molecular beam epitaxial growth of In2 Te3. The unique structure
of In2Te3, with 1/3 of the In sublattice sites vacant, is of fundamental interest for molecular
beam epitaxial growth dynamics. We show that thin film (50OA-7000A) single crystal
In2 Te3 can be grown successfully on InSb(100) homoepitaxial layers at substrate
temperatures of 300 - 3500 C and Te/In flux ratios of 3/2 - 5/2. Epitaxy has been monitored
by reflection high energy electron diffraction and the stoichiometry of the grown layers
assessed by Auger spectroscopy and energy dispersive X-ray analysis. Raman studies of
the layers are presented and compared with a bulk In2Te3 standard. Crystal structure has
been determined by X-ray diffraction using Weissenburg and oscillation photographs,
confirming that the layers have a f.c.c. crystal structure with a lattice parameter of 18.50A,
in excellent agreement with the bulk value. Bandgap measurements have been performed
on the layers by photoreflectance. We report a value for the cx- In2 Te3 bandgap of 1.19 eV
and 1.31 eV at 300K and 77K respectively. Molecular beam epitaxial growth of InSb and
CdTe on epitaxial In2Te3 films for fabrication of InSb/ln2 Te3 /InSb and InSb/
In2 Te3 /CdTe multilayers has been studied. Auger depth profiling of the resulting layers
• ' I I |
shows severe intermixing into the In2 Te3. These results ae supported by thermodynamic
considerations of the InSb-In2Te3 interface.
Introduction
In the course of a study into the molecular beam epitaxial growth (MEE) of InSb/CdTe
multilayers and superattices, it has been foundl,2 that there is a strong chemical reaction at
the InSb/CdTe interface during growth at substrate temperatures compatible with the
growth of high succuraI and electronic quality of both InSb and CdTe (275-300 C).
Recent studies 3 ,4.5 ,6 by soft X-ray photoemission and Raman spectroscopy of the
heterojunction formed from the MBE growth of CdTe on InSb under stoichiometric flux
conditions and growth temperatures of 200-300*C reveal that the InSb and CdTe react to
form a complex interface identified to consist predominantly of n2Te3 and elemental Sb.
In2 Te3 is a defect zinc blende type semiconductor with a bandgap of approximately 1 eV,
lying between that of InSb (E1=0.l8eV) and CdTe (E5 l.44eV). Thermodynamic
considerations of the stability of In2Te3 with respect to InSb and CdTe7 ,8 led us to
examine the possibility of MBE growth of In2Te3 for use as a barrier between the CdTe
and InSb to facilitate fabrication of InSbfCdTe multilayers, and the interesting possibility of
exploiting the large difference in badgaps between InSb and In2Te3 to grow InSb/ln2Te 3
multilayers for novel quantum well structures.
In this paper we present results of the study of the MBE growth of In2 Te3 and
characterization of the grown layers. We show that thin film single crystal In2 Te3 can be
grown on InSb (100) homoepitaxial layers. Reflection high-energy electron diffraction
(RHEED) was used to monitor growth of the .n 2 Te3, and stoichiometry of the grown
layers was assessed by Auger spectroscopy and energy dispersive X-ray analysis. Raman
spectroscopy of the In2Te3 ME grown layer was compared with a ln2Te3 bulk sample to
verify chemical composition. The crystal structure of the In2Te3 layers was determined by
X-ray diffraction techniques, confirming a f.c.c. structure with a lattice parameter of
18.50X, indicating that the MBE grown In2Te3 is in the alpha phase (a-n2Te3)9 where
the In sublartice vacancies are ordered. The bandgap of the In2Te3 layers has been
determined by photoreflectance, yielding a value of 1.2eV at 300K and 1.31eV at 77K.
As a preliminary investigation into InSb/n2Te 3/CdTe and InSb/In2Te3 multilayer
structures, we have studied the MBE growth of InSb and CdTe on 500A thick In2 Te3
epilayers. Auger depth profiling has been used to examine the layers and their interfaces.
Results show a severe intermixing of both the CdTe and InSb into the In2Te3 layers. The
experimental results are supported by thermodynamic considerations of the InSb-In2 Te3
interface. Points of interest are discussed.
Experimental
The In2Te3, InSb and CdTe layers were grown in a Varian 360 MBE system equipped
with a quadropole mass analyzer and in rim RHEED and flux monitoring facilities. Base
pressure during growth was below 5XI0- 10 Toir. The InSb(100) substrates were solvent
cleaned and mounted onto molybdenum beating blocks using a colloidal suspension of
graphite in alcohol. Immediately before growth the native oxide was removed from the
substrate surfaces by heating at 410 *C in an Sb4 flux. A i000A InSb buffer was grown
on all substrates to ensure a consistent high quality InSb (100) surface present for the
growth of the Indium telluride.
Separate effusion cells containing high purity In, Sb, CdTe, Cd and Te were used for the
growth of the InSb, CdTe and In2Te3 layers. A relative measure of the flux from a given
cell was obtained by interposing an ion gauge flux monitor into the molecular beam and
3
relating the measutred beam equivalent pressure to the cell temperature, molecular weight
and ionizatio efficiency of the beam species.
Indium telluride growth was studied for substrate temperatures, Ts within the range
300ls<350*C and Te/In flux ratios (JTe/Jn) between I< JTe/In <5/2. The Te/In flux
ratio range was resticted to help avoid the possibility of growth of the compounds InTe
and n2TeS. The In cell setting was kept constant throughout. Growth rates were
estimated to be 0.2 pm/hr as calculated from the measured In flux, assuming a unity
sticking coefficient for the In.
Following growth the samples were analyzed by Auger spectroscopy, and the atomic and
weight percentages of the layers were assessed by energy dispersive X-ray analysis (EDX)
using a standard facility with an accelerating voltage of 15 KV and a beam current of
approximately 8X10-10 amps. All layers examined by EDX were -7000A thick to ensure
then would be no contributions from the InSb substrate appearing in the EDX spectra. A
sample of bulk In2Te3 supplied by CERAC/PURE Inc was used for standardization
purposes for both the Auger and EDX analysis. Selected layers were further characterized
and studied by Raman spectroscopy, X-ray diffraction techniques and photoreflectance.
During the course of our studies, it became apparent that the In2Te3 was oxidizing over a
period of time when exposed to atmosphere. To avoid oxidation during transportation for
the Raman studies, a A Sb cap layer was deposited at room temperacture on the In2Te3 . It
has been reported that Sb deposited under these conditions may be either amorphous or
crystalline, dependant on the substrate and thickness I0 . RHEED studies during deposition
of the Sb indicated that the Sb was amorphous. The Raman studies employed an Argon
ion laser as an exciting source (X=514.5 inm, P-4OmW) with the spectra taken in a back
scattering geomeuy. A bulk In2 Te3 sample4 grown by the horizontal Bridgeman technique
at UC Cardiff was used as a standard for comparison. X-ray analysis of the layers used a
single crystal diffractometer together with Weissenberg and oscillation photographs
4
utilizing Cu Ka radiation. Photoreflectanc measmements were used to measure the band-
gap and were performed with a standard facility 1 that used a Ge or Si detector and a
6328A He-Ne pump beam to produce the electric field modulation. Measurements were
conducted at 300 and 77K in the vicinity of the dc gap, E.
InSb and CdTe were grown by MBE on thin film (500A) In2Te3 epilayers at 300C. The
InSb and CdTe growth, together with the experimental details of the Auger depth profiling,
has been described elsewhere 1 .
Results
Initiating growth of the Indium telluride resulted in an immediate conversion from the Sb
stabilized (2X2) InSb RHEED pattern 1 2 to a characteristic streaked pattern shown in
Figur 1. For flux ratios JTe/In <3/2 this pattern became spotted and eventually vanished,
indicative of a disordered growth, after approximately 300A growth for all substrate
temperatures studied. For growth with flux ratios JTe/ 7in2.3/2 and a substrate temperature
of 300*C, the pattern remained streaked throughout the entire growth. Growth at substrate
temperatures greater than 300*C required Te/In flux ratios greater than 3/2 to preserve the
suteaked two-imensional growth pattern.
Energy dispersive X-ray spectra of samples grown at Ts - 300°C and flux ratios,
ITe/JIn<3/2 and JTe.Jin23/ 2 (Figure 2, (a) and (b) respectively) are compared with the
spectrum obtained from a bulk n2Te3 reference standard. The spectra for layers grown at
300*C and flux ratios JTe/in. 3 /2 were identical to that obtained from the bulk In2 Te3
standard, indicating that the stoichiometry of these layers was identical to that of the In2 Te3
bulk standard. Layers grown with flux ratios JTeIn<3 /2 were identified to be In rich.
5
Auger spectra of the UBE grown layers were compared with the Auger spectrum obtained
for the In2Te3 bulk standard. The spectra obtained from samples grown at 300*C with flux
ratios JTvan 3/2 matched the spectrum obtained from the bulk standard. Samples grown
with flux ratios JTe/lin<3/2 were identified as In rich, correlating and supporting the
results obtained from the EDX studies. An Auger spectrum from a stoichometric In2Te3
layer is shown in Fig 3. High magnification SEM imaging of the samples showed a clear
distinction between the stoichiometric and non-stoichiometric layers. The stoichiometric
layers had a smooth surface morphology whilst non-stoichiometric layers were rough and
sructured with island-type precipitates. Further examination by EDX identified these
precipitatesas indium.
Raman studies of an I IOOA thick MBE layer grown under JTe/n>3/2 and Ts=300 *C
growth conditions were compared with the spectrum obtained from a bulk In2Te3 sample.
Although the optical constants of In2Te3 am not well documented, the penetration depth of
the laser light used in the studies was considered to be less than 1 iooA, ensuring that there
would be no contributions from the InSb buffer and substrate appearing in the Raman
spectrum. The spectra obtained for the MBE layer and In2Te3 bulk am shown in Fig.
4(a) and (b) respectively. Both spectra exhibit prominent peaks labeled AB,C and D for
the In2Te3 layer and A',B',C and D' for the bulk materiaL The position of these features
are in excellent agreement. Slight interference with A,B and C is caused by an additional
background from the amorphous Sb, which has a characteristic broad band which peaks at
152 cm-I 13. Additional features at 70 cm- I can also be seen, although difficult to
separate from the increasing background due to Rayleigh scattering from a slightly rough
sample surface. Whilst little is known about the Raman spectmum of many of the indium
tellurides, InTe has been reported14, and the spectrum differs substantially from Fig. 4 (a)
and (b).
6
Further characteizaon of the M3E grown indium telluride and analysis of the crystal
strucre were obtained from X-ray diffraction, Weissenberg and oscillation photographs.
Fig 5(a) and (b) show 20 diffractometer traces of a 1IOOk thick In2TC3 layer grown on
an InSb(100) substrate for InSb(200) and InSb(400) reflections respectively. Jn2Te3 (600)
and In2Te3(1200) reflections are observed in these traces suggesting that In2Te3<100>
lies along InSb<100> direction with an out of plane lattice parameter of 18.50K This
lattice parameter is in good agreement with a value of 18.486 reported 9 for bulk In2 Te3 .
To confirm that the periodicity was 1 8.OA in the plane of the epitaxial film, Weissenberg
and oscillation photographs were taken. Figure 6 shows the oscillation photographs of
In2Te3/InSb(100) taken about (011). oscillation axis. First and second layers of InSb are
observed with a repeat distance of 4.58A (6.48/42 A). Weak and medium intensity layers
are seen for In2Te3. Zero to sixth layers are identifiable with a repeat distance of 13.08 A
(18.50/42 A), suggesting that the in plane lattice parameter of the film is 18.50k All the
reflections can be indexed based on a f.c.c. lattice with lattice parameter of I 8.50A.
Photoreflectance spectra, measured in the vicinity of Eg are presented in Fig 7. The line
shapes are qualitatively similar to those observed in other semiconductors I 1 with a sharp
oscillation near the transition energy; the broad feature below it is most likely due to
absorption modulation of the light that is tansmitted through the film and reflected from the
In2Te3/InSb interface 15 . The maximum in the spectrum is assumed to represent the
transition energy and found to be 1.19±0.02 eV and 1.31±0.015 eV, at 300 and 77K
respectively.
The ability to grow thin film In2 Te3 made it possible to study the subsequent growth of
InSb and CdTe on these layers. An Auger depth profile of a 500A In2Te3 layer grown on
a InSb homoepitaxial layer is shown in Fig. 8. The interface is seen to be abrupt and
estimated to be less than IO,, with no evidence of intermixing between the InSb and
In2Te3 layers. A 800-s InSb growth was attempted on a 500A In2 Te3 layer with Ts=3 00
7
°C and a growth rate of 0.6 A/s to give a nominal growth of 500A. The RHEED pattern
showed no change from the characteristic In2Te3 pattern (Fig. 1) upon initiating, and
during growth of the InSb. The Auger depth profile of the resulting layer is shown in Fig.
9(a). There is no evidence of InSb layer growth. Comparison with Fig. 8 indicates that
the layer is In rich with respect to In2Te3 and the interface broader. A 300-s CdTe growth
was attempted on a 500k In2Te3 layer at Ts=300*C under stoichiometric flux conditions
(JCd/JTeml) and nominal growth rate of 1.85k/s. The RHEED pattern remained
unchanged from the In2Te3 pattern during growth of the CdTe. The resulting Auger depth
profile is shown in Figure 9(b). No CdTe layer growth is evident. However, Cd has
been incorporated through approximately 400, of the indium teiluride layer, suggesting the
formation of an alloy. A similar CdTe growth was attempted, with a Cd/Te flux ratio
JCd/JTe=3/1. The RHEED pattern remained unchanged from the In2Te3 pattern during
growth. The resulting profile is shown in Fig. 9(c). In this case the Cd has mixed
uniformly throughout the indium telluride layer. Initial analyses of the Auger data using
sensitivity factors suggest that the resulting layer has ratios Cd:In:Te of 1:2:4
Discussion
We have shown that single crystal In2Te3 can be grown epitaxially on InSb(100) at
substrate temperatures 300:STs-350 and flux ratios 3/2.JTe/Jin<5i/2. For a given
growth temperature, there is a minimum Te/[n flux ratio required for epitaxial growth (3/2
at 300°C), below which the layers are non-epitaxial and indium rich. The minimum Te/In
flux ratio increases as the growth temperature is increased, suggesting that the ratio in the
Te/In sticing coefficients decrease with increasing growth temperature.
The Weissenberg and oscillation photographs indicate that the In2Te3 layers are in the a-
phase, where the cation vacancies are distributed uniformly throughout the In sublatice. It
8
is of fundamental interest for MBE growth dynamics to ascertain whether the ordered
vacancy stuicture is formed on the growth surface during epitaxy, or if the vacancy
distribution on the growth surface is random and there is a subsequent rearrangement of the
In sublatice within the layer. A detailed analysis of the RHEED pattern during growth may
help to clarify these points.
We report a value for the bandgap of the a-phase of In2Te 3 to be 1.19±0.02eV and
1.31-0.015eV, at 300 and 77K, respectively as measured by photoreflectance. As far as
we are aware, this is the first direct measurement of the band-gap of a-In2Te3 . A value
of 1.12eV has been deduced from electrical measurementsl& Other 300K band gap values
reported range from L.OeV for the f-phase, determined from absorption measuM ts17 ,
to 1.2eV determined from the reflectivity1 s; pseudopotential calculations yield a band gap
of 1.3eV for the J-phase1 9 .
We have shown that an attempt to grow an epitaxial layer of either InSb or CdTe on In2 Te3
layers at growth temperatures compatible with InSb/In2Te 3 and CdTe/Ln2Te 3fInSb
multilayer growth (300*C) results in severe intermixing in the In2Te3 layer with no InSb
or CdTe layer growth. RHEED studies indicate that the layers formed are crystalline. For
InSb growth on n2Te3 Auger depth studies show In incorporation into the In2Te 3 layer.
In contrast an abrupt interface is observed for the growth of In2Te3 on InSb.
Consideration of the therim is of the In2 Te3-InSb interface can help to explain the
heterogeneity of the two growths. For this, we ignore the fact the reactants may be
dimers or teurmer (i.e. Te2 , Sb 4 ), assuming that the molecular beam species degenerate
into the monomer upon contact with the growth surface. As such the deduced change in
enthalpy should only be considered approximate. For the growth of In2Te3 on InSb we
can consider the following reaction where In and Te are present on an InSb surface.
InSb(,) + 21n(s)+ 3Te(,) " InSb(,) + In2 Te3() (1)
9
for which the change in enthalpy AM is -45.8 kca120. Our growth conditions were
care~fy controlled by tailoring JTdJIn to ensure In2Te3 formation rather than InTe or
In2 TeS, and we will not consider these other possible reactions here. For the growth of
InSb on In2Te3 we consider In and Sb present on an In2Te3 surface.
In2 Te3(,) + In(,) + Sb(,) -* In2Te3(,) + InSb(s) AH=-7.520 (2)
where the In and Sb react to form InSb. However, if the Sb has a short residence lifetime
with respect to the In, the growth surface may be Sb deficient and we should consider the
following possible reactions.
In2T3(,) + In(s) 3InTe(,) AH,- S.S20 (3)
In2 Te3(,) + 4I1(,) > 3In2Te(,) " - 11.220 (4)
For reactions (3) and (4) the elemental In converts In2Te3 to InTe or In2Te respectively.
The experimental results indicating the inability to grow InSb on In2 Te3, with the
incorporation of In into the In2Te3 layer suggests that the MBE growth of InSb on In2Te3
may be described by a reaction such as (3) or (4).
Whilst these thermodynamic considerations are non-rigorous, they suggest that InSb
growth may be possible on In2Te3 if sufficient Sb is present on the In2Te3 growth
surface to bond with the available In. Once InSb growth has been initiated on the In2Te3 ,
the resulting interface should be stable, since no chemical equation with a negative change
in enthalpy can be written with InSb and In2Te3 as the reactants. This thermodynamic
consideration is supported by the abrupt interface observed for the growth of In2Te3 on
InSb (Figure 8).
For CdTe growth on In2Te3 there is severe Cd interdiffusion. Increasing the Cd flux
results in a uniform incorporation of the Cd throughout the indium telluride layer.
Preliminary Auger data suggests that the layer so formed may be a ternary alloy such as
Cdln2Te4. Further analysis by Raman and X-ray will help to clarify these points. The
10
readiness of Cd to inwffuse into the In2 Te3 layer supports the notion that the interfacial
layer formed during growth of CdTe on InSb under conventional MBE growth conditions
is complex and is unlikely to consist of merely a thin layer of In2 Te3 and elemental Sb.
While these studies suggest that the successful growth of InSb/In2Te3 and CdTeIn2 Te3
multilayeus is unlikely, the results shed additional insight into the MBE growth of mixed
systems. Recent studies 1 ,8 of the growth of CdTe on InSb have shown that simple
thermodynamic considerations 8 are applicable to the ME growth of this II-V=/II-V mixed
system. The results presented in this paper also suggest that simple thermodynamic
concepts can be successfully applied to the ME growth of a mixed M-V/l.VI system.
We believe that these thermodynamic concepts combined with considerations of the
surface residence lifetimes of the incident beam species will prove to be of fundamental
importance in predicting and optimizing growth conditions for the MBE of other mixed
multilayer systems.
Acknowledgments
The authors are grateful to J. Bratton for technical support and KJ. Mackey for many
helpful discussions. T.D. Golding would like to thank Prof. M. Pepper for assistance, and
acknowledges a CASE award with the GEC Hirst Research Centre, London.
I1
References
1 T.D. Golding, M.Martinka, and J.H.Dinan, J.Appl.Phys. 64 1873 (1988).
2 T.D.Golding, J.H.Dinan, M.Martinka, A.G.Cullis, G.M.Williarns, S.Barrat,C.F.McConville, C.R.Whitehouse, J.E.Macdonald and K.M.Conway, to be published.
3 K.J.Mackey, P.M.G.Allen, W.G.Herrenden-Harker, R.H.Williams, C.R.Whitehouseand G.M.Williams, Appl.Phys.Lett 49 (6) 354 (1986).
4 D.R.T.Zahn, K.J.Mackey, R.H.Williazns, H.Munder, J.Geurts and W.Richter,Appl.Phys.Lett. 50 742 (1987).
5 D.R.T. Zahn, T.D.Golding, K.J.Mackey, W.Richter, M. Martinka, .J.H. Dinan andR.H.Williams, to be published.
6 D.R.T.Zahn, T.D.Golding, K.J.Mackey, T.Eickhoff, J.Geurts, W.Richter, M.Martinka, J.H.Dinan and R.H.Williams, to be published.
7 B.Legendre, B.Gather, and R.Blachnik, Z.Metallkd. 71, 588 (1980)
8 K.J.Mackey, D.R.T.Zahn, P.M.G.Allen, R.H.Williams, W.Richter and R.S .Williams,J.Vac.Sci.Technol. B5,1233 (1987).
9 B.Grzeta-Plenkovic, S.Popovic, B.Celustka, Z.Ruzic-Toros, B.Santic and D.Soldo,J.Appl.Cryst. (1983) 16, 415.
10 D.R.T. Zahn, N.Esser, W.Pletschen, J.Geurts and W.Richter, Surf. Sci 168, 823(1986).
11 D.E. Aspnes, Handbook on Semiconductors, Vol.11, ed. M.Ballcanski (North-Holland,New York, 1980), p.109.
12 A.J. Noreika, M.H. Francoinbe and C.E.C. Wood, J.Appl.Phys. 52 7416 (198 1).
13 M.Wihl, P.J. Stiles and J1. Tauc, Proc. 1 Ith Intern. Conf. on the Physics ofSemiconductors, Warsaw 1972, p.4 84.
14 M.A. Nizarnetdinova, Phys. Status Solidi b 97, K9 (1980).
15 C. Vazquez-Lopez, H. Navarro, R. Aceves, M.C. Vargas and C.A. Menezes,J.Appl.Phys. 58, 2066 (1985).
16 V.P. Zhuze, A.1. Zaslavaskii, V.A. Petrusevic, V.M. Sergeiva and A.I. Slelykh, Proc.Int. Conf. on the Physics of Semiconductors, Prague, 1960, p.87 1.
17 S.Sen and D.N. Bose, Solid State Commun. 50, 39 (1984).
18 D.L. Greenaway and M. Cardona, Proc. Int. Conf. Physics of Semiconductors, Exeter,1962, p.666.
19 G.Guizzetti, F.Meloni and A.Baldereschi, J.Phys.Soc.Japan 49 19809 Suppl. A p93-96.
20 Change of enthalpy calculated from heats of formation obtained from Ref.7
13
7",-A
6
Figure captions
Figure-Reflection high energy electron diffraction pattern observed along [011] azimuth duringgrowth of In 2 Te3 with Ts=300'C and JTe/JIn = 3/2.
Figure 2Energy dispersive X-ray spectra of 7000A thick indium telluride layers grown by MBE ata growth temperature of 300C and with flux ratios (a) JTe/JIn < 3/2 and (b) JTe/JIn 3/2.The spectra are compared with the spectrum obtained from a In2 Te3 bulk standard.
Figure 3Auger spectrum of MBE grown In2 Te3 layer.
Figure 4Raman spectra (1=514.5nm, T=80K) of (a) 1 100A thick In 2Te3 layer capped with 50A ofSb. (b) In2Te3 bulk standard.
Figure 5Diffraction traces of 1 100A thick In2Te3 layer grown on InSb(100) substrate for (a)InSb(200) reflection. (b) InSb(400) reflection.
Figure 6Oscillation pattern of In2Te3 /InSb(100) taken about (01 1/) oscillation axis. All reflectionscan be indexed based on a f.c.c. lattice with lattice parameter of 18.50A.
Figure 7Photoreflectance spectra of I 100A In2Te3 layer. The transition energy is 1.19±0.02eVand 1.31±0.015eV, at 300K and 77K respectively.
Figure 8Auger depth profile of 500A In2Te3 layer grown on an InSb homoepitaxial layer.
Figure 9Auger depth profiles of the attempted growth at 300C of (a) InSb, (b) CdTe with JCd/JTe= 1, (c) CdTe with JCd/JTe = 3. on 500A thick In2Te 3 layers. The estimated nominalthickness of the InSb and CdTe growths was 500A.
14
-STANDARD
In Te
In Te
C"
co '
0m0n
cmm
en S
cm0 o
CQ
cmca
3p3
NP
a
cm 8
LD
eca
I-
A&
Uka
Ilo cm
I,-
50 150 250
i5 (cm 1 1
b
S A'
i-
U, m
U16O 502'
'I-. c' -
-- i I II I I
a800
In2Te3 (600)
t= 600
InSb (200)
g
00
2 400
026 28 30 32
28 (degrees)
b16000
o
InSb (400)~12000at
S8000
LU
S4000 Z-In 2TO3 (12,00)
- X33
56 58 s0 62
20 (degrees)
Eg (3000K) Eg (77*K)
0.9 10111.2 1.3 1.4 1.5
ENERGY (eV)
a
IL
SPUTTER TIME (MIN.)
ToV
SPUTTER TIME IMI.)
Molecular-beam epitaxial growth of InSb/CdTe heterojunctions formultilayer structures
T. D. Golding," M, Martinka, and J. H. DinanU. S Army Center for Night Vision and Llectro-Optics. AMSEL-RD-NV-IT Fort Belvoir.Virginia 22060-5677
(Received 10 March 1988; accepted for publication 27 April 1988)
We have used the technique of molecular-beam epitaxy to grow layers of CdTe on InSb andInSb on CdT e and have performed a detailed analysis of the layers and their interfaces usingAuger depth profiling and reflection high-energy electron diffraction. We show that significantimprovements in interfacial quality can be obtained by the proper choice of fluxes duringgrowth. The use of a Cd/Te flux ratio of 3:1 (Jcd/JT, = 3) during the growth of CdTe hasenabled epitaxy at a substrate temperature of 300 "C. Interfaces formed with this flux ratio areabrupt, in sharp contrast to those formed under stoichiometric flux conditions (Jcd/JT, = I).Subsequent growth of InSb at a substrate temperature of 300 "C on thin CdTe epilayers (400and 800 A) is examined as a function of the InSb growth rate and Sb/In flux ratio. Quality ofthe interfaces shows a progressive improvement with increasing InSb growth rate.
I. INTRODUCTION growth of CdTe on InSb homoepitaxial layers, and of InSb/here is at present considerable interest in the InSb/ on CdTe epilayers for potential applications in two-dimen-
CdT," material system. The mixed III-V/II-VI heterostruc- sional electron gas (2DEG) and electron tunneling devices.tur" has a near perfect lattice match (Aa/a < 0.05%), Auger depth profiling has been used to examine the layers
- combined with the large difference in band gaps and their interfaces. We show that significant improvements
bet ,,cen CdTe (E, = 1.44 eV) and InSb (E, = 0.18 eV), in interfacial quality can be obtained by the proper choice of
offL.rs great potential for the fabrication of quantum-well ]a- fluxes during growth. CdTe layers have been grown on InSb
sers and detectors spanning the photon energy range 0.2-0.5 for substrate temperatures of 200 and 300 "C under stoichio-
eV (2.5-6/um).The small effective mass of electrons in InSb metric (Jcd/Jr, = 1) and Cd-enhanced (Jcd/J, = 3) flux
preritses low-temperature HEMT structures with excep- conditions. The use of an enhanced Cd flux has enabled epi-
tionially high electron mobilities' and should allow quantiza- taxy of CdTe on InSb at 300 "C, a temperature which is com-
tion effects to be seen in significantly wider wells than with patible with the subsequent growth of electrically active
the more familiar material systems. InSb. Growth of InSb epilayers on thin layers (400 and 800Attempts at experimental realization of multilayer A) of CdTe at a substrate temperature of 300 "C is examined
structures have met with only limited success. Although as a function of InSb growth rate and Sb/In flux ratio. Inter-
good structural properties have been reported for CdTe lay- faces are intermixed, but show a progressive improvement
ers grown by molecular-beam epitaxy (MBE) on InSb sub- with increasing InSb growth rates.strates2
.3 and homoepitaxial layers' and for InSb layers
grown by MBE on (100) oriented CdTe substrates,s recent II. EXPERIMENTstudies"' of the growth of CdTeon InSb under typical MBE CdTe and InSb epitaxial layers were grown in a Variangrowth conditions have shown that the interfaces are com- 360 MBE system equipped with a quadrupole mass analyzerplex due to a chemical reaction between InSb and CdTe. A and in situ RHEED and flux monitoring facilities. Base pres-study of the growth of InSb/CdTe multilayers has been re- sure during growth was below 5 x 10- s Tor. Prior to load-ported, 7 where reflection high-energy electron diffraction ing, InSb (100) substrates were solvent cleaned and mount-(RHEED) patterns indicated that individual layers grown ed onto molybdenum support blocks using a colloidalat a substrate temperature of 220-240 C were ordered. Al- suspension of graphite in alcohol. Immediately beforethough structurally good InSb epilayers have been reported' growth the native oxide was removed from substrate sur-for substrate temperatures in the range 225-275 "C, electri- faces by heating at 410 "C in an Sb, flux. A single effusion cellcally active layers have only been reported" for substrate containing high-purity CdTe was used to provide a stoichio-temperatures in excess of 270 C. To be useful for electronic metric beam of Cd and Tez (Ref. 2) and was supplementeddevice applications which involve tunneling, the thickness of by a cell containing Cd for Cd-enhanced (Ucd /Jt > 1) stud-individual layers of a multilayer device must be limited to ies. A relative measure of the flux from a given cell was ob-several hundred angstroms. To facilitate growth, a substrate tained by interposing an ion gauge flux monitor into the mo-temperature must be found which results in high structural lecular beam and relating the measured beam equivalentand electronic quality of both CdTe and InSb. pressure to the cell temperature, molecular weight, and ioni-
In this paper, we present results of a study of the MBE zation efficiency of the beam species. Normal settings for the
'Permanent address- Cavendish Laboratory. University of Cambridge, CdTe cell gave a total flux of 2.4 X 10" atoms/cm /s whichCambndge. England, corresponded to a homoepitaxial growth rate at T, = 220 "C
1873 J. AppI. Phys 64 (4), 15 August 1968 002-8979/68/161873-05S02 40 ( 1988 Arncarn inattute of PhysS 1673
of approximately 1.7 A/s (0.6pm/h). The Cd cell was set tosupply its maximum safe flux of 2.6 x 10" atoms/cm2/s giv-
ing a flux ratio value Jcd/JT. = 3 during the Cd-enhancedgrowth experiments. Separate effusion cells containing high-purity In and Sb were used for growth of InSb. All InSbhomoepitaxial and heteroepitaxial layers were grown at T,= 300 "C with growth rates in the range 0.3-2.0 A/s (0.1- r
0.7 Mm/h). Practical considerations related to the In celltemperature prevented growth of InSb with rates in excess of g - 0
0.7 pm/h. J4/J, was kept in the range 1.1-3.5 to keepwithin the (2 X 2) Sb-stabilized surface reconstruction" asindicated by RHEED.
Immediately following growth, each sample was trans-ferred in air to a Physical Electronics 560 scanning Auger 1 u L M, 3 a' •2microscope configured with a double pass cylindrical-mir- Sa,' TO, 1100.1
ror analyzer and a differentially pumped and gettered argon 1b)ion gun. Depth profiles of the constituent species, as well asof carbon and oxygen, were obtained by alternate sputteringand data acquisition periods of 30 and 80 s, respectively. Nocarbon or oxygen recontamination was seen within the lay-ers during the profile. To reduce artificial broadening of thegrowth interface, the sputter beam of 4-kV argon ions was Toincident at 40' from the sample surface normal with the sput-ter rate reduced to 20-30 A/min by expanding to a 6 X 6 mmraster. The growths and sputter depths were limited to a few U --
thin compositional periods to reduce roughness due to pref-erential ion etching of (100) InSb as observed in electronmicrographs. The electron beam was used in spot modc andincident on the surface at 40" from the surface normal. An I0 4 a D t is X) •4 a do
electron energy of 3 kV was used with a moderate and con- ,.-rfp, " IN I
stant current of 0. 16pA. Auger peaks overlapping the MNNtransitions were minimized by collecting high energy resolu- FIG. 1. Aulerdepth proilesof 300-s growths ofCdTe on InSb homoepitx.tion, AE/E = 0.4%, pulse count data from selectively con- ial layers with .4,/Jr. - I for substrate temperatures of (a) 200 and (b)tracted energy "windows." 300"C.
The Auger analysis was calibrated in the usual manner.The analyzer energy scale was referenced to the known ener-gy, 2 keV, of electrons backscattered from a sample surface (2 x 2) InSb RHEED pattern vanished immediately afterwhile at its focal point, and the beam current was measured CdTe growth was initiated, indicating a disordered growth.from the sample stage while biased to + 130 V. The current The resulting layer is deficient in Cd and rich in In and Te.
S.. - density of the argon beam was 600pA/cm- with FWHM of Layer thickness is estimated to be 150 A. These results are550 pm and the sputter rates were referenced to thin-film similar to those recently reported by Mackey et aL.,' wherestandards and to bulk (110) CdTe. These agree well with indium telluride compounds were identified at the InSb/our known MBE growth rates. Finally, Auger signal CdTe interface. It has been suggested that the lack of Cdstrengths from bulk (100) InSb and (I 11) CdTe were used incorporation in layers grown at 300 C may be attributed toto judge stoichiometric conditions within the MBE layers. icroaini aesgona 0 Cmyb trbtdt
jhoiy a Cd deficiency on the growth surface, thereby allowing Te
to react with the InSb. To overcome this Cd deficiency wegrew layers of CdTe on InSb using an enhanced Cd flux. For
Ill RESULTS 7, = 300 "C, the additional Cd flux has a dramatic effect on
Auger depth profiles of 300-s growths of CdTe under the quality of layers and interfaces. An Auger depth profilestoichiometric beam conditions (jcd/JT, = I) on 1000-A of a 300-s growth of CdTe using Cd-enhanced flux condi-InSb homoepitaxial layers at substrate temperatures 7, tions (Jcd/JT, = 3) isshown in Fig. 2(a).The (2x2) InSb= 200 and 300"C are shown in Fig. 1. For growth at 7, RHEED pattern converted toa streaked (2X 1) patternim-= 200 "C, the (2 x 2) lnSb RHEED pattern convened to a mediately after CdTe growth was initiated. The depth pro-
streaked (2 X I) CdTe pattern immediately after CdTe files show that the interfacial width for this sample is asgrowth was initiated. The epilayer is stoichiometric with a abrupt as that grown under stoichiometric flux conditions atthickness of approximately 500 A which corresponds closely T, = 200 *C. To establish that the high-quality epitaxy ofto that calculated from flux measurements. Width of the thin CdTe layers grown at elevated temperatures is due tointerface is estimated to be 120-180 A, a value which is near the enhanced Cd flux and not to the total Cd flux or growththe Auger resolution limit. For growth at 7', - 300 "C, the rate, the temperature of the CdTe cell was set to give a Cd
1674 J. App Phys.. Vol. 64. No. 4, IS August 1966 Gm. Martnla. an Dwan 1874
and I min for growth rates of 0.2. 0.4, and 0.7ym/h, respec-tively. Auger analysis reveals that the degree of intermixing
d, is independent of J/J,. but dependent on the lnSb growthrate, with the highest growth rate resulting in the least inter-mixing. The decrease in In and Sb in the CdTe underlayer as
S. the InSb growth rate is increased is evident in Fig. 3. For allgrowth rates, however, In and Sb are present throughout the400-A CdTe layer leading to a deterioration and broadening
Wof the initial (CdTe on InSb) interface. To prevent this~~J • ,=. fteiiil(deo nb nef Ce.T lyren this
broadening of the initial interface, an 800-A CdTe layer wasgrown. The depth profile of this structure, grown with ourmaximum InSb growth rate of 0.7 pm/h, is shown in Fig. 4.
.- :There is a distinct asymmetry between the first and secondI a 6 a 3 1 . interfaces. Growth of the InSb epilayer has induced a Cd loss
Sr" Nresulting in an In- and Te-rich interfacial region and incor-poration of In and Sb throughout the CdTe layer. In a sepa-rate study, small shifts in energy and shape of Auger peaks
/a
-.. ' TI\. -
toal dyeru in (ate ad'rteol ~ ~ ~ ~ PM TO IalC fuMi ()
component equal to the total Cd flux used during the Cd- 2enhanced growth. An Auger depth profile of a 300-s growthof CdTe at T, - 300C with .///IT. = I and total flux of7.2 x 10l atoms/cm2/s (corresponding to a growth rate of1.5 pm/h) is shown in Fig. 2 (b). Again, the layer formed isrich in In and Te with severely degraded interfacial quality.
The ability to grow thin layers of CdTe on InSb at T,300 C made it possible to study the subsequent growth of-.
InSb on these layers. Layer and interface quality of the films I'-
grown was examined as a function of the InSb growth rate \, Nrand Sb/In flux ratio. Epitaxy of InSb on CdTe epilayers was to.not achieved for growth rates less than 0.15pm/h for"s4.lJt. in the range 1.1-3.5. Immediately after initiation ofgrowth of InSb, the streaked (2 X 1) CdTe RHEED patternbecame spotty, an indication of three-dimensional nuclea- 'Ition. This spotted pattern remained unchanged throughout .1ithe growth period. Auger depth analysis revealed complete ' I .;..degradation of the 400-A CdTe underlayer with severe inter-mixing throughout. For growth rates of 0.2 pm/h or higher, d is X" U a* n I2 2 , iS , n
the (2 x I) CdTE RHEED pattern immediately became S •,=.i,
spotty, but gradually changed to the (2 x 2) Sb-stabilized FIG. 3. Auger depth profile of an InSb epilayer grown with a substratepattern characteristic of homoepitaxial InSb. The time taken temperature of 300 "C on a 400.O-thick CdTe layer with In and Sb fluxes
for the RHEED pattern to evolve was approximately 3, 2, cowrresponding to growth rates of (a) 0.7, (b) 0.4. and (c) 0.2 m/h.
t75 J. App. Phy.. Vol. 64. NO. 4. 15 Augu t 1968 Goklkin. Martirka, and Drnan 1675
It
I
*are in progress to assess the 2DEG transport characteristics.. It would be of interest to extend our growth studies to in.
- elude higher values of Jcd/I.. The ability to grow CdTe1A .layers on InSb at 200 "C with J /J., = 3 clearly suggests
4that higher flux ratios could be beneficial at higher growth, temperatures.
* ., ,,, ! /.For the growth of InSb on CdTe, the dependence of4 i! /"interfacial abruptness and epitaxy on the InSb growth rate
.msuggests a competitive process between the intermixing anddeterioration in the CdTe and the nucleation and growth ofInSb. It is possible that broadening of the interface profile iscaused by nonuniform sputtering of the InSb. however, the
Iobserved Cd deficiency indicates that deterioration of theSt o ,a 11 ,, so W. i W CdTe has occurred during the InSb growth, most likely due
S ,l,,,. Tugto formation of indium telluride compounds. Results similarto those in Fig. 4 have also been obtained for a structure
FIG. 4. Auger depth profile of an 1nSb/CdTe/InSb structure pown at a grown at 275 6C, indicating that atomically abrupt interfacessubstrate temperature of 300 C with an equivalent growth rate for lnSb of0.7 /*m/h. The CdTe layer was grown to a thickness of = S0 A under Cd- may not be obtainable for values of T, compatible withenhanced flux conditions (c1JT., 3). growth of electrically active InSb.
V. CONCLUSIONfor the elements at each interface indicated similar electronic In conclusion, we have found evidence of a strong chem-environments at each interface and the presence of excess ical reaction which occurs at the CdTe/InSb interface dur-antimony. ing MBE growth at a substrate temperature of 300 *C. For
growth of CdTe on InSb, we have presented a technique forIV. DISCUSSION suppressing this reaction by use of Cd-enhanced flux condi-
We have shown that an attempt to grow CdTe on InSb tions. For growth of InSb on CdTe layers at a substrate tern-at T, = 300 "C under stoichiometric flux conditions results perature of 300 "C, we have found that epitaxy is stronglyin a severely intermixed and disordered layer. The use of an dependent on the InSb growth rate. Our results extend theenhanced Cd flux during growth has a dramatic effect, re- substrate temperature range to 225<7T,<300":C, withinsuiting in a CdTe layer which is epitaxial and an interface which single-crystal layers of both CdTe and InSb can bewhich is abrupt. This suggests that the residence time of Cd grown at a fixed substrate temperature. We have shown thaton the InSb surface is less than that of Te, resulting in a Cd multilayer growth is possible at substrate temperatures corn-deficiency at the growth surface and allowing Te to react patable with electrically active InSb, but that thin periodwith InSb to form indium telluride compounds, as suggested (200 A) InSb/CdTe structures for tunneling devices mayby not be feasible due to breakdown of the CdTe barrier. How-
2 InSb,,t + 3 Teti) -ln2Te3(,) + 2 Sbt,), ever, the question of whether the chemical profiles presented23T T 2bare adequate for 2DEG or tunneling devices an be answered
t-for which the change in enthalpy AHis - 30.9 kcal. Use of only after electrical transport measurements have beenan enhanced Cd flux clearly reduces or eliminates the Cd made."
deficiency, ensuring that sufficient Cd is available on theInSb growth surface to bond with the incident Te and inhibitaccess of Te to the In. Thus, perhaps, the enhanced Cd fluxneed only be applied during the initial stages of growth to ACKNOWLEDGMENTSallow nucleation of the CdTe, after which the growth can be The authors are grateful to P. Boyd for electron micros-continued under stoichiometric beam conditions. copy studies, J. Bratton for technical support, and S. Greene
We believe that our data for growth at T, = 300 'C of and Dr. M. Pepper for assistance and discussions. T. D.CdTe on InSb have implications for growth at 200 *C. Corn- Golding acknowledges a CASE award with the GEC Hirstparison of the interfacial width of layers grown on InSb at Research Centre, London.200 "C with and without enhanced Cd flux indicates that anenhanced flux improves interfacial quality at this "typical"MBE growth temperature. We speculate that the formationof In,Te seena in samples grown under similar conditions is 'R. 0. vanWelzenis and B. K.Radley. Solid State Electron. 27. 113 (1984)
of " sR. F. C. Farrow, G. R Jones. G. M. Williams. and i. i. Young, Appi.reduced or eliminated by use of enhanced Cd flux during Phys. Lett. 39. 954 (1981).growth. The relatively low mobilities reported ' for a 2DEG 'T. H. Myers, Yawcheng Lo. J. F. Scheizmna. and S R. Jost. J. Appi Phys.at an lnSb/CdTe interface may result from indium loss with- 53. 9232 (1982).in the lnSb due to indium telluride compound formation at 'S. Wood. J. Gregg. R. F. C. Farrow, W J Takei, F. A. Shirland. and A. .
Noreika. J. Appi. Phys. S. 4225 (1984).the interface. Samples have been grown at , = 200 "C using K. Sugoyama. J. Cryst. Growth 60. 450 (1932).an enhanced Cd flux and magnetotransport measurements *K. J. Mackey. D. R. T. Zahn. P M.G. Allen. R. H Williams. W Richter.
1876 J. Appi. Phys.. Vol. 64, No. 4. 15 August 1983 o0k"g. Marmnka. and'lmnn 1876
and R. S. Williams, J. Vac. Sci. Technol. B S, 1233 (1987). 7416 (1931).'M. Kimata. A. Ryoji, and T. Aoki. J. Cryst. Growth 81. 50 (1987). 'Y. D. Zheng, Y. H. Chang, B. D. McCombe. R. F. C. Farrow, T. Temo-'G. M. Williams. C. R. Whitehouse., T. Martin. N. 0. Chew. A. 0. Cullis, font, and F. A. Shirland. AppI. Phys. Lett. 49. 1137 (1986).
, T. Ashley, D. E. Sykes. and K. Mackey. presented at 4th MBE Interna. "G. M. Williams, C. R. Whitehouse. N. 0. Chew. G. W. Blackmore. anduoW Conference, York, England, 1986 (unpublished). A. 0. Cullis, J. Vac. Sci. Technol. B 3. 704 (1935)."A. J. Noreika. M. H. Francombe. and C. E C. Wood. J. Appl. Phys. 52,
.9
,67 .. ApI ty. ol 4 o 41"-..198Glug.Mraia adOa a
%A
a-06
Qa:
20.L
0
Z L
02
L 0
LU
0
LU
0
NON-LINEAR OPTICALPROPERTIES.
GARY WOOD.
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
0%Rockwell IntemationScience Center
SC4466T
zA.
oj jW V.oI 0.
EaW
0 1411 LIDl00 ka 1
__~o I-. _ _ _ _ _ _ j__ ____I_00 _ I 1 0.zI- 0 ,MII.
2' EwZOU
LL 4W &W 0
Ij ILL~g
I- I z-j~~ iiI0 MiPwu - t.L zE
U>1 0 4.
W9 22A/ejwti
A0
.90%*
&a0
0j
0w LU L
= j -u zz ZZ
U-j LU
z. i- Cl)
-o LUj0) -) -L *m
LUZ W C) N J CDu
~0 .z Z U.o)
U L) WxCl)
LU 0
0 o~z
U4)
0 0 C
o E z (4,Zz E.J
0 4Uz
=)4 z10 0
0n zP 4
o 0J
w cIo0i~ 4 D
0i -z IoI 0 0
0 gC.U)z 0 U
4L <z x I-rZ m -CL -
0 0 00
00
0
o.-
30 u 0
ooA + -i wLL ecc00c w m 4.-L
V ~ CC M~
UJ r w z
zo~~c 57~JE->w
z C.) WOi UA z LCDccaIr dOII
0,
0m Cm)
0c0 0
0
Om~z E.
Zz I- C.00 z CD (
z am T2. E .)
aw _. wZ co m (G
o4 4)Z U)~0w2 CL C)11 1
Zi 0 Z
z z CU
Z IL0j 0 0
0
ct
b.
0 0 c
cmczNE
SOFft
be NzL Oz 0 xm
w0
GU 0
0
U'cc
-L
41 z CC.
-4 PI
q-
0 .5 A
Co C1 V) Wo ) uoj
w 4c
.j 0
0 4c
m w N m(n'e9 ..oL0 (C)X
dom
C00
COC
p 4..
0 C-
0, 0 6U. E 0
z0 000 2D
W0 E w0Uw z) C4so
0000~ (Oj()
Cm co
< 0
U) 6)
ui- j
0~~( CL >, 3>
0 ICc. c rUl UJ ui '4
z 0
0
Cv
z C
0 z
Z>- 0
z wL ao (0 0
0Zo u
C U,wh0
CID z
ENEwn 0 U;
ZU CYEco
0 0
Icc
00
CL I m
cc- 0
z o
C'C,
I .
w c CLwl -
Iwo 0z U0
00N N c
N N 0 0)-Z~
0(0
uju
) U) N o
Z., 0
0 0
toC U) In
0z 00
oo z 0 X OP 0 L
z Z
0 0 0 Z
> Z.
ccLLJ 4)
0
0
0~ ow
0
Om z
LLJ z4n Lu
0 'p.0
KU 0 z0ID 00
6. o<c- N C
uCl, V p
LUi U, T
000; 2 1-a - V- 0J
Ch mI ~<ZA X Z
x ~4 - 5c
P 0 s P
z x0
CYE o
NN CCK~
9- - V
0 z
-I. - T-
w Z
OC) 0 ZU3 00
t to N j-K 9
LU K o 0 0
to go toI. c
wn 0- x0 xzo0. g o ~INE 9- 0;
K 9- 9-
C- K.iA x x- --
0 4
a.
0
ViLU
z>m0
U
oLL0
Zw
IL.
I--z
0LU
LASER DIODE ARRAYS.
DAVID CAFFEY
DECEMBER 7,1988.CECOM CENTER FOR NIGHT VISION
AND ELECTRO-OPTICS.
II"
4L4
040
CD w
.9 IM
LU I
0"0
wEZL 0
00
wwLA-
CL >-AaMz0
C
w wcc M
-cnoLUZa: U
00
000
Im n0000-ICC)
I-r
ui0o
UA U) ClN Co UC an CD toO 4 o C
V) M~3~liGo zj
AI' t ". U :'M
7 ~ ' .1
aR
-lob
Cgo - - 4
&luftAlCf
860. - .aDLLP*. .o '
840 -
.. x.. Theoretical PlotE 820" STC Paignton Resul-- S/ ltno
cc'a/
780 I
7607I I iiJ" -
20 40 60 80 100 120
Well Width (A)
Shift in Emicted Wavelength as a Function
of Quantum Well Thickness, LL z
V.
b0
E
0
0 a
>.aa
a,
a CB
i--
(uJua
'~p
4X
CL~~ 4--50 0
0-4
EXC 61
cc
75-
05~
I * -~ .1
5.- ~ ~ -, 54 4- 5-.
I I I I I ~ . i. 4tu
.1 .. .' ~' ;& P. 4>~U'5. .5-'545i'
A(3~JRF~IC~ C)F ,. 1 Nd :YA~ Lu
0.0.
80080 4- -80
WAEENT
ADVANCED SOUD STATE LASER PUMP STUDY
10-1299a
* W AWSW RilSES. 1W0 im DtitATWI. 10 ma
0 11"T Slik TESI6RATURE -2.4
a HOKII'AL SCALE 1 rn PER DI~iSiON
Figure 14. Primary array wavelength
10-12999
813
812
I811I-A
k,>z1
SLP 02nmOas
FJ15 8102
CURRENT CURRENT
Figure 15. Primary array wavelength versusheat sink temperature
M~~o~w1IDii~a~ T4M~6 C~~ew - Lbf'eV
61001
161
E CD
owIj
0 >a U
DIODE BAR FABRICATION
METAL"Tg0• IiN114 SLon
OF0T WMWAIM I aI " M ''Bpft
LASM P Ar
D-
OMOAL
D~O~s
I0 h~ta1S -oI 0
bi/
In,
1~1 4-4
//
/ ~A 'LI
f~ 0/
b2
~jf
/ aU,I 0
III.. N-
I U~uI // I
/I Ia /1
TYPE I SURFACE EMITTERS WITH LASINGPARALLEL TO THE SURFACE
OUTPUT LIGHT OUTPUT LIGHT
(a) ILASER STRUCTURE
ETCHED DEFLECTORS
LASER WITH ETCHED DEFLECTORS FOR OUTPUT COUPLING
OUTPUT LIGHT
til It I Itd I 2"d ORDER
-' ~ GRATING____ ___ ___ ___ ___ ___ ___ ___ ___ ___ _M_ ACTIVE RIN wON
(b)
LASER WITH A GRATING FOR OUTPUT COUPLING
9r.,.
4*0
CSI2~v ~rhw-
Cog)'
Cl)~~ ~ P _.41
c4
ct a
02~~vo
42 'Emj0
0.WI-n(CLg
rA >= LA-
zE
wow
.... .....
4" 44
7 p
It, - -0
~Z'f OR
A4
0p
46
LAPE
a,,p
>6
4'4
-Ile
let,
2 02
4* 1
rocu
lCL 44 -
V) '4.
'rrt 'k-.
z
IL0
:-
4. LIST OF ATTENDEES
OPTOELECTRONIC WORKSHOP
SUPERLArrICE DISORDERING
December 7, 1988
NAME ORGANIZATION PHONE
Rudolf Buser NVEOC 703-664-3067Paul Amirtharaj NVEOC 703-664-5043Frederick Carlson NVEOC 703-664-5036Ronald Graft NVEOC 703-664-5780John Dinan NVEOC 703-664-5825Jan Davis NVEOC 703-664-Unchul Lee NVEOC 703-664-5780Michael Martinka NVEOC 703-664-5043John Pollard NVEOC 703-664-5780Gary Wood NVEOC 703-664-5767
Ronald Antos UR, Optics 716-275-4179Susan Houde-Walter UR, Optics 716-275-7629Brian Olmsted UR, Optics 716-275-5101
