applied physics reviews the physics and technology of

APPLIED PHYSICS REVIEWS

The physics and technology of gallium antimonide: An emergingoptoelectronic material

P. S. Duttaa) and H. L. BhatDepartment of Physics, Indian Institute of Science, Bangalore-560 012, India

Vikram KumarSolid State Physics Laboratory, Lucknow Road, Delhi-110 054, India

~Received 22 February 1996; accepted for publication 20 January 1997!

Recent advances in nonsilica fiber technology have prompted the development of suitable materialsfor devices operating beyond 1.55mm. The III–V ternaries and quaternaries~AlGaIn!~AsSb! latticematched to GaSb seem to be the obvious choice and have turned out to be promising candidates forhigh speed electronic and long wavelength photonic devices. Consequently, there has beentremendous upthrust in research activities of GaSb-based systems. As a matter of fact, thiscompound has proved to be an interesting material for both basic and applied research. At present,GaSb technology is in its infancy and considerable research has to be carried out before it can beemployed for large scale device fabrication. This article presents an up to date comprehensiveaccount of research carried out hitherto. It explores in detail the material aspects of GaSb startingfrom crystal growth in bulk and epitaxial form, post growth material processing to device feasibility.An overview of the lattice, electronic, transport, optical and device related properties is presented.Some of the current areas of research and development have been critically reviewed and theirsignificance for both understanding the basic physics as well as for device applications areaddressed. These include the role of defects and impurities on the structural, optical and electricalproperties of the material, various techniques employed for surface and bulk defect passivation andtheir effect on the device characteristics, development of novel device structures, etc. Severalavenues where further work is required in order to upgrade this III–V compound for optoelectronicdevices are listed. It is concluded that the present day knowledge in this material system is sufficientto understand the basic properties and what should be more vigorously pursued is theirimplementation for device fabrication. ©1997 American Institute of Physics.@S0021-8979~97!04109-1#

l EN

TABLE OF CONTENTS

I. Importance of gallium antimonide. .. . . . . . . . . . . . . 5822II. Compound preparation and crystal growth. . . . . . . 5823

A. Phase equilibria. . . . . . . . . . . . . . . . . . . . . . . . . .5823B. Bulk growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5824

1. Czochralski technique. . . . . . . . . . . . . . . . . . 58252. Bridgman technique. . . . . . . . . . . . . . . . . . . . 58253. Vertical gradient freeze technique. . . . . . . . . 58264. Travelling heater method. . . . . . . . . . . . . . . . 58265. Liquid phase electro-epitaxy. . . . . . . . . . . . . 58266. Growth under microgravity. . . . . . . . . . . . . . 58267. Growth under hypergravity. . . . . . . . . . . . . . 58278. Bulk growth of GaSb based ternaries. . . . . . 5827

C. Epitaxial growth. . . . . . . . . . . . . . . . . . . . . . . . . .58281. Liquid phase epitaxy. . . . . . . . . . . . . . . . . . . 58282. Vapour phase epitaxy. . . . . . . . . . . . . . . . . . . 58293. Molecular beam epitaxy. . . . . . . . . . . . . . . . . 5829

a!Present address: Department of Mechanical Engineering, Aeronauticagineering and Mechanics, Rensselaer Polytechnic Institute, Troy,12180. Electronic mail: [email protected]

J. Appl. Phys. 81 (9), 1 May 1997 0021-8979/97/81(9)/5821/

Downloaded¬26¬Apr¬2003¬to¬128.113.123.115.¬Redistribution¬subjec

4. Metal-organic molecular beam epitaxy. . . . . 58305. Plasma assisted epitaxy. . . . . . . . . . . . . . . . . 5830

III. Structural properties. . . . . . . . . . . . . . . . . . . . . . . . .5830A. Lattice parameter. . . . . . . . . . . . . . . . . . . . . . . . .5830B. Density. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5830C. Crystal structure. . . . . . . . . . . . . . . . . . . . . . . . . .5830

IV. Thermal properties. . . . . . . . . . . . . . . . . . . . . . . . . .5830A. Heat capacity and Debye temperature. . . . . . . . . 5830B. Elastic moduli and phonon dispersion. . . . . . . . . 5831C. Thermal expansion. . . . . . . . . . . . . . . . . . . . . . . .5832D. Thermal conductivity. . . . . . . . . . . . . . . . . . . . . .5832

V. Electronic and transport properties. . . . . . . . . . . . . 5834A. Band structure. . . . . . . . . . . . . . . . . . . . . . . . . . . .5834B. Effective masses of electrons and holes. . . . . . . 5835C. Electron transport. . . . . . . . . . . . . . . . . . . . . . . . .5835D. Hole transport. . . . . . . . . . . . . . . . . . . . . . . . . . . .5837E. Magnetophonon effect. . . . . . . . . . . . . . . . . . . . .5839F. Electron and hole transport in ternaries. . . . . . . 5839

VI. Optical properties. . . . . . . . . . . . . . . . . . . . . . . . . . .5840A. Dielectric constant. . . . . . . . . . . . . . . . . . . . . . . .5840B. Photoconduction. . . . . . . . . . . . . . . . . . . . . . . . . .5840

n-Y

582150/$10.00 © 1997 American Institute of Physics

t¬to¬AIP¬license¬or¬copyright,¬see¬http://ojps.aip.org/japo/japcr.jsp

uth

re-

-vticsns

if

aer

mte

ionsse

n

h asial

avew

ofic-aSbin

onc-r-eenicesen

ari-4

heateroystec-ng,

likeasm-

nd

C. Photoelectric threshold and work function. . . . . 5841D. Nonlinear optical effect. . . . . . . . . . . . . . . . . . . . 5841E. Radiative recombination and stimulated

emission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5841VII. Defects and impurities. . . . . . . . . . . . . . . . . . . . . .5841

A. Extended defects. . . . . . . . . . . . . . . . . . . . . . . . . .5841B. Native defects. . . . . . . . . . . . . . . . . . . . . . . . . . . .5842C. Shallow dopant impurities. . . . . . . . . . . . . . . . . . 5843D. Deep level impurities. . . . . . . . . . . . . . . . . . . . . .5844E. Magnetic impurities. . . . . . . . . . . . . . . . . . . . . . .5845F. Isotopic effects. . . . . . . . . . . . . . . . . . . . . . . . . . .5845G. Self- and impurity diffusion. . . . . . . . . . . . . . . . . 5845H. Ion bombardment induced defects. . . . . . . . . . . . 5846

VIII. Surface and bulk defect passivation. . . . . . . . . . 5846A. Wet chemical treatment. . . . . . . . . . . . . . . . . . . . 5846B. Hydrogen plasma passivation. . . . . . . . . . . . . . . 5847C. a-Si:H passivation. . . . . . . . . . . . . . . . . . . . . . . .5848

IX. Device aspects. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5849A. Wafer preparation. . . . . . . . . . . . . . . . . . . . . . . . .5849B. Dry etching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5850C. Atomically clean surfaces. . . . . . . . . . . . . . . . . . 5850D. Fabrication techniques. . . . . . . . . . . . . . . . . . . . .5850

1. Ohmic contacts. . . . . . . . . . . . . . . . . . . . . . . .58502. Schottky contacts. . . . . . . . . . . . . . . . . . . . . .58533. Oxides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58534. Homojunctions. . . . . . . . . . . . . . . . . . . . . . . .58545. Heterojunctions. . . . . . . . . . . . . . . . . . . . . . . .5854

E. Device structures. . . . . . . . . . . . . . . . . . . . . . . . .58561. Metal/a-Si:H/GaSb structures. . . . . . . . . . . . 58562. Injection lasers. . . . . . . . . . . . . . . . . . . . . . . .58573. Photodetectors and solar cells. . . . . . . . . . . . 58584. Transistors. . . . . . . . . . . . . . . . . . . . . . . . . . . .58595. Quantum wells, quantum dots and

superlattices. . . . . . . . . . . . . . . . . . . . . . . . . .5860X. Concluding remarks and future outlook .. . . . . . . . 5863

I. IMPORTANCE OF GALLIUM ANTIMONIDE

Historically, the research and development of varioIII–V compound semiconductors is associated withwavelength of the optical fiber loss minima.1 The shift in thefiber loss minima towards higher wavelengths from 0.8mmover the past 2 decades has shifted the material of intefrom time to time.1 Even though the present day optical communication systems are tuned to 1.55mm, the next generation systems may have to be operated well above this walength. This is because recent developments in the opfiber research have shown potentiality for certain classenonsilica fibers for optical communication applicatiowhose loss minima fall in the 2–4mm range.2 For example,the heavy metal fluoride glasses are speculated to have mmum attenuation at 2.55mm with a loss, one to two orders omagnitude lower than the present day silica fibers. Thisalso important since, at longer wavelengths, loss due to Rleigh scattering is significantly reduced. Consequently, thhas been an upthrust in research activities in new matesystems for sources and detectors operating in the 2–4mmregime. Among compound III–V semiconductors, galliuantimonide~GaSb! is particularly interesting as a substra

5822 J. Appl. Phys., Vol. 81, No. 9, 1 May 1997


se

st

e-alof

ni-

isy-reial

material because its lattice parameter matches solid solutof various ternary and quaternary III–V compounds whoband gaps cover a wide spectral range from;0.3 to 1.58eV,3 i.e., 0.8–4.3mm, as depicted in Fig. 1. Also, detectioof longer wavelengths, 8–14mm, is possible with intersub-band absorption in antimonide based superlattices.4 Thesehave stimulated a lot of interest in GaSb for basic researcwell as device fabrication. Some of the important materproperties of GaSb are listed in Table I.5

From device point of view, GaSb based structures hshown potentiality for applications in laser diodes with lothreshold voltage,6,7 photodetectors with high quantumefficiency,8 high frequency devices,9,10 superlattices with tai-lored optical and transport characteristics,11 booster cells intandem solar cell arrangements for improved efficiencyphotovoltaic cells and high efficiency thermophotovolta~TPV! cells.12 Interestingly, the spin-orbit splitting of the valence band is almost equal to the energy band gap in Gleading to high hole ionization coefficients. This resultssignificant improvement in the signal-to-noise ratio atl. 1.3 mm in GaAlSb avalanche photodetectors grownGaSb.8 GaSb is also predicted to have a lattice limited eletron mobility greater than GaAs making it of potential inteest in the fabrication of microwave devices. InGaSb has bproposed as an ideal material for transferred-electron devby Hilsum and Rees10 with a low threshold yield and a largvelocity peak-to-valley ratio, using a Monte Carlo simulatiobased on the three-level model.

GaSb-based devices are promising candidates for a vety of military and civil applications in the 2–5 and 8–1mm regimes:13 to mention a few, infrared~IR! imaging sen-sors for missile and surveillance systems~focal plane arrays!,fire detection and monitoring environmental pollution. Tabsorption wavelengths of several industrial gases and wvapour lie in the near IR range for which GaSb based allare suitable. Gas purity monitoring and trace moisture detion in corrosive gases like HCl in semiconductor processidetecting microleaks of toxic gases such as PH3, in situmonitoring of plasma etching, detecting hazardous gasesHF and H2S in chemical plants, monitoring green house gfluxes, measurements of flame species in microgravity co

FIG. 1. Band gap as a function of lattice constant for III–V compounds atheir ternary and quaternary alloys~from Ref. 5!.

Dutta, Bhat, and Kumar


e

ic1eto

ersth

eiayo

-tenrytoas.pa-

Sb isuede-ofothual

ing

y

velhistals

ifi-ess-ica-d oningit-heinguresfu-

his.

and

inedas

-ev-edThe705C

ive

ure

-

of

chf Ga

bustion and humidity determination are a few areas whGaSb based alloys might find potential applications.13 Sb-based alloys can also find several biological and medapplications in the near IR regime. IR detectors in the 8–mm regime based on GaAlSb/AlSb and InAs/InGaSb suplattices and InAsSb are believed to be potential competifor the present day HgCdTe detectors.4

GaSb has also proved to be a model material for sevbasic studies.14 Because of the band structural propertieGaSb has proved to be an ideal material for studyingAuger recombination processes.15 Due to low vapour pres-sures and low melting points, GaSb and InGaSb serveappropriate model materials to study the effects of convtion and diffusion on the solutal distribution under terrestrand microgravity conditions.16 Sulfur doped GaSb is the onlIII–V binary compound which reveals high concentrationdonor related deep traps~commonly known asDX centers!at atmospheric pressure.14 Hence it is the most suitable material for studying the behaviour of such metastable cenwithout the complication of high pressure or alloy broadeing effects, encountered in other III–V binary and ternaalloys. Because of high concentration of native acceppresent in the as-grown unintentionally doped GaSb, it isinteresting system to study impurity compensation effect14

Technological and material aspects of GaSb have singly been studied until now, compared to other III–V com

TABLE I. Material properties of GaSb~compiled from Ref. 5!.

Lattice constant (Å) 6.0959Density (gm cm23) 5.6137Melting point ~K! 985Debye temperature~K! 266Coefficient of thermal expansion (1026 °C21)~at 300 K!

7.75

Thermal conductivity at 300 K (W cm21 K21) 0.39Direct energy gap at 300 K~eV! 0.725Direct energy gap at 0 K~eV! 0.822Temperature dependence ofminimum energy gap (3 1024 eV K21) a 4.2b 140

Spin-orbit splitting energy,D0 ~eV! 0.80Effective mass of electrons~in units ofm0! 0.0412Effective masses of holes~in units ofm0!Heavy hole mass 0.28Light hole mass 0.05Spin-orbit split mass 0.13Wave number of LO phonons (cm21) 233.0Wave number of TO phonons (cm21) 224.0Refractive index~near band-gap energy! 3.82Dielectric constante0 15.69e` 14.44

Elastic compliances (310212 cm2 dyn21)S11 1.582S12 20.495S44 2.314

Deformation potential constantsa ~eV! ~for direct gap! 28.28b ~eV! 22.0d ~eV! 24.7

J. Appl. Phys., Vol. 81, No. 9, 1 May 1997


re

al4r-rs

al,e

asc-l

f

rs-

rsn

r-

pounds, such as GaAs, InSb, InP, GaP, etc. Undoped Gaalwaysp type in nature irrespective of the growth techniqand conditions. Work over the last 3 decades has beenvoted mainly for understanding the nature and the originthe residual acceptors which are the limiting factors for bfundamental studies and device applications. The residacceptors with concentration of;1017 cm23 have beenfound to be related to gallium vacancies (VGa) and galliumin antimony site~GaSb! with doubly ionizable nature.17 At-tempts have been made to reduce their content by growthe crystals from nonstoichiometric melts.18 Recent studieson epitaxial layers of GaSb grown by liquid phase epitax19

and molecular beam epitaxy~MBE! with excess antimony20

have shown the possibility of reducing substantially the leof natural acceptors and increasing the hole mobility. Tstimulated the renewed interest in growth of GaSb cryswith reduced residual acceptors.

At present, GaSb technology is in its infancy and signcant progress has to be made both in materials and procing aspects before it can be employed for device appltions. Current research and developments are focusseareas of high quality materials growth, better understandof electronic and photonic properties and fabrication of suable device structures. In this article, an overview of tbasic physics of the material, preparation and processtechnologies, and developments in practical device structand their properties is presented. Certain avenues whereture efforts should be concentrated in order to exploit tIII–V compound for optoelectronic devices are suggested

II. COMPOUND PREPARATION AND CRYSTALGROWTH

A. Phase equilibria

As early as in 1926, Goldschmidt synthesized GaSbdetermined its lattice constant.21 Since then, it was followedby several workers and the lattice constant was redetermmore precisely.22 The phase diagram of this compound hbeen determined simultaneously by Koster and Thoma23 andGreenfield and Smith.24 Later on, the liquidus has been reevaluated in different regions of the phase diagram by seral researchers.25–36 The phase diagram and the calculatsolidus of GaSb are shown in Figs. 2 and 3, respectively.melting point of GaSb has been reported to lie betweenand 712 °C.27 A melting point depression of less than 50 °is observed for compositions6 30 at. % on either side of thestoichiometric composition.27 Brice and King37 showed thatthe liquid–solid–vapour equilibrium temperature is sensitto pressure—712 °C being the maximum~refer to Fig. 4!.The heat of formation and fusion are2 4.976 0.22 and 6.06 0.36 kcal/g atom, respectively. The dissociation pressat the melting point is about 1022 mm Hg. Above 370 °C,Sb starts volatilizing from the melt. Thus, GaSb will decompose to yield Sb2v) and GaSb dissolved in liquid Ga.38,39Atthe maximum melting point, the partial vapour pressureSb is ;331026 Torr.38 With this partial pressure, up to;231015 Sb atoms per second could be lost from easquare centimetre of solid surface. The partial pressure ois less than 1029 Torr at the maximum melting point.38 Thus,

5823Dutta, Bhat, and Kumar


sina

auaeclatto

ra-pertheob-ee-

the

0surensitu-ot aa

the

otnn

e

int

in 10 h run,'1023 mol of Sb would be lost which amountto ;0.1% change in the Sb/Ga ratio. Hence, usually dursynthesis of GaSb the Sb/Ga ratio in the melt is taken1.001.

On the basis of electrical resistivity measurements, Mnomura and Drickamer40 observed that GaSb undergoesphase transition around 80–100 kbar at room temperatThe transition was presumed to be structural in nature,though melting could not be discounted. A study of the effof pressure on the melting point of GaSb by Jayaraman, Kment, and Kennedy41 in the range 0–65 kbar indicates ththe melting temperature decreases by 5 °C per kbar

FIG. 2. Phase diagram of GaSb~from Ref. 27!.

FIG. 3. Calculated solidus of GaSb~from Ref. 209!.



gs

i-

re.l-te-

a

triple point near 56.5 kbar and 385 °C. The melting tempeture of the high pressure form of GaSb increases 3.4 °Ckbar. The triple point observed by these workers verifiesspeculation made by Minomura and Drickamer that theserved transition was solid–solid, but is not in good agrment quantitatively. Later Jamieson42 verified that the roomtemperature transition was indeed solid-solid and thathigh pressure form is tetragonal Sn type~metallic! witha55.348 Å andc52.937 Å. The transition pressure of 9kbar observed by Jamieson tends to confirm the prestransition observed by Minomura and Drickamer. Ozoliet al.43 produced GaSb with excess amounts of the constent elements and concluded that the lattice parameter is nfunction of melt composition. Hence the compound hasvery narrow homogeneity range.

B. Bulk growth

The thermophysical properties of molten GaSb atmelting point are listed in Table II.44 These properties indi-cate that the growth of low dislocation density crystals is nvery difficult.39 Bulk GaSb crystals have been mainly growby Czochralski technique~CZ!. There are a few reports oBridgman~BG! technique, travelling heater method~THM!,vertical gradient freeze~VGF! technique and liquid phaselectro-epitaxy~LPEE!. A brief account of crystal growthusing these techniques is given below.

FIG. 4. Partial vapour pressure of Sb as a function of temperature~fromRef. 37!.

TABLE II. Thermophysical properties of molten GaSb at the melting po~from Ref. 44!.

Density 6.03 g/cm3

Thermal conductivity of liquid GaSb 0.171 W/cm KThermal conductivity of solid GaSb 0.0781 W/cm KViscosity 0.0231 g/cm sSpecific heat capacity 0.328 J/g KVolume expansion coefficient 9.5831025 K21

Latent heat of fusion 131.16 J/gEmissivity 0.5Thermal diffusivity 0.087 cm2/sPrandtl number 0.044



o

ere

ufre

wth

id

hr t

inedike

itint

m

ee

feththg

aofotiheeioanbnavthSCZ

h2em

by

.3–

essofwn

wth1.1ff-ers,oni-ity

the

es, itith

elt-

.ld

ses

on-

nohetheipcanific

heyen--l–

d then-

heItfter

1. Czochralski technique

Single crystals of GaSb are usually grown by meansthe CZ technique39,45–60 ~see also references therein!. Themain problem encountered during growth by early workwas the appearance of a thin oxide film on the GaSb msurface. The scum has been identified as Ga2O3.

38 This filmimpairs the seeding of the melt and promotes twinning ding the subsequent growth. To ensure a melt completelyof oxide film, a double-crucible technique46 was used. Oneof the disadvantages of the double-crucible techniquethe rather high axial temperature gradient resulting information of many growth twins and dislocations.47 Theelimination of the oxide was later solved by means of liquencapsulation of the melt by B2O3. However, the viscosityof the molten B2O3 encapsulant was found to be too higaround the melting temperature of GaSb. Later, in ordelower the viscosity, 3.2 mol % of Na3AlF6 was added toB2O3.

48 An encapsulant of a mixture of NaCl and KCl~1:1!with a melting point of about 645 °C and a low viscositythe temperature range of 710–720 °C has also been us49

Growth without any encapsulant but with pure inert gas lN2 or Ar, pure hydrogen and 95% N2:5% H2 have also beenused to protect the melt surface against oxidation.50 A com-parative study on the growth of GaSb single crystals wand without encapsulation had shown that high quality, twfree crystals were obtained when growth was carried ouhydrogen ambient.51 Moreover, crystals grown by thismethod show higher purity by an order of magnitude in coparison with liquid encapsulated Czochralski~LEC! growncrystals using a B2O31Na3AlF6 encapsulant. However, thlimitation of this method is the slow reduction kinetics of thGa2O3 by H2. To overcome this problem, Moet al.53 haveadopted an upper–lower crucible technique forin situ syn-thesis and growth of GaSb single crystals. The specialture of this method is that GaSb synthesis is done inlower crucible in the presence of pure hydrogen andscum-free melt is transferred to the upper crucible usinsluice gate. Recently, Watanabe and co-workers54 have de-veloped a new technique of pulling bulk crystal from Gsolution fed with a GaSb source. Oliviera and Carvalh60

have suggested that chemical cleaning of the materialslowed by high temperature baking in vacuum can be safactorily used to grow crystals from scum-free melts. Tdiameter of crystals grown by the various workers mentionabove were in the range of 3–5 cm. The pulling and rotatrates employed by them were in the range of 6–12 mm/h20–55 rpm, respectively. Typical GaSb crystals grownthe CZ technique show microfacets, impurity striations atwins,55,56although as discussed later twin-free crystals hbeen grown under special circumstances. Very recently,Firebird Semiconductors Ltd., Canada have grown Gasingle crystals up to 85 mm in diameter employing thetechnique.57 In addition to growingp- andn-type GaSb, theyclaimed to have produced the first commercial higresistivity GaSb. The resistivity of the crystals at 300 K is31021 V cm and 4.23103 V cm at 77 K. These values arabout twice and 40 000 times, respectively, those of norundoped material.



f

slt

r-e

ase

o

.

h-in

-

a-eea

l-s-

dndydeeb

-

al

2. Bridgman technique

Vertical Bridgman growth of GaSb has been reportedrelatively few workers.14,61–68Usually crystals are grown inquartz ampoules sealed at 1026 Torr. In earlier works, theampoule lowering rate employed were in the range of 01.8 mm/h.61,62 Unlike CZ grown crystals, Bridgman grownones do not exhibit facets and impurity striations and possbetter quality. However, this technique has the limitationthe maximum diameter of single crystal that can be groand has often led to polycrystals.61 Roy and Basu62 couldgrow undoped single crystals upto 1.1 cm diameter. Growith ampoule diameters less than 0.8 cm or more thancm led to polycrystals.62 They attributed the occurence opolycrystallinity to nonuniform heat conduction and wall efect in the case of larger and smaller ampoule diametrespectively. Sometimes twins were observed near the ccal tip at the bottom of the ampoule. The dislocation densin the grown crystals was of the order of 105 cm22

. Highdislocation density occurs due to radial heat loss fromampoule.63 By using proper shielding around the ampoule63

or by enclosing the furnace in vacuum,14,64or by submergingthe heater in the melt,69 the radial heat flow and thereby thdislocation density can be reduced. In more recent workhas been demonstrated that high quality single crystals wvery low dislocation density (,50 cm22) could be grown byemploying a planar melt–solid interface shape.14,64–67 Acritical ratio of the furnace temperature gradient at the ming temperatureG of GaSb to the ampoule lowering rateVhas been found to result in planar interface during growth64

It is worth pointing out that low defect content crystals coube grown even with much higher lowering rates~4 mm/h!than that used previously~1.8 mm/h!. This was facilitated bythe recent theoretical modelling of crystal growth procesin single zone vertical Bridgman furnace.66 Extensive studieshave further shown that the critical value ofG/V for planarinterface lies in a narrow range for various experimental cditions as depicted in Fig. 5.14,67

In Bridgman grown crystals without seed, there ispreferential orientation along the growth axis. One of tproblems which was encountered by the authors duringBridgman growth is the improper initial nucleation at the tof the ampoule due to nonwetting of the melt with the siliampoules. This often led to polycrystallinity of the growingots. The problem has been circumvented by a specdesign of the ampoule tip.64

Full encapsulation of liquid semiconductors during tBridgman growth helps to improve the crystal quality bdecreasing the nucleations of grains and the dislocation dsity. Duffar et al.68 have recently grown high quality 2-in.diam GaSb single crystals in silica crucibles with the LiCKCl eutectic~58% LiCl–42% KCl! as full encapsulant. Theencapsulant prevents the contact between the sample ancrucible. The wetting of both silica crucible and the semicoductor by this eutectic is extremely good~contact anglespractically zero!. Moreover, the salts do not contaminate tmelt and is of high thermal stability and chemically inert.can be easily removed from the surface of the crystal agrowth.



rte

o/va

litucethogmcaS

loeth-

rspoceeelinn

l atthen-

erstureol-–13ow-allype-icalaln-inedin-tedg-

ed

an

torSbtheth-theinedrt istionm-edthecal

ra-ent

asity

e

ingle

oes.

As far as horizontal Bridgman growth is concerned, velittle work seems to have been carried out. The only reporwork has been from Lewadowskyet al.70 who employed thehorizontal Bridgman technique to grow single crystalsGaSb. The boat sliding rate employed by them was 11 mmCrystals up to 1.3 cm in diameter and 15 cm in length habeen grown. The quality of the bulk crystals was as goodthat of epitaxially grown high purity layers.

3. Vertical gradient freeze technique

The VGF technique has been used to grow high quadevice grade single crystals of III–V compound semicondtors with very low dislocation density. In the VGF techniquthe crucible and the furnace are kept stationary andgrowth is achieved by slowly cooling the melt in an apprpriate temperature gradient. One of the principal advantaof this technique is the much reduced axial and radial teperature gradients with concomittant advantages of reduconvective flow and thermal stresses. Garandet, Duffar,Favier71 used this technique to grow single crystals of Gaof 1 cm diameter and 5 cm length on^111&-oriented seedswith dislocation density below 100 cm22

. The growth ratewas about 3 mm/h. They could grow practically zero discation GaSb single crystals in silica crucibles by complencapsulation with LiCl–KCl eutectic by the VGF tecnique.

4. Travelling heater method

Only a few reports exist on the growth of GaSb by vetical and horizontal THM.72,73 The main advantage of thitechnique is low growth temperature which reduces vapressure problems and stoichiometric native defect contration. In THM, a solvent zone placed between a solid sand the feed material is heated and moved by a travelheater. In this way crystallization takes place at the adva

FIG. 5. Furnace temperature gradientG vs ampoule lowering rateV forplanar melt–solid interface. The experimental values ofG andV are shownby open squares~from Refs. 14 and 67!.



yd

fh.es

y-,e-es-edndb

-e

-

rn-dgc-

ing seed/solvent interface and dissolution of feed materiathe solvent/feed phase boundary. The solute transport inliquid zone may be established by diffusion and/or by covection. As discussed by Benz and Muller,73 the dislocationdensity in the crystals grown by this technique can be ordof magnitude lower than in the seed. The growth temperaused was in the range of 500–560 °C. Both Ga and In svents have been used with zone length in the range of 3mm. The main disadvantage of the THM technique is its lgrowth rate~;few mm/day!. For GaSb, growth rates between 0.7 and 5 mm/day have been achieved. It is principlimited by the transport rate of the slowest constituent scies through the solution zone. Moreover, the morphologinstability of the growing interface caused by constitutionsupercooling limits the maximum growth rate for inclusiofree single crystals. The usage of Ga-rich melts resultsdendritic growth and compositional inhomogeneity causdue to fluctuations in the freezing rate at the solid–meltterface. Several workers used uniform rotation, acceleracrucible rotation, rf heating or stirring by alternating manetic fields to increase the growth rate~see references inRefs. 72 and 73!; however, the growth rate merely increasby a factor of 2.

5. Liquid phase electro-epitaxy

The LPEE technique was used earlier by Gevorkyet al.74 and recently by Bischopink and Benz75 to grow bulkGaSb crystals. The growth is carried out in vertical reacmade of graphite and pyrolytic boron nitride. The Gasource material and the solution zone were stored inmovable upper part of the growth cell which served as caode. The GaSb substrate was placed in the lower part ofgrowth cell which forms the anode. When the feed iscontact with the solution zone, the electrical circuit is closand the growth starts. The mechanism of material transpoprincipally based on Peltier cooling at the substrate–soluinterface and on the electromigration of the dissolved copounds in the solution zone. The growth is mainly governby electromigration. The Peltier cooling only enhancesgrowth rate during the early stages of growth. The electriresistivity of the entire growth system~growth cell, solutionzone, substrate and electrical contacts! was such that theamount of joule heating was very low. The growth tempeture was varied in the range of 550–575 °C. The currdensity was in the range of 5–10 A/cm2. The growth ratewas linearly proportional to the current density and wfound to vary from 0.4 to 0.8 mm/day. When current densin excess of 5 A/cm2 was used, local pits of 50–100mmdimensions were found on thep-type substrate. This is duto the high Peltier heating inp-type material, which entailslocal dissolution of the substrate. No significant differencethe growth rates ofp- andn-type materials was seen provinthat the material transport by Peltier cooling is negligibagainst electromigration.

6. Growth under microgravity

In space at zero gravity, the molten semiconductor dnot stick to the quartz wall leading to good surface quality76



tiamsi-ioae

inasieahe00lso

ruricnasonm

as

is

ve

icmoibotu

th

lefaer

od-roofia

andretheein

thied

is-oru-

the

inbyisnder-

ing-l isle,e-hase-rlyofub-ralthisis

oreti-

The heat transfer in space differs from that in the terresconditions. The vacuum gap between the ingot and thepoule wall allows only a limited amount of heat to be dispated from the surface. The heat flows in the axial directwhich results in relatively stress free crystals. Lendvet al.76 were able to grow high quality bicrystals using thBridgman technique under microgravity conditions~1025

g). The translational rate of the ampoule was 0.188 mm/mA crystal with diameter of 8 mm and length of 39 mm wgrown. While the dislocation density, resistivity and carrconcentration were same for both the terrestrial and spgrown crystals, the mobility was found to be higher in tlatter. The dislocation density was of the order of 15

cm22. The hole mobility at 77 K was 2700 and 200

cm2/V s for the space grown and terrestial grown crystarespectively. The resistivity and carrier concentration werethe order of 0.12V cm and 1016 cm23 at 77 K, respectively,for both the crystals. In the past few years, a large numbespace experiments has been carried out to study the impsegregation phenomena in GaSb and InGaSb under mgravity conditions.16 In space, the gravity driven convectioin the melt is reduced and diffusion controlled heat and mtransport conditions can be achieved. Under such conditiit is possible to grow binary and ternary alloys with uniforaxial impurity or solutal concentrations, respectively.

7. Growth under hypergravity

Growth of crystals under hypergravity conditions halso been investigated by some workers in recent times.72,77

The high gravitational fields are known to influence the dtribution of dopants and crystal morphology.72 The effect ofenhanced acceleration on gravitational convection was intigated by Muller and Neumann.72 They had grown GaSb byTHM at an acceleration of 20g using a horizontal centrifugearrangement. Inclusion-free crystals~at Tm5550 °C! couldbe grown at growth rate as high as 14 to 20 mm/day, whis nearly one order of magnitude faster than that at norg. The increase in growth rate is due to the enhanced cvection in the solution zone. Doping striations were seenthe crystals similar to that in the case of crystals grownTHM at normalg. This suggests that the striations do nresult from unsteady convection caused due to temperaand solute gradients. The striations may be caused byinterface instability mechanism which is independent ofgravity driven convection. Regel and Shumaev77 employedthe Bridgman technique coupled with a centrifuge acceated to 5.2g for the growth of GaSb with a cooling rate o1.5 cm/h. The solid–melt interface was found to be planNo striations were found in the grown crystals; howevtwins were found in the upper end of the crystals.

8. Bulk growth of GaSb based ternaries

In recent years, a few groups have investigated the psibility of growing bulk crystals of ternaries like AlGaSb anInGaSb.78–81 The LPEE, Bridgman, THM and VGF techniques have been employed for growth either under micgravity or in terrestial conditions. The main motivationthese studies is to grow crystals with uniform radial and ax



al-

ny

.

rce

,f

ofityro-

ss,

-

s-

haln-nytreane

r-

r.,

s-

-

l

compositions. For this purpose, the effects of convectiondiffusion on the composition of the grown crystals westudied. A great deal of effort has been made to eliminateproblem of initial and final solutal transients by growing thcrystals with preestablished solutal concentration profilesthe polycrystalline feed. The morphological instability wiincreasing Al and In content in the melt has been studboth theoretically and experimentally.

The binary compunds GaSb and InSb are totally mcible, both in liquid and in solid state, thereby allowing fthe physical properties of the InGaSb alloy to vary continously with the InSb composition.82 The equilibrium phasediagram in the GaSb–InSb quasibinary section ofGaInSb system is shown in Fig. 6.83 During the solidificationof InGaSb, there is a rejection of solute molecules~InSb! bythe solid into the liquid. The rejected material accumulatesfront of the interface and spreads into the liquid phase,diffusion and mixing induced by convection, which in turndependent on the growth conditions. Joullie, Allegre, aBougnot84 studied the structural and compositional propties of directionally solidified InxGa12xSb with homoge-neous starting charge withx50.8. Polycrystalline ingotswere obtained with grain structure continuously changalong the growth direction due to variation in crystal composition by segregation. The homogeneity of the materiahighly dependent on the crystallization rate. For exampwhile crystallization up to 1 mm/h gives radially homogneous samples, the rates above this result in a second p~in the form of inclusions! in the nearly homogeneous matrix. One of the technical difficulties encountered by eaworkers is the long time required for reaching equilibriumthis compound when prepared from powdered starting sstances. Extremely long mixing times of the order of sevehundreds of hours was found to be insufficient to preparecompound in homogeneous form. Another major problem

FIG. 6. Pseudobinary phase diagram of InSb–GaSb. The lines are thecally calculated curves and the symbols are experimental data~from Ref.83!.



gengn

tFu

to

i-ry

t-arura

raa

br

d

bye30

hel ofn

nt

ure

icalthatlt isr-ra.os-or-s.es

lu-lu-lu-of

heo

ers.thand

ae

the sticking of the material to quartz ampoule~due to unre-acted indium and its oxide! and ampoule cracking durinhomogeneization and growth which limits the utility of thgrown crystals. Using a critical high temperature bakicycle prior to homogeneization, the problem of sticking acracking has been solved and homogeneous mixtureIn0.2Ga0.8Sb could be prepared.

85 Garandet and co-workers71

could grow reasonably good quality single crystals up5% of InSb with uniform radial composition by the VGtechnique. Beyond 5% InSb, a large number of cracks, sgrains and high dislocation density were observed duechemical misfit. Presently, CZ-grown bulk Ga0.95In0.05Sb isunder commercial production at the Firebird SemiconducLtd., Canada.57

The AlxGa12xSb ternaries have growth properties simlar to those of InxGa12xSb. The GaSb–AlSb pseudobinadiagram is shown in Fig. 7.86 Single crystals ofAl xGa12xSb withx50.06 to 0.16 have been grown with saisfactorily uniform radial and axial compositions. Due towide separation between the liquidus and solidus, the sttural quality of the crystals is poor when the Al concenttions are high.

Further efforts are required to improve the structuquality of the bulk ternary crystals. If succeeded, this csubstitute for epitaxially grown ternaries in devices.

C. Epitaxial growth

Epitaxial growth of GaSb has been largely carried outliquid phase epitaxy~LPE!. A few reports exist on vapouphase epitaxy~VPE!, chemical vapour deposition~CVD!,metal-organic chemical vapour deposition~MOCVD!, MBE,metal organic molecular beam epitaxy~MOMBE! andplasma assisted epitaxy~PAE!. The details are discussebelow.

1. Liquid phase epitaxy

There exist quite a few reports on growth of GaSbLPE technique.87–96 Ga-, Sb-, Sn- and Bi-rich melts havbeen used for the growth in the temperature range of 3

FIG. 7. Pseudobinary phase diagram of GaSb–AlSb. The different linestheoretically calculated curves by different groups and the symbols areperimental data~from Ref. 86!.



dof

o

b-to

rs

c--

ln

y

–

680 °C. By carrying out growth at low temperatures, tnative acceptor concentration could be reduced to a leve1016 cm23

. The Sb-rich melt is the most effective solutiofor reducing the native acceptor concentration;19 however,growth from Sb-rich solutions is limited by the eutectic poiin the Ga–Sb phase diagram atT5588.5 °C for 0.884 atomfraction of Sb. Woelk and Benz89 have grown undopedpepilayers from Ga- and Sb-rich solutions in the temperatrange of 330–470 and 635–680 °C, respectively.

In recent studies on surface morphology and electrand optical properties of GaSb layers, it has been foundthe growth temperature range of 500–550 °C with Ga meoptimal for obtaining high quality layers with excellent suface morphology.95 Typical morphology of an epilayegrown from Ga-rich melts at 550 °C is shown in Fig. 8Even though the epilayers grown at low temperatures psess low native defect concentration, they exhibit poor mphology and hence are not suitable for device application96

Similarly, the Sb-rich melt, even though it efficiently reducthe native defect content, results in faceted growth~Fig. 8b!which is again undesirable for device applications. Theminescence efficiency for layers grown from Sb-rich sotions was less compared to those grown from Ga-rich sotions. This has been attributed to higher contaminationimpurities which is expected for high growth rates in tcase of Sb-rich solutions. By carrying out growth from a twphase solution technique, abruptp–n junctions with epilayerthickness;1 mm have been obtained.95

Chandvankar and Arora90 have carried out dissolutionstudies of GaSb in Sn solution and grown undoped epilayBi melts have been used by a few workers in the growtemperature range of 390–550 °C. Recently, Gladkov

rex-

FIG. 8. Typical surface morphologies of GaSb epilayers~a! grown on~100!substrate at 550 °C from Ga-rich melt,~b! on 7° off ~111! substrate at660 °C from Sb-rich melt~from Refs. 95 and 96!.



-

alebac

r

md

o

ro

-

nahth

enmakpachera

ibndai-

tuane

rs

yhaea

echC

on-

etz

thetris-nelh-h-ngofallowhighon-beow

res

n

dhis

nieveke

.500

aveesdue

ntter

yer.

are

BE

o

veis-

co-workers91 employed Bi-rich melts for the growth of undoped GaSb. A native acceptor concentration of;1016

cm23 could be obtained. Photoluminescence studies revethe incorporation of approximately 0.015 at % Bi in thgrown epilayers. This resulted in reduction of band gap0.8 meV. Nevertheless, Bi was found to be electrically intive.

For n-type layers, Te doping has been carried out.88,92,93

Capasso and co-workers88 could achieve very low net donoconcentration in the range of 1014–1015 cm23 from undopedGa-rich solutions in the 300–375 °C range and by Te copensation using Ga-rich melt at 500 °C. Miki anco-workers94 were also able to grow undopedn-type layersat 400 °C from Ga-rich melt with net donor concentration;1015 cm23 and mobility as high as 7700 cm2/V s. In ourstudies we have observed that undoped layers grown fantimony-rich melts always exhibitp-type conductivity irre-spective of the growth temperature.96 In contrast, a type conversion from p to n was observed in layers grown fromGa-rich melts below 400 °C.96

Secondary ion mass spectrometry~SIMS! analysis ofGaSb grown by LPE technique usually shows unintentioimpurities, C, Si and O, which are acceptors in GaSb. Tcarbon comes from the graphite boat that is used forgrowth and Si from the quartz tube. Oxygen is usuallydonor in III–V, but may form an acceptor complex. Oxygincorporation can take place from microleaks in the systeThe concentration of these impurities depends on the bout duration of the melt prior to growth. Although pileuimpurities are detected at the epilayer–substrate interfthere is no correlation between hole concentrations and timpurities, implying that the background carrier concenttion originates from native defects.

2. Vapour phase epitaxy

Vapour growth of GaSb is difficult because the equilrium vapour pressure of antimony is extremely low ahence the transport rate or the growth rate of GaSb is usuvery small. A few workers97–101have succeeded in the eptaxial growth of GaSb by VPE. Arizumi and co-workers100

carried out epitaxial growth of GaSb on GaAs substratesopen tube system by reacting a mixture of SbH3 and HClgases with metallic Ga. They have also used a closedsystem with polycrystalline GaSb as the source materialHCl, SbCl5 and I2 as transporting agents. The growth ratwere found to range from 0.5 to 10mm/h when the initialamount of the transport agent in the charge was from 1028 to1026 mol/cm3

. The surface morphology of the grown layewas found to depend on the substrate temperature andamount of vapour etching done prior to growth. Smooth laers were obtained for growth temperatures higher t600 °C with 30 min of vapour etching prior to growth. Thminimum hole concentration in the undoped layers wfound to be 1.531016 cm23 with a mobility larger than 700cm2/V s. Layers grown using HCl gas as the transport agusually show nonuniform distribution of acceptors whipresumably come from water vapour contained in the H



ed

y-

-

f

m

leea

.e-

e,se-

-

lly

in

beds

the-n

s

nt

l

gas. When carefully dried HCl gas was used, the hole ccentration decreases to the level of 1016 cm23

, which corre-sponds to the native acceptor concentration.

CVD growth of GaSb has been reported by Jakowet al.102MOCVD growth of GaSb has been carried out103–110

using trimethylgallium~TMGa! or triethylgallium ~TEGa!and a host of Sb-containing metal-organic precursors intemperature range of 450–625 °C. From these studies,dimethylaminoantimony~TDMASb! has been found to be aexcellent precursor for the MOVPE growth over a widrange of growth temperatures.104 The optical and electricaproperties of layers grown by the CVD and MOCVD tecniques are similar to that of layers grown by the VPE tecnique. The main source of background impurities duriMOCVD growth comes from the carbon during pyrolysisTMGa or TEGa, yielding high concentration of unintentionacceptors. TESb and TMSb decompose insufficiently attemperatures and need to be thermally precracked attemperatures. However, in spite of the residual carbon ctamination, the dominant acceptors have been found tothe native defects. The TDMASb is expected to produce lcarbon contamination due to the absence of direct Sb—Cbonds and effective decomposition at low temperatu~300 °C!.

3. Molecular beam epitaxy

The main difficulty in growing GaSb by MBE is agaithe low vapour pressure of antimony.111 As a result, duringcrystal growth Sb will have a low surface mobility and tento aggregate together forming clusters and precipitates. Tleads to vacant Sb sites. Thus, antisite defects like GaSb areformed. Therefore, to improve the quality of MBE growlayers, an Sb-rich environment is needed. One can achthis by using proper orientation of the substrate li~311!B, ~111!B, etc. Longenbach and Wang20 used~311!Boriented substrates to reduce the nativep-type centres in thegrown layers. Usually the growth rate varied from 0.6 to 2mm/h for the growth temperatures in the range of 500–6°C. Very low acceptor concentration~'1015 cm23! could beobtained using this approach. Undoped GaSb epilayers hshown C, O and Si impurities. The origin of these impuritican vary from one growth system to another and can beto different sources.

Generally, Te doping is accomplished in MBE growlayers by the use of a PbTe cell since this provides becontrol of temperature than the use of elemental Te.112 Thereappears to be a trace incorporation of Pb in the grown laDonor concentrations of above 1018 cm23 are readily ob-tained. Other compounds that may be used for Te dopingGaTe, SbTe, GeTe and SnTe.113–118Selenium doping from aPbSe cell119 and sulfur doping from a Ag2S electrochemicalcell120 have also been carried out. Thep-type conductivitymay be accomplished by Ge, Sn or Be doping in the Mprocess.3 The covalent radius of Sb~1.41 Å! is larger thanthat of Ga~1.25 Å! and therefore group IV atoms tend toccupy the Sb sites. Sn~1.40 Å! is a larger atom than Ge(1.22 Å) and consequently, it is more difficult to achieheavy doping with Sn. Silicon being a small atom gets d



io

ied

reoneybathehethth

faIII

otielf-at

o

trritsloer

heopfoe

nothdautrof

a

b

37-

e-

iced Sbndentrenotents

. Theeds-actthe.ing

ons-

-e-

eenace

e inf73

e

eni-rp-

tributed amongst the two sites and the compensation rathigh. Hence, Si is not generally used as ap-type dopant.

4. Metal-organic molecular beam epitaxy

MOMBE growth of GaSb and InAsSb has been carrout using TEGa, TEIn, TESb and TEAs.121 The maximumgrowth rate of GaSb is observed at a substrate temperatu500 °C. Recently, MOMBE growth and etching of GaSbflat and high index surfaces using TDMASb, as well asemental antimony Sb4 , and TEGa have been reported bYamamotoet al.122 When only TDMASb is supplied, GaSsurfaces are etched for all misoriented substrates and noxide could be removed at 540 °C. On the other hand, wTEGa is simultaneous supplied in addition to TDMASb, tsurface reaction is changed from etching to growth onn53,4,5 surfaces. A higher growth rate is obtained on~n11! surface~n53,4,5! using TEGa and Sb4 . A growth tem-perature in the range of 450–570 °C was used. Good surmorphology is obtained in the 540–570 °C and high V/ratios. Growth rate is typically 1mm/h around 500 °C. Adecrease in growth rate for T.550 °C is observed due treevaporation of ethyl Ga molecules. The optical properof the layers are extremely good with the full at hamaximum~FWHM! of 9 meV for the band-to-band peak77 K. Unintentionally doped GaSb epilayers showp-typewith the hole mobility of 820 cm2/V s and hole concentrationof 1.231016 cm23 at 77 K.

5. Plasma assisted epitaxy

The low temperature epitaxial growth of GaSb in hydrgen plasma was reported by Satoet al.123 This process hasseveral advantages, such as cleaning effect of the subssurface by sputtering or by plasma etching, efficient impudoping, high growth rate, good surface morphology acroslarge area and reduced stoichiometric defects due togrowth temperatures. Growth was carried out in the tempture range of 340–440 °C with H2 pressure of 531022 Torr.The ratio of Ga-to-Sb supply ranged from 1:3 to 1:6. Tplasma power was varied in the range of 0–100 W. Antimum plasma power exists for each growth temperaturewhich the mobility of the grown layers was found to bmaximum. The growth rate was in the range of 1–2mm/h.The hole concentration of the undoped layers was foundbe of the order of 1016 cm23

. Impurity analysis has not beecarried out on the grown epilayers. Surprisingly, in spiteseveral advantages of this technique, growth usingmethod has not been pursued rigorously. As is discusseSec. VIII, plasma exposure leads to surface degradationthis will certainly affect the device properties. To bring opossible technological application of this technique, futuwork should be in the direction of solving the problemsurface degradation during growth.

III. STRUCTURAL PROPERTIES

A. Lattice parameter

From the powder x-ray data, the lattice parameter25.15 °C was found to be 6.095 93 Å.124 The temperaturedependence of the lattice parameter upto 680 °C is given



is

of

l-

iven

ee

ce

s

-

ateyawa-

-r

to

fisinnd

e

t

y

a5a01a1T1a2T21a3T

31a4T4,

whereT is in °C. The values of the constantsa0 , a1 , a2 ,a3 and a4 are 6.0958 82 Å, 3.496331025 Å °C21

, 3.345631028 Å °C22

, 24.6309310211 Å °C23 and 2.6369310214 Å °C24

, respectively.

B. Density

The density of GaSb at 300 K is measured to be 5.61g cm23

.124 There is very little variation of density with tem

perature. At 900 K, it is found to be 5.60 g cm23.125

C. Crystal structure

GaSb crystallizes in zinc-blende structure, which blongs to the space groupF43m in the Hermann–Mauguinnotation, or Td

2 in the Schoenflies notation.126 The zinc-blende structure is identical to that of the diamond lattexcept that each Ga atom has four tetrahedrally arrangeneighbours and vice versa. However unlike the diamostructure, the zinc-blende structure does not possess a cof inversion, and opposite directions in the crystal arenecessarily equivalent. This leads to interesting arrangemof the atoms in the~110!, ~100! and~111! planes. The~100!surfaces are stepped and contain both Ga and Sb atomsnature of chemical bonds in III–V compounds are of mixcovalent-ionic type. The ionicity of GaSb is 0.33. The preence of a slight ionic component in the bonds and the fthat there are equal numbers of Ga and Sb atoms on~110! planes results in the~110! cleavage of the compoundThe zinc-blende lattice structure and partial ionic bondimpart to the crystal a polarity along the111& axis. The~111! planes can be prepared with either Ga or Sb atomsthe surface. The~111! plane composed of Ga atoms is deignated as~111!A. The (1,1,1) plane is composed of Sbatoms and is designated as~111!B. These two surfaces exhibit striking differences in their chemical, electrical and mchanical properties.

From high pressure Raman studies at 300 K, it has bfound that above 7.65 GPa, a white tin structure with spgroupD4h

192I (41 /a)md results.127

IV. THERMAL PROPERTIES

A. Heat capacity and Debye temperature

Very few measurements of heat capacity are availablthe literature for GaSb. Piesbergen128measured the values ocp , cv andU for GaSb in the temperature range of 12–2K. For temperatures in the range 20–700 °C,cp can be ex-pressed as:129

cp50.043 511~4.63531025T! cal/g deg,

whereT is the temperature in °C.The value ofcp at 298 K is 0.060 58 cal/g deg. Th

temperature dependence of specific heatcv is shown in Fig.9.

The plot of Debye temperature (U) versus temperatureis shown in Fig. 10.130 As can be seen from the figure, thU runs through a minimum at low temperatures. This mimum is in agreement with the results from the lattice abso



gy

onat

ryu

–Vrmrne-

12ro-om

-

309eeir

fro ata;

tion bands in the infrared, which give a relatively low enerfor the transverse acoustic phonons.131,132 The decrease inU at higher temperatures is believed to be due to anharmeffects in the lattice vibrations. In order to give an estimof this contribution, a parameterU` is calculated fromThirring expansion. This value for GaSb is 316 K. At velow temperatures, where only the low frequencies contribto atomic heat, the value ofU is calculated to be 266K.133–135

B. Elastic moduli and phonon dispersion

The measurements of the elastic properties of the IIIcompounds have so far been mainly confined to the detenation of the three second-order constants. The second-oelastic moduli versus temperature plots for GaSb obtaiusing ultrasonic technique136 are shown in Fig. 11. The values of elastic moduli,c11, c44 andc12 ~in 1011 dyn cm22! at296 K are 8.834, 4.322 and 4.023, respectively.

The phonon dispersion relations are shown in Fig.The experimental points are obtained from inelastic neutscattering experiments.137 The continuous curves are obtained from the parameter shell model calculation. Fr

FIG. 9. Comparison of heat capacity of GaSb calculated fromg(n) with theexperimental data~from Ref. 130!.

FIG. 10. Comparison of the Debye temperature for GaSb calculatedg(n) with experimental data~from Ref. 130!.



ice

te

i-derd

.n

first-order127 and second-order138 Raman scattering experiments at 300 K, the phonon wave numbers~in cm21! ob-tained are indicated below

nTO~G!:223.6, nLO~G!:232.6,

nTA~L !:46, nTA~X!:56,

nTA~W!:75, nLA~L !:155,

nLO~L !:204, nLO~X!:210,

nTO~L,X,S!:218.

Fig. 13 shows the Raman spectra of GaSb taken for the 5Å line at various pressures.127 With increasing pressure, thLO and TO phonons shift to higher frequencies and th

m

FIG. 11. Temperature dependence of second-order elastic moduli~from Ref.136!.

FIG. 12. Phonon dispersion relation for GaSb. Symbols: experimental dcontinuous lines: theoretical calculation~from Ref. 137!.



id7.la

snt

ture

n intionit

ere

ionra-re

re

edri-

09

b

nt of

nt of

peak heights also vary. The intensities of phonon peakscrease with pressure, go through a maximum and thencrease. The transition to the metallic phase occurs atGPa. The shifts of the phonon frequencies with relativetice constant (2Da/a) and pressure are shown in Fig. 14.127

FIG. 13. First-order strokes Raman spectra for GaSb taken with the 53laser line at different pressures~from Ref. 127!.

FIG. 14. Dependence of the TO and LO phonon frequencies of GaSlinear lattice compression~lower scale! and pressure~upper scale!. Solidlines are least squares fits to the experimental data~from Ref. 127!.



n-e-65t-

C. Thermal expansion

The linear thermal expansion coefficienta for GaSb inthe temperature range 25–340 K is shown in Fig. 15.139 Ascan be seen from the figure,a is negative for temperatureless than 0.2U. For cubic structures, the volume coefficieof thermal expansionb is equal toa.

The thermal expansion of GaSb above room temperahas been investigated by Bernstein and Beals.140 The tem-perature dependence of the relative expansion is showFig. 16. As can be seen from the figure, a sharp deviafrom linearity is observed in the interval 300–400 °C, andwas impossible to make measurements beyond 436 °C.

The temperature dependence of Gruneisen parametgfor GaSb is shown in Fig. 17.141Near the Debye temperaturg is independent of temperature. At low temperaturesg hasa region of negative values which coincides with the regin which the expansion coefficient is negative. The tempeture at whichg changes sign coincides with the temperatuat whicha also changes sign.

D. Thermal conductivity

The thermal conductivity as a function of temperatufor four GaSb samples~p and n type! is shown in Fig.18.142,143The shape of the curves is similar to that obtaintheoretically for III–V compounds by considering the cont

Å

on

FIG. 15. Temperature dependence of linear thermal expansion coefficieGaSb from 4.2 to 340 K~from Ref. 139!.

FIG. 16. Temperature dependence of linear thermal expansion coefficieGaSb above room temperature~from Ref. 140!.



linsodpen0a.-nrysc-hee

ion.sted

do-pe

of

ofhas

the-

val-.

g

sti-ra-t of

butions from various scattering processes like crystalboundaries, impurities, three-phonon, four-phonon, renance and electron–phonon scattering; however, it shoulnoted that the experimental data showed a change in slolow temperatures for thep-type samples. As can be seefrom the figure, even though then-type samples contain 1times more impurity than thep-type samples, they possesshigher value ofk for most of the low temperature regionTheoretically calculated thermal conductivity at low temperatures~up to 10 K! in the boundary scattering regiooverestimates the experimental data by almost a facto100. An attempt was made to fit the data using an analdue to Ziman144 which treats scattering of phonons by eletrons in a degenerate band. While the magnitude of the tmal conductivity is correctly predicted, this scattering donot account for the change in slope in thep-type material nor

FIG. 17. Temperature dependence of Gru¨neisen parameter for GaSb~fromRef. 141!.

FIG. 18. Thermal conductivity of GaSb forn- and p-type samples.~1! n5 431018 cm23, ~2! n5 1.43 1018cm23, ~3! p5 131017 cm23, ~4! p5 23 1017 cm23 ~from Ref. 143!.



e-beat

ofis

r-s

can it account for the dependence on impurity concentratFurther, by using the electron–phonon mechanism suggeby Keyes,145 an attempt was made to fit the data.146 In thiscase, the scattering is due to the strain sensitivity of thenor ~or acceptor! ground-state energy. The change in slonear 5 K could be obtained in the analysis.

The high temperature lattice thermal conductivityGaSb is shown in Fig. 19.147 The decrease ink with increas-ing free carrier concentration is attributed to scatteringphonons by electrons. Some optical phonon scatteringalso been identified. Steigmeier and Kudman147,148 haveevaluated the influence of optical mode scattering onlattice thermal conductivity of group IV and III–V semiconductors. Using the values ofM1/M2 , M , U and k, theycalculated the Gruneisen anharmonicity parameter. Theues of each of these quantities for GaSb are listed below

M1 /M2 ~atomic mass ratio!:1.75,

M ~mean atomic mass!:95.7,

U ~Debye temperature in K!:265.5,

k ~at 300 K in W/cm deg!:0.390,

g ~at T5U!:0.86.

The large value ofg for GaSb implies significant scatterinby the optical mode.

The thermal conductivity of GaSb has also been invegated in the presence of magnetic field in the low tempeture region, however it has been found to be independenthe magnetic field.130

FIG. 19. High temperature thermal conductivity of GaSb~from Ref. 147!.



it

anin

oan

lythe-

ngax-

sa

eK

iv

-

ature

he

V. ELECTRONIC AND TRANSPORT PROPERTIES

A. Band structure

Fig. 20 shows the band structure of GaSb obtained wa nonlocal pseudopotential calculation.149 The symmetrysymbols are in double group notation. The conduction bis characterized by three sets of minima. The lowest mmum is atG. The next higher minima are theL points at thesurface of the Brillouin zone and at theX points. The valenceband has the structure common to all zinc-blende semicductors. The energies of the symmetry points of the bstructure relative to the top of the valence band~in eV! aregiven in Table III. The first column gives the theoreticalcalculated values, the second column is obtained fromangle resolved photoelectron spectroscopy experim~ARPES! at 300 K150 and the third column is the electroreflectance data~EL! at 10 K.151

Using photoluminescence spectroscopy152at 2 K, the ex-citonic gapEgx has been found to be 0.8099 eV. Assumian exciton binding energy of 1.4 meV, the direct band gEg,dir (G8v2G6c) was evaluated to be 0.8113 eV. The etrapolated band gap at 0 K was found to be 0.822 eV fromthe electroreflectance experiment153 as shown in Fig. 21. Theband gap at 300 K is 0.725 eV. The values of constantaand b used in the equation for evaluating the band gapvarious temperatures from the band gap at 0 K are 4.231024 eV K21 and 140, respectively.

The direct gap energies calculated from interband dirFaraday rotation are 0.74 and 0.82 eV at 296 and 77respectively.154

The pressure dependence of the absorption edge is gbelow:154

dEgdP

51231026 eV/kg/cm2 for P<18 000 kg/cm2,

dEgdP

57.331026 eV/kg/cm2

for 18 000 kg/cm2<P<45 000 kg/cm2,

FIG. 20. Band structure of GaSb~from Ref. 149!.



h

di-

n-d

ent

p

t

ct,

en

dEgdP

5negative for P>45 000 kg/cm2.

At normal pressures, the~000! band lies lowest, fol-lowed by the~111! and then~100! bands. The minima assigned to the above three pressure ranges are~000!, ~111!and ~100!, respectively.

The critical point and spin-orbit splitting energies~ineV! as measured from modulation spectroscopy155 at 27 Kare given below

FIG. 21. Energies of electroreflectance peaks as a function of temper~from Ref. 153!.

TABLE III. Energies of symmetry points of band structure relative to ttop of valence band~in eV! ~from Refs. 149–151!.

Symmetrypoints Theory ARPES EL

E(G6v) 212.00 211.64E(G7v) 20.76 20.82 20.756E(G8v) 0E(G6c) 0.86 0.822E(G7c) 3.44 3.191E(G8c) 3.77 3.404E(G8c) 7.9E(L6v) 210.17 210.06E(L6v) 26.25 26.60E(L6v) 21.45 21.55 21.530E(L4,5v) 21.00 21.10E(L6c) 1.22 1.095E(L4,5c) 4.43 4.36E(L6c) 4.59 4.49E(X6v) 29.33 29.62E(X6v) 26.76 26.90E(X6v) 22.61 23.10E(X7v) 22.37 22.86E(X6c) 1.72E(X7c) 1.79E(S3,4v

min ) 23.6423.90



th

0s.ti

K

wa

b

s

rau

V

byn

t

th

eervGvaoe

so-,

ith

eenthe

s are

ec-nt

ighpho-aSben--

ived

inass

kythe

ject

t in

E01D05~G7v2G6c!51.569,

E15~L4,5v2L6c!52.185,

E11D15~L6v2L6c!52.622,

EL5~G8v2L6c!50.871.

The intraconduction band energy difference betweenlowest conduction band minimum (G6c) and the lowestLband minimum (L6c) was found to be 61 and 82 meV at 30K employing Hall and magnetoresistance measurement156

The two values have been obtained by assuming the effecmass of electron inL band to be 0.22m0 and 0.43m0 , respec-tively. The energy difference between theG-band minimumand theX-band minimum was found to be 430 meV at 10using the electroreflectance technique.151

The camel’s back structure of conduction band edgeestimated from thek • p theory using the GaP data.157 Thevalues of various camel’s back parameters are tabulatedlow

D:178 meV,

DE:25.1 meV,

km :0.127~2p/a!,

mt :0.250m0 ,

mi :1.2m0 .

The electrong factor at 30 K calculated using stresmodulated magnetoreflectance has been evaluated to27.8.158

Molar magnetic susceptibility as a function of tempeture in the range of 4.2–900 K has been evaluated and foto lie in between240 and238 cm3/mol.154

The energy–wave-vector relation for holes in III–compounds, including GaSb, contains a linear-k term owingto lack of inversion symmetry in their crystal structure,the spin-orbit splitting. As shown by Dresselhaus, Kip, aKittel159 because of the linear-k term, the light and heavyhole bands are split into two nondegenerate bands, andenergy maxima of the valence bands are not atk(0,0,0). Themaxima of the heavy hole bands are shifted in@111# direc-tion and that of the light hole bands in@100# direction. Froman analysis of transport data the following values fordifference of energies at the top of the bands and atk 5 0have been found.160,161

DE@111#:20 meV;DE@100#:5 meV;

DE@111#2DE@100#:7.5 meV.

The influence of the linear-k term on the shape of thisoenergetic surfaces inp-GaSb has been deduced by Robet al.162 from galvanometric measurements. They hashown that the nonquadratic band model employed forand Si is insufficient to account for all the observed galnometric phenomena. They determined the anisotropy cficients of the light and heavy hole ellipsoids along the@100#and @111# directions to be 1.66 and 3, respectively.

The valence band parametersA, B andC used in thenonquadraticE 2 k relation calculated usingk • p theory163



e

ve

s

e-

be

-nd

d

he

e

tee-f-

are 11.7, 8.19 and 11.07, respectively. From cyclotron renance measurements onp-GaSb in the range 12–20 KStradling164 determined these values to be 1160.6, 661.5and 1164, respectively, which are in good agreement wthe theoretically predicted values.

B. Effective masses of electrons and holes

The effective masses of electrons and holes have bevaluated by cyclotron resonance technique and fromdensity of state analysis of transport data.165,166In Table IV,the effective masses for electrons (me) in the G, L andXconduction bands and that for holes (mp) in heavy and lighthole valence bands along with the density of state masselisted.

The density of state effective mass obtained from eltron concentration~transport! measurements can be differefrom that obtained from reflection measurements.154 The dif-ference is due to the anisotropy of the upper subband. Hresolution magneto-absorption measurements made atton energies just above the intrinsic absorption edge in Ghave revealed an oscillatory spectrum. The value of theergy gap found isEg50.8136 0.001 eV and electron effective mass,m*5~0.47 6 0.003!m0 .

154 The optically deter-mined hole mass is considerably smaller than the one derfrom electrical measurements.

Due to close proximity of theL band to the conductionband minimum, appreciable population of electrons existstheL band above room temperature. Since the effective mof electron is more in theL band, it will affect the electronmobility above room temperature. Moreover, for Schottdiodes the Richardson constant~and hence the barrier heigh!will also get affected by the population of electrons in tL band.

C. Electron transport

The transport properties of GaSb have been the subof investigation for the past 3 decades.167–184This is becauseof some special features of its band structure. Transpor

TABLE IV. Effective masses of electrons and holes~in terms of free elec-tron mass,m0! ~compiled from Refs. 165 and 166!.

NotationNumericalvalue Remarks

me(G) 0.0412 from cyclotron resonance of hotelectrons in the temperaturerange of 1–30 K

me(G) 0.0396 same data as above but by takinginto account the nonparabolicityand polaron effect

me(L)' 0.11 from transverse conductivityme(L) i 0.95 from longitudinal conductivitymde(L) 0.226 density of state massme(X)' 0.22 from transverse conductivityme(X) i 0.51 from longitudinal conductivitymp(h) 0.28 from conductivity datamp( l ) 0.05 from conductivity datamdp 0.82 density of state mass



omnsree

-0nh

adansirevi

lin

ottsng

edpingpre-isTherib-, theture,ob-ableue ofce,hethird

ticp-e.ncereheserhisher--c-onngeint.

ofa.eV

ritynronshatatim-oma-ec-esSb.ndllestSe-eforeivityity-

eda-

n-type GaSb is complicated due to the contributions frG, L andX conduction bands. Experimental data on traport coefficients can be consistently explained by a thband model,166 the X bands contributing to transport abov180 °C. The room temperature electron mobility~incm2/V s! for a sample withn51.4931018 cm23 was foundto be 3750, 482 and 107 at theG-, L- andX-band minima,respectively. Furthermore, for theL-band at room temperature, the mobility was found to be 500–800 and 800–16cm2/V s at 120 K.155 Fig. 22 shows the temperature depedence and the contributions of various scattering mecnisms to the electron mobility.166

Sagar168 discussed the results of Hall coefficient mesurements onn-type material on the basis of the two-banmodel. The purer samples have higher Hall coefficient atgiven temperature. The special features are the increaHall coefficient with increasing temperature for the pusamples and the appearance of a maximum for the headoped crystals. Since the effective mass at the~000! minimais smaller than that at the~111! minima, for a constant totanumber of electrons, an increasing Hall coefficient withcreasing occupation ration2 /n1 , i.e., with increasing tem-perature, is observed. A maximum is reached when the tnumber of electrons starts to increase at the onset ofintrinsic region, since thenRH drops rapidly. Measurementby Strauss169 show a characteristic difference in the dopi

FIG. 22. Temperature dependence of electron mobility inG andL band forGaSb withn 5 1.4931018 cm23 ~from Ref. 166!.



-e

0-a-

-

yng

er-

-

alhe

dependence of the Hall coefficient for Se- and Te-dopsamples. The difference is most pronounced at heavy doand disappears for very pure samples. One possible intertation is that impurity-band conduction is dominating in thregion, and that this depends on the nature of the donor.two subbands of the conduction band of GaSb both contute strongly to the magnetoresistance. In purer samplesmagnetoresistance decreases with decreasing temperawhereas in heavily doped samples, the opposite trend isserved. The influence of the second band becomes noticeas soon as the electron concentration surpasses the val1.2531018 cm23

. Pressure dependence of the resistanHall coefficient and the thermoelectric power confirmed tpresence of three subbands in the conduction band. Thesubband appears only at 25 000 kg/cm2

. From theZ coeffi-cients ~piezoresistance! measurements under hydrostapressure, the~000! subband has been found to displace uward relative to the~111! subband with increasing pressurIts contribution to the conductivity therefore decreases. Sithe density of state in the~000! band is considerably smallethan in the~111! band, there are only few electrons in th~000! band; however, they contribute considerably to tconductivity because of their high mobility. When theelectrons are brought into the~111! band by pressure, theismall number does not add much to the contribution of tsubband to the conductivity and piezoresistance. From tmoelectric power measurements onp-GaSb in the temperature region from 280 K to above the melting point the effetive mass of the holes has been found to strongly dependtemperature and to be doping dependent. An abrupt chaof the thermoelectric power is observed at the melting po

In GaSb, the donor states below the~000! minimum fuseinto the conduction band; the ones lying below the~111!minima, however, keep a finite ionization energy becausethe larger effective mass of the electrons in these minimThese discrete impurity levels lie a few hundredths of anbelow the~111! minima but still above the~000! conductionband at constant energy. Thus the separation of the impulevels from the~111! conduction band is not a true ionizatioenergy that must be overcome to make the donor electelectrically active. The only property of these states is tthey are empty at low electron concentration and filledhigh concentration. In the first case they act as chargedpurities, in the second as neutral ones. This can cause anlies in the dependence of the electron mobility on the eltron concentration. This model can explain the differencbetween the electrical properties of Se- and Te-doped Ga

The pressure variation of the resistivity of S-, Se- aTe-doped~n-type! GaSb has been studied to 50 kbar. Athree types exhibit a saturation in resistivity at the highpressures attained, although the resistivity of S- anddoped samples increases several orders of magnitude bsaturation, in contrast to Te-doped samples, whose resistincreases only by a factor of 14. The saturation in resistivis due to theX1 minima becoming the lowest conductionband edge at these pressure.

Even though numerous investigators have explaintheir Hall data onn-GaSb by considering the effects of reltively low lying L and X conduction bands,168 there are,



opexintio

leon

reatra

tly

e

n-

n-

nsth

ththnopeonghkewnumda-thn

enewoneenathdi

t-n

ndocnd

bewllip-nd-

ndbe-of.ongk-ss

ofsti-en-u-

ptorcal-ob-

lesltsare

Teo-tialthe

s

however, some peculiarities in the electrical propertiesn-GaSb occurring at lower doping levels which do not apear to be the result of multiple-band conduction. Forample, the mobility at low temperatures is observed tocrease monotonically with increasing electron concentrafor Te-doped GaSb.169 Also, the diffusion of lithium intoTe-doped material has resulted in significant increase in etron mobilities coupled with modest increase in electrconcentration.170,171Long and Hager,172 in their investigationof near resonance scattering in GaSb at low temperatupointed out the basic importance of compensation in regto the observed increase in mobility with electron concention. This behaviour is in contrast to that found inn-typeGaAs and InAs, where compensation is significanlower.173 Later, Baxter and co-workers174 illustrated the ef-fect of compensation on Coulomb scattering from the wknown Brooks–Herring treatment.175 The ionized impuritymobility is in general an increasing function of carrier cocentrationn and inversely proportional toNI , the ionizedimpurity concentration. In the absence of compensation,NI

is equal ton; however, with high concentration of compesating centres,NI varies much more slowly withn and thevariation of mobility is entirely controlled throughn. Fur-ther, to fit the mobility data for low electron concentratiothe required hole concentration is 2–3 times more thanobserved in undoped GaSb.174 Baxter and co-workers174

have explained this with the double ionizable nature ofnative defect. The higher hole concentration needed to fitdata arises from a deep lying acceptor level which isappreciably ionized at room temperature in the undosamples. Such a level would, of course, be completely iized in n-type material and account for the apparently hiestimates for the compensating acceptors. Several worhave adopted the approach of compensating the crystalsTe to shift the Fermi level toward the conduction band ahence observe the deeper level; but, as evidenced from lnescence measurements176 and pointed out by Johnson anco-workers,177 incorporation of donor dopant Te is accompnied by the formation of an additional acceptor state nearsingly ionized state of the native defect, which is often cofused with the energy levels of the native acceptor.178–182

This would also explain the high estimates for the compsating acceptors. Shubnikov–de Haas oscillations have bobserved in Te-doped GaSb diffused with lithium as shoin Fig. 23.183 The oscillation has been attributed to electrpopulation in theL band.183 Recently, we have studied thhole transport properties of undoped and tellurium compsatedp-GaSb in the temperature range of 4.2–300 K withaim to clarify these anomalies, paying special attention tobehaviour of the deeper energy levels. This aspect iscussed below.

D. Hole transport

In p-type III–V compounds, the dominant factors limiing mobility have been found to be acoustic, nonpolar apolar optical phonons and ionized impurity scatterings.163,184

In p-GaSb, owing to the close proximity of the heavy alight hole bands, intervalley and intravalley scatteringscur. Thus, the contributions from both the light hole a



f---n

c-

s,rd-

ll

at

eetd-

rsithdi-

e-

-enn

-nes-

d

-

heavy hole bands to various transport properties shouldtaken into account.159,160The hole transport properties at lotemperatures can be explained consistently by the multiesoidal model to take into account the shift of the heavy alight hole bands away fromk 5 0, whereas at high temperatures a warped sphere model~as in Si and Ge! isadequate.159,160 Analyses of transport data between 77 a300 K have to take into account the intermediate regiontween the two limiting cases in which the effective massheavy holes varies with temperature as shown in Fig. 24185

Fig. 25 shows the temperature dependence of mobility alwith the contributions of heavy and light holes without taing into account of the variation of heavy hole effective mawith temperature.159

By taking into account the temperature dependenceheavy hole density of states effective mass, we have invegated the transport properties of undoped and Te compsatedp-GaSb with the knowledge of defect levels from lminescence studies.186 Evidence for self-compensationisseen on Te doping by the formation of an Te-related accecomplex. Excellent agreement between the theoreticallyculated and experimentally measured mobilities has beentained, by including the Te-related acceptorVGa GaSb TeSbapart from the doubly ionizable native defectVGaGaSb. Themobility as a function of temperature for undoped sampgrown from stoichiometric and nonstoichiometric mealong with Te compensated samples to various degreesshown in Fig. 26. As depicted in Fig. 26, with increase inconcentration the mobility decreases and a shift in the mbility peak to higher temperature is observed. The parmobilities for various scattering mechanisms along with

FIG. 23. ~a! Transverse and~b! longitudinal magnetoresistance oscillationat various temperatures for Te-dopedn-GaSb samples after Li diffusion~from Ref. 183!.



pleraourinlehe

n-ehed-

s

oith

f

ndat

therer

lec-heentsen-

ci

iesl

total mobility for the undoped and Te compensated samare shown in Figs. 27 and 28, respectively. At low tempetures, the largest contribution to hole scattering comes frionized impurities for both the samples. At room temperatnonpolar and polar optical phonons and acoustic scatterall make significant contributions for the undoped sampwhereas ionized impurity scattering is still dominant for tTe compensated sample.

Johnsonet al.187 have also studied the effect of compesation in MBE grownp-GaSb epilayers by varying thSb/Ga ratio in the flux. They found that the layers with tlowest residual acceptor concentrations are not those thatplay the highest hole mobility, confirming the effect of compensation.

Fig. 29 shows the intrinsic carrier concentration asfunction of temperature. As can be seen, GaSb achieveintrinsic carrier concentration ofni'1017 cm23 in the neigh-bourhood of 600 K.188 The temperature dependence of mbility above room temperature for a typical sample w

FIG. 24. Heavy hole density of states effective mass as a function of rerocal temperature~from Ref. 185!.

FIG. 25. Hole mobility as a function of temperature:~a! for two samples~1!831017 cm23, ~2! 3 3 1018 cm23; ~b! contributions from heavy~h! andlight ~l! holes~from Ref. 159!.



s-megs,

is-

aan

-

NA'1017 cm23 is shown in Fig. 30. At a temperature oapproximately 630 K, the sample converts fromp to n typedue to rapidly increasing intrinsic carrier concentration athe larger mobility of the electrons. The exact temperaturewhich this conversion takes place depends, among othings, on the acceptor concentration and is lower for lowdensities of acceptors.

For GaSb there is no exhaustion region where the etron population of the impurities remains constant while tFermi level varies with temperature. Hence the employmof data from luminescence experiments is all the more estial to interpret the Hall data accurately and reliably.

p-

FIG. 26. Hole mobility as a function of temperature for (s) undopedp-GaSb grown from stoichiometric melt, (h) undopedp-GaSb grown fromGa-rich melt, (n) undoped p-GaSb grown from Sb-rich melt, and(d,¹,3) Te-compensatedp-GaSb to various degrees~from Ref. 186!.

FIG. 27. Temperature variation of theoretically calculated partial mobilitand effective mobility~solid curves! for undopedp-GaSb. The experimentadata are represented by (s) ~from Ref. 186!.



c-fneve

lec-ithan-

sti-

ar-

ure

rndinp

n-Alea-ereare

thefor

een

peds in

tied

tur

E. Magnetophonon effect

Oscillations in the ohmic longitudinal magnetoresistan~OLMR! have been observed inn-GaSb with carrier concentration in excess of 1017 cm23 in the temperature range o20–80 K.154 These oscillations were interpreted as magtophonon and spin-magnetophonon oscillations. Howe

FIG. 28. Temperature variation of theoretically calculated partial mobiliand effective mobility~solid curves! for a typical heavily Te-compensatep-GaSb. The experimental data are represented by (s) ~from Ref. 186!.

FIG. 29. Plot of intrinsic carrier concentration as a function of temperafor undopedp-GaSb~from Ref. 188!.



e

-r,

for the degenerate case oscillations due to transitions of etrons between the Fermi level and the Landau levels, wabsorption or emission of an LO phonon, can occur at qutized magnetic fields.

F. Electron and hole transport in ternaries

The electrical properties of InGaSb have been invegated in detail by Joullie and co-workers84 As-grown un-doped InxGa12xSb is n type in nature abovex50.5 andptype for x below 0.5. The compositional dependence of crier concentration is shown in Fig. 31.84 The Hall mobility asa function ofx is shown in Fig. 32.84 As can be seen fromthe figure, the electron mobility steadily decreases from pInSb value tillx50.5 ~to then–p transition!, after which thehole mobility falls to pure GaSb value.

In the regime of lowx in Al xGa12xSb, the material hasdirect gap withG being the conduction band minimum. Fox.0.52, the material is indirect with the conduction baminimum atX.114 Donor related deep traps were detectedTe-doped AlxGa12xSb forx.0.2. The concentration of deetraps increases steeply withx in the range 0.3,x,0.5 andsaturates forx.0.5.114 The deep trap concentration also icreases linearly with donor concentration for the samecomposition. In the temperature-dependent Hall effect msurements both shallow donor and deep donor levels wobserved. The characteristics of deep traps in Te:AlGaSbsimilar to theDX centre in AlGaAs. This implies that thedeep traps in Te:AlGaSb arise from the same origin asDX centre in AlGaAs. The shallow donor predominatesx< 0.3, while the deep donor predominates forx> 0.4 inconcentration. Considering the energy separation betwtheG andL band, and the effective mass inL band, the deepelectron traps should be detected even in GaSb if theDXcentre is simply a donor associated with theL band.DXcentres are observed in S:GaSb, but not in Te- and Se-doones as is discussed later in Sec. VII. The deep trap

s

e

FIG. 30. Mobility as a function of temperature for undopedp-GaSb aboveroom temperature~from Ref. 188!.



f

d

rge

gap

of

ncesch-ex

heped

inon--ofgertherethe

un

ion

ton

Te:AlGaSb cannot be explained by the band crossing oGand L bands, since electron traps were not detectedx50.2 where theG and L bands cross. The sulfur relateshallow states in GaSb are linked to theG minimum. The

FIG. 31. Carrier concentration as a function of alloy composition indoped InxGa12xSb at~a! 77 K and~b! 300 K ~from Ref. 84!.

FIG. 32. Room temperature Hall mobility as a function of alloy compositin undoped InxGa12xSb ~from Ref. 84!.



at

deep donors in S:GaSb have two states which exhibit lalattice relaxation.

The compositional dependence of the direct energyin GaxIn12xSb at 300 K is given by:5

EG50.17210.165x10.413x2.

For the AlxGa12xSb, the compositional dependencesdirect and indirect energy gaps are given below:5

EG50.7311.10x10.47x2 ~direct gap!,

EX51.0510.56x ~ indirect gap!.

VI. OPTICAL PROPERTIES

A. Dielectric constant

The dielectric constantse~0! ande(`) were found to be15.69 and 14.44, respectively, at 300 K using reflectatechnique and oscillator fit.189 The real and imaginary partof the dielectric constant measured by ellipsometric tenique are shown in Fig. 33. The values for refractive indn, extinction coefficientk, and reflectanceR calculated fromthese data are given in Table V for various wavelengths. Ttemperature dependence of refractive index for undoGaSb with hole concentration of 1017 cm23 is given by:190

S 1nD S dndTD5~8.260.2!31025 °C21.

B. Photoconduction

Very little work has been carried out on photoeffectsthis material. Spectral sensitivity measurements of photocductivity were carried out on onep-type specimen by Frederikse and Blunt.191 Their results indicate energy gaps0.65 eV at room temperature and 0.77 eV at 85 K. Habegand Fan192 have reported observations of long wavelengphotoresponse due to impurities. They concluded that thare levels situated at 34, 62, 76 and 103 meV above

-

FIG. 33. Real and imaginary parts of the dielectric constants vs phoenergy~from Ref. 189!.



edti

on

-nirhurero

r

LThtionees

hiitsnKin

r-20wseako aonaentnsoftheatanas

ec-r-

n byis-

ion

on-

theSbfgler

-a-e

leen-etalod

87445

valence band edge. Lukes193 found that the spectral responsfell rapidly as the wavelength was decreased, presumablyto high surface recombination. It is estimated from the raof photoconductivity~PC! to photoelectromagnetic~PEM!effects that the lifetime;1 ms.194

C. Photoelectric threshold and work function

Data available in the literature concerning work functiand photoelectric properties are meagre.195 The values ofband gapEg , work function f, photoelectric thresholdF t , direct transition thresholdFd and electron affinityx aregiven below for p-GaSb ~110! with NA51.231017 cm23

and resistivity of 0.07V cm.

Eg50.70 eV, f54.76 eV, F t54.76 eV,

Fd55.24 eV, x54.06.

D. Nonlinear optical effect

The unique contribution of III–V compounds to nonlinear optics has been in the observation of reflected harmoThis is mainly because many of the III–V compounds aopaque to the second harmonic of ruby laser light and texperimentally there is no confusion between a weakflected harmonic and scattered radiation from a much stger transmitted harmonic. Bloembergenet al.196 calculatedthe relative value ofx36 of GaSb with respect to KDP. Fol51.06 mm, x36 of GaSb is 1300 times that of KDP~5631029 esu! and atl50.0694mm, it is 400 times larger.

E. Radiative recombination and stimulated emission

Johnson and co-workers177 have studied absorption, Pand injection luminescence as a function of temperature.band–band peak increases with the free hole concentraIn p-type samples, compensated with Se or Te, severalbands appear, some of which are characteristic of pure STe. At high donor doping levels the band–band peak shifthigher energies due to the Burstein–Moss effect.197

In alloyed diodes prepared onp-type crystals, Calawa198

has observed a deep level band at 77 K which neither swith current nor with magnetic field, clearly indicating thatis not a band–band transition. The peak intensity increalinearly with current, becoming sublinear at high curredensities. The external quantum efficiency is 8.5% at 77A new deep level is seen in junction regrowth layers conta

TABLE V. Optical constants of GaSb~from Ref. 189!.

hv ~eV! e1 e2 n k R

1.5 19.135 3.023 4.388 0.344 0.392.0 25.545 14.442 5.239 1.378 0.482.5 13.367 19.705 4.312 2.285 0.483.0 9.479 15.738 3.832 2.109 0.443.5 7.852 19.267 3.785 2.545 0.484.0 21.374 25.138 3.450 3.643 0.5834.5 28.989 10.763 1.586 3.392 0.6515.0 25.693 7.529 1.369 2.751 0.5855.5 25.527 6.410 1.212 2.645 0.5926.0 24.962 4.520 0.935 2.416 0.610



ueo

cs.es-n-

en.worto

fts

est.-

ing tin. Diodes fabricated by diffusion of zinc into modeately heavily dopedn-GaSb show a broad band around 7meV at 77 K. This band shifts to higher energy and narroas the current increases and at high currents a new pappears at 790 meV. The shifting peak was attributed ttransition from an acceptor impurity band to a donor levelthe n side of the junction and the high current peak tovalence band donor level transition. Peaks shifts with currwere also observed in heavily doped Cd:Te grown junctiothat apparently lased at higher currents. The low densitystates in the conduction band and the close proximity ofnext higher lying minima, which can become populatedhigh carrier concentrations, or at high injection levels clead to profound changes in the recombination statisticsthe Fermi level rises.

Increase in transmissivity and extension of infrared sptrum of undopedp-GaSb by lithium diffusion has been caried out by Hrostowski and Fuller.199 The room temperatureabsorption spectra of GaSb before and after compensatioLi diffusion are shown in Fig. 34. The increase in transmsion is attributed to reduction in free carrier absorption.

Lasing has been achieved in GaSb at 77 K by injectand by electron bombardment.200 The complete details of thestructures, doping levels, lasing threshold and operation cditions are given in the respective reports~see Refs. in Ref.200!.

VII. DEFECTS AND IMPURITIES

A. Extended defects

Several studies have been carried out to investigatedefects generated during the growth of bulk Gacrystals.45,201–207Influence of crystallographic orientation othe seed on the formation of dislocations and twins in sincrystals ~grown with encapsulant! has been studied fo^111&, ^112& and ^115& orientations.201 It has been foundthat the orientation115& is the most favourable one for lowering the dislocation density and for depressing twin formtion. Also the twin formation is greatly influenced by thpolarity of the ^111& seed.203 In the case of a seed with~111!B plane oriented into the melt, twin-free GaSb singcrystals can be obtained; however, with the opposite oritation ~Ga plane! of seed, twinning always occurred. Thinfluence of different orientations of the seed on crysgrowth and twinning for the case of the Czochralski meth

FIG. 34. Room temperature absorption spectra of undopedp-GaSb~i! be-fore and~ii ! after lithium diffusion~from Ref. 199!.



n

eieftra

.c

ente

en

ll

ehethd

inehaphallr

--toisseSa

o-mseorrie

he

acatu

rgynsand

eVngeher at

atownh

akis

t asion

asestalna-ith

ion.ed

io-

in a hydrogen atmosphere has also been investigated207and itis concluded that the tendency to polycrystalline growth atwinning increases in the following order:^111&, ^112& and^100&. Characterization of extended defects in heavily Tdoped 111& CZ-grown GaSb single crystals has been carrout by Doerschel and Geissler.204 The nature and density othe defects are correlated with the local tellurium concention. Prismatic dislocation loops with~110! and ~111! habitplanes~perfect or faulted loops! are the dominant defectsExtended dislocation clusters consisting of complete dislotions with a Burgers vector of the type (a/2)^110& have alsobeen observed. Since the dislocations do not lie in~111!glide planes they move by climb, especially in~110! planes.Often, large multi-layered planar defects with displacemfringe contrast are also observable near dislocation clusThey have been identified as Ga2Te3 precipitates lying in~111! lattice planes. The prismatic dislocation loops are gerated at a tellurium concentration of 331018 cm23 and thedislocation clusters and large planar defects occur at a terium concentration of about 1019 cm23

.

B. Native defects

Undoped, pure GaSb is usuallyp type in nature withacceptor concentration of;1017 cm23

.208 Because of its

small band gap, GaSb begins to show complications duintrinsic conduction at relatively low temperatures. By tyear 1963, there was fairly convincing evidence thattroublesome background aceptors were related to nativefects rather than chemical impurities. Experiments involvion pairing between Li and the unknown acceptor showthe acceptor to be doubly ionizable chemical defect tcould have escaped detection by mass-spectrograanalysis.17 The residual acceptors can be made electricineffective by ion-pairing with lithium. According to Baxteand co-workers,17 the hole density decreases from 1017 to~2–3!31015 cm23 when lithium is diffused into GaSb. Subsequently, Reid and co-workers18 performed a series of experiments in which GaSb crystals were grown from nonsichiometric melts. As the Sb concentration in the meltincreased above 60%, one obtains significant reductionthe level of background acceptors, indicating that the acctor defect involves either excess of Ga or a deficiency ofin the crystal. They suggested that the most likely candidwas a Ga on a Sb site. The model of aVGaGaSb complex asthe dominant acceptor is in agreement with the resultsthermodynamical investigations;209 however, Zeeman experiments show that the defect has tetrahedral local symtry and thus it seems difficult to reconcile theconclusions.210 Both singly and doubly ionisable acceptmodels have been proposed to account for the behaviouthis defect in relation to the electrical and optical propertof GaSb.

The results of theoretical calculations which give tconcentration of each of the native defects likeVGa, VSb,GaSb and SbGa for crystals grown from Ga and Sb melts asfunction of temperature are shown in Figs. 35 and 36. Asbe seen from the figure, with decrease in growth temperathe native acceptor concentration decreases.



d

-d

-

a-

trs.

-

u-

to

ee-gdticy

-

inp-bte

f

e-

ofs

nre

The low temperature photoluminescence~PL! spectrumof undoped GaSb exhibits about 20 transitions in the eneinterval of 680–810 meV, but only a part of these transitiohas been associated with specific impurities or defects,these are listed in Table VI.211–213The dominant transition isat 777 meV. In S:GaSb, the 777 meV is shifted by 4–6 mto lower energy and the S-related peak appears in the raof 731–733 meV. The S donor level is at 81 meV below tconduction band. The Te- and Zn-related peaks appea740 and 775 meV, respectively.

Typical PL spectra at 4.2 K for GaSb grown by LPEvarious temperatures from Ga- and Sb-rich melts are shin Fig. 37. The highest purity samples which exhibit higluminescence efficiency are the LPE-grown ones.95 More-over, for MBE-grown GaSb layers on GaAs, the PL pepositions are slightly different than usually seen in bulk. Thhas been attributed to strain present in the epilayer.211

Efforts have been made by several workers in the paswell by our group to reduce the native defect concentratduring growth from nonstoichiometric melts18,214 and bypostgrowth annealing experiments.214 Postgrowth annealingsin Sb- and Ga-rich atmospheres do not effectively decrethe native defect concentration. On the other hand, crygrowth from Ga- or Sb-rich melts results in decrease intive defect concentration up to two orders of magnitude wincreased mobility.18,214However, the higher mobility valuesobserved in these crystals should be treated with cautAnomalously high hole mobility in GaSb has been attributto sample inhomogeneity.215 Our recent investigation oncompositional analysis of as-grown GaSb from nonstoichmetric melts shows the presence of metallic inclusions~Ga

FIG. 35. Temperature dependence of vacancy concentration in GaSb~fromRef. 209!.



.ityityitropioai

odbecatt lun

enwsg.e

eosrens,

S,uldtra-

reds

ondlu-

nGe

e-ultf Sf

ther

and Sb! which would show high mobility in the crystalsThus while concluding about the high crystalline qualthrough carrier mobility, one should rule out the possibilof metallic inclusions. These metallic inclusions would limthe usage of wafers prepared from GaSb ingots grown fnonstoichiometric melts for device fabrication. Thus the otion of reducing native defect concentration by nonstoichmetric melt growth does not seem advantageous for obting high quality wafers with uniform properties.

Theoretical calculations for the electronic structurethe divacancies at different charge states in GaSb baseself-consistent tight-binding theory has been carried outXu.216 The energy positions and localizations of the deflevels have been calculated for the predicted charge stEach divacancy at a charge state introduces seven defecels, three at the edges of the lower gap and four in or arothe fundamental band gap.

C. Shallow dopant impurities

To grow n or p-type GaSb various dopants have beused.217–219 Usually, Te, Se and S are used to gron-GaSb and Zn, Ge and Si forp-GaSb. The binding energieof various impurity levels in GaSb are shown in Fi38.220–227The donor states of tellurium in GaSb can be dscribed satisfactorily by the hydrogenic model.224 The bind-ing energies calculated in accordance with this model aragreement with the experimental values. The energy ptions of selenium and sulfur impurity levels in GaSb aaffected by the deviation from the hydrogenic model alothe Te→ Se→ S series and by the corresponding increain the binding energy.228–231As clearly shown in the figure

FIG. 36. Temperature dependence of antisite concentration in GaSb~fromRef. 209!.



m--n-

fonytes.ev-d

-

ini-

ge

Ge, Li, Si, Ga, Zn, Sb and Cu are shallow acceptors andTe and Se are shallow donors in GaSb. However, it shobe noted that Fe is a deep acceptor. With carrier concention of the order of 1017 cm23

, the highest resistivity is ob-tained by Cu~0.8 V cm!.217 Mn doping also gives rise top-type conductivity.

The solubilities of Se and Te at the melting point a1.531018 cm23

. Above these concentrations, compounform between Se or Te and Ga.154 Investigation of impurityvariations by cathodoluminescence~CL! imaging inGaSb:Te has been carried out by Chin and Bonner.231Withinthe resolution of the CL technique, no evidence of a secphase is found in GaSb crystals doped with Te to the sobility limit.

The formation ofp-GaSb by Ge doping is surprising. Iview of the atomic covalent radius, it is expected thatwould substitute Ga making GaSbn-type ~radii: Ga, 1.25 Å;Sb, 1.41 Å; Ge, 1.22 Å!. Even the electronegativities of G~2.0!, Sb ~1.8! and Ga~1.8! would not suggest the substitution of Ge in Sb site. Sulfur doping seems to be most difficdue to its high evaporation rate. The most suitable way odoping is done by using Sb2S3 in a flowing atmosphere ohydrogen and sulfur vapour.232 Following the idea of latticehardening by isoelectronic dopants as observed in o

TABLE VI. PL transitions observed in bulk GaSb at 4.2 K~compiled fromRefs. 211–213!.

Energy~meV! Transition

830 (e,h) recombination~77 K!812 band gap810 free exciton808 excitonic transition807 excitonic transition805 exciton bound to

(VGaGaSb)0

803 Exciton bound to(VGaGaSb)

0

801 D1-Si-acceptor800 Exciton bound to

(VGaGaSb)0

797 Excitonic transition796 Exciton bound to

(VGaGaSb)0

795 Excitonic transition792 Exciton bound to

(VGaGaSb)0

781 C-(VGaGaSb)0

777 D1-(VGaGaSb)0

775 C-Zn acceptor765 LO phonon replica of

the 796 meV transition758 Acceptor B746 LO phonon replica

of 777 meV740 C-(VGaGaSbTeSb)

0

728 LO phonon replica of758 meV

717 Exciton bound to(VGaGaSb)

2

710 C-(VGaGaSb)2

682 LO phonon replica of710 meV transition



a

etaS

eeap

opn

aps ofIn-

lowde-

on-

eri-r.ofE-er-ant-enti-c

byed

tanceent

ndbleusik

ed

p traps arefor

III–V compounds, indium and nitrogen doping of GaSb hbeen carried out using elemental indium and GaN.233 How-ever, lattice hardening did not occur and the dislocation dsity remained same as in the case of the undoped crysThe segregation coefficients of various impurities in Gaare listed in Table VII.154

D. Deep level impurities

The shallow donors Te, Se and S also revealed dimpurity levels in the band-gap intrinsic of the dopant spcies. The deep level properties of Te, Se and S doped Gon MBE-grown epilayers115 have been evaluated by deelevel transient spectroscopy~DLTS! and in bulk crystals234

by DLTS and thermally stimulated capacitance spectrosc~TSCAP!. The trap levels reveal various concentrations a

FIG. 37. Typical PL spectra of undopedp-GaSb at 4.2 K for~a! bulksubstrate,~b!–~e! LPE-grown epilayers from Ga melts,~f!,~g! epilayersgrown from Sb melts. The growth temperatures~in °C! are~b! 580,~c! 550,~d! 500, ~e! 450, ~f! 600 and~g! 630 ~from Ref. 95!.

FIG. 38. Impurity levels~in meV! of GaSb~from Ref. 220!.



s

n-ls.b

ep-Sb

yd

activation energies. These are listed in Table VIII. The trdensities in Te- and Se-doped samples are two ordermagnitude lower than the shallow donor concentration.terestingly, the DLTS spectrum of S:GaSb exhibitsDX-likenature with trap concentration comparable to that of shaldonor concentration. These electron levels are commonlytected inn-GaSb, but not in undopedp-GaSb, suggestingthat the level is not a simple native defect, but may be cnected with the impurity used forn doping; however, theexact origin of this level is not clear and is subject of furthinvestigation. Pooleet al.115 proposed for S:GaSb a coordnate model for a large lattice relaxation of sulfur donoThey115 attributed the difference in deep level propertiesvarious donors to residual sulfur contamination in the MBgrown layers. However, the difference in deep level propties of the bulk samples also indicates that it is dopintrinsic.234 By numerically deconvoluting the DLTS spectrum of S:GaSb, we have observed two peaks with differactivation energies.234 Such a splitting has been found prevously in a Si-relatedDX centre in GaAs under hydrostatipressure~30 kbar!.235 It is obvious that the splitting in GaSbcannot be due to alloy broadening as strongly arguedMooney, Caswel, and Wright to explain the multiple peakDLTS spectra due to aDX centre in AlGaAs.236Our TSCAPmeasurements also showed large nonexponential capacitransient with multiple peak nature of S:GaSb in agreemwith the DLTS measurements.234 Since GaSb doped withsulfur showsDX centres even at atmospheric pressure aunder nondegenerate doping conditions, it is most suitafor fundamental investigation of this metastable level thavoiding alloy broadening effects. Very recently, Hub

TABLE VII. Distribution coefficients (Cs /Cl) of dopants in GaSb~fromRef. 154!.

Dopant Distribution coefficient

Zn 0.3Cd 0.02In 1Si 1Ge 0.32Sn 0.01As 2–4S 0.06Se 0.4Te 0.4

TABLE VIII. Electrical properties of deep levels in Te-, Se- and S-dopGaSb~from Refs. 115 and 234!. HereEe , Ec , s, NT andND are the meanemission energy, mean capture energy, capture cross-section, total deedensity and total donor concentration respectively. The bracketed valuefor the vertical Bridgman grown bulk GaSb and the unbracketted onesthe MBE grown GaSb layers.

Dopant Ee ~meV! Ec ~meV! s (cm2) NT /Nd

S 280~287! 200 ~220! 6310218 (7310219) 1 ~1!Se 315~320! 195 ~180! 2310220 (9310221) 0.045~0.030!Te 310~312! 190 ~174! 3310220 (2310220) 0.015~0.010!



to

eaithveliteig

h.tioityVt-r

Setela

Echve

rov

rti

ctr-

gMte

esetiontrixaSb

fance

ngby

b-

tsand

m.if-ra-atfheion

n es

et al.237,238 have also inferred theDX-like nature of the Sdonor in GaSb using Hall, photo-Hall and persistent phoconductivity measurements.

TheDX-like centre dominates in Hall measurements btween 77 and 300 K. However, if a deep level is the msource of free carriers in a wide temperature range,evaluation of capacitance in DLTS studies of such a lebecomes complicated due to transient nonexponentiaThis nonexponentiality can be avoided if samples are msured where the concentration of shallow impurities is hin comparison with that of deep levels~10 times or more!.Since theDX centre is difficult to observe inp-type material,it is necessary to preparen-GaSb crystals doped both witsulfur and with some shallow donor impurity like tellurium

The donor states in S:GaSb tied to various conducband minima and slow relaxation of electrical conductivin these samples below 120 K have been studied byet al.224,228Low-temperature conductivity exhibits the lighinduced Mott transition, indicating the existence of sulfurelated shallow states linked to theG minimum. On this basisa phenomenological model of the deep donors in GaSb:proposed, invoking the existence of two sulfur-related destates with different photosensitivity and thermally activacapture cross section. Both deep states seem to exhibitlattice relaxation.

DLTS measurements made onp–n1 junctions of GaSbreveal the presence of a hole trap with activation energy0.33 eV, concentration of 531014 cm23 and capture crosssection of 1.3310215 cm2.239 Kuramochiet al.240 have alsoobserved deep levels in undoped GaSb grown by MBThree hole traps were detected in undoped GaSb whoseacteristics are shown in the Table IX. The shallowest le~A! shows low uniform concentration of 1014 cm23 and isconsidered to originate from a native defect like VGa–GaSb.The concentration of the trap C decreased from 1015 to1014 cm23 away from thep–n1 interface and is believed tobe connected with the diffusion of Te from then1 substrateto the undoped layer during growth. The trap B gives veweak signal and is similar to that reported by Polyaket al.239

E. Magnetic impurities

Recently, there has been some interest in the propeof magnetic impurity centres in GaSb.241–243Mn, Fe and Gddoped GaSb crystals have been investigated. PL and elespin resonance~ESR! spectroscopic studies indicate the fomation of a centre~VSb 1 Mn, 3d5 1 h! in Mn:GaSb.241

Mn:GaSb exhibits a negative magnetoresistance. The matude of the magnetoresistance increases with increase inconcentration. This effect was attributed to the spin scat

TABLE IX. Activation energyEa , capture ratecp and capture cross sectiosp for deep levels observed in undoped GaSb~from Ref. 240!.

Trap Ea ~eV! cp(3104 s21) s(310220 cm2)

A 0.25 1.8~150 K! 23 ~150 K!B 0.3–0.35 undefined undefinedC 0.63 1.0~292 K! 7.1 ~292 K!



-

-nely.a-h

n

ul

-

ispdrge

of

.ar-l

y

es

ron

ni-nr-

ing of holes by the magnetic moments of the manganions. On the other hand, the existence of the polarizamagnetic moment of the Fe and Gd ions in the GaSb madid not give rise to a negative magnetoresistance in Fe:Gand Gd:GaSb.

F. Isotopic effects

The values of line widths~dH in gauss! and second mo-ments~DH2 in gauss2! for various isotopic composition oGaSb have been studied by the nuclear magnetic reson~NMR! technique and are listed below.244

Ga69Sb:5.1 G and 6.5 G2,

Ga71Sb:5.6 G and 6.2 G2,

GaSb121:4.7 G and 6.7 G2,

GaSb123:5.1 G and 8.4 G2.

The antishielding factors for charged impurities arisidue to the distortion of the core electrons were evaluatedOliver245 and are 175, 321 and 49 for Ga69Sb, Ga71Sb andGaSb121, respectively.

The experimental values of quadrupole transition proabilities are 12 and 1.55 s for Ga71Sb and GaSb123, respec-tively.

G. Self- and impurity diffusion

The self-diffusion and the impurity diffusion coefficienin GaSb have been evaluated using radiotracer, SIMSp–n junction depth measurement techniques.246–251The re-sults are summarized in Table X. The diffusion of indiuduring growth depends on the composition of the crystal154

Indium diffuses faster in Sb- than Ga-rich crystals. The dfusion of tin into crystal depends on the carrier concenttion. The isoconcentration diffusion coefficient of zinc560 °C isD51.8310211 cm2 s21

. A strong dependence othe Zn distribution coefficient on the Zn concentration in tmelt in GaSb is observed. With increasing Zn concentrat~as well as with increasing Ga content!, KZn decreases. TheLi diffusion at 665 °C varies from 10211 to 231026 cm2/s at1016 and 731017 cm23

, respectively.

TABLE X. Self- and impurity diffusion coefficients and activation energiin GaSb~compiled from Refs. 246–251!.

Element D0 (cm2 s21) Q ~eV! Temperature~°C!

Ga 3.23103 3.15 650–700Sb 3.43104 3.45 650–700

8.731023 1.13 470–570

In 1.231027 0.53 400–650Sn 2.431025 0.80 320–570Te 3.831024 1.20 400–650Cd 1.531026 0.72Zn 4.031022 1.6



toel

f

vlini

inewsecdvhevdntelk

ctmbe

eo

nstath

k

urpl

heariacm

tiont to

inrenttrib-en

t toev-sti-

s-

ed

hese

ftertatece

es,ces

si-. 41.

-

ttoak-

H. Ion bombardment induced defects

Optical and electrical studies were made onn- andp-type gallium antimonide irradiated with 4.5 MeVelectrons.252Hall-effect measurements showed that accepwere produced in then-type samples and that donors as was acceptors were produced in thep-type samples. Infraredabsorption and photoconductivity studies gave evidencefour defect levels: Ev10.075 eV, Ev10.48 eV,Ec2~0.12–0.20! eV, Ec2~0.47–0.50! eV. The absorptionassociated with the levelEv10.075 eV is similar to thatobserved in donor-compensated GaSb. The leEc2~0.47–0.50! eV is an acceptor level, while the leveEc2~0.12–0.20! eV is probably a donor level. A decreaseconductivity under bombardment by neutrons was foundGaSb.

VIII. SURFACE AND BULK DEFECT PASSIVATION

Defect passivation is interesting both for understandthe physics of the material as well as from the point of viof technological applications. Development of GaSb badevices has been hindered due to high leakage surfacerents and fast degradation of the bare surface exposeambient. This is due to high surface density and Fermi lepinning, and residual native oxide layer on the surface. Tmakes surface passivation an essential step in GaSb dtechnology. Usually, chemical treatment with compounlike ~NH4!2Sx and RuCl3 results in surface reconstructiodue to which the surface recombination radecreases.239,253–256On the other hand, defects in the buare passivated by plasma hydrogen exposure.257–259 How-ever, in spite of its effectiveness in passivating bulk defeplasma exposure has the drawback of degrading the sasurfaces.260–262The problem of surface degradation cansolved by a protective cap layer likea-Si:H on the samplesurface during the plasma exposure which is permeablhydrogen.263 The details of various passivation treatmentsGaSb carried out hitherto are discussed below.

A. Wet chemical treatment

Wet chemical treatment using ammonium sulphide,254 orRuCl3 ,

255 has been found to reduce the surface state deties due to which the device characteristics improve dracally. To demonstrate the effect of various chemical trements prior to sulphurization on the surface state density,DLTS spectra of a metal–oxide–semiconductor~MOS!structure on GaSb~see Fig. 39! are extremely useful.254 Thelowest density of state~as depicted from the DLTS peaheight! was obtained for the sample etched using HCl:H2O~1:1! and treated by sulfur. The strength of room temperatPL signal was also found to be the highest for this samThis indicates that the reduced surface recombination rate~asevidenced by the increase in PL intensity! is mainly due tothe reduction in surface state density~as electrically mea-sured!. There can be an additional contribution due to tincrease in band bending brought in by the surface trments. Downward band bending near the surface givesto a field which tends to repel electrons that may approthe surface from the bulk regions. Since the sites for reco



rsl

or

el

n

g

dur-toelisices

s,ple

ton

si-i-t-e

ee.

et-seh-

bination are presumed to be near the surface, recombinacan be effectively decreased by inhibiting carrier transporthe surface. Decrease in ideality factor of diodes~fabricatedon passivated surfaces! is observed as a result of reductionsurface recombination component. Also the reverse curin diodes decreases by a factor of 10–20 times and is atuted to the formation of a very thin dielectric layer betwethe semiconductor and the metal.

There has been no specific measurement carried oudetermine the surface recombination velocity in GaSb. Nertheless, the surface recombination velocity can be emated by using the well known equation

Sp5spv thNst,

whereNst is the surface state density,v th is the carrier ther-mal velocity ands is the recombination cross section. Asuming the values ofv th and s to be 107 cm/s and 10215

cm2, and substituting the values ofNst as 831012 and 531010 cm22 eV21 for unpassivated and sulfur passivatsamples,254 the calculatedSp are 53102 and 83104 cm/s forthe passivated and unpassivated surfaces, respectively. Tare comparable to theSp in InP, but lower than in GaAs.

The C–V characteristics of a typical Al/oxide/n-GaSbMOS at 1 MHz fabricated on untreated and RuCl3 treatedGaSb are shown in Fig. 40. The decrease in loop width apassivation also confirms the reduction in the interface sdensity at the oxide–semiconductor junction. The surfatrap densities were found to be of the order of 1010 and1012 cm22 eV21 for the treated and the untreated samplrespectively. Usually, oxides grown on passivated surfaexhibit higher resistivity and higher breakdown voltage.255

The effect of passivation on the forwardI –V character-istics of Au/n-GaSb Schottky diode fabricated on unpasvated and Ru passivated samples is demonstrated in FigFor the unpassivated sample~curve 1!, two distinct exponen-tial regions withn of 1.8 and 1.3 at low and high bias respectively can be seen. The passivated diodes exhibitI –Vcharacteristics with no surface recombination componen~nclose to 1! as shown in the figure by curve 2. Also duereduction in surface channel conductivity, the reverse le

FIG. 39. DLTS spectra of GaSb MOS structures onn-GaSb.~1! Etched in0.03% Br2–CH3OH, ~2! etched in Br2–CH3OH and S treated,~3! etched inCH3COOH:HNO3 :HF ~40:18:2!, ~4! etched in CH3–COOH:HNO3 :HF andS treated,~5! etched in HCl:H2O ~1:1! and S treated~from Ref. 254!.



imaftasos

nett

r-isseshasSbary

dis-n

mic.

ex-ofisn-aceellea-otinedronxida-ttergthet isdiesandd do

isingis-the

rvede-sstionudeplebent

verfaced by

age current in diodes decreased by 30–40 times. Theproved surface property after passivation is due to thesorption of the Ru31 ions followed by the partial removal othe surface states from the band gap due to the electrosinteraction. Similar effects have been observed by ParkinHeller, and Miller in GaAs.264The fact that the PL intensitieincrease with surface treatment for both then- and p-typesamples indicates that the surface Fermi level is unpinand there is an increase in band bending. This showsamphoteric nature of Ru on GaSb surfaces depending onconductivity type, i.e., positively charged onp-GaSb and

FIG. 40. C–V plots of MOS structures onn-GaSb. Solid curve is for theuntreated samples and the dotted curve is for the treated sample~from Ref.255!.

FIG. 41. ForwardI –V characteristics of~1! untreated and~2! Ru-treatedn-GaSb samples. The dotted line corresponds to the idealI –V characteristic~from Ref. 255!.



-d-

ticn,

dhehe

negatively charged onn-GaSb. However, the fact that suface pinning is observed after Schottky barrier formationdue to the reaction of the metals with GaSb which caurepinning. Apart from surface passivation, Ru treatmentalso shown increase in room temperature mobility in Gapolycrystalline samples as a result of grain boundpassivation.255

The surface instability mechanism in GaSb has beencussed by Schirmet al.265 Low temperature processes caproduce on GaSb a nonequilibrium Ga2O3–Sb2O3 surfaceoxide layer as follows:

2GaSb13O2→Ga2O31Sb2O3 .

The only stable phases that can exist in thermodynaequilibrium with GaSb are Ga2O3 and elemental antimonyAny Sb2O3 should react with GaSb to give rise to Ga2O3

plus free Sb as:

2GaSb1Sb2O3→Ga2O314Sb .

This reaction spontaneously occurs~DG5212 kCal/mol! even at room temperature. Thus a GaSb surfaceposed to air will form a native oxide layer consistingSb2O3, Ga2O3 and a fraction of a monolayer of free Sb. Thfractional Sb monolayer is quite sufficient to drastically icrease the surface recombination velocity and surfleakage.261 Similar surface degradation mechanism is wknown in GaAs. Alkaline sulphides remove native oxidlayer from the surface of GaSb and stabilize it by the formtion of stable binary compounds of Ga–S and Sb–S. Peret al.256 carried out physicochemical analysis of sulfurizGaSb surfaces through ellipsometric, PL and Auger electspectroscopy measurements. They conclude that a deotion plus a sulfurization process leads to a surface of bequality giving good rectifying Schottky diodes. Pollutinoxygen and carbon agents are found to be removed fromsurface using this process. In addition, the sulfur treatmenshown to stabilize the cleaned surface. These and our stuclearly show that the passivated surfaces are chemicallythermally robust, stable against atmospheric exposure annot hinder the long term device characteristics.

B. Hydrogen plasma passivation

Even though passivation of bulk defects by hydrogennow well established, avoiding surface degradation durplasma exposure is not very trivial. In our studies, the extence of defects near the surface has been deduced fromabnormal behaviour ofC–V characteristics.262 During theC-V measurements, capacitance transient was obsewhich indicated slow emission rate of carriers from the dfects as shown in Fig. 42. A typical defect layer thicknehas been found to range from a few angstroms to a fracof a micron. The trap densities are comparable in magnitto the carrier concentration. The defects introduce multienergy levels in the band gap. This defect layer couldremoved in a controlled manner by a slow etchaHCl:H2O2:NaK-tartrate~8:1:1!.262,266,267

The effect of atomic hydrogen treatment in a microwaplasma crossed beam and in a direct plasma on the suproperties of GaSb and InGaAsSb has been investigate



ctr,

eoa

n-tnN

nith

taeooeveZn

ogcoanrodi

witiod

en

theatedhepas-in-. Inonve-efi-PLm-

essto

00

ed

Polyakov et al.260 Both the crossed beam and direH-plasma exposures led to surface degradation. Howevethe H-plasma exposure is followed by a N2 plasma exposureat 450 °C, the diode leakage currents decreased and thminescence intensity improved by more than an ordermagnitude. Formation of a passivating wide band-gap Glayer at the surface of GaSb depleted of Sb~Ga-rich layer!by preliminary intense H2 treatment is thought to be resposible for the effect. Both processes, the Sb depletion andGaN formation, seem to be temperature activated, sincether treatment in H2 plasma at 250 °C and subsequent2plasma treatment at 450 °C nor H2 treatment at 450 °C andN2 treatment at 250 °C led to a decrease in reverse curreIt would be interesting to perform these experiments wNH3 plasma.

Improvement in luminescence intensity and fundamenchanges of the radiative recombination after hydrogplasma treatment have been observed in undoped and dGaSb.259,267,268PL measurements indicate that passivationacceptors is more efficient than that of donors and, in geral, the passivation efficiency depends on the doping leThe passivation efficiency is the highest for heavilydopedp-GaSb. DLTS measurements carried out onn-GaSbindicated decrease in deep trap concentration after hydrnation. Extended defects like grain boundaries and dislotions are also found passivated. The most striking effectserved after hydrogenation was the efficient minority doppassivation in these compensated samples. The depth pof net donor concentration for ann-GaSb sample co-dopewith Te and Zn before and after hydrogenation is shownFig. 43. Since the material isn type, one would expect the Hto be in a negative charge state and thus not interactacceptors. However, the increase in net donor concentranear the surface as depicted in Fig. 43 indicates the mefficient passivation of Zn acceptors compared to the Tenors in the co-dopedn-type samples.

The effective diffusion coefficient of hydrogen has bedetermined inn1 andp1 GaSb.257 Thermal reactivation of

FIG. 42. Capacitance transients for different reverse bias voltages at 3~from Ref. 262!.



if

lu-fN

heei-

ts.

lnpedfn-l.

e-a-b-tfile

n

thonreo-

passivated shallow levels occurs around 200 °C. Onother hand, deep levels and extended defects are reactivabove 250 °C, indicating the higher thermal stability of tpassivated nonradiative centres. The results of hydrogensivation have been explained by assuming that hydrogentroduces a deep donor close to the valence band edgeGaSb there is very little effect of hydrogen passivationthe electrical properties of the material, but drastic improments are observed in the optical properties. This is bencial for optoelectronic devices like lasers wherein theefficiency can be increased without affecting the shallow ipurity level concentration.

C. a-Si:H passivation

Our recent studies clearly demonstrated the effectivenof a-Si:H treatment in achieving bulk passivation due

K

FIG. 43. Depth profiles of net donor concentration for Te–Zn-codopn-GaSb samples~1! before,~2! after H treatment,~3!–~5! after reverse biasand unbiased annealings~from Ref. 259!.



FIG. 44. Panchromatic CL images ofp-GaSb~a! as grown~10 keV!, ~b! treated bya-Si:H ~10 keV! and ~c! treated bya-Si:H ~20 keV! ~from Ref. 263!.

m

a

ltd-thd

a

nebte-Zhinuionnerucalsabtd

nwaron

thepri-

ani-

ageral

therectI.theess-rck,lu-nat-in-n,HF,stm.blentsof-y

iza-t asV

han

u-Sb,

plasma hydrogen in addition to a defect-free surface ternated bya-Si:H.263,268,269Efficient passivation of point andextended defects can be carried out bya-Si:H. Typical depthresolved CL images recorded before and after passivationshown in Fig. 44. Effects of passivation is seen up tomm in the bulk. With 10 keV beam, defects typical of megrown crystals like cellular dislocation network, twin bounaries, precipitates or impurity clusters can be seen inunpassivated sample~Fig. 44a!. On passivation, the extendedefects are found passivated upto a thickness of 2.8mm ~seeFig. 44b!. However, with 20 keV beam~corresponding to adepth of;3 mm!, the defects in the bulk are seen againthey are not passivated~Fig. 44c!. The CL and PL spectrarecorded on passivated samples exhibited higher lumicence efficiency than the unpassivated ones. Most notaefficient passivation of minority dopant in Te compensap-GaSb is observed. This is similar to that observed in Teco-dopedn-GaSb during hydrogen plasma passivation. Tpassivation efficiency is found to improve with increasea-Si:H deposition temperature. The passivation of variorecombination centres in the bulk is attributed to formatof hydrogen-impurity complexes by diffusion of hydrogeions from the plasma.a-Si:H acts as a protective cap layand prevents surface degradation which is usually encotered by bare exposure to hydrogen plasma. Thus for tenological applications,a-Si:H treatment can be used assingle step process through which both surface as welbulk defects can be passivated. Moreover, the procesglow discharge is highly suitable for large scale device frication. The effect ofa-Si:H buffer layer on the currentransport properties of GaSb Schottky diodes is discussethe following section.

IX. DEVICE ASPECTS

In this section, we focus our attention to the importaaspects of device fabrication technologies starting fromfer preparation. The electrical and optical properties of vaous device structures along with their possible applicatiare discussed.



i-

re3

e

s

s-ly,dne

s

n-h-

asof-

in

t-i-s

A. Wafer preparation

The successful preparation of a surface depends onproper cutting and abrasion processes followed by approate chemical etching. Gatos, Lavine, and Warekois270 havemeasured the depth of damaged layer in GaSb after mechcal polishing. With abrasive particle size of 20mm, the depthof damage was 6 and 11mm for the ~111!A and ~111!Bsurfaces, respectively. The difference in the depth of damfor the two surfaces is attributed to a distorted tetrahedstructure on the~111!A surface.

A number of chemical etchants have been used for eidamage-free surface preparation or for defcharacterization.271 These etchants are listed in Table XGaSb requires an oxidising agent to break the bonds. Ofnumerous oxidising agents that have been employed succfully, only HNO3 and H2O2 have found widespread use fovarious reasons involving etch rate, purity, mode of attaetc. Tervalent antimony that can exist in strong acid sotions is easily hydrolyzed to antimonyl ions, which upodilution may precipitate as antimonyl oxy-salts. Whentacked by strong oxidizing agents, antimonides form thesoluble trioxide or pentoxide. To keep antimony in solutioit is necessary to have a suitable complexing agent likeHCl, tartaric acid, lactic acid, citric acid or oxalic acid. Moof these complexing agents do not interfere with galliuAlthough specific ratios of components are given in TaIX, the proportions can be varied with suitable adjustmein etching time and temperature. For example, if the ratioHNO3 to HCl is very large, insoluble antimony tri- or pentoxide forms while all other ratios yield soluble antimonchlorides. Also ratios of HNO3 to organic acids should bekept low because strong acidic solutions suppress the iontion of the organic acids which is necessary for them to accomplexing agents. Due to the polar nature of the III–compounds, the etch rates on~111!A and ~111!B are differ-ent. Faust and Sagar272 found that the HNO3:tartaric acidetchant attacks the antimony face of GaSb more rapidly tthe gallium face.

Electrolytic etching and polishing have proven invalable in the manufacture of semiconductor devices. For Gaelectrolytic etching can be carried out using H2SO4:H2O



a

tovim

y

tunceb

atocca

rs

s

thhod

e, iwetyeigyis-irio

lvhsTb

va

ain-on-fore

hing.smaec-av-

mi-tedom-

eatdif-edredorible

s of

herp-m-nt ory-

atn ofEDf COaled

xy-Thein

ace

8ning

~1:1!,271 perchloric acid:acetic acid~1:4! ~Ref. 271! or HF inglycerine, alcohols or glycols.273

Further, molten gallium at 500 °C has been used forloy etching~100! surfaces of GaSb.271Concentrated HCl canbe used as a reagent for removing the alloying element.

For obtaining atomically flat surfaces of GaSb priorMBE growth, a large number of chemical etchants preously used for other III–V compounds has also been eployed; however a few have been found to be suitable.274–276

Smooth GaSb ~100! surfaces can be produced bHCl:H2O2:NaK ~tartrate! ~830 mM:300 mM:83 mM!,HF:HNO3:CH3COOH ~1:10:15! and Br:methanol~5:1000!succeeded by heating in ultrahigh vacuum at temperabetween 480–510 °C under an Sb flux. On the other ha2% Br2 :methanol etching produces surface defects. Exlent reproducible atomically flat surfaces are producedetching in CH3COOH:HNO3:HF ~40:18:2! at room tempera-ture for 40 s followed by HNO3:HCl ~1:30! at 5 °C for 1min. Also, HBr:HCl:HNO3 ~0.9 M:0.8 M:0.04 M! inCH3COOH solvent followed by annealing in Sb flux550 °C for 15 min exhibited a flat surface with no tracesO2, least carbon contamination and a stoichiometric surfa

Controlled mesa etching for device structures can beried out using NaK ~tartrate!/HCl/H2O2 in variousproportions.266 Etching rates in the range 0.1–1mm/min canbe obtained using different acid compositions.

For large scale device fabrication, high quality wafewith uniform properties are necessary. A recent study277 ofspatial CL imaging of GaSb wafers obtained from bulk agrown crystals indicates that the luminescence centres~ordefects! are distributed inhomogeneously. On annealingwafers in Ga, Sb or vacuum, the centres get distributedmogeneously and the properties become uniform throughthe wafer. Amongst these, annealing in Ga has been founbe the most effective treatment.

B. Dry etching

Unlike in bulk wafer preparation, device applications rquire considerable advances in processing technologycluding precision dry etching techniques. Compared tochemical etching, dry etching provides improved uniformigreater repeatability, more anisotropic profiles, improvetch depth control and higher aspect ratios. Owing to hvolatility of group III and group V chlorides, the most widelused plasma chemistry for III–V etching employs Cl dcharges. Anisotropic dry etching of various III–V and theternary and quaternary alloys has been achieved in highdensity electron cyclotron resonance~ECR! generatedplasma by CH4/H2/Cl2/BCl3/Ar. The use of CH4/H2 for re-active ion etching~RIE! of III-V compounds offers severaadvantages when compared to other chemistries. Ethough this has been demonstrated to be a universal etcfor various III–V alloys, to the best of our knowledge it hanot been specifically tried on GaSb and related systems.only work on dry etching of GaSb has been carried outPeartonet al.278 employing RIE with Cl2 and SiCl4 . Theetching rate is faster in Cl2/Ar than in SiCl4/Ar discharges,presumably because of more active chlorine species aable in the former case. Average etch rates up to 8mm/min



l-

--

red,l-y

fe.r-

-

eo-utto

-n-t,dh

n

enant

hey

il-

have been achieved. After the RIE process, chlorine conting residues are left on the surface. The depth of this ctamination is less than 20 Å and is always more prevalentCl2 than for SiCl4 . However, the amount of surface residuis less than that obtained with conventional CCl2F2 dis-charge.

The dependence of etched feature depth on the etctime for SiCl4/Ar and Cl2/Ar discharges is shown in Fig45a. The etching rate as a function of total pressure, plapower density and Ar% are shown in Figs. 45b–d, resptively. The roles of these parameters on the etching behiour have been discussed in detail by Peartonet al.278

C. Atomically clean surfaces

Several methods have been used for preparing atocally clean surfaces like evaporated layers from the heabulk, cleavage, heat treatment under high vacuum, ion bbardment and annealing, etc.279 Low energy electron diffrac-tion ~LEED! measurements on GaSb have indicated that htreatment alone does not produce surfaces showing sharpfraction patterns. Also, no diffraction patterns were obtainfrom the ion bombarded surfaces, indicating a disordestructure. Subsequent annealing at 400 °C for 30 minmore after ion bombardment resulted in sharp reproducdiffraction patterns.

The only adsorption studies made on clean surfaceGaSb are that of oxygen, CO and CO2. The effects of ex-posing the cleaved-sputtered annealed~110!, ~111!A,~111!B and~100! surfaces to oxygen have been studied. T~110! surface was extremely inert with respect to the adsotion of oxygen. For the other surfaces, the oxygen presuably was adsorbed on the surface in a random arrangemea polycrystalline oxide was formed. The adsorption of oxgen on the~111! Ga face proceeded at a faster rate~stickingcoefficient'1024! than on the~111!B ~sticking coefficient'1025!. Oxygen could be removed from the GaSb by hetreatment at 350–400 °C. At these temperatures, diffusiooxygen into the bulk has been shown by combined LEand secondary emission measurements. No adsorption ooccurs at room temperature on ion bombarded and annesamples. Initial sticking coefficients of 1026 and 1025 forthe CO2 adsorption on~111!B and ~111! Ga surfaces ofGaSb, respectively, have been obtained. The sorption of ogen on GaSb has also been studied at 78 K and 26.2 °C.amount of oxygen that is irreversibly absorbed in 1300 mat 78 K and 26.2 °C under oxygen pressure of 400mTorr are7.631014 and 2.631015 atoms/cm2, respectively. A low tem-perature ~;250 °C! treatment using microwave ECR H2plasma has been found to be effective in removing surfoxides from GaSb.261

D. Fabrication techniques

1. Ohmic contacts

Ohmic contacts top-GaSb doped in the range of310162131019 cm23 are provided by thermal evaporatioof several metals and alloys of Au, Ag and In after annealin the range of 250–350 °C for 10–30 min.280–282The spe-cific contact resistivityrc varies in the range of 1024–



J. Appl. Phys., Vol.

Downloaded¬26¬A

TABLE XI. Chemical etchants for GaSb~compiled from Ref. 271!. All formulas involving acetic acid are forglacial acetic acid~99.5%!.

No. Composition Ratio Remarks

1 HF:HNO3:CH3COOH ~2:18:40! polishing effect2 HNO3:HF:CH3COOH ~6:2:1! polycrystalline

material cleaning

3 HF:CH3COOH:KMnO4 ~0.05 M! ~1:1:1! impurity striations~AB etchant! ~high etch rate!

4 HCl:HNO3:H2O ~1:1:2! dislocations on~111! A~slow etch rate!

5 HCl:H2O:HNO3 ~6:6:1! dislocations6 HCl:30%H2O2:H2O ~1:1:1! solid–melt

interface shape

7 HF:HNO3:H2O ~1:3:6! twins8 HCl:H2O2 ~1:1! reveal dislocations, striations

and microdefects

9 HCl white spots at dislocations~boiling and illuminatedby 500 W lamp for 30 min!

10 HNO3:HF:H2O ~1:1:1! dislocations and grain boundaries~for 15 s!

11 HNO3:HF:CH3COOH:H2O ~9:5:1:10! grain boundaries and dislocations12 HCl:H2O2 ~2:1! white spots at

dislocations on~100!dislocations on~111! A

13 HCl:H2O2:tartaric acid 0.7:7.0:0.25 dislocations on~100!~molar ratio!

14 HNO3:HF:CH3COOH:H2O ~5:3:3:11! dislocations on~111!15 KOH:H2O ~9:11! dislocations on~111!

~for 2 min! ~weight ratio!

16 HF:HNO3CH3COOH ~1:9:20! dislocations on~111!~for 1 min!Br2–CH3COOH ~1:19!~for 4 min!

17 45% aqueous KOH dislocations on~111!~for 2 min!CH3COOH:HNO3:HF ~20:9:1!~for 1 min!5% Br2–CH3OH~for 11 min!

18 2% Br2–methanol dislocations on~111! A19 HF:HNO3:CH3COOH:Br2 ~3:5:3:0.06! dislocations on~111! A

~CP4 etchant!diluted by1% Br2–methanol

20 HF:HNO3:H2O ~13:20:17! stained cross-sectional~at 40 °C! view for p–n junctions

21 H2O:HF:H2O2 ~20:2:1! stain etching on~110!22 HF:HNO3:H2O ~3:7:10! p region etches more rapidly

leaving a step at thep–n junction.

23 HNO3:tartaric acid ~1:3! dislocations on~111! A ~etch rate: 6 mg/cm2/min!24 HNO3:HF ~1:1! polishing effect~etch rate: 40 mg/cm2/min!25 HNO3:HF:H2O ~1:1:1! dislocations on~111! A ~etching time: 30 s!26 HNO3:HF:HAc ~2:1:1! dislocations on~111! A ~etching time: 15 s!

~5:3:3! dislocations on~111! A ~etching time: 30 s!

27 HNO3:HF:HAc:Br ~25:15:15:3! dislocations on~111! A ~etching time: 10 s!28 HNO3:HCl ~1:1! polishing effect~etch rate: 2 mg/cm2/min!

585181, No. 9, 1 May 1997 Dutta, Bhat, and Kumar

pr¬2003¬to¬128.113.123.115.¬Redistribution¬subject¬to¬AIP¬license¬or¬copyright,¬see¬http://ojps.aip.org/japo/japcr.jsp

FIG. 45. ~a! Etched depth as a function of time and, average etching rate as a function of~b! pressure,~c! power density and~d! Ar % for RIE of GaSb inSiCl4 (s) and Cl2 (d) discharges~from Ref. 278!.

f

r f

actsis-ed

1026 V cm2. Tadayonet al.283 fabricated metal contacts o

Cr/Au, Ti/Pt/Au and Au onp-GaSb by MBE. For Au con-tacts,rc in the range of 1.4–7.831028 V cm2 have beenobtained which are the lowest values ever reported so fa



or

p-GaSb. A critical surface preparation prior to metal contwas found to be necessary for achieving low contact retivities. Prior to metallization, oxide layers were removusing ~1:1! HCl:H2O ~for 30 s!, followed by ~1:1! buffered



ahateieie

yanrnis

t

Aftetulik

°

aittAaina

Au

ctoagt

derexichcthileav

thtl

Vcts.

the

dif-

icat-ncesilytomace

atovel

ddi-izedfor

r-lityenttheen-g atsur-ndase

n-it-ostionox-ion-s for

free°C.to

-ta-h-ostyntaland

HF:H2O ~for 30 s!. Without this surface preparation, thercwas found to be much higher~in the 1026 V cm2 range!.283

The effect of specific chemical etchants used for the surfetching prior to metallization on the leakage currentsbeen investigated by us~unpublished!. It has been found thawhen the bulk substrates are dipped in HF for a long timthe Schottky diodes fabricated show capacitance transwhich indicates a high conducting surface layer with carrconcentration of the order of 1017 cm23 or more. This hasbeen attributed to the increase in surface state densitreflected from the photoluminescence peak intensityDLTS peak height of GaSb MOS structure. The preevapotion surface preparation for ohmic contact to the emitter acollector regions of heterojunction bipolar transistor consing of oxygen plasma descum, 1 buffered HF:1H2O dip fol-lowed by de-ionized water~DI! rinse, and a final 1HCl:1H2O etch with DI rinse has been adopted by Anastasiouobtainrc of the order of 1027 V cm2.284

To check the stability of the contacts, Tadayonet al.283

annealed the contacts in forming gas for 60 s. For thecontact, therc decreases to 200 °C and increases thereaThese diodes were further annealed at various temperafor several hours. At higher annealing temperatures250 °C,rc degraded to;1026 V cm2 within 60 s. On theother hand, for low annealing temperatures such as 100the contact retained their lowrc even after 10 h. Milnes andco-workers280 found that the Ag contacts were stable forleast 100 h at 350 °C and may be expected to have lchange over many thousands of hours at 70 °C. Thebased contacts however began to increase in resistance30 h at 250 °C. Ag is more stable than Au because thetermetallic compounds of Ag with Ga and Sb requirehigher temperature for formation than the equivalentcompounds.280

Ohmic contacts onn-GaSb usually show higher contaresistivity compared to thep-GaSb ones. For contacts tn-GaSb, Au, AuTe or AuSn have been used with alloyingabout 250 °C resulting in contact resistivities in the ran0.5–131024 V cm2

. Au–Ge–Ni contact with 300 °C heatreatment for 5 min gives resistivities of the order of 1025

V cm2.

2. Schottky contacts

Rectifying metal–semiconductor contacts are basicvices in the technology of semiconductors. However, fabcation of ideal GaSb rectifying contacts has provedtremely difficult due to the oxidation of GaSb surfaces whis more serious than for Si, GaAs and InP. Schottky contaon GaSb assume all the more importance for evaluatingcarrier concentration in epilayers due to lack of seminsulating substrates which have greatly obstructed the etrical characterization by Hall technique. Several metals hbeen used to fabricate Schottky contacts onn-GaSb.285–308

The barrier heights~at 300 K! are 0.65 eV~for Al !, 0.6 eV~for Au, In, Pd!, 0.5 eV~for Ga, Ni! and 0.42 eV~for Sb, Cr,Ag!. Silver has been found to chemically react with Sb atinterface. Even though the Schottky barrier height moslies in a small range of values~0.5–0.6 eV!, it depends onthe preparation technique. For instance,in situAl barriers on



ces

,ntr

asda-dt-

o

ur.rese

C,

tleufter-

te

-i--

tse-c-e

ey

MBE-grown n-GaSb have given values of 0.54–0.56 ecompared to 0.65 eV for thermally evaporated contaFrom the barrier height measurements for several metals~Cs,Ga, Sb, Au! and oxygen coverages on~110! cleavedn-GaSb, Spiceret al.293 found that the Fermi level at thesurface tends to pin slightly~;0.1 eV! above the valenceband edge. There has been no independent check ofFermi level pinning position for metals onp-GaSb as thesecontacts always show near ohmic nature and hence it isficult to carry out reliableI –V or C–V measurements. Withdecrease in temperature, the barrier height increases inding that the surface Fermi level tends to track the valeband edge. Pinning is less pronounced for metals that eareact with GaSb~like Sb! by forming chemical bonds even alow temperatures. Au selectively removes the Sb atoms frthe interface and creates a highly nonstoichiometric interfregion which becomesp-like to pin Ep at the valence bandmaximum.295 On the other hand Tan and Milnes found thAu contacts on GaSb deteriorate only after annealing ab350 °C for tens of hours.307 The mechanism of Fermi levepinning after metal deposition on GaSb~110! has been theo-retically explained by a defect model by Nishida.309 Thepresence of atomic Ga at surface Sb vacancy sites, in ation to surface Ga vacancies, gives electronic states localnear the top of the valence band which can be responsiblethe pinning.309 The ideality factor of the diodes usually vaies in the range of 1.3–2.4. The high value of the ideafactor is attributed to the generation–recombination currfrom a near midgap center. The reverse currents ofSchottky barriers are governed by tunneling via midgap cters and surface leakage, or by band-to-band tunnelinhigher voltages. As discussed in the previous subsection,face passivation prior to metal deposition is beneficial adrastically improves the device characteristics like decrein ideality factor and leakage currents, etc.

3. Oxides

Deposition of good oxides with low interface state desity, high-resistivity, high breakdown field strength and suable for high temperature processing is problematic for mIII–V compound semiconductors, and GaSb is no exceptto the rule. The interface state characteristics for nativeides of GaSb depend on the oxidizing process. The relatship of free Sb to interface states is not as established aAs at an oxide/GaAs interface. Wilmsen310 in a study ofnonaqueous anodic oxides concluded that there was noSb to interface unless the temperature exceeded 300Teare and Fischer311 evaluated the interface state densitybe of the order of 1012 cm22 for anodic oxides grown fromKOH on ~111!-oriented GaSb. A detailed study of the composition of thermal oxide on GaSb was carried out by Kimuraet al.312 by the x-ray photoelectron spectroscopic tecnique. They found that the oxide layer consists almentirely of Ga2O3. A small amount of oxidized Sb generallexists at the crystal surface and a large amount of elemeSb is located at the oxide–GaSb interface. Basu, Basu,Barman313 have carried out wet and dry O2 oxidations toproduce oxide layers of Ga2O3 and Sb2O3. They report in-terface state densities as low as 1010 cm22 eV21

, electrical



frsude

reth

f-io

a

nkof

f-hi

e

.wif

a

too

ult

loter

eage

ea

iorsi

tion

Theof, forms

of.

ressesave

plit-

ro-k-

terloftch-althenbci-

d in

ion,itsil-ace-

lsoasendureb.

p.esinin-the

re-esedhed

resistivities of 1010 V cm2 and breakdown field strengths o106 V/cm. A treatment with ruthenium trichloride furtheimproves the performance. The presence of Rb on theface also enhances the oxidation rate of GaSb by six orof magnitude with the formation of Ga2O3 and Sb2O5 atroom temperature.p-GaSb surfaces are found to be moreactive than then-GaSb ones suggesting the nature ofdopant to play an important at the interface.265 Very littlework has been carried out on the deposition of SiO2 ,Si~ON!x or Al oxide, and the results are not encouraging.311

4. Homojunctions

There exists very limited work in literature on the difused homojunction structures. In the early 1960s, diffusof In, Sn and Te was carried out~see Table X!. Later in1980, Capassoet al.314 fabricatedp–n junctions using theLPE technique. Surface leakage was found to be the mproblem. Recently, Polyakovet al.239 have reported im-provements in the electrical properties ofp–n junctions aftersulfur passivation. Usually, for one-sided abrupt junctiowith no fringing field, one would expect avalanche breadown at about 20 V for a doping concentration1016 cm23 and 2 V for a doping of 531017 cm23. However,zinc diffusion inn-GaSb~Te doped: 331017 cm23! at 680°C for 4 h in flowing N2 has led to breakdown voltages o20–30 V with a widep-type region of acceptor concentration of 531015 cm23.315 The occurrence of the large widtof this region may be due to reduction in native defectsundoped GaSb at high temperatures (>600 °C), resulting inhigh resistivity of the material. Such situations have beencountered during MBE growth of GaSb at 600 °C,316 andduring LPE growth from Sb-rich solution.317

GaSb light emitting diodes~LED! operating at roomtemperature were fabricated by a Zn diffusion technique318

The peak electroluminescence emission wavelengthfound to depend both on the LED drive current and the dfusion conditions used to produce thep–n junction.318

Longer diffusion periods resulted in diodes which emittedshorter wavelengths.

Ion implantation in GaSb is much difficult comparedGaAs, GaP or InAs since at moderate to high doses, amphization of the top layer takes place which is very difficto anneal.319–321Amorphous layer upto a depth of 1mm hasbeen observed. Further, it is seen that this layer devesmall crystallites, dislocation loops and exhibit swelling afthe annealing treatment.319 The threshold damage density foamorphization in GaSb is about 431019 keV/cm3 and theamorphous layer depth is about 65 nm for a 1015 cm22 doseof Si at 20 keV.320 Efforts have been made to minimize thdamage by varying the dose rate and accelerating voltSeveral species like Si, Mg, Ne, S, Se, Ar and Cs have bused for implantation.321 Least swelling (, 0.2mm) is ob-served for a Mg1 dose of 331014 cm22 and 300 keV.

Recently, we have observedp- to n-type conversion inthe near surface region of GaSb as a result of argon-ion bmilling.322 Electron beam induced current~EBIC! measure-ments have been employed for detecting the type conversEnhancement in luminescence intensity is observed ingions where ion beam treatment occurs. The type conver



r-rs

e

n

in

s-

n

n

as-

t

r-

psr

e.en

m

n.e-on

is proposed to occur due to a combined effect of generaof native donors~like SbGa or VSb! and gettering of nativeacceptors originally present in the as-grown samples.underlying mechanism that actually leads to the formationthe donor centres is yet to be understood. Neverthelesstechnological applications, the ion milling technique seeto be promising for the fabrication of reliablep–n junctionsfor photodetectors and other optoelectronic devices.

GaSb homojunctions normally have ideality factorabout 2.0 rising from traps near the centre of the band gap253

The hole diffusion length inn-type bulk GaSb with dopingconcentration of 1017 cm23 is usually in the range of 1–1.5mm as measured by EBIC.253 The reverse characteristics aaffected by large surface leakage currents. Auger procethat might influence recombination–generation in GaSb hbeen studied as a function ofp-type doping323,324and someusual characteristics have been observed due to large soff in the valence band.

5. Heterojunctions

The degree of lattice matching is important in hetejunctions except for strained layers below the critical thicness. The critical thickness of GaSb~001!/AlSb structures hasbeen shown to be 100–170 Å.325,326For a lattice mismatch ofas little as 1023

, the interface state density may be greathan 1012 cm22 which is sufficient to dominate the electricaproperties. Moreover the thermal expansion coefficientsthe compounds are not the same and therefore lattice maing at the growth temperature may involve strain at normdevice operation temperatures. LPE has been used forgrowth of AlGaSb, GaAlAsSb and GaInAsSb oGaSb.327–333The LPE growth of GaAlAsSb and InGaAsSover a wide range of compositions is limited by the misbility gap.330,331

The phase diagram of AlGaAsSb has been describegreat detail by Lazzariet al.330 The limitations of the solidphase composition have been discussed. The first limitatwhich prevents the As introduction into the solid phase, islow solubility in the liquid phase below 600 °C. The solubity is further reduced by the addition of Al in the melt. Asconsequence of these low solubility values, GaSb lattimatched AlxGa12xAsySb12y structures withx.0.4 couldnot be grown below 600 °C. The AlGaAsSb system asuffers from the existence of an extremely wide solid phmiscibility gap. The binodal and spinodal curves at 500 a615 °C are shown in Fig. 46. The dashed lines in the figgive the composition of AlGaAsSb lattice matched on GaSAt 500 °C this line is~for a large part! located inside thethermodynamically metastable region of the miscibility gaAt 615 °C the miscibility gap is reduced and it becompossible to obtain AlGaAsSb alloys in the stable domainthe whole range of composition. The intersection of the bodal curve and the lattice matched composition line givesmaximum values ofx and y of the AlxGa12xAsySb12y

lattice-matched alloys which can be grown in the stablegion. This figure shows that high growth temperatur(.600 °C) are needed to obtain stable lattice-matchAlGaAsSb alloys in the entire range of composition. TLPE growth of device quality epilayers of AlGaSb an



derath

lityo

riliaity

free

s-aao

l-

ps

fo

tunS

oreaSbe

e

e

eeon

esneth

ofetheex

s inofm

thAws

Vp-asof

AlGaAsSb with high Al content (x50.82) has been carrieout by us.333 The pregrowth dissolution cycles affect thmorphology of the layers drastically. The optimum tempeture to grow perfectly latticed-matched AlGaAsSb wix50.82 has been found to be around 650 °C.

Due to the existence of a large solid phase miscibigap, extending the operation wavelengthGa12xInxAsySb12y/GaSb lasers beyond 2mm becomes dif-ficult. It has been argued that growth inside an unstablegion can still occur because the substrate tends to stabepitaxial layers which are nearly lattice matched. Lazzet al.have discussed the growth limitations by the miscibilgap in the LPE of GaInAsSb on GaSb.331 Fig. 47 shows thebinodal curves calculated with the hypothesis of stress-solid phase~broken curves! and by taking into account thsubstrate-induced strain energy~solid lines!. One can see areduction in the total size of the miscibility gap of the sytem, and the existence of an extra miscible region for epitial layers with little or no mismatch. In that case, there issolid phase stabilization by the substrate, and growthGa12xInxAsySb12y alloy lattice matched with GaSb is alowed in the whole domain of indium concentration 0, x, 1 at 615 °C. It should be noted that the miscibility gaare almost the same for strained~100! and strained~111!epitaxial layers.

MBE and MOCVD techniques have also been usedthe growth of GaSb, InAs and AlSb based structures.334–339

For lasers and waveguides where optical confinemenneeded, knowledge of the refractive indices of the compomaterials is required. The relative dielectric constant of Gais 14.44 and that of AlSb is 10.3. Hence~AlGa!Sb alloys canprovide satisfactory confinement. The electron affinityGaSb is 4.06 eV. Hence if large energy barriers are requiAl compounds have to be used. The band-gap energiesrefractive indices of the quaternary compounds AlGaAsGaInAsSb and InPAsSb lattice matched to GaSb have bcalculated by Adachi.340

The lattice-matching condition for thAl xGa12xAsySb12y on GaSb substrate is given by:

y50.0396x/~0.442610.0318x! for ~0<x<1.0!.

The lattice matching condition for thGaxIn12xAsySb12y on GaSb substrate is given by:

y5~0.383520.3835x!/~0.421010.216x!

for 0<x<1.0.

The lattice matching condition for the InPxAsySb12x2y

on GaSb substrate is given by:

x50.628120.6895y for 0<y<0.91.

Fig. 48a shows the band gap energies~E0 , EgX andEg

L!and the refractive indices as a function of thex compositionfor Al xGa12xAsySb12y lattice matched to GaSb. One can sfrom the figure that there is a direct–indirect transitiaroundx'0.45. It is worth noting that theE0 value in thedirect gap region of this system is very close to the lowindirect gap energy (Eg

L). This degeneracy of the conductioband at theG and L points is known to result in a largthreshold current, owing to an injected electron loss in



-

f

e-zeri

e

x-

f

r

isdb

fd,nd,en

t

e

L valleys. The calculated refractive indices as a functionthe photon energy withx composition increments of 0.1 arshown in Fig. 48b. As can be seen from the figure,AlGaAsSb/GaSb system provides a large refractive indstep within a whole range of thex composition for the lightconfinement in an active region.

Fig. 49a shows the band gap energies~E0 , EgX andEg

L)and the refractive indices as a function of thex compositionfor GaxIn12xAsySb12y lattice matched to GaSb. UnlikeAlGaAsSb, the absorption at the fundamental optical gapthis system is expected to be direct within the whole rangethex composition. Moreover, the conduction band minimuat theG point is much lower than those at theL andX points,especially for smaller values ofx. The calculated refractiveindices as a function of the photon energy wix-composition increments of 0.1 are shown in Fig. 49b.noticable feature found in this figure is that the system shoa refractive index anamoly, i.e., the smallerE0-gap materialhas a smaller value of the refractive index. Most of the III–alloys like AlGaAs, InGaAsP and AlGaAsSb show the oposite trend. The origin of the refractive index anomaly hbeen discussed by Adachi. Since the refractive indices

FIG. 46. Solid phase miscibility gap of AlGaAsSb system at~a! 500 °C and~b! 615 °C: ~—! binodal curves;~-.-.-.! spinodal curves;~••••••••! GaSblattice-matched alloys~from Ref. 330!.



omon

r-sre

enigio-

tive

ra-de-ob-hencehtsur.

mi-toaous

AlGaAsSb alloys are usually greater than thoseGaInAsSb, cladding layers of AlGaAsSb make a strong cobination with GaInAsSb active layer to confine radiatifrom such an injection laser.

Fig. 50a shows the band-gap energies~E0 , EgX andEg

L!and the refractive indices as a function of thex compositionfor InPxAsySb12x2y lattice matched to GaSb. This quatenary has direct gaps as the fundamental absorption edgethe full range of they composition. The direct gap values amuch lower than the lowest indirect gap energiesEg

L . Thecalculated refractive indices as a function of the photonergy withy-composition increments of 0.1 are shown in F50b. The refractive index anomaly occurs only in the regof smallery composition (y<0.2). The long wavelength indices do not differ as much with they composition. It is

FIG. 47. Miscibility gap of the Ga12xInxAsySb12y system at 615 °C:~- - -!binodal curve without strain effect; ~—! binodal curve withGa12xInxAsySb12y/GaSb strain effect;~s! typical compositions of LPE~100! epitaxial layers;~n! typical compositions of LPE~111!B epitaxiallayers~from Ref. 331!.



f-

for

-.n

evident from this figure that InPAsSb lasers require an aclayer of largery and cladding layers of smallery composi-tions in order to obtain sufficient optical confinement.

E. Device structures

1. Metal/a-Si:H/GaSb structures

Metal–semiconductor rectifying junctions have been tditionally used for both materials characterization and invice structures. As discussed in the previous subsection,taining ideal Schottky contacts on GaSb is difficult. Tsurface Fermi level in GaSb is pinned close to the valeband edge leading to extremely low Schottky barrier heig(;0.1 eV) onp-GaSb. This results in near ohmic behavioof the metal/p-GaSb structure, thus limiting its applicabilityThe problems of low barrier height onp-GaSb and highleakage currents have been alleviated by modifying the seconductor surface using thin interfacial layers priormetallization.341 The interfacial layer can be an insulator,larger band gap crystalline semiconductor or an amorph

FIG. 48. ~a! Band-gap energies as a function of thex-composition forAl xGa12xAsySb12y lattice matched to GaSb and~b! the calculated refractiveindices as a function of photon energy withx increments of 0.1~from Ref.340!.

s
FIG. 49. ~a! Band-gap energies as a function of thex composition for GaxIn12xAsySb12y lattice matched to GaSb and~b! the calculated refractive indices aa function of photon energy withx increments of 0.1~from Ref. 340!.


s
FIG. 50. ~a! Band-gap energies as a function of they composition for InPxAsySb12x2y lattice matched to GaSb and~b! the calculated refractive indices aa function of photon energy withy increments of 0.1~from Ref. 340!.
ato

re

iawthen

ont i

re

e-s

he

asThf t

ie

l/

Sorbe

,re,ialto

ffi-

t2.3om

nrs,ustersher--re

o-

sig..226as

nechas

semiconductor. Amorphous semiconductors present madvantages compared to the other two and hence seemsthe best alternative.

The I –V characteristics of the Mg/n-GaSb and Mg/a-Si:H/n-GaSb diodes for two interfacial layer thickness ashown in Fig. 51a. In the forward bias region, an increasethe turn-on voltage after the deposition of the interfaclayer can be seen. The turn-on voltage increases furtherincrease in the film thickness. In the reverse bias region,diodes with interfacial layer exhibit excellent reverse currsaturation, extremely low reverse current~;tens of nA! andhigh breakdown voltage (;10 V).

The I –V characteristics of various diodes fabricatedp-GaSb are shown in Fig. 51b. Systematic improvementhe diode characteristics is seen with increase ina-Si:H layerthickness. Good forward characteristics with reduction inverse leakage current and breakdown voltage as high asV are observed. SuchI –V characteristics would not havbeen possible inp-GaSb Schottky diode but for the interfacial layer. Moreover, the reverse current in these structurelower than thep–n junctions fabricated previously.239 It isworth mentioning that even though the current flow in tforward bias regime will reduce due to series resistancethea-Si:H layer, the same is not responsible for the increin the turn-on voltage as depicted in Figs. 51a and 51b.increased turn-on voltage has been explained in terms ospace charge limited current~SCLC! in the thin a-Si:Hlayer.341

The current in these diodes are controlled by the barrat thea-Si:H/GaSb and metal/a-Si:H junctions appearing inseries with an effective resistance of thea-Si:H layer. Froma technological point of view, since a rectifying metaa-Si:H contact dominates theI –V characteristic in reversebias and low forward bias due to high barrier heights, Gabased metal–semiconductor field-effect transist~MESFETs! with extremely low gate leakage currents canfabricated with interfacial layer ofa-Si:H. Here the depletion



nybe

inlithet

n

-1.5

is

ofeehe

rs

bs

layer extends through thea-Si:H layer into the bulk GaSballowing modulation of the channel in FETs. Furthermodue to its ability to reduce leakage current, thin interfaclayers ofa-Si:H can be potential surface passivants. Duethe extremely low dark currents, thea-Si:H/GaSb structurescan also be employed as stacked solar cells of high eciency.

2. Injection lasers

The ~AlInGa!~AsSb! set of alloys have shown excellendevice performance in the wavelength range of 0.8–mm.342–351Optical pumping has been used to achieve rotemperature laser oscillation at 2.07mm in a GaInAsSb/AlGaAsSb double heterostructure on GaSb;343 however, thenormal mode of interest is the pumping by injectiocurrent.344 In high-injection-current devices such as lasewhere substantial power is involved and temperatures mbe kept low, the thermal resistance of the quaternary laybecomes important. The quaternary alloys have higher tmal resistivity than GaSb (3.3 cm K/W). Initially LPEgrown 2.2mm laser structures exhibited room temperatuthreshold current densities of about 8 kA/cm2

. With im-proved growth conditions and using a different alloy compsition, reduction in threshold current density to 3.5 kA/cm2 at290 K has been achieved.345

Caneauet al.345,346 had grown GaInAsSb active layerwith AlGaAsSb confinement layers by LPE as shown in F52 and demonstrated room temperature emission at 2mm for pulsed operation with threshold current densitylow as 1.7 kA/cm2. Chiu, Zyskind, and Tsang347 alsoachieved similar threshold current density with MBE-growInGaAsSb on~100! GaSb for emission wavelength in th2–2.5mm range. Room temperature cw operation of suLPE-grown lasers with confining layers more rich in Al hbeen demonstrated by Bochkarevet al.348 Eglash, Choi, andTurner349 and Eglash and Choi350 reported MBE-grown



u-r

fahla-

nise

i

r

tter-sedonithderesrge

a-ni-rs

3.8

an

GaInAsSb/AlGaAsSb lasers on GaSb with a threshold crent density of 1.7 kA/cm2 and differential quantum efficiencies of 18% per facet for pulsed operation at room tempeture. GaSb-based diode lasers emitting at 2 and 4mm withthreshold current density as low as 260 A/cm2 and cw outputpower of 190 mW/facet at room temperature have beenricated by using cladding layers rich in Al content. Higspeed modulation of a GaSb/AlGaSb multiquantum-wellser diode up to 1 GHz forl51.66mm has been demonstrated by Toba and Nosu.351

3. Photodetectors and solar cells

The ratio of ionisation coefficients of holes and electro(kp /kn) is large and is a key factor for high speed, low noavalanche photodetectors~APDs!. The high ratio of ionisa-tion coefficients as shown in Fig. 53 can be observed

FIG. 51. ~a! I –V characteristics of~i! Mg/n-GaSb,~ii ! Mg/a-Si:H ~100 Å!/n-GaSb,~iii ! Mg/a-Si:H ~150 Å!/n-GaSb;~b! I –V characteristics of~i! Mg/p-GaSb, ~ii ! Mg/a-Si:H ~100 Å!/p-GaSb, ~iii ! Mg/a-Si:H ~150 Å!/p-GaSb,~iv! Mg/a-Si:H ~200 Å!/p-GaSb~from Ref. 341!.



r-

a-

b-

-

s

n

Ga12xAl xSb structures with x in the range of0.04–0.06.352–364 Minimum in the threshold energy foimpact ionisation nearD'Eg composition cannot explainenhancement of hole ionisation coefficient. Increased scaing rate is responsible for this enhancement which is cauby a minimum change of momentum in the hole ionisatinear the threshold and the mixing of a S-like state wthe valence-like state induced by a composition disorin this ternary compound. The dark current of such diodtends to be largely due to generation in the space charegion and by tunnelling through deep centres~at high re-verse voltages!. Reduction of deep impurity centres and ntive defect concentration during growth has helped in mimizing the tunnel current. Avalanche multiplication factoof 30–50 with external quantum efficiencies~without antire-flection coating! of over 50% at 1.3mm and a sensitivity of0.6 A/W have been obtained. Ap1n Al0.053Ga0.947Sb homo-junction avalanche photodiode for 1.7mm detection hasdemonstrated a gain of 30 with an excess noise factor ofdB which is better than that of an InGaAs APD.358 Thestructure and reverseI –V characteristics of such an APDare shown in Fig. 54. A detector fabricated fromn-Ga0.82In0.18As0.17Sb0.83 layer with Zn diffusedp

1 face has

FIG. 52. Structure and emission spectrum for a~GaIn!~AsSb! laser ~fromRef. 346!.



thto

-

ente1in

o

tu

tsvethayisuficinexeseffich-r

-

m

e

e

larltyt.ed

Sby

e

Sb

ionhe

l

exhibited rise and fall times of about 50 and 500 ps anddetector response to pseudomodulation at bit rates upGbit/s.363

IR detectors in the 8–14mm regime are presently fabricated by the mercury based compound HgCdTe whichextremely difficult and hazardous to manufacture. At presthe Sb-based III–V are under intense investigation for detors in this range. Detection of longer wavelengths, 8–mm, is possible with intersubband absorptionGa12xAl xSb/AlSb superlattices.4 The novel type II strained-layer superlattice system of InAs–InxGa12xSb also showspromise as material for infrared detection in the 8–14mmrange.365 In this system, the two constituent layers are nlattice matched. The internal strain effects caused by the mmatch are used to reduce the band gap of the InAs quanwells. Calculations show that the InxGa12xSb–InAs has aneffective mass greater than and absorption comparable toof Hg12xCdxTe ~MCT!. The quality of the superlattices ihighly dependent on growth conditions. Substantial improments in structural quality can be achieved by growingsuperlattice structures on GaSb substrates; however, Gsubstrates are less expensive and provide the possibilitmonolithic integration with readout circuitry. Hence, itdesirable to have the superlattice structure on a GaSb blayer grown on GaAs substrate. Unfortunately, the lattconstant of GaSb is 7% larger than that of GaAs, causlarge concentrations of dislocations at the interface. To ovcome the lattice mismatch problem, thick stress-relabuffer layers of GaSb on~100!-oriented GaAs substratehave been grown with a short-period heavily strained suplattice at the interface of GaAs/GaSb. An absorption coecient of 2000 cm21 at 10mm has been measured from sua superlattice with an 11mm energy gap. This value is comparable to that of bulk MCT, the current industry standafor IR detectors in the 8–14mm range.365

p1-InSb/p-InAs12xSbx/n1-InSb heterojunction photo

diodes operating at room temperature in the 8–13mm regionhave recently been reported by Kimet al.366 The voltageresponsivity-area product of 331025 V cm2/W has been ob-tained at 300 K for thel510.6mm optimized device. Thisis close to the theoretical limit set by the Auger mechanis

FIG. 53. Hole and electron ionization coefficient ratiob/a in Ga12xAl xSbas a function ofD/Eg . ~D: spin orbit splitting energy! ~from Ref. 8!.



e2

ist,c-4

tis-m

hat

-eAsof

feregr-d

r--

d

,

with a detectivity at room temperature of'1.53108

cm Hz1/2/W.GaSbp1n solar cells have been proposed as units to b

placed behind GaAs solar cells. At 503 light concentrationand air mass zero~AM0! the GaSb cells exhibit 6.5% effi-ciency which under favourable conditions might boost thefficiency of a GaAs/GaSb tandem stack to over 30%.12 Inorder to use a single load, the current in the tandem socells have to be matched. However, because of the difficuin obtaining matching conditions over a large range of inpuillumination, the applications of GaSb solar cells are limitedReduced series resistance in future GaSb cells by improvcontact metallization and a denser grid design will allowtandem solar energy conversion efficiencies over 35%. GaTPV cells coupled with a gas-fueled burner have alreaddemonstrated 27% efficiency at a loading of 13 W/cm2 in ageometry amenable to automotive applications.57 Further im-provements can be achieved by using thGa0.9In0.1Sb alloy.

Charge storage and charge transfer in a AlGaAsSb/Gaheterojunction charge coupled device~CCD! have been dem-onstrated by Liuet al.367A charge filling time as long as 450s has been observed at 77 K. Room temperature operathas not been possible due to high leakage current in tdevice.

4. Transistors

The performance of bipolar and FETs of~AlInGa!~AsSb! has been evaluated by severa

FIG. 54. AlGaSb avalanche photodetector structure and dark current~fromRef. 358!.



el

unnae

ent

.tteth

pes

s-

nosn

7OD

a

ye

eo

s

Tsac-c-z

ET-s/-theval-e

s.s–an-es

twos,der

workers.368–377 The GaSb/InAs heterojunctions are of thbroken gap variety as the valence band edge of the GaSbabove the conduction band edge of the InAs~see Fig. 55!.Therefore near-ohmic behaviour and tunnelling occursless Al is included in the GaSb to create a staggering jution interface. Such a structure as shown in Fig. 56 withAlSb emitter, an InAs base and a GaSb collector has bfabricated by Chui and Levi.368 The best collector efficiencyobserved was about 0.9. Tairaet al.369,370fabricated two hotelectron transistors~HET! with a p-GaSb emitter/n-InAsbase/p-GaSb collector and with GaSb emitter/InAs basgraded Ga0.9In0.1Sb collector. In the latter structure, a curregain of 6 was obtained at 300 K. Furukawa and Mizuta372

fabricated a heterojunction bipolar transistor~HBT! withn-Al0.5Ga0.5Sb emitter/p1-GaSb base/n-GaSb collector. Fig.57 shows the structure andI –V characteristics of the HBTAt 300 K, the hole mobility in the base was three times bethan that for similar doping concentration of GaAs andgain was as high as 160 for a 1003100mm2 size emitter.The AlGaAs/GaAsSb strained system has desirable proties for high-current-gain high-frequency HBT applicationGaAsSb has been shown to facilitatep-type ohmic contactsto p–n–p AlGaAs/GaAs HBTs with specific contact resitivity as low as 531027 V cm2.281 A prototype n–p–nAlGaAs/GaAsSb/GaAs DHBT exhibited a stable curregain of 5 and a significant collector current density53104 A/cm2. GaSb p-channel modulation-doped FET~MODFETs! with AlSb0.9As0.1/AlSb barrier layers have beestudied by Luo and co-workers.373 For 1 mm gate lengthdevices the transconductances were 50–86 mS/mm at 7The schematic cross section and band diagram of the MFET are shown in Fig. 58. Longenbach and co-workers374

proposed a complementary heterojunction FET~HFET!structure with InAs as then channel and GaSb as thepchannel. The structure is shown in Fig. 59. ComplementFET pairs of such types have been studied by Kanjiet al.375

FETs using InAs active layer and GaSb or AlSb barrier lahave been fabricated by Kop’evet al.376 and Werkinget al.377 With a gate length of 1.7mm, the transconductancof an InAs channel FET has been found in the range460–500 mS/mm at room temperature.

FIG. 55. Energy gap lineups of InSb–GaSb–AlSb–InAs~from Ref. 3!.



ies

-c-nen

/

re

r-.

tf

K.-

ry

r

f

Quantum wells ~QWs! of InAs with ~AlGa!~AsSb!barriers combine the high intrinsic mobilitie(>30 000 cm2/V s) and the high drift velocities in InAs~'43107 cm/s at 2.5 kV/cm!, with the high electron con-finement barrier~1.35 eV! of the InAs–AlSb interface. Thesetwo sets of properties are promising for InAs based HFEwith superior speed compared to other materials. Whiletual device performance still falls short of theoretical expetations, with 0.5mm gate length, cutoff frequency of 93 GHhas been demonstrated.378

The three terminal resonant interband tunnelling F~RITFET!, based on type II InAs/AlSb/GaSb RITD integrated into the drain or source region of an InGaAs/AlGaAGaAs FET~see Fig. 60! demonstrates the possibility of combining the best tunneling diode structure with the best ofFET technology to achieve the highest possible peak toley ~P/V! current ratio and gain for future quantum threterminal devices.379

5. Quantum wells, quantum dots and superlattices

Way back in 1980, Chang and Esaki380 demonstratedtunneling actions in InAs–GaSb superlatticeSemiconductor–semimetal transition is observed in InAGaSb superlattices. The critical layer thickness for this trsition is calculated to be 85 Å, below which the superlattic

FIG. 56. ~a! Schematic diagram of the epitaxial layer sequence andlevel transistor mesa structure for hot electron transport measurement~b!conduction band diagram of the AlSb/InAs/GaSb heterostructure unbase/emitter biasVbe and collector/base biasVcb ~from Ref. 368!.



tallia

owr-inecS

e

feb//a

eiapo

ele

ghin-ors,

ts

2

s/ry

behave as semiconductors and above which semimeMore recently, several researchers have fabricated tunnedevices based on the InAs–AlSb–GaSb QW structuresstudied their transport properties.381–400Ten different tunnelstructures are realizable by employing this system as shin Fig. 61.381–393The optical properties of QWs and supelattices involving AlGaSb have been found to be interestfor device applications specially for long wavelength dettion. Forbidden Auger process in strained InGaSb/AlGaQWs has been studied by Jiang and co-workers.396 Carret al.397 made an analysis of the quantum-confined starkfect in GaSb/AlGaSb multiple QWs~MQWs!. A normal-incidence modulation is proposed which uses the Stark efto induceG–L transitions in asymmetrically stepped AlSInAs/GaSb/AlSb QWs.398 The calculations indicate that onoff ratios can be achieved in this structure operatingT<150 K with electric fields on the order of;100 kV/cmfor any infrared wavelength within the range of 3–20mm.Electron transfer fromG to L valleys, and hence the devicswitching, can be achieved efficiently under a moderate b~see Fig. 62!. Infrared devices based on this system cantentially operate at both long-wave~8–12 mm! and mid-wave ~3–5 mm! infrared ranges, due to the largeL valleyoffsets of InAs and GaSb with AlSb.

In GaSb, due to the proximity of theL conduction bandto theG, for ~001! grown QWs or superlattices, the effectivmass tensor is nondiagonal allowing normal incidence e

FIG. 57. Hot electron emitter transistor,~a! conduction band arrangemenof emitter and base~G andL valley minimum values of Al0.5Ga0.5Sb are notso different and they are shown by single broken line!, ~b! common-emitterI –V characteristics for the AlGaSb/GaSb HBT at 300 K. Base current ismA/step ~from Ref. 372!.



ls.ngnd

n

g-b

f-

ct

t

s-

c-

tronic intersubband transitions.L valleys can be populatedeither through confinement by adding Al to GaSb or throuhigh doping and strong far infrared absorption at normalcidence is observed. This is attractive for infrared detect

0FIG. 58. GaSbp-channel MODFET.~a! schematic cross section;~b! banddiagram at zero bias~from Ref. 373!.

FIG. 59. ~a! Band diagram of a complementary HFET structure of InAAlGaSb/GaSb,~b! cross section of InAs/AlGaSb/GaSb complementaHFET integrated circuit~from Ref. 374!.



ei-rbes

rm

cr afa

s

noawtico

dlendan20

lyineir

vity

tord byc-in-ereare

llin-

modulators and second harmonic generation~SHG!, etc. Sur-face emitting SHG at normal incidence usingL valley inter-subband transitions in AlSb/GaSb/GaAlSb/AlSb steppQWs has been demonstrated.399 Second harmonic susceptbility of at least 931028 m/V has been achieved undedouble resonance conditions, which is comparable to theobtainable withG valley processes in GaAs/GaAlAs systemfor 45° propagation angles. Besides the advantage of noincidence geometry, the largeL valley conduction band off-set between GaSb and AlSb enables doubling of frequenspanning the entire midwave infrared spectral region. FoAl xGa12xSb/GaSb QW with barrier composition ox50.31, the confinement inducedG–L crossover occurs atwell thickness of 12 monolayers.

In GaSb/AlSb quantum wells, ‘‘quasidirect’’ transitionfrom theX conduction band to theG valence band in theGaSb is possible provided its thickness is only a few molayers. PL characteristics at 1.4 K suggest that the interbselection rules fork conservation are relaxed in such narroQWs.400 Optically induced femtosecond electromagnepulses from GaSb/AlSb strained-layer superlattices wereserved by Zhanget al.401

GaSb-AlSb MQWs are ‘‘good relaxed systems’’ anstructures grown on GaAs substrates are of good optoetronic quality. The difference in optical index of GaSb aAlSb is about a factor of 2 better than that between InPInGaAsP quaternary alloy. It is known that it requires

FIG. 60. ~a! Schematic cross section of the resonant interband tunneFET and~b! experimentalI –V characteristics of the structure at room temperature~from Ref. 379!.



d

st

al

iesn

-nd

b-

c-

d

pairs of InP/InGaAsP layers to acheive a reflectivity of on95%. GaSb–AlSb MQWs have one excitonic absorption lpeaking at about 1.5mm at room temperature and a 10 paBragg reflector shows 97% reflectivity.402 This will proveuseful for microcavity based devices such as vertical casurface emitting lasers.

Magnetic field induced semimetal and semiconductransition in InAs–GaSb superlattices has been observeKawai et al.403 with closely overlapped subbands of eletrons and holes. The transition is manifested in a sharpcrease in the magnetoresistance in the quantum limit, whthe ground Landau levels associated with the subbands

g

FIG. 61. Tunnel structures of InAs/GaSb/AlSb~from Refs. 381–393!.

FIG. 62. Schematic conduction band alignments associated withG andLvalley stepped QWs for an AlSb/InAs/GaSb/AlSb structure~from Ref. 398!.



hneo-cia

nAl–

sicceaensidtuu

edanvneara

e

la

cunhkibtS

o

nelicesi

t tfabonrlo

ereialcialbele

ro-edanbyandSbar-ugha-ls,ofo-lkvente-onb-E,o

boutherehisas

westhetionto

e-o-iceects

eto

nceinac-riedtoms.mayAs.tial.

of

crossed at the Fermi level, resulting in carrier depletion. TInAs–GaSb system is known to exhibit unusual magtotransport properties arising from coexisting twdimensional electrons and holes which result from the peliar band offsets at the interface between the two materand with an electron mobility of 3.53105 cm2/V s in thepresence of holes. The holes disappear at a critical Ithickness around 60 Å resulting from the semimetasemiconductor transition.404

Sb/GaSb~111! heterojunction and multilayers structureare potential systems with indirect narrow gap superlattwhere spacial quantization effects induce a positive valenconduction band energy gap in the Sb semimetal layersthe large characteristic lengths in the Sb layers can beploited for the study of size quantization effects. The traport and optical properties of such structures have beenvestigated by Goldinget al.405 Sb/GaSb quantum wells ansuperlattices may be ideal system for studying quantransport and electron–hole correlation effects. Taking bvalues for the electron and hole densities (;1019 cm23) andlow temperature mobilities (>105 cm2/V s), one obtainsmean free paths of nearly 10mm. One also expects that thAuger recombination lifetime should be orders of magnitulonger than in diret gap semiconductors with the same bgap. The large mass anistropy, multiple conduction andlence minima, long lifetime and small absorption coefficieacross the indirect gap are all highly favourable for nonlinoptical applications, such as optical switches devices opeing in the infrared.

Self-assembled nanoscale quantum dots of GaSb wgrown on GaAs~001! by MBE.406–409In situ scanning tun-nelling microscopy measurements taken after 1–2 monoers growth of GaSb reveal that the surface is a networkanisotropic ribbonlike platelets. These platelets are a presor to quantum dot~QD! nucleation. Transmission electromicroscopy measurements indicate that the QDs are coently strained. The growth of these QDs occurs via StransKrastanov mode. QDs of GaSb capped by GaAs exhstrong luminescence near 1.1 eV. This line is attributedradiative recombination of 0-D holes located in the Gadots and electrons located in the surrounding regions.

X. CONCLUDING REMARKS AND FUTURE OUTLOOK

The usage of GaSb based systems as an alternative tpresent day devices operating in the 1.55mm and for sourcesand detectors in the 2.5mm regime relies on the productioof high quality material with low background doping levand defect density. While the basic material quality is dtated by the crystal growth conditions, the physical propties of the material are profoundly influenced by the proceing cycles and the conditions under which the deviceoperated. Hence in making a good device, it is importanunderstand the material issues that are related to devicerication and device operation and to achieve synergiestween material preparation, processing and device functi

A comparison of the problems encountered by eaworkers during the developmental stages of other III–V



e-

u-ls

s

e,–ndx--n-

mlk

eda-trt-

re

y-ofr-

er-i–itob

the

-r-s-sob-e-s.yr

II–VI optoelectronic materials and GaSb indicates that this relatively less hinderance for bringing out the full potentapplications of the latter system. There are several spebasic properties of GaSb and related materials that canimmediately exploited for large scale commercially reliabdevices.

High quality device grade bulk substrate can be pduced more suitably by vertical Bridgman with submergheater410 and vertical gradient freeze methods rather thconventional Czochralski technique. Materials producedthese techniques have shown the lowest defect contentare commercially viable. The stacking fault energy of Gais highest amongst III–V and the critical resolved shestress~CRSS! is relatively higher; hence, obtaining low defect density crystals is relatively easier. Further, even tholot of efforts have been made to grow bulk crystals from Gor Sb-rich melts with low native background acceptor leveit is disadvantageous from the structural quality pointview. Metallic inclusions that get introduced during nonstichiometric melt growth can limit the usefulness of the bucrystals as substrates. Instead, the best ways to circumthe high background doping level, which may affect the dvice performance, can be either by postgrowth Li diffusior by epitaxial growth of GaSb buffer layer on the bulk sustrate. High quality epilayers can be grown by LPMOCVD or MBE with native acceptor concentrations tworders of magnitude lower than the bulk.

Even though there exists substantial understanding athe physics of the material, there are several avenues wfurther work has to be carried out in order to upgrade tIII–V system for optoelectronic devices. Some of the arethat need immediate attention are discussed below.

Doping of impurities has been restricted to only a feelements. A rigorous study on doping of various impuritishould be carried out, first to see the effects of dopants onnative defects and second to obtain carrier concentrawhich can be varied from the intrinsic concentration limitvery high levels'1020 cm23. Various doping levels may bepossible with different impurities. Also, the spectrum of dfect levels for various impurities has to be determined. Cdoping may lead to interesting material properties and devapplications. Isoelectronic dopants may reduce native defconsiderably.

Ion implantation and diffusion of impurities need to bworked on more carefully and attempts should be madesolve the problems encountered previously. Luminescefrom impurity levels in the band gap may prove promisingexploiting this material as IR sources. Hence, optical charterization of various luminescence centres should be carout in detail. Attempts to increase carrier mobilities havebe made by suitably defect complexation mechanisDX-centre based basic research and device applicationsprove useful using S:GaSb rather than conventional AlGaA theoretical analysis to this effect in GaSb is also essen

High resistivity substrate~with intrinsic resistivity limit!is of utmost importance. Even though some indicationshigh resitivity GaSb through high temperature annealing313



isb-

tinssit-

onrin

tai

layeescait

b

eticzaveeoin

oheaDie

.eltenel

as.Sfeouri

uatie

akaAre

ra-heportm-ld bea-a-era-ethe

i

ap

3–edac-er.teds.for-

e

e in

ub-nd

nd

d onh

stcan

re-asaSb

ve-for-uresan-the

de-is-

or by Te compensation411 have been given, such a studyyet to be pursued rigorously. Since the band gap of GaScomparatively low, the intrinsic resistivity limit is of the order of 103 V cm. Unlike in wide band gap III–V like GaAsand InP where high resistivity can be achieved by creamidgap levels, such an approach has not been succeuntil now in GaSb. Sulfur can be an interesting dopant asa deep trap~DX centre! which can reduce the carrier concentration by three orders of magnitude from 1017 to 1014

cm23 with decrease in temperature from 300 to 77 K. Ibeam milling can also be adopted to reach the level of intsic resistivity.322

Both hydrogen plasma passivation and proton implantion can be used to produce current confining regionsdouble heterostructures lasers with AlGaAsSb claddingers.n- to p-type conversion of AlGaAsSb has been observby proton implantation.412Hydrogen passivation seems to bmore advantageous than proton implantation if one aimproducing narrow stripe lasers for high frequency applitions because the isolation is provided by high resistivregions rather than byp–n junction formation.

The problem of high leakage current in devices caneliminated by effective passivants like S, Ru anda-Si:H,which have been found to be robust. Further work is neccsary to employ these surface passivation techniques forbatch on-line device fabrication cycles. Also, basic devtechnologies like wafer preparation, oxide growth, metallition, etc., have to be standardized. Long term stability omany thousands of hours at the highest operating tempture is a matter of considerable importance for contacts. Ctact technology is not yet fully developed. Extended testis needed to develop firm data on the 100 °C lifetime.

Until now, most of the GaSb based work has beenbulk crystals. Inspite of promising device capabilities, tepitaxial technology is highly under-utilized. Material prepration using sophisticated epitaxial techniques like MOCVand MBE should be pursued and their properties studInitial problems encountered in epitaxial techniques needbe solved by knowledge gathered from GaAs and InPtypical example is of LPE growth of GaSb from Sb-rich mwherein the problem of bad surface morphology usuallycountered may be solved by the doping of rare earthments to the melt.

Preliminary efforts in bulk growth of GaSb based ternries ~InGaSb and AlGaSb! have shown encouraging resultIf succeeded, this will drastically reduce the cost of Gabased optoelectronic devices and make it a much moresible technology compared to the existing ones. Rigorcharacterization of such material systems should be carout for further advancement.

From device point of view, GaSb based ternary and qternary heterostructures exhibit interesting optical properfor sources and detectors in the 1.3–2.5mm regime. In spiteof several important properties, Sb based compounds hlagged behind in the past due to several reasons. Lacsuitable substrates and their large lattice mismatch to Gand InP is one. The existence of large miscibility gaps p



is

gfulis

-

-n-d

at-y

e

s-hee-rra-n-g

n

-

d.toA

-e-

-

ba-sed

-s

veofs-

cludes the growth of high quality materials at lower tempetures by equilibrium and nonequilibrium techniques. On tother hand, because of their low band gaps, the transproperties are expected to improve considerably at low teperatures and therefore these materials and devices wouimportant for cryogenic applications. However, epitaxial mterials grown by MBE exhibit type conversion and degradtion of the transport properties as the measurement tempture is lowered from 300 K. In terms of the final devicstructure, it is envisaged that GaSb layers would formbuffer and contact layers; Ga12xInxAsySb12y , the active lay-ers for lasers and detectors in the 2–4mm regime andGa12xAl xAsySb12y , the cladding layers. Since the Fermlevel is pinned close to the valence band edge inp-GaSb, itforms an excellent ohmic contact material on wide band gIII–V compounds.

The present day APDs based on InGaAs/InP for 1.1.55mm light wave communication systems can be replacby AlGaSb/GaSb APDs, which have lower excess noise ftor and higher gain–bandwidth product than the formGain–bandwidth product of 90 GHz has been demonstrawhich is the largest ever reported for long wavelength APD

In comparison to the present day HgCdTe detectors8–14mm regime, the InAs/Ga12xInxSb superlattices are expected to hold several advantages:

~i! a higher degree of uniformity, which is crucial for thfabrication of large infrared detector arrays;

~ii ! smaller leakage currents due to the tunable increaseffective mass available in a superlattice;

~iii ! reduced Auger recombination rates, due to the sstantial splitting of the light and heavy hole bands athe increase in electron effective mass;

~iv! better understood device processing techniques; a~v! compatibility with GaAs-based readout electronics.

Furthermore, the performance of infrared detectors baseInAs/Ga12xInxSb superlattices is not limited by the higthermal generation rates which preclude largeD* in multi-quantum well infrared detectors~such as GaAs/AlGaAs!.The other candidate, InAs12xSbx , also exhibits several im-portant advantages over Hg12xCdxTe like better stability,higher electron and hole mobilities, high quality and low cosubstrates like GaAs. The performance of these detectorsbe further improved by optical immersion with lenses ppared directly from the GaAs substrate similar to what hbeen done for HgCdTe devices. Recent advances in Gand InGaSb TPV technology413 indicate a favourable futuremarket trend.

As far as the sources are concerned, further improment may be expected in GaInAsSb injection laser permance since many features found in GaAs laser stucthave not yet been thoroughly explored. This includes qutum wells and graded index confining layers to improvelinewidth and efficiency.

Intraband and interband negative resistance tunnelvices with high peak-to-valley ratios and FETs have proming features for high speed applications.



ldt

idntaonhacn

hiopishtu

neththc-

non

J.-

G

on.

so

di

and

l.

rg.

R,

pl.

-

tz,

A.

v.

I.

om/

ech-

st.

It is clear that considerable developmental work woube needed to bring the performance of transistors up tolevel of GaAs. Interesting phototransistor possibilities~at1.55mm! might justify further exploratory studies.

The growth of Sb on GaSb may be pertinent to a wvariety of experimental studies and practical applicatiosince it is one of the few examples of thin-film single-cryssemimetals that exhibit heteroepitaxial growth with semicductors. Such structures would allow the fabrication of higquality interconnects between circuit elements, as wellthree-dimensional device integration through buried contaand ground planes, and hybrid devices such as a resotunnelling transistor with a metal base.

Summarizing, while the present day knowledge in tmaterial is sufficient to understand the basic physical prerties, rigorous efforts toward device studies seem to besufficient. The future efforts should be concentrated so amove from materials science to the device arena, where ttechnological development can rapidly be brought to marity.

ACKNOWLEDGMENTS

Thanks are due to Geeta Rajagopalan, who has dowonderful job of typing and compiling the manuscript wigreat patience and K. Venugopalan for coordinating allart work. This work is supported by Council for Scientifiand Industrial Research~India!, Defence Research and Development Organization~India! and University Grant Com-mission~India!.

1R. A. Laudise, J. Cryst. Growth65, 3 ~1983!.2Special Issue On Low Loss Fibers, IEEE Lightwave Technol.LT-2~1984!.

3A. G. Milnes and A. Y. Polyakov, Solid-State Electron.36, 803 ~1993!.4H. Xie, J. Piao, J. Katz, and W. I. Wang, J. Appl. Phys.70, 3152~1991!.5V. Swaminathan and A. T. Macrander,Materials aspects of GaAs and InPbased structures~Prentice–Hall, Englewood Cliffs, NJ, 1991!, Chap. 1.

6G. Motosugi and T. Kagawa, Jpn. J. Appl. Phys.19, 2303~1980!.7M. B. Z. Morosini, J. L. Herrera-Perez, M. S. S. Loural, A. A. G. VoZuben, A. C. F. da Silveira, and N. B. Patel, IEEE J. Quantum ElectrQE-29, 2103~1993!.

8O. Hildebrand, W. Kuebart, K. W. Benz, and M. H. Pilkuhn, IEEEQuantum Electron.QE-17, 284 ~1981!; see also O. Hildebrand, W. Kuebart, and M. H. Pilkuhn, Appl. Phys. Lett.37, 801 ~1980!.

9K. Segawa, H. Miki, M. Otsubo, and K. Shirata, Electron. Lett.12, 124~1976!.

10C. Hilsum and H. D. Rees, Electron. Lett.6, 277 ~1970!.11L. Esaki, J. Cryst. Growth52, 227 ~1981!.12L. M. Frass, G. R. Girard, J. E. Avery, B. A. Arau, V. S. Sundaram, A.Thompson, and J. M. Gee, J. Appl. Phys.66, 3866~1989!; see also D. J.Williams and L. M. Fraas, inProceedings of the 2nd NREL ConferenceTPV Generation of Electricity, Colorado Springs, CO, edited by J. PBenner, T. J. Coutts, and D. S. Ginley~AIP, New York, 1996!, p. 134.

13See the NETSCAPE homepages: DRA highlights, diode laser gas senetc.

14P. S. Dutta, Ph.D. thesis, Indian Institute of Science, Bangalore, In1995.

15G. Benz and R. Conradt, Phys. Rev. B16, 843 ~1977!.16G. Muller, J. Cryst. Growth99, 1242 ~1990!; also see the NETSCAPEhomepage: http://www.space.ca/cfzf/bkgnd.html

17R. D. Baxter, R. T. Bate, and F. J. Reid, J. Phys. Chem. Solids26, 41~1965!.



he

esl--stsant

s-n-toeir-

a

e

.

.

rs,

a,

18F. J. Reid, R. D. Baxter, and S. E. Miller, J. Electrochem. Soc.133, 713~1966!.

19C. Anayama, T. Tanahashi, H. Kuwatsuka, S. Nishiyama, S. Isozumi,K. Nakajima, Appl. Phys. Lett.56, 239 ~1990!.

20K. F. Longenbach and W. I. Wang, Appl. Phys. Lett.59, 2427~1991!.21V. M. Goldschmidt, Skrifter Norshe Vide. Akads. Oslo I: Mat. Naturv. KVIII ~1926!.

22G. Giesecke and H. Pfister, Acta Crystallogr.11, 369 ~1958!.23W. Koster and B. Thoma, Z. Metallkd.46, 291 ~1955!.24I. G. Greenfield and R. L. Smith, Trans. AIME203, 351 ~1955!.25G. A. Kalyuzhnaya, L. K. Polushina, and D. N. Tretyakov, Russ. J. InoChem.9, 813 ~1964!.

26R. N. Hall, J. Electrochem. Soc.110, 385 ~1963!.27Constitution of Binary Alloys, edited by M. Hansen~McGraw–Hill, NewYork, 1958!, p. 755.

28Yu. T. Zakharov and M. S. Mirgalovskaya, Izv. Akad. Nauk SSSNeorg. Mater.3, 1739~1967!.

29D. R. Lide, inCRC Handbook of Chemistry and Physics, 72nd ed.~CRC,New York, 1991!, p. 5.61.

30W. P. Allred, inCompound Semiconductors, edited by R. K. Willardsonand H. L. Goering~Reinhold, New York, 1962!, Vol. I, pp. 187–193.

31M. Kumagawa, H. Oka, N. V. Quang, and Y. Hayakawa, Jpn. J. ApPhys.19, 753 ~1980!.

32S. Tohno and A. Katsui, J. Electrochem. Soc.128, 1614~1981!.33Y. Ohmori and K. Sugii, J. Cryst. Growth56, 39 ~1982!.34K. K. Ozolis and E. Y. Kokorish, Growth of Crystals4, 150 ~1966!.35D. Effer and P. J. Etter, J. Phys. Chem. Solids25, 451 ~1964!.36K. Wade and A. J. Banister,The Chemistry of Aluminium, Gallium, Indium and Thallium~Pergamon, New York, 1973!, p. 1082.

37J. C. Brice and M. J. King, Nature~London! 199, 897 ~1963!.38K. B. McAfee, Jr., D. M. Gay, R. S. Hozack, R. A. Laudise, G. Schwarand W. A. Sunder, J. Cryst. Growth76, 263 ~1986!.

39W. A. Sunder, R. L. Barns, T. Y. Kometani, J. M. Parsey, Jr., and R.Laudise, J. Cryst. Growth78, 9 ~1986!.

40S. Minomura and H. G. Drickamer, Phys. Chem. Solids23, 451 ~1962!.41A. Jayaraman, W. Klement, Jr., and G. C. Kennedy, Phys. Rev.130, 540

~1963!.42J. C. Jamieson, Science139, 845 ~1963!.43G. V. Ozolins, G. K. Averkieva, N. A. Goryunova, and A. F. Ievins, SoPhys. Crystallogr.8, 207 ~1963!.

44A. S. Jordan, J. Cryst. Growth71, 551 ~1985!.45F. Moravec, J. Cryst. Growth128, 457 ~1993!.46W. P. Allred and R. T. Bate, J. Electrochem. Soc.108, 258 ~1961!.47S. Kondo and S. Miyazama, J. Cryst. Growth56, 39 ~1982!.48A. Katsui and C. Uemura, Jpn. J. Appl. Phys.19, L318 ~1980!.49S. Miyazama, S. Kondo, and M. Naganuma, J. Cryst. Growth49, 670

~1980!.50B. Cockayne, V. M. Steward, G. T. Brown, V. R. MacEwan and M.Young, J. Cryst. Growth58, 267 ~1982!.

51F. Moravec and Y. Tomm, Cryst. Res. Technol.22, K30 ~1987!.52V. Sestakova and B. Stepanek, Thermochim. Acta209, 277 ~1992!.53P. G. Mo, H. Z. Tan, L. X. Du, and X. Q. Fan, J. Cryst. Growth126, 613

~1993!.54A. Watanabe, A. Tanaka, and T. Sukegawa, J. Cryst. Growth128, 462

~1993!.55M. Kumagawa, Y. Asaba, and S. Yamada, J. Cryst. Growth41, 245

~1977!.56M. Kumagawa, Jpn. J. Appl. Phys.16, 2297 ~1977!; see also M. Kuma-gawa, J. Cryst. Growth44, 291 ~1978!.

57See the NETSCAPE homepage: www.canadian-content.caboutus.html/firebird/index.html

58T. Migauchi and H. Sonomura, Jpn. J. Appl. Phys. 4, 317~1965!.59F. Moravec, V. Sestakova, B. Stepanek, and V. Charvat, Cryst. Res. Tnol. 24, 275 ~1989!.

60C. E. M. Oliveira and M. M. G. de Carvalho, J. Cryst. Growth151, 9~1995!.

61M. Harsy, T. Gorog, E. Lendvay, and F. Koltai, J. Cryst. Growth53, 234~1981!.

62U. N. Roy and S. Basu, Bull. Mater. Sci.13, 27 ~1990!.63M. Harsy, F. Koltai, and T. Gorog, Acta Phys. Hung.57, 245 ~1985!.64P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, J. CryGrowth141, 44 ~1994!.



st

st

st

s,–

hy

I.

T.y,

ro

ys

tro

aik

nf

tu

st

.

. J

ss

B.

pl.

T.

F.

C.pl.

th

st.

m.

-

n

pl.

.

ys.

iou,

65P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, J. CryGrowth141, 476 ~1994!.

66P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, J. CryGrowth154, 213 ~1995!.

67P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, J. Ind. InSci. 76, 235 ~1996!.

68T. Duffar, J. M. Gourbil, P. Boiton, P. Dusserre, and N. EustathopouloCryst. Growth~in press!; P. Boiton, Ph.D. thesis, Universite MontpellierIII, 1996.

69A. G. Ostrogorsky, US Patent No. 5,047,113~1991!.70W. Lewadowski, N. Gajewska, B. Pastuszka, and M. Bugajski, Acta P57, 303 ~1985!.

71J. P. Garandet, T. Duffar, and J. J. Favier, J. Cryst. Growth96, 888~1989!.72G. Muller and G. Neumann, J. Cryst. Growth63, 58 ~1983!.73K. W. Benz and G. Muller, J. Cryst. Growth46, 35 ~1979!.74V. A. Gevorkyan, L. V. Golubev, V. N. Karyaev, Yu. V. Shartsev, andL. Shul’pina, Sov. Phys. Tech. Phys.24, 1216~1979!.

75G. Bischopink and K. W. Benz, J. Cryst. Growth128, 466 ~1993!.76E. Lendvay, M. Harsy, T. Gorog, I. Pozsgai, F. Koltai, J. Gyulai,Lohner, G. Mezey, E. Koltai, F. Paszti, V. T. Hrjapov, N. A. Kultchiskand L. L. Regel, J. Cryst. Growth71, 538 ~1985!.

77L. L. Regel and O. V. Shumaev, J. Cryst. Growth119, 70 ~1992!.78G. Bischopink and K. W. Benz, J. Cryst. Growth108, 465~1991!; see alsoG. Bischopink and K. W. Benz, J. Cryst. Growth128, 470 ~1993!.

79J. P. Garandet, T. Duffar, and J. J. Favier, J. Cryst. Growth106, 426~1990!.

80K. J. Bachmann, F. A. Thiel, H. Schreiber, Jr., and J. J. Rubin, J. ElectMater.9, 445 ~1980!.

81J. F. Yee, M. C. Lin, K. Sarma, and W. R. Wilcox, J. Cryst. Growth30,185 ~1975!.

82J. C. Wooley and B. A. Smith, Proc. Phys. Soc. London, Sec. B72, 214~1958!.

83I. Ansara, M. Gambino, and J. P. Bros, J. Cryst. Growth32, 101 ~1976!.84A. Joullie, J. Allegre, and G. Bougnot, Mater. Res. Bull.7, 1101~1972!.85C. Marin, P. S. Dutta, E. Dieguez, P. Dusserre, and T. Duffar, J. CrGrowth ~in press!.

86A. Joullie, P. Gautier, and E. Monteil, J. Cryst. Growth47, 100 ~1979!.87C. Heinz and W. Schmidt Auf Altenstadt, J. Cryst. Growth67, 393

~1984!.88F. Capasso, M. B. Panish, and S. Sumski, IEEE J. Quantum ElecQE-17, 273 ~1981!.

89C. Woelk and K. W. Benz, J. Cryst. Growth27, 177 ~1974!.90S. S. Chandvankar and B. M. Arora, J. Cryst. Growth80, 69 ~1987!.91P. Gladkov, E. Monova, and J. Weber, J. Cryst. Growth146, 319 ~1995!.92A. Hojo and I. Kuru, J. Appl. Phys. Suppl.43, 226 ~1974!.93C. Pickering, J. Phys. C13, 2959~1980!.94H. Miki, K. Segawa, and K. Fujibayashi, Jpn. J. Appl. Phys.13, 203

~1974!.95P. S. Dutta, K. S. R. Koteswara Rao, H. L. Bhat, K. Gopalakrishna Nand Vikram Kumar, J. Cryst. Growth152, 14 ~1995!.

96P. S. Dutta, H. L. Bhat, and Vikram Kumar, Bull. Mater. Sci.18, 865~1996!.

97N. Holonyak, Jr., D. C. Jillson, and S. F. Bevacqua, Met. Soc. Co1961, p. 49.

98R. B. Clough and J. J. Tietjen, Trans. AIME245, 583 ~1969!.99M. Kakehi, R. Shimokawa, and T. Arizumi, Jpn. J. Appl. Phys.9, 9

~1970!.100T. Arizumi, M. Kakehi, and R. Shimogawa, J. Cryst. Growth9, 151

~1971!.101N. Kitamura, M. Kakehi, J. Shen, and T. Wada, Jpn. J. Appl. Phys.20,995 ~1981!.

102W. Jakowetz, W. Ruhle, K. Breuninger, and M. Pilkuhn, Phys. StaSolidi A 12, 169 ~1972!.

103H. M. Manasevit and K. L. Hess, J. Electrochem. Soc.126, 2031~1979!.104J. Shin, A. Verma, G. B. Stringfellow, and R. W. Gedridge, Jr., J. CryGrowth151, 1 ~1995!.

105C. B. Cooper, R. R. Saxena, and M. J. Ludowise, J. Electron. Mater11,1001 ~1982!.

106S. K. Haywood, A. B. Henriques, N. J. Mason, R. J. Nicholas, and PWalker, Semicond. Sci. Technol.3, 315 ~1988!.

107F. Pascal, F. Delannoy, J. Bougnot, L. Gouskov, G. Bougnot, P. Groand J. Kaoukab, J. Electron. Mater.19, 187 ~1990!.

108G.-Y. Wei and R.-W. Peng, J. Electron. Mater.23, 217 ~1994!.109T. F. Keuch and E. Veuhoff, J. Cryst. Growth68, 148 ~1984!.



.

.

.

J.

s.

n.

t.

n.

,

.,

s

.

.

e,

110C. H. Chen, C. T. Chiu, L. C. Su, K. T. Huang, J. Shin, and G.Stringfellow, J. Electron. Mater.22, 87 ~1993!.

111M. Kodama and M. Kimata, J. Cryst. Growth73, 641 ~1985!.112S. M. Newstead, T. M. Kerr, and C. E. C. Wood, J. Appl. Phys.66, 4184

~1989!.113S. Subbanna, G. Tuttle, and H. Kroemer, J. Electron. Mater.17, 297

~1988!.114Y. Zhu, Y. Takeda, and A. Sasaki, J. Appl. Phys.64, 1897~1988!.115I. Poole, M. E. Lee, I. R. Cleverley, A. R. Peaker, and K. E. Singer, ApPhys. Lett.57, 1645~1990!.

116S. J. Eglash, H. K. Choi, and C. W. Turner, J. Cryst. Growth111, 601~1991!.

117T. H. Chin, J. A. Ditzenberger, H. S. Luftman, W. T. Tsang, and N.Ha, Appl. Phys. Lett.56, 1688~1990!.

118J. H. Chen and A. Y. Cho, J. Cryst. Growth111, 619 ~1991!.119T. D. McLean, T. M. Kerr, D. I. Westwood, C. E. C. Wood, and D.Howell, J. Vac. Sci. Technol. B4, 601 ~1986!.

120I. Poole, M. E. Lee, K. E. Singer, T. E. F. Frost, T. E. Kerr, C. E.Wood, D. A. Andrews, W. J. M. Rothwell, and C. J. Davies, J. ApPhys.63, 395 ~1988!.

121T. Kaneko, H. Asahi, Y. Itani, Y. Okuno, and S. Gonda, J. Cryst. Grow111, 638 ~1991!.

122K. Yamamoto, H. Asahi, T. Hayashi, K. Asami, and S. Gonda, J. CryGrowth164, 117 ~1996!.

123Y. Sato, K. Matsushita, T. Hariu, and Y. Shibata, Appl. Phys. Lett.44,592 ~1984!.

124M. E. Straumanis and C. D. Kim, J. Appl. Phys.36, 3822~1965!.125V. M. Glazov, S. N. Chizhevskaya, and S. B. Evgen’ev, Zh. Fiz. Khi43, 373 ~1969!.

126F. Donald Bloss,Crystallography and Crystal Chemistry: An Introduction ~Holt, Rinehart and Winston, New York, 1971!.

127K. Aoki, E. Anastassakis, and M. Cardona, Phys. Rev. B30, 681~1984!.128U. Piesbergen, inSemiconductors and Semimetals, edited by R. K. Wil-lardson and A. C. Beer~Academic, New York, 1966!, Chap. 3, Vol. 2.

129N. M. Kochetkora and T. N. Rezukhina, Chem. Abstracts56, 4167E~1962!.

130M. G. Holland inSemiconductors and Semimetals, edited by R. K. Wil-lardson and A. C. Beer~Academic, New York, 1966!, Chap. 1, Vol. 2.

131S. J. Fray, F. A. Johnson, and R. H. Jones, Proc. Phys. Soc. Londo76,939 ~1960!.

132W. Cochran, S. J. Fray, J. E. Quarrington, and N. Williams, J. ApPhys.32, 2102~1961!.

133S. K. Joshi and S. S. Mitra, Proc. Phys. Soc. London76, 295 ~1960!.134J. K. D. Verma, B. D. Nag, and P. S. Nair, Z. Naturforschung19a, 1561

~1964!.135E. F. Steigmeier, Appl. Phys. Lett.3, 6 ~1963!.136W. F. Boyle and R. J. Sladek, Phys. Rev. B11, 2933~1975!.137M. K. Farr, J. G. Taylor, and S. K. Sinha, Phys. Rev. B11, 1587~1975!.138P. B. Klein and R. K. Chang, Phys. Rev. B14, 2498~1976!.139S. I. Novikova and N. Kh. Abrikosov, Sov. Phys. Solid State5, 1558

~1963!.140L. Bernstein and R. J. Beals, J. Appl. Phys.32, 122 ~1961!.141S. I. Novikova, inSemiconductors and Semimetals, edited by R. K. Wil-lardson and A. C. Beer~Academic, New York, 1966!, Vol. 2, Chap. 2.

142M. G. Holland, Phys. Rev.134, A471 ~1964!.143M. G. Holland, Proc. Int. Conf. Phys. Semicond. Paris, 1964, p. 713144J. M. Ziman, Philos. Mag.1, 191 ~1956!.145R. W. Keyes, Phys. Rev.122, 1171~1961!.146A. Griffin and P. Carruthers, Phys. Rev.131, 1976~1963!.147E. F. Steigmeier and I. Kudman, Phys. Rev.141, 767 ~1966!.148E. F. Steigmeier and I. Kudman, Phys. Rev.132, 508 ~1963!.149J. R. Chelikowsky and M. L. Cohen, Phys. Rev. B14, 556 ~1976!.150T. C. Chiang and D. E. Eastman, Phys. Rev. B22, 2940~1980!.151D. E. Aspnes, C. G. Olson, and D. W. Lynch, Phys. Rev. B14, 4450

~1976!.152W. Ruhle, W. Jakowetz, C. Woelk, R. Linnebach, and M. Pilkuhn, PhStatus Solidi B73, 225 ~1973!.

153A. Joullie, A. Z. Eddine, and B. Girault, Phys. Rev. B23, 928 ~1981!.154O. Madelung,Physics of III–V Compounds~Wiley, New York, 1964!.155C. Alibert, A. Joullie, A. M. Joullie, and C. Ance, Phys. Rev. B27, 4946

~1983!.156C. D. Kourkoutas, P. D. Bekris, G. J. Papaioannou, and P. C. EuthymSolid State Commun.49, 1071~1984!.

157A. A. Kopylov, Solid State Commun.56, 1 ~1985!.



em

D

ate

v-

ys

i

ue

C

ona-

m

c.

es.

l.

and

th

g.

Sci.

6thy C.

er-

v-

v,

ett.

yst.

and

J.

ini,

lid

r.

s.

.

.

-

158M. Reine, R. L. Aggarwal, and B. Lax, Phys. Rev. B5, 3033~1972!.159G. Dresselhaus, A. F. Kip, and C. Kittel, Phys. Rev.98, 368 ~1955!.160P. C. Mathur and S. Jain, Phys. Rev. B19, 3152~1979!.161M. W. Heller and R. G. Hamerly, J. Appl. Phys.57, 4626~1985!.162J. L. Robert, B. Pistoulet, D. Barjon, and A. Raymond, J. Phys. ChSolids34, 2221~1973!.

163J. D. Wiley, in Semiconductors and Semimetals, edited by R. K. Wil-lardson and A. C. Beer~Academic, New York, 1975!, Vol. 10.

164R. A. Stradling, Phys. Lett.20, 217 ~1966!.165D. A. Hill and C. F. Schwerdtfeger, J. Phys. Chem. Solids35, 1533

~1974!.166H. J. Lee and J. C. Woolley, Can. J. Phys.59, 1844~1981!.167M. E. Lee, I. Poole, W. S. Truscott, I. R. Cleverley, K. E. Singer, andM. Rohlfing, J. Appl. Phys.68, 131 ~1990!.

168A. Sagar, Phys. Rev.117, 93 ~1960!.169A. J. Strauss, Phys. Rev.121, 1087~1961!.170T. O. Yep and W. M. Becker, J. Appl. Phys.37, 456 ~1966!.171R. D. Baxter and F. J. Reid, Bull. Am. Phys. Soc.10, 599 ~1965!.172D. Long and R. J. Hager, J. Appl. Phys.36, 3436~1965!.173D. Long, J. D. Zook, P. W. Chapman, and O. N. Tufte, Solid StCommun.2, 191 ~1964!.

174R. D. Baxter, F. J. Reid, and A. C. Beer, Phys. Rev.162, 718 ~1967!.175H. Brooks, inAdvances in Electronics and Electron Physics~Academic,New York, 1955!, Vol. 7, p. 156.

176A. I. Lebedev and I. A. Strel’nikova, Sov. Phys. Semicond.13, 229~1979!.

177E. J. Johnson, I. Filinski, and H. Y. Fan, inReport of the InternationalConference on the Physics of Semiconductors, Exeter, 1962~The Instituteof Physics and The Physical Society, London, 1962!, p. 375.

178K. Nakashima, Jpn. J. Appl. Phys.20, 1085~1981!.179M. A. Habegger and H. Y. Fan, Phys. Rev.138, 598 ~1965!.180I. K. Lazareva and V. M. Stuchebnikov, Sov. Phys. Semicond.4, 550

~1970!.181A. S. Kyuregyan, I. K. Lazareva, V. M. Stuchebnikov, and A. E. Yunoich, Sov. Phys. Semicond.6, 208 ~1972!.

182M. D’Olne Campos, A. Gouskov, L. Gouskov, and J. C. Pons, PhStatus Solidi A2, 779 ~1970!.

183T. O. Yep and W. M. Becker, Phys. Rev.144, 741 ~1966!.184D. Kranzer, Phys. Status Solidi A26, 11 ~1974!.185J. Kolodziejczak, S. Zukotynski, and H. Stramska, Phys. Status Solid14,471 ~1966!.

186P. S. Dutta, V. Prasad, H. L. Bhat, and Vikram Kumar, J. Appl. Phys.80,2847 ~1996!; see also P. S. Dutta, B. Mendez, J. Piqueras, E. Diegand H. L. Bhat,ibid. 80, 1112~1996!.

187G. R. Johnson, B. C. Cavenett, T. M. Kerr, P. B. Kirby, and C. E.Wood, Semicond. Sci. Technol.3, 1157~1988!.

188H. N. Leifer and W. C. Dunlap, Jr., Phys. Rev.95, 51 ~1954!.189D. E. Aspnes and A. A. Studna, Phys. Rev. B27, 985 ~1983!.190M. Cardona, inProceedings of the International Conference Semicductor Physics, Prague~Czech Acadamy of Science, Prague and Acdemic Press, New York 1961!, p. 388.

191H. P. R. Frederikse and R. F. Blunt, Proc. IRE43, 1828~1955!.192M. A. Habegger and H. Y. Fan, Bull. Am. Phys. Soc.8, 245 ~1963!.193F. Lukes, Czech. J. Phys.6, 359 ~1956!.194C. Hilsum and A. C. Rose-Innes,Semiconducting III–V Compounds

~Macmillan, Pergamon, New York, 1961!.195G. W. Gobeli and F. G. Allen, inSemiconductors and Semimetals, editedby R. K. Willardson and A. C. Beer~Academic, New York, 1966!, Vol.2, Chap. 11.

196N. Bloembergen, R. K. Chang, J. Ducuing, and P. Lallemand, Bull. APhys. Soc.9, 269 ~1964!.

197E. Burstein, Phys. Rev.93, 632~1954!; and T. S. Moss, Proc. Phys. SoLondon, Sec. B76, 775 ~1954!.

198A. R. Calawa, J. Appl. Phys.34, 1660~1963!.199H. J. Hrostowski and C. S. Fuller, J. Phys. Chem. Solids4, 155 ~1958!.200F. Stern, inSemiconductors and Semimetals, edited by R. K. Willardsonand A. C. Beer~Academic, New York, 1966!, Vol. 2, Chap. 14.

201I. Hirai and T. Obokata, Jpn. J. Appl. Phys.21, 956 ~1982!.202Y. Ohmori, K. Sugii, S. Akai, and K. Matsumoto, J. Cryst. Growth60, 79

~1982!.203A. Katsui and C. Uemura, Jpn. J. Appl. Phys.21, 1106~1982!.204J. Doerschel and U. Geissler, J. Cryst. Growth121, 781 ~1992!.



.

.

.

z,

.

-

.

205A. Van Mau, M. Averous, and G. Bougnot, Mater. Res. Bull.7, 857~1972!.

206M. I. Aliev, G. I. Safaraliev, and S. G. Abdinova, Phys. Status Solidi22,741 ~1967!.

207F. Moravec, V. Sestakova, B. Stepanek, and V. Charvat, Cryst. RTechnol.24, 275 ~1989!.

208Y. J. Van Der Muelen, J. Phys. Chem. Solids28, 25 ~1967!.209M. Ichimura, K. Higuchi, Y. Hattori, T. Wada, and N. Kitamura, J. AppPhys.68, 6153~1990!.

210W. Ruhle and D. Bimberg, Phys. Rev. B12, 2382~1975!.211M. Lee, D. J. Nicholas, K. E. Singer, and B. Hamilton, J. Appl. Phys.59,2895 ~1986!.

212P. S. Dutta, K. S. R. Koteswara Rao, K. S. Sangunni, H. L. Bhat,Vikram Kumar, Appl. Phys. A61, 149 ~1995!.

213D. J. Nicholas, M. Lee, B. Hamilton, and K. E. Singer, J. Cryst. Grow81, 298 ~1987!.

214P. S. Dutta, H. L. Bhat, K. S. Sangunni, and Vikram Kumar, inSeventhInternational Workshop on Physics of Semiconductor Devices, NewDelhi, India ~Narosa, New Delhi, 1993!, p. 459.

215M. S. Mirgalovskaya, G. V. Kukuladze, and V. A. Kokoshkin, NeorMater.4, 694 ~1968!.

216H. Xu, Phys. Rev. B46, 12 251~1992!.217V. Sestakova and B. Stepanek, J. Cryst. Growth146, 87 ~1995!.218R. N. Hall and J. H. Racette, J. Appl. Phys.32, 856 ~1961!.219V. Sestakova, B. Stepanek, J. Sestak, P. Hubik, and V. Smid, Mater.Eng. B21, 14 ~1993!.

220I. I. Burdiyan, Sov. Phys. Semicond.7, 449 ~1973!.221W. Jakowetz, D. Barthruff, and K. W. Benz, in Proceedings of theInternational Symposium on GaAs and related compounds, edited bHilsum ~London Inst. Phys., London, 1976!, p. 41.

222W. Ruhle, W. Jakowetz, and M. H. Pilkuhn, in Proceedings of the Intnational Conference on Luminescence, Leningrad, 1972, p. 444.

223A. S. Kyuregyan, I. K. Lazareva, V. M. Stuchebnikov, and A. E. Yunoich, Sov. Phys. Semicond.6, 208 ~1972!.

224A. Ya Vul’, L. V. Golubev, T. A. Polyanskaya, and Yu. V. ShmartseFiz. Tekh. Poluprovodn.3, 786 ~1969!.

225B. B. Kosicui, W. Paul, A. J. Strauss, and G. W. Iseler, Phys. Rev. L17, 1175~1966!.

226G. W. Iseler and A. J. Strauss, Bull. Am. Phys. Soc.12, 404 ~1967!.227J. F. Chen and A. Y. Cho, J. Appl. Phys.70, 277 ~1991!.228A. Ya. Vul, G. L. Bir, and Y. V. Shmartsev, Sov. Phys. Semicond.4,2005 ~1971!.

229B. B. Kosicki and W. Paul, Phys. Rev. Lett.17, 246 ~1966!.230K. Hoo, W. M. Becker, and R. Y. Sun, Solid State Commun.18, 313

~1976!.231A. K. Chin and W. A. Bonner, Appl. Phys. Lett.40, 248 ~1982!.232B. Stepanek, V. Sestakova, P. Hubik, V. Smid, and V. Charvat, J. CrGrowth126, 617 ~1993!.

233B. Stepanek and V. Sestakova, J. Cryst. Growth123, 306 ~1992!.234P. S. Dutta, K. S. R. Koteswara Rao, K. S. Sangunni, H. L. Bhat,Vikram Kumar, Appl. Phys. Lett.65, 1412~1994!.

235M. Mizutsa, M. Tachikawa, H. Kukimoto, and S. Minomura, Jpn.Appl. Phys. 124, L143 ~1985!.

236P. M. Mooney, N. S. Caswel, and S. L. Wright, Jpn. J. Appl. Phys.62,4786 ~1987!.

237P. Hubik, J. J. Mares, J. Kristofik, A. Baraldi, C. Ghezzi, and A. ParisSemicond. Sci. Technol.10, 455 ~1995!.

238P. Hubik, V. Smid, J. Kristofik, B. Stepanek, and V. Sestakova, SoState Commun.86, 19 ~1993!.

239A. Y. Polyakov, M. Stam, A. G. Milnes, and T. E. Schlesinger, MateSci. Eng. B12, 337 ~1992!.

240E. Kuramochi, N. Kondo, Y. Takanashi, and M. Fujimoto, Appl. PhyLett. 63, 2664~1993!.

241E. I. Georgitse, L. M. Gutsulyak, V. I. Ivanov-Omskii, V. F. Masterov, VA. Smirnov, and K. F. Shtel’makh, Sov. Phys. Semicond.26, 50 ~1992!.

242E. I. Georgitse, L. M. Gutsulyak, V. I. Ivanov-Omskii, V. F. Masterov, VA. Smirnov, and Sh. U. Yuldashev, Sov. Phys. Semicond.25, 959~1991!; 25, 1180~1991!; 25, 1219~1991!.

243E. I. Georgitse, L. M. Gutsulyak, V. I. Ivanov-Omskii, V. M. Pogorletskii, and A. N. Titkov, Sov. Tech. Phys. Lett.17, 614 ~1991!.

244R. L. Mieher, inSemiconductors and Semimetals, edited by R. K. Wil-lardson and A. C. Beer~Academic, New York, 1966!, Vol. 2, Chap. 7.

245D. J. Oliver, J. Phys. Chem. Solids11, 257 ~1959!.246F. H. Eisen and C. E. Birchenall, Acta Metall. 5, 265~1957!.



i

ys

, J

ne

. JJ.

row

ev

.

ac

ys

E

dt

al

E

pp

Sc

.at

,

de

e

J.

ac.

. E.

hys.

C.

tate

. 1

ett.

ry,

d

.

nd

au,

.

s.

i, J.

. T.pl.

G.

D.

247B. I. Boltaks and Yu. A. Gutorov, Sov. Phys. Solid State1, 930 ~1960!.248V. A. Uskov, Sov. Phys. Semicond.8, 1573~1975!.249D. Mathiot and G. Edelin, Philos. Mag. A41, 447 ~1980!.250J. Bougnot, L. Szepessy, and S. F. Du Cunha, Phys. Status Solid26,K127 ~1968!.

251S. F. Du Cunha and J. Bougnot, Phys. Status Solidi A22, 205 ~1974!.252R. Kaiser and H. Y. Fan, Phys. Rev.138, A156 ~1965!.253P. Barman, U. N. Roy, and S. Basu, Mater. Lett.10, 203 ~1990!.254P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, Appl. PhLett. 65, 1695~1994!.

255P. S. Dutta, K. S. R. Koteswara Rao, H. L. Bhat, and Vikram KumarAppl. Phys.77, 4825~1995!.

256M. Perotin, P. Coudray, L. Gouskov, H. Luquet, C. Llinares, J. J. BonL. Soonckindt, and B. Lambert, J. Electron. Mater.23, 7 ~1994!.

257A. Y. Polyakov, S. J. Pearton, R. G. Wilson, P. Rai-Choudhury, RHillard, X. J. Bao, M. Stam, A. G. Milnes, T. E. Schlesinger, andLopata, Appl. Phys. Lett.60, 1318~1992!.

258V. Sestakova, B. Stepanek, J. J. Mares, and J. Sestak, J. Cryst. G140, 426 ~1994!.

259P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, Phys. RB 51, 2153~1995!.

260A. Y. Polyakov, A. G. Milnes, Xiaolei Li, A. A. Balmashnov, and N. BSmirkov, Solid State Electron.38, 1743~1995!.

261Z. Lu, Y. Jiang, W. I. Wang, M. C. Teich, and R. M. Osgood, Jr., J. VSci. Technol. B10, 1856~1992!.

262P. S. Dutta, K. S. Sangunni, H. L. Bhat, and Vikram Kumar, Appl. PhLett. 66, 1986~1995!.

263P. S. Dutta, A. K. Sreedhar, H. L. Bhat, G. C. Dubey, Vikram Kumar,Dieguez, U. Pal, and J. Piqueras, J. Appl. Phys.79, 3246~1996!.

264B. A. Parkinson, A. Heller, and B. Miller, J. Electrochem. Soc.126, 954~1979!.

265K. M. Schirm, P. Soukiassian, P. S. Mangat, Z. Hurych, L. Soonckinand J. J. Bonnet, J. Vac. Sci. Technol. B10, 1867~1992!.

266J. G. Buglass, T. D. Mclean, and D. G. Parker, J. Electrochem. Soc.133,2565 ~1986!.

267P. S. Dutta, H. L. Bhat, and Vikram Kumar, in Proceedings of MRS FMeeting, Boston, MA, 1995.

268P. S. Dutta, H. L. Bhat, and Vikram Kumar, inProceedings of the EighthInternational Workshop on Physics of Semiconductor Devices, NewDelhi, India ~Narosa, New Delhi, 1995!, p. 467.

269U. Pal, J. Piqueras, P. S. Dutta, H. L. Bhat, Vikram Kumar, andDieguez, in Proceedings of MRS Fall Meeting, Boston, MA, 1995.

270H. C. Gatos, M. C. Lavine, and E. P. Warekois, J. Electrochem. Soc.108,645 ~1961!.

271J. W. Faust, Jr., inCompound Semiconductors, Preparation of III–VCompounds, edited by R. K. Willardson and H. L. Goering~Reinhold,Chapman and Hall, London, 1962!, Vol. 1, Chap. 50.

272J. W. Faust, Jr. and A. Sagar, J. Appl. Phys.31, 331 ~1960!.273R. W. Haisty, J. Electrochem. Soc.108, 790 ~1961!.274L. J. Gomez Zazo, M. T. Montojo, J. L. Castano, and J. Piqueras, J. APhys.136, 1480~1989!.

275M. Kodama, J. Hasegawa, and M. Kimata, J. Electrochem. Soc.132, 659~1985!.

276F. W. O. DeSilva, M. Silga, C. Raisin, and L. Lassabatere, J. Vac.Technol. B8, 75 ~1990!.

277B. Mendez, P. S. Dutta, J. Piqueras, and E. Dieguez, Appl. Phys. Lett67,2648~1995!; see also P. S. Dutta, C. Marin, E. Dieguez, and H. L. BhJ. Cryst. Growth160, 207 ~1996!.

278S. J. Pearton, U. K. Chakrabarti, W. S. Hobson, and A. P. KinsellaVac. Sci. Technol. B8, 607 ~1990!.

279D. Haneman, inCompound Semiconductors, Preparation of III–V Com-pounds, edited by R. K. Willardson and H. L. Goering~Reinhold, Chap-man and Hall, London, 1962!, Vol. 1, Chap. 47.

280A. G. Milnes, M. Ye, and M. Stam, Solid State Electron.37, 37 ~1994!.281J. B. B. Oliveira, C. A. Olivieri, J. C. Galzerani, A. A. Pasa, and F. C.Prince, J. Appl. Phys.66, 5484~1989!.

282C. H. Heinz and W. Schmidt, Electron. Lett.18, 859 ~1982!.283B. Tadayon, C. S. Kyono, M. Fatemi, S. Tadayon, and J. A. MitteredJ. Vac. Sci. Technol. B13, 1 ~1995!.

284K. I. Anastasiou, IEEE Trans. Electron DevicesED-40, 878 ~1993!.285C. H. Heinz, Int. J. Electron.54, 254 ~1983!.286N. Koshizaki, M. Kudo, M. Owari, Y. Nihei, and H. Kamada, Jpn.Appl. Phys.19, L349 ~1980!.



.

.

t,

.

th

.

.

.

.

,

l

.

l.

i.

,

J.

r,

287J. R. Lince and R. S. Williams, Thin Solid Films137, 251 ~1986!.288R. Veresegyhazy, I. Mojzes, and E. Lendvay, Thin Solid Films138, L55

~1986!.289F. A. Padovani and R. Stratton, Solid State Electron.9, 695 ~1966!.290G. W. Gobelli and F. G. Allen, Phys. Rev. A137, 245 ~1965!.291D. Mao, A. Kahn, and L. Soonckindt, Phys. Rev. B40, 5579~1989!.292C. A. Mead, Solid State Electron.9, 1023~1966!.293W. E. Spicer, P. W. Chye, P. R. Skeath, C. Y. Su, and I. Lindau, J. VSci. Technol.16, 1422~1979!.

294P. W. Chye, I. Landau, P. Pianetta, C. M. Garner, C. Y. Su, and WSpicer, Phys. Rev. B18, 5545~1978!.

295P. W. Chye, I. Landau, P. Pianetta, C. M. Garner, and W. E. Spicer, PRev. B17, 2682~1978!.

296C. A. Mead and W. G. Spitzer, Phys. Rev. Lett.22, 1433~1980!.297R. Chin, R. A. Milano, and H. D. Law, Electron. Lett.16, 626 ~1980!.298Y. Nagao, T. Hariu, and Y. Shibata, IEEE Trans. Electron DevicesED-28, 407 ~1981!.

299T. D. Mclean, T. M. Kerr, D. I. Westwood, S. R. Blight, D. Page, andE. C. Wood, inGallium Arsenide and Related Compounds 1985, Inst.Phys. Conf. Series Vol. 79, edited by M. Fujimoto~Institute of Physics,Bristol, 1985!, p. 349.

300I. Poole, M. E. Lee, and K. E. Singer, Semicond. Sci. Technol.6, 881~1991!.

301S. A. Walters and R. H. Williams, J. Vac. Sci. Technol. B6, 1421~1988!.302Y. K. Su, N. Y. Li, F. S. Juang, and S. C. Wu, J. Appl. Phys.68, 646

~1990!.303I. Poole, M. E. Lee, M. Missous, and K. E. Singer, J. Appl. Phys.62,3988 ~1987!.

304Y. K. Su, N. Y. Li, and F. S. Juang, Solid State Electron.34, 426~1991!.305P. K. Vasudev, B. L. Mattes, E. Pietras, and R. H. Bube, Solid SElectron.19, 557 ~1976!.

306S. A. Walters and R. H. Williams, Mater. Sci. Eng. B9, 51 ~1991!.307S. S. Tan and A. G. Milnes, Mater. Sci. Eng. B21, 94 ~1993!.308P. S. Dutta, H. L. Bhat, K. S. Sangunni, and Vikram Kumar, inSeventhInternational Workshop on Physics of Semiconductor Devices, NewDelhi ~Narosa, New Delhi, 1993!, p. 564.

309M. Nishida, Surf. Sci.99, L384 ~1980!.310C. W. Wilmsen, J. Vac. Sci. Technol.19, 279 ~1981!.311S. W. Teare and C. W. Fischer, J. Appl. Phys.68, 2983~1990!.312N. Kitamura, T. Kikuchi, M. Kakehi, and T. Wada, Jpn. J. Appl. Phys23, 1534~1984!.

313S. Basu, N. Basu, and P. Barman, Mater. Sci. Eng. B9, 47 ~1991!.314F. Capasso, M. B. Panish, S. Sumski, and P. W. Foy, Appl. Phys. L36, 165 ~1980!.

315C. Heinz, Electron. Lett.22, 276 ~1986!.316A. Y. Polyakov, M. Stam, R. G. Wilson, Z. Q. Fang, P. Rai-ChoudhuR. J. Hillard, and A. G. Milnes, J. Appl. Phys.72, 1316~1992!.

317A. N. Baranov, T. I. Voronina, N. S. Zimogorova, L. M. Kanskaya, anY. P. Yakovlev, Sov. Phys. Semicond.19, 1030~1985!.

318A. Krier, M. K. Parry, and D. S. Lanchester, Semicond. Sci. Technol6,1066 ~1991!.

319S. J. Pearton, A. R. Von Neida, J. M. Brown, K. T. Short, L. J. Oster, aU. K. Chakrabarti, J. Appl. Phys.64, 629 ~1988!.

320K. S. Jones and S. J. Santana, J. Mater. Res.6, 1048~1991!.321R. Callec, P. N. Farannec, M. Salvi, H. L’Haridon, and M. GauneAppl. Phys. Lett.59, 1872~1991!.

322G. Panin, P. S. Dutta, J. Piqueras, and E. Dieguez, Appl. Phys. Lett67,3584 ~1995!.

323A. N. Titkov, G. N. Iluridze, I. F. Mironov, and V. A. Cheban, Sov. PhySemicond.20, 14 ~1986!.

324D. Yevick and W. Bardyszewski, Appl. Phys. Lett.51, 124 ~1987!.325H. J. Gossmann, G. P. Schwartz, B. A. Davidson, and G. J. GualtierVac. Sci. Technol. B7, 764 ~1989!.

326H. J. Gossmann, B. A. Davidson, G. J. Gualtieri, G. P. Schwartz, AMacrander, S. E. Sluzky, M. H. Grabow, and W. A. Sunder, J. ApPhys.66, 1687~1989!.

327Y. Takeda, S. Noda, and A. Sasaki, J. Appl. Phys.57, 1261~1985!.328S. C. Chen, Y. K. Su, and F. S. Juang, J. Cryst. Growth92, 118 ~1988!.329H. Luquet, L. Gouskov, M. Perotin, A. Jean, A. Rjeb, T. Zarouri, andBougnot, J. Appl. Phys.60, 3582~1986!.

330J. L. Lazarri, J. L. Leclerc, P. Grunberg, A. Joullie, B. Lambert,Barbusse, and R. Fourcade, J. Cryst. Growth123, 465 ~1992!.



er

i-

J.s

. J

er,

an

te

ar,

A.

us

. A

.

hi,

S.

-

O

.

pl.

A.

to

n,

m

e

th

on

ron

.

.

ys.

ice

hu,

.

.TO-

ce

pl.

o,

tt.

tt.

tt.

e-

ki,

.

B.

331J. L. Lazzari, E. Tournie, F. Pitard, A. Joullie, and B. Lambert, MatSci. Eng. B9, 125 ~1991!.

332A. Joullie, F. J. Hua, F. Karouta, and H. Mani, J. Cryst. Growth75, 309~1986!.

333P. S. Dutta, H. L. Bhat, Vikram Kumar, N. V. Sochinskii, and E. Deguez, J. Cryst. Growth160, 177 ~1996!.

334K. F. Longenbach, S. Xin, and W. I. Wang, J. Appl. Phys.69, 3393~1991!.

335S. K. Haywood, A. B. Henriques, D. F. Howell, N. J. Mason, R.Nicholas, and P. J. Walker, in14th International Symposium on GaAand Related Compounds~Inst. of Physics, Bristol, 1989!, Vol. 91, p. 271.

336S. K. Haywood, R. W. Martin, N. J. Mason, R. J. Nicholas, and PWalker, J. Cryst. Growth107, 422 ~1991!.

337M. Lakrimi, R. W. Martin, N. J. Mason, R. J. Nicholas, and P. J. WalkJ. Cryst. Growth110, 677 ~1991!.

338A. Giani, J. Bougnot, F. Pascal-Delannoy, G. Bougnot, J. Kaoukab,G. G. Allogho, M. Bow, Mater. Sci. Eng. B9, 101, 121~1991!.

339Y. K. Su, F. S. Juang, N. Y. Li, K. J. Gan, and T. S. Wu, Solid StaElectron.34, 815 ~1991!.

340S. Adachi, J. Appl. Phys.61, 4869~1987!.341P. S. Dutta, A. K. Sreedhar, H. L. Bhat, G. C. Dubey, Vikram Kumand E. Dieguez, Appl. Phys. Lett.67, 1001~1995!.

342Y. Horikoshi, inSemiconductors and Semimetals~Academic, New York,1985!, Vol. 22C, p. 93.

343J. P. van der Ziel, T. H. Chiu, and W. T. Tsang, J. Appl. Phys.60, 4087~1986!.

344C. Caneau, A. K. Srivastava, A. G. Dentai, J. L. Zyskind, and M.Pollack, Electron. Lett.21, 815 ~1985!.

345C. Caneau, A. K. Srivastava, A. G. Dentai, J. L. Zyskind, C. A. Burrand M. A. Pollack, Electron. Lett.22, 992 ~1986!.

346C. Caneau, J. L. Zyskind, J. W. Sulhoff, T. E. Glover, J. Centanni, CBurrus, A. G. Dentai, and M. A. Pollack, Appl. Phys. Lett.51, 704~1987!.

347T. H. Chiu, J. L. Zyskind, and W. T. Tsang, J. Electron. Mater.16, 57~1987!.

348A. E. Bochkarev, L. M. Dolginov, A. E. Drakin, R. G. Eliseev, and B. NSverdlov, Sov. J. Quantum Electron.18, 1362~1988!.

349S. J. Eglash, H. K. Choi, and G. W. Turner, J. Cryst. Growth111, 669~1991!.

350S. J. Eglash and H. K. Choi, Appl. Phys. Lett.57, 1292~1990!.351N. Toba and K. Nosu, Electron. Lett.23, 188 ~1986!.352A. N. Baranov, Appl. Phys. Lett.59, 2360~1991!.353H. Kuwatsuka, T. Mikawa, S. Miura, N. Yasuoka, M. Ito, T. Tanahasand O. Wada, J. Appl. Phys.70, 2169~1991!.

354D. Kasemset, IEEE J. Quantum Electron.QE-17, 1595~1981!.355L. Gouskov, B. Orsal, M. Perotin, M. Karim, A. Sabir, P. Coudray,Kibeya, and H. Luquet, Appl. Phys. Lett.60, 3030~1992!.

356K. Brennan, K. Hess, and Y. C. Chang, J. Appl. Phys.57, 1971~1985!.357H. Luquet, M. Perotin, L. Gouskov, C. Llinares, H. Archidim, M. Lahbadi, M. Karim, and B. Mbow, J. Appl. Phys.68, 3861~1990!.

358S. Miura, T. Mikawa, H. Kuwatsuka, N. Yasuoka, T. Tanahashi, andWada, Appl. Phys. Lett.54, 2422~1989!.

359Y. Jiang, M. C. Teich, and W. I. Wang, J. Appl. Phys.67, 2488~1990!.360S. M. Bedair, J. Electrochem. Soc.122, 1150~1975!.361I. A. Andreev, A. N. Baranov, M. Z. Zhingarev, V. I. Korol’kov, M. PMikhailova, and Y. P. Yakovlev, Sov. Phys. Semicond.19, 987 ~1985!;Appl. Phys. Lett.60, 3030~1992!.

362M. Mebarki, T. Belatoui, A. Joullie, B. Orsal, and R. Alabedra, J. ApPhys.68, 4106~1990!.

363J. E. Bauers, A. K. Srivastava, C. A. Burras, J. C. Dewinter, M.Pollack, and J. L. Zyskind, Electron. Lett.22, 137 ~1986!.

364A. M. Fox, A. C. Machiel, J. F. Ryan, and T. Kerr, Appl. Phys. Lett.51,430 ~1987!.

365See the NETSCAPE homepage of the Naval Research LaboraU.S.A., by Bennet and co-workers on GaSb/GaAs epitaxy.

366J. D. Kim, S. Kim, D. Wu, J. Wojkowski, J. Xu, J. Piotrowski, E. Bigaand M. Razeghi, Appl. Phys. Lett.67, 2645~1995!.

367Y. Z. Liu, I. Deyhimy, J. S. Harris, Jr., R. J. Anderson, J. Appelbauand J. M. Pollard, Appl. Phys. Lett.36, 458 ~1980!.

368T. H. Chiu and A. F. J. Levi, Appl. Phys. Lett.55, 1891~1989!.369K. Taira, K. Funato, F. Nakamura, and H. Kawai, J. Appl. Phys.69, 4454

~1991!.370K. Taira, F. Nakamura, K. Funato, and H. Kawai, Inst. Phys. Conf. S112, 489 ~1990!.



.

.

d

,

.

.

ry,

,

r.

371K. Taira, F. Nakamura, I. Hase, H. Kawai, and Y. Mori, J. Cryst. Grow107, 952 ~1991!.

372A. Furukawa and M. Mizuta, Electron. Lett.24, 1378~1988!.373L. F. Luo, K. F. Longenbach, and W. I. Wang, in International ElectrDevices Meeting Abstracts, IEEE CH2865, 1990, p. 515.

374K. F. Longenbach, R. Beresford, and W. I. Wang, IEEE Trans. ElectDevicesED-37, 2265~1990!.

375Y. Kanji, D. Kiyomi, H. Taniguchi, M. Yaho, and M. Ihoue, Inst. PhysConf. Ser.120, 173 ~1991!.

376P. S. Kop’ev, S. V. Ivanov, N. N. ledenstov, B. Y. Mel’tser, M. YNadtochil, and V. M. Ustinov, Sov. Phys. Semicond.24, 450 ~1990!.

377J. Werking, G. Tuttle, C. Nguyen, E. L. Hu, and H. Kroemer, Appl. PhLett. 57, 905 ~1990!.

378C. R. Bolognesi, E. J. Caine, and H. Kroemer, IEEE Electron DevLett. EDL-15, 16 ~1994!.

379S. Tehrani, J. Shen, H. Goronkin, G. Kramer, R. Tsui, and T. X. ZIEEE Electron Device Lett.EDL-16, 557 ~1995!.

380L. L. Chang and L. Esaki, Surf. Sci.98, 70 ~1980!.381D. Z. Y. Ting, E. T. Yu, D. A. Collins, D. H. Chow, and T. C. McGill, JVac. Sci. Technol. B8, 810 ~1990!.

382E. T. Yu, D. A. Collins, D. Z. Y. Ting, D. H. Chow, and T. C. McGill,Appl. Phys. Lett.57, 2675~1990!.

383D. A. Collins, D. H. Chow, E. T. Yu, D. Z. Y. Ting, J. R. Soderstrom, YRajakarunannayake, T. C. McGill, and T. J. Watson, presented at NAARW Workshop, El Escorial, Spain, 14–18 May, 1990.

384J. R. Soderstrom, D. H. Chow, and T. C. McGill, IEEE Electron DeviLett. EDL-11, 27 ~1990!.

385J. F. Chen, M. C. Wu, L. Yang, and A. Y. Cho, J. Appl. Phys.68, 3040~1990!.

386J. F. Chen, M. C. Wu, L. Yang, A. Y. Cho, and T. C. McGill, IEEEElectron Device Lett.EDL-11, 532 ~1990!.

387J. F. Chen, L. Yang, M. C. Wu, S. N. G. Chu, and A. Y. Cho, J. ApPhys.68, 3451~1990!.

388L. Yang, M. C. Wu, J. F. Chen, Y. K. Chen, G. L. Snider, and A. Y. ChJ. Appl. Phys.68, 4286~1990!.

389L. Yang, J. F. Chen, and A. Y. Cho, J. Appl. Phys.68, 2997~1990!.390G. Tuttle, H. Kroemer, and J. H. Eglash, J. Appl. Phys.65, 5239~1989!.391L. F. Luo, R. Beresford, and W. I. Wang, Appl. Phys. Lett.55, 2023

~1989!.392L. F. Luo, R. Beresford, and K. F. Longenbach, J. Appl. Phys.68, 2854

~1990!.393K. F. Longenbach, L. F. Luo, and W. I. Wang, Appl. Phys. Lett.57, 1554

~1990!.394M. Iwata, Y. Nakahara, and I. Hirosawa, Inst. Phys. Conf. Ser.106, 459

~1989!.395G. P. Schwartz, G. J. Gualtieri, and W. A. Sunder, J. Cryst. Growth102,147 ~1990!.

396Y. Jiang, M. C. Teich, and W. I. Wang, Appl. Phys. Lett.57, 2922~1990!.

397E. C. Carr, T. H. Wood, C. A. Burrus, and T. H. Chiu, Appl. Phys. Le53, 2305~1988!.

398H. Xie, W. I. Wang, and J. R. Meyer, J. Appl. Phys.76, 92 ~1994!.399H. Xie, W. I. Wang, J. R. Meyer, and L. R. Rammohan, Appl. Phys. Le65, 2048~1994!.

400B. Brar, H. Kroemer, and J. Eglish, J. Cryst. Growth127, 752 ~1993!.401X. C. Zhang, B. B. Hu, S. H. Xin, and D. H. Auston, Appl. Phys. Le57, 753 ~1990!.

402B. Lambert, Y. Toudic, Y. Rouillard, M. Baudet, B. Guenais, B. Dveaud, I. Valiente, and J. C. Simon, Mater. Sci. Eng. B21, 185 ~1993!.

403N. J. Kawai, L. L. Chang, G. A. Sai-Halasz, C. A. Chang, and L. EsaAppl. Phys. Lett.36, 369 ~1980!.

404H. Munekata, E. E. Mendez, Y. Iye, and L. Esaki, Surf. Sci.174, 449~1986!.

405T. D. Golding, J. A. Dura, W. C. Wang, J. T. Zborowski, A. Vigliante, HC. Chen, and J. R. Meyer, J. Cryst. Growth127, 777 ~1993!.

406B. R. Bennett, R. Magno, and B. V. Shanabrook, Appl. Phys. Lett.68,505 ~1996!.

407B. R. Bennett, B. V. Shanabrook, and R. Magno, Appl. Phys.68, 958~1996!.

408B. R. Bennett, P. M. Thibado, M. E. Twigg, E. R. Glaser, R. Magno,V. Shanabrook, and L. J. Whitman, J. Vac. Sci. Technol. B14, 2195~1996!.



hs

J

n

.

R.

lec-d D..

409F. Hatami, N. N. Ledentsov, M. Grundmann, J. Bohrer, F. Heinric

dorff, M. Beer, D. Bimberg, S. S. Ruvimov, P. Werner, U. Gosele,

Heydenreich, U. Richter, S. V. Ivanov, B. Ya. Meltser, P. S. Kopev, a

Zh. I. Alferov, Appl. Phys. Lett.67, 656 ~1995!.410A. G. Ostrogorsky and G. Muller, J. Cryst. Growth137, 64 ~1994!.411A. G. Milvidskaya, A. Y. Polyakov, G. P. Kolchina, A. G. Milnes, A. V



-

.

d

Govorkov, N. B. Smirnov, and I. V. Tunitskaya, Mater. Sci. Eng. B22,279 ~1994!.

412A. Y. Polyakov, S. J. Eglash, A. G. Milnes, M. Ye, S. J. Pearton, andG. Wilson, J. Cryst. Growth127, 728 ~1993!.

413Proceedings of the 2nd NREL Conference on TPV Generation of Etricity, Colorado Springs, CO, edited by J. P. Benner, T. J. Coutts, anS. Ginley~AIP, New York, 1996!, pp. 3, 128, 134, 251, 339, 423, 488



applied physics reviews the physics and technology of

Documents