JOURNAL OF ELECTRONIC MATERIALS
	ABSTRACTS Volume 25, Number 9, September 1996 This Month Featuring: Regular Issue Papers. View September 1996 Table of Contents.

REGULAR ISSUE PAPERS

KEY WORDS Defects, silicon efficiency, solar cell

The precipitation rate of intentionally introduced iron during low-temperature heating is studied among a variety of single-crystal and polycrystalline silicon solar cell materials. A correlation exists between the iron precipitation rate and the carrier recombination rate in dislocation-free as-grown material, suggesting that diffusion-length-limiting defects in as-grown material are structural defects which accelerate iron precipitation. Phosphorous diffusion gettering was found to be particularly ineffective at improving diffusion length after intentional iron contamination in materials with high iron precipitation rates. We propose that intragranular structural defects in solar cell silicon greatly enhance transition metal precipitation during cooling from the melt and become highly recombination-active when decorated with these impurities. The defects then greatly impair diffusion length improvement during phosphorus gettering and limit carrier lifetimes in as-grown material.

The Structural, Optical, and Electrical Properties of Vacuum Evaporated Cu-Doped ZnTe Polycrystalline Thin Films
L. FENG,¹ D. MAO,² J. TANG,² R.T. COLLINS,² and J.U. TREFNY²
1--Department of Physics, Colorado School of Mines, Golden, CO 80401. Permanent Address: Department of Materials Science, Sichuan University, Chengdu, China. 2--Department of Physics, Colorado School of Mines, Golden, CO 80401.

KEY WORDS II-VI semiconductor, polycrystalline thin films, ZnTe

We have studied the structural, optical, and electrical properties of thermally evaporated, Cu-doped, ZnTe thin films as a function of Cu concentration and post-deposition annealing temperature. X-ray diffraction measurements showed that the ZnTe films evaporated on room temperature substrates were characterized by an average grain size of 300Å with a (111) preferred orientation. Optical absorption measurements yielded a bandgap of 2.21 eV for undoped ZnTe. A bandgap shrinkage was observed for the Cu-doped films. The dark resistivity of the as-deposited ZnTe decreased by more than three orders of magnitude as the Cu concentration was increased from 4 to 8 at.% and decreased to less than 1 ohm-cm after annealing at 260°C. For films doped with 6-7 at.% Cu, an increase of resistivity was also observed during annealing at 150-200°C. The activation energy of the dark conductivity was measured as a function of Cu concentration and annealing temperature. Hall measurements yielded hole mobility values in the range between 0.1 and 1 cm²/V.s for both as-deposited and annealed films. Solar cells with a CdS/CdTe/ZnTe/metal structure were fabricated using Cu-doped ZnTe as a back contact layer on electrodeposited CdTe. Fill factors approaching 0.75 and energy conversion efficiencies as high as 12.1% were obtained.

Dry Etching of InGaAlP Alloys in Cl₂/Ar High Ion Density Plasmas
J. HONG,¹ J.W. LEE,¹ E.S. LAMBERS,¹ C.R. ABERNATHY,¹ C.J. SANTANA,¹ S.J. PEARTON,¹ W.S. HOBSON,² and F. REN²
1--University of Florida, Gainesville, FL 32611. 2--AT&T Bell Laboratories, Murray Hill, NJ 07974.

KEY WORDS AlInP, dry etching, InGaP, plasma chemistry

Etch rates above 1 µm min^-1 are achieved for InGaP and AlInP under electron cyclotron resonance conditions in low pressure (1.5 mTorr) Cl₂/Ar discharges. Much lower rates were obtained for AlGaP due to the greater difficulty of the bond breaking that must precede formation and desorption of the etch products. The etched surface morphology and stoichiometry are strong functions of the plasma composition (i.e., the ion/neutral flux ratio). Under optimal conditions, there are no detectable chlorine related residues, in sharp contrast to reactive ion etching with this plasma chemistry.

Low-Pressure Pyrolysis Studies of a New Phosphorus Precursor: Tertiarybutylbis(dimethylamino)phosphine
C.W. HILL,¹ R.W. GEDRIDGE, JR.,² T.J. GROSHENS,² G.B. STRINGFELLOW,¹ and L.P. SADWICK¹
1--Departments of Materials Science and Engineering and Electrical Engineering, University of Utah, Salt Lake City, UT 84112. 2--Chemistry and Materials Branch (Code 474220D), Research and Technology Division, Naval Air Warfare Center Weapons Division, China Lake, CA 93555.

KEY WORDS Phosphorous precursor, pyrolysis studies, tertiartybutylbis(dimethylamino) phosphine

The low pressure decomposition of tertiartybutylbis(dimethylamino) phosphine, (t-Bu)P(NMe₂)₂, (TBBDMAP), has been studied on quartz and deposited GaP and InP surfaces. This new phosphorus precursor has been found to pyrolyze on quartz surfaces at much lower temperatures than the related compounds tertiarybutylphosphine, (t-Bu)PH₂, (TBP) and tris(dimethylamino)phosphorus, P(NMe₂)₃, (TDMAP). In contrast to the results obtained for TDMAP, GaP and InP surfaces decrease the decomposition temperature of TBBDMAP only slightly. The TBBDMAP reaction products were dimethylamine, methylmethyleneimine, and isobutylene, consistent with previous pyrolysis studies of TBP and TDMAP.

Phase Equilibria in InAsSbP Quaternary Alloys Grown by Liquid Phase Epitaxy
M.R. WILSON,¹ A. KRIER,² and Y. MAO²
1--Department of Chemistry, University of Durham, South Road, Durham, DH1 3LE, UK. 2--School of Physics & Chemistry, University of Lancaster, Lancaster, LA1 4YB, UK.

KEY WORDS InAsSbP, liquid phase epitaxy, miscibility gap, phase diagrams, spinodal decomposition

The quaternary alloy InAs_1-x-ySb_xP_y, lattice-matched to InAs, is a promising material for the production of infrared light sources for the detection of gases in the 2-4 µm region of the spectrum. In this work, thermodynamic phase equilibrium calculations have been carried out to determine the compositions required for liquid phase epitaxial growth and the extent of the miscibility gap in the solid material. For high band gap materials, the desired growth temperature is found to be intermediate between a low temperature required to grow P-rich solids and higher temperatures required to avoid spinodal decomposition. Conventional LPE growth at an intermediate temperature of 583°C is found to produce good material with high luminescence efficiency and excellent optical characteristics. Problems with phosphorus loss from the melt are also discussed and lower growth temperatures are found to considerably reduce this problem. Growth in the metastable region, between the binodal and spinodal lines has been achieved with the production of phosphorus-rich solids with concentrations up to y = 0.445.

Fabrication of Dry-Etched Mirrors in GaAs-Based and In-P Based Lasers Using Chemically Assisted Ion-Beam Etching at Low Temperatures
R.E. SAH, J.D. RALSTON, J. DALEIDEN, E.C. LARKINS, S. WEISSER, J. FLEISSNER, and W. BENZ
Fraunhofer-Institut für Angewandte Festkörperphysik, Tullastr. 72, 79108 Freiburg, Germany.

KEY WORDS Dry-etched mirrors, GaAs-based lasers, InP-based lasers, ion-beam etching

We have fabricated dry-etched mirrors in high-speed InGaAs/GaAs/AlGaAs pseudomorphic multiple quantum well ridge-waveguide lasers at 60°C and in InGaAs/InP bulk lasers at 5°C using enhanced chemically assisted ion-beam etching (CAIBE) technique. The technique allows the etching of laser structures with good surface morphology and excellent anisotropy without cold traps in the etching system. Characteristics of the dry-etched facet lasers match those of cleaved devices. The low sample temperatures for etching allowed the use of standard photoresists as etch masks.

The Modification of Electrophotographic and Mechanical Properties of Organic Photoconductors by Ultra-Violet Irradiation
C.K.H. WONG,¹ Y.C. CHAN,¹ Y.W. LAM,¹ D.P. WEBB,¹ K.M. LEUNG,² and D.S. CHIU²
1--City University of Hong Kong, Department of Physics & Materials Science, Tat Chee Avenue, Kowloon, Hong Kong. 2--City University of Hong Kong, Department of Electronic Engineering, Tat Chee Avenue, Kowloon, Hong Kong.

KEY WORDS Charge transport, electrophotography, organic photoconductor

Experiments have shown that UV treatment has a hardening effect on the surface of a drum-based organic photoconductor (OPC) and improves the retentivity. The dark decay and the photoinduced discharge rates were reduced linearly with the increase of time of UV irradiation. The dark decay rate and photoinduced discharge rates were reduced to 60% and more than 70% of the initial rate, respectively, while there was an increase of the activation energy of the charge transport layer. An initial increase in the residual potential was also observed from 20 to about 70 V. After 40 min of UV irradiation, the residual potential was seen to drop to 50 V again, and at the same time a slight increase of the hardness of the photoreceptor surface was detected from 13 to 16 in Vickers scale (Hv). The phenomena is most likely explained by a photochemical change which reduces the density of charge transport sites, builds up the number of deep traps and changes the molecular distribution. The decrease in charge transport site density and buildup of deep traps reduces the conductivity and hence the dark decay and photoinduced discharge rates. The fragmentation of charge transport molecules may result in a change of activation energy as well as the increase in the residue potential. Moreover, the redistribution of molecules leads to the change in molecular density and hardness.

Photoluminescence and Raman Scattering Characterization of Silicon-Doped In _0.52Al_0.48As Grown on InP (100) Substrates by Molecular Beam Epitaxy
S.F. YOON, Y.B. MIAO, K. RADHAKRISHNAN, and S. SWAMINATHAN
School of Electrical and Electronic Engineering, Nanyang Technological University, Nanyang Avenue, Singapore 2263 Republic of Singapore.

KEY WORDS InAlAs, molecular beam epitaxy (MBE), photoluminescence, Raman scattering

Growth of In_0.52Al_0.48As epilayers on InP (100) substrates by molecular beam epitaxy at different silicon doping levels is carried out. The doped samples show an inverted S-shaped dependence of the PL peak energy variation with the temperature which weakens at high doping levels due to a possible reduction in the donor binding energy. There is a reduction in both the AlAs-like and InAs-like longitudinal-optic (LO) phonon frequencies and a broadening of the LO phonon line shape as the doping level is increased. The PL intensity also showed in increasing degrees at higher doping levels, a temperature dependence which is characteristic of disordered and amorphous materials.

Al-Ga-As-Dopant Phase-Equilibria for Liquid Phase Epitaxy
C. ALGORA and M.A. MARTÍNEZ
Instituto de Energía Solar, E.T.S.I. Telecomunicación (U.P.M.), Ciudad Universitaria s/n, 28040 Madrid, Spain.

KEY WORDS Al-Ga-As-dopant, liquid phase epitaxy, phase equilibria

Calculated phase data for the Al-Ga-As-dopant liquid phase epitaxial system are presented. Modifications to interaction parameters are made. Firstly, the activity coefficient of the ternary solid phase has been taken as unity, i.e., solid solution AlGa-AlAs is assumed ideal, which permits the elimination of discontinuities when aluminum atomic fraction in the liquid goes to zero. Accordingly, the Al-As interaction parameter must be recalculated and a value of 2122-14T cal/mole (from 973 to 1273K) has been obtained. With these changes in the interaction parameters, the model allows a fairly accurate fit to the solidus and liquidus data, and former problems associated with high temperatures and high and low aluminum atomic concentrations in the liquid have been resolved. Finally, the theory is extended to the case of a fourth component which acts as a dopant. Tin is taken as an example. Interaction parameters for this dopant have been calculated permitting a close adjustment to the experimental values. In order to achieve a more precise fitting an aluminum dependence in the Al-Sn interaction parameter is suggested.

Incorporation of Group V Elements in Ga_xIn_1-xAs_yP_1-y Grown by Gas Source Molecular Beam Epitaxy
TSUEN-LIN LEE, JIN-SHUNG LIU, and HAO-HSIUNG LIN
Room 419, Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, Republic of China.

KEY WORDS GaInAsP, gas source molecular beam epitaxy, incorporation

A simple growth model has been successfully developed for the determination of the As to P incorporation ratio, i.e., mole fraction y, in growing Ga_xIn_1-xAs_yP_1-y quaternary alloys by gas source molecular beam epitaxy. The model covers the whole composition range with only two fitting parameters, k_In and k_Ga, whose physical meanings are the product of As to P desorption time constant ratio and incorporation rate constant ratio for InAs_yP_1-y and GaAs_yP_1-y, respectively. The best fitting values of k_In and k_Ga from our experimental results are 28 and 3, respectively, at a growth temperature of 480°C. The temperature dependency of the parameters were also studied. The activation energies of k_In and k_Ga are +30 and -330 meV, respectively. The significant differences between the parameters may be due to the different bond energies of binary alloys.

Formation of Ohmic Contacts to p-ZnTe
J.T. TREXLER, J.J. FIJOL, L.C. CALHOUN, R.M. PARK, and P.H. HOLLOWAY
Department of Materials Science and Engineering, University of Florida, Gainesville, FL 32611-6400.

KEY WORDS Au/p-ZnTe contacts, ohmic contacts, ZnTe

At formation and temperature stability of sputter deposited gold ohmic contacts to molecular beam epitaxiallygrown p-type ZnTe (doped with nitrogen to a free hole concentration of 3 x 10¹⁸ cm³) have been studied using current-voltage (I-V), Auger electron spectroscopy, secondary ion mass spectrometry, and optical and scanning electron microscopy. The I-V characteristics of 1500Å Au/p-ZnTe contacts were measured as-deposited and after heat treatments at 150, 200, 250, and 350°C for 15 min intervals up to 90 min. As deposited, the contacts were poor Schottky contacts, but became ohmic after 15 min at all temperatures. There was an increased resistance at t >15 min for T¾250°C, and a very large resistance increase upon heat treatment for all times at 350°C. The interface between the metallization and ZnTe was initially very planar, and remained planar upon formation of the ohmic contact. Upon heating at T250°C, Au diffused into ZnTe. The ohmic behavior of the Au/p-ZnTe contacts is attributed to this diffusion which created a highly doped near-surface region in the ZnTe. Microscopy showed that Au also migrated across the ZnTe surface forming an extended reaction zone (100 µm) around the dot contact at T250°C.

Microstructural Analysis of a Au/Pt/Pd/Zn Ohmic Contact to an AlGaAs/GaAs Heterojunction Bipolar Transistor
P. JIAN,¹ D.G. IVEY,¹ S. EICHER,² and T.P. LESTER²
1--Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta, Canada T6G 2G6. 2--Bell Northern Research, Ltd., P.O. Box 3511, Station C, Ottawa, Ontario, Canada K1Y 4H7.

KEY WORDS AlGaAs/GaAs, Au/Pt/Pd/Zn ohmic contacts, transmission electron microscopy

Gold-based ohmic contacts, incorporating Pt, Pd, and Zn layers, to AlGaAs/GaAs heterojunction bipolar transistors (HBTs) have been characterized using transmission electron microscopy (TEM). The metallization was deposited onto a 30 nm graded emitter layer of n-type Al_xGa_1-xAs, which was on a 30 nm emitter layer of n-type Al_0.3Ga_0.7As, with the aim of contacting the underlying 80 nm thick graded base layer of p-type Al_xGa_1-xAs. Metal layers were deposited sequentially using electron beam evaporation and the resultant metallizations were annealed at temperatures ranging from 250-500°C for up to several minutes. A minimum contact resistance of 8.5 x 10^-7 -cm² was achieved, which corresponded to the decomposition of ternary phases at the metallization/semiconductor interface, to binary phases, i.e., PdGa and PtAs₂. Long term stability tests were done on the optimum contacts. Anneals at 270°C for up to four weeks in duration produced virtually no change in microstructure, with the exception of some outward diffusion of Ga and As.

The Scaling of CVD Rotating Disk Reactors to Large Sizes and Comparison With Theory
ALAN G. THOMPSON,¹ R.A. STALL,¹ P. ZAWADZKI,¹ and G.H. EVANS²
1--EMCORE Corporation, 394 Elizabeth Avenue, Somerset, NJ 08873. 2--Sandia National Laboratories, Livermore, CA 94550.

KEY WORDS AlGaAs, CVD, InGaAlP, MOCVD, rotating disk reactors

An important advantage of the rotating disk reactor (RDR) geometry for chemical vapor deposition is the one-dimensional nature of the transport process over a wide range of operating conditions. Due to geometric complexity, models for most chemical vapor depositon (CVD) reactors have been used to fit existing data and are not very useful as predictors of performance for changed dimensions, particularly larger ones. Previous RDR modeling results have shown good agreement with experiment for smaller sizes (typically single wafer) and have also resulted in a set of scaling "laws" for scaling to larger sizes. In this work, we report the use of these scaling laws to design larger RDRs and to determine optimal process conditions. We then compare the experimental results to those predicted by the model. Previously, our largest RDR utilized an 180 mm diameter disk, which holds 6 x 50 mm or 3 x 76 mm wafers. The new systems have disks of 300 and 420 mm diameter, holding 17 and 38, 50 mm wafers, respectively. In general, excellent agreement was obtained when the new systems were run at conditions predicted by the scaling laws. As the disk diameter increases, the model calls for a reduced rotational rate, a lower operating pressure and a main carrier gas flow that increases sublinearly with the disk area. Experimental growths were made with the III-V materials AlGaAs and InGaAlP, on GaAs substrates of 50, 76, and 100 mm diameters. Uniformities and electrical and optical properties similar to those obtained in smaller RDR systems were obtained under conditions predicted by the model. We also discuss the trade-offs involved where one parameter is fixed by process needs, necessitating that others must be adjusted accordingly. It should be stressed that these adjustments are all calculated from the model and are designed to maintain the ideal flow and thermal environments obtained with the smaller systems. We find that typical metalorganic chemical vapor deposition process parameters are easily accommodated for these disk sizes. We believe that this is the first report of a truly scalable CVD reactor geometry with both theoretical and experimental justification. Using the RDR technique therefore enables the grower to develop a process on a small system and transfer it to a larger system for manufacturing, without having to reinvent the process. It also enables the CVD system manufacturer to design and build new systems that will perform in a predictable manner, or to adapt new process conditions to an existing system, without requiring extensive re-engineering efforts.

Heterojunction Diodes in 3C-SiC/Si System Grown by Reactive Magnetron Sputtering: Effects of Growth Temperature on Diode Rectification and Breakdown
Q. WAHAB, M. KARLSTEEN, O. NUR, L. HULTMAN, M. WILLANDER, and J.-E. SUNDGREN
Department of Physics, Linköping University, S-581 83 Linköping, Sweden.

KEY WORDS Heterojunction diodes, magnetron sputtering, silicon carbide, wide bandgap semiconductors

3C-SiC/Si heterojunction diodes were prepared by reactive magnetron sputtering of pure Si in CH₄-Ar discharge on Si(111) substrates kept at temperatures (T_s) ranging from 800 to 1000°C. A good diode rectification process started for films grown at T_s ¾900°C. Heterojunction diodes grown at T_s = 850°C showed the best performance with a saturation current density of 2.4 x 10^-4 A cm^-2. Diode reverse breakdown was obtained at a voltage of -110 V. The doping concentration (Nd) of the 3C-SiC films was calculated from 1/C² vs V plot to be 3 x 10¹⁵ cm^-3. Band offset values obtained were -0.27 and 1.35 eV for the conduction and valence band, respectively. X-ray diffraction analysis revealed the film grown at T_s = 850°C to be single-phase 3C-SiC. The full width at half maximum of the 3C-SiC(111) peak was only 0.25 degrees. Cross-sectional transmission electron microscopy showed the film to be highly (111)-oriented with an epitaxial columnar structure of double positioning domain boundaries.

Efficient 2.0-2.6 µm Wavelength Photoluminescence from Narrow Bandgap InAsP/InGaAs Double Heterostructures Grown on InP Substrates
D. GARBUZOV,¹ D.-S. KIM,¹ S.R. FORREST,¹ R. MENNA,² M. LANGE,³ G.H. OLSEN,³ and M. COHEN³
1--Advanced Technology Center for Photonics and Optoelectronic Materials, Department of Electrical Engineering and the Princeton Materials Institute, Princeton University, Princeton, NJ 08544. 2--David Sarnoff Resarch Center, Princeton, NJ 08543. 3--Sensors Unlimited, Princeton, NJ 08540.

KEY WORDS 2.0-2.6 µm wavelength, efficient photoluminescence, graded composition buffer layer, InAsyP1-y/InxGa1-xAs double heterostructure

Photoluminescence spectra and efficiency have been measured for several strained InAs_yP_1-y/In_xGa_1-xAs (0.28 < y ¾ 0.62; 0.66 ¾ x < 0.83) double heterostructures grown by vapor phase epitaxy on InP substrates with graded InAsP buffer layers. Luminescence peak positions between the wavelengths of 1.99 and 2.57 µm at a temperature of 295K are consistent with bandgap luminescence from the In_xGa_1-xAs active regions. Despite a high density of dislocations in the buffer layers, internal radiative recombination efficiencies of from 25 to 50% for the structures are found at 295K.

Comparative Measurements of the Electron Emission Behavior of Si₃N₄-InGaAs Interfaces Prepared by Remote and Direct PECVD
P.J.M. PARMITER and J.G. SWANSON
Department of Electronic and Electrical Engineering, King's College London, Strand, London WC2R 2LS, United Kingdom.

KEY WORDS Interfaces, PECVD, Si3N4/InGaAs

The effects on the insulator-semiconductor interface of two different deposition methods of silicon nitride on In_0.53Ga_0.47As have been electronically studied using capacitance-voltage (CV) and deep level transient spectroscopy (DLTS) measurements. The CV data showed very similar behavior on both types of sample, but the DLTS results were surprisingly different. The behavior of samples fabricated using direct plasma enhanced chemical vapor deposition (PECVD) can be explained by electronic emission from interface states to the conduction band minimum. However, the remote PECVD method gave a DLTS peak corresponding to electron emission with a low activation energy 0.13 eV. As with the direct PECVD samples, the thermal activation energy decreased as the bias was made more positive, but in this case the change in emission energy was only a fraction of the shift in the Fermi level position. A model is proposed involving the concepts of lattice relaxation and hole capture to explain this behavior.

Growth, Characterization and Modeling of In_xGa_1-xP Stripes by Selective-Area MOCVD
J.F. KLUENDER, A.M. JONES, R.M. LAMMERT, J.E. BAKER, and J.J. COLEMAN
Microelectronics Laboratory and Materials Research Laboratory, University of Illinois at Urbana-Champaign, Urbana, IL 61801.

KEY WORDS InGaP, MOCVD, oxide stripes

A computational diffusion model is used to predict the lateral thickness and composition profiles of In_xGa_1-xP stripes grown by selective-area, atmospheric pressure metalorganic chemical vapor deposition. Standard profilometry is used to measure the thickness profiles of InP and GaAs stripes grown on SiO₂ patterned InP and GaAs substrates, respectively. The model is used to find self-consistent empirical diffusion parameters for the In and Ga components which yield fits to the measured thickness data. The InP and GaAs data is then used to predict the growth thickness profile of InGaP by a weighted sum of the predicted profiles of the InP and GaP binary constituents. InGaP composition profiles are calculated by taking the ratio of the InP deposited volume to the InGaP deposited volume predicted by the model at each of the simulation points. Predicted thickness profiles are verified by standard profilometry, and composition profiles are verified by secondary ion mass spectrometry imaging using a fast resistive anode encoding detector. It is found that the measured thickness and composition profiles agree well with the profiles predicted by the model, thus verifying that the model can be used for the InGaP material system. The derived empirical parameters are used to predict the thicknesses and compositions of selectively grown InGaP quantum wells as a function of oxide width.

Phase Modulated Ellipsometry Used for Composition Control during MBE Growth of CdHgTe: An Analysis of Instrumental Factors and an Assessment of the Material Produced
R.H. HARTLEY,¹ M.A. FOLKARD,¹ D. CARR,¹ P.J. ORDERS,¹ G. SHEN,¹ V. KUMAR,¹ T.A. STEELE,¹ I.K. VARGA,¹ B.A. JOHNSON,¹ K. FUELOOP,¹ P. CAPPER,² D. DUTTON,² S. BARTON,² I. GALE,³ and F. GRAINGER³
1--CMTEK Ltd., P.O. Box 1500, Salisbury 5108, South Australia. 2--GEC-Marconi Infrared, P.O. Box 217, Southampton, SO15 0EG, United Kingdom. 3--Philips Research Laboratories, Redhill, Surrey, RM1 5AH, United Kingdom.

KEY WORDS CdHgTe, ellipsometry, molecular beam epitaxy (MBE)

We have found phase modulated ellipsometry (PME) to be a sensitive analytical technique capable of providing real time information on composition, epilayer thickness, growth rate, interdiffusion and surface roughness. To fully exploit the benefits of PME, the instrument must be carefully calibrated and the many factors affecting its performance need to be understood and allowed for. In this paper we examine how the accuracy of the determination of composition of molecular beam epitaxially grown CdHgTe alloy films is affected by misalignment of the optical components, the presence of vacuum windows, signal conditioning prior to analog to digital conversion, temperature changes, and modulator settings. We conclude by presenting results which demonstrate the quality of CdHgTe layers grown on (211)B CdZnTe and (211)B GaAs substrates using our techniques.

Extremely Strong and Sharp Photoluminescence Lines from Nitrogen Atomic-Layer-Doped GaAs, AlGaAs and AlGaAs/GaAs Single Quantum Wells
TOSHIKI MAKIMOTO and NAOKI KOBAYASHI
NTT Basic Research Laboratories, 3-1 Morinosato Wakamiya, Atsugi-shi, Kanagawa, 243-01 Japan.

KEY WORDS AlGaAs/GaAs, atomic-layer doping, -doping, GaAs, isoelectronic traps, photoluminescence (PL), secondary ion mass spectrometry (SIMS)

We have performed nitrogen atomic-layer doping into GaAs, AlGaAs, and AlGaAs/GaAs single quantum wells using atomic nitrogen cracked by a hot tungsten filament. While the atomic-layer-doped GaAs layers show a series of sharp and strong photoluminescence lines relating to excitons bound to nitrogen atoms at 8K, atomic-layer-doped AlGaAs layers show several broad nitrogen-related lines. For the atomic-layer-doped single quantum well at the center of the GaAs layer, the quantum well luminescence itself disappears and a dominant and sharp luminescence is observed at a longer wavelength. It is found that the As pressure during the atomic-layer doping greatly affects the luminescence characteristics.

Film Morphology and Reaction Rate for the CVD of Tungsten by the WF₆-SiH₄ Reaction
O.H. GOKCE,¹ J.T. SEARS,² and T. SAHIN³
1--Department of Physics, New Jersey Institute of Technology, Newark, NJ 07102. 2--Department of Chemical Engineering, Montana State University, Bozeman, MT 59717. 3--Applied Materials, 3050 Bowers Avenue, Santa Clara, CA 95054.

KEY WORDS Chemical vapor deposition (CVD), crystal orientation, film morphology, reaction rate, SiH4, tungsten, WF6

Low pressure chemical vapor deposition (LPCVD) of tungsten (W) by SiH₄ reduction of WF₆ on Si(100) surfaces was studied in a single-wafer, cold-wall reactor over a temperature range of 137-385°C and a pressure range of 1-10 Torr at a SiH₄/WF₆ ratio of 1.0. Rate data were obtained in the absence of gas-phase mass transport limitations and were measured using gravimetric techniques. The amount of tungsten that was deposited varied between 5.79 x 10^-5 and 1.70 x 10^-2 g/cm² (~300-88,000Å based on a tungsten density of 19.3 g/cm³), and the rates were between 1.02 x 10^-4 and 1.74 x 10^-3 g/cm² min (~500-9,000Å/min). The apparent overall activation energy increased with pressure; 0.12 eV/atom at 1 Torr, and 0.40 eV/atom at 10 Torr for short reaction times (0.5-1.5 min). The overall rate was dependent on reaction time (film thickness). Better film morphologies were obtained at higher temperatures and lower pressures. A W(110) preferential orientation was observed at the Si-W interface. Tungsten orientation switched from (110) to (100) as the films grew thicker. Higher apparent activation energies observed at higher pressures were attributed to gas phase reactions and/or by-product readsorption. The interdependence of rate and film morphology was attributed to a reconstruction of W(100) surfaces on which reactant diffusion/surface reaction is favored.

Co-Pyrolysis of DIPSbH and TMIn
Y.S. CHUN,¹ G.B. STRINGFELLOW,¹ and R.W. GEDRIDGE, JR.²
1--Department of Materials Science & Engineering, University of Utah, Salt Lake City, UT 84112. 2--Naval Air Systems Command, Naval Air System Command Headquarters, 1421 Jefferson Davis Hwy., Arlington, VA 22243.

KEY WORDS Co-pyrolysis, DIPSbH, TMIn

The reaction mechanisms for the pyrolysis of diisopropylantimonyhydride (DIPSbH, (C₃H₇)₂SbH) alone and for the co-pyrolysis of DIPSbH and trimethylindium (TMIn, (CH₃)₃In) in D₂ and H₂ ambients have been studied in an isothermal flow-tube, "ersatz" reactor using mass spectrometry to analyze the reaction products. The rate limiting step in the pyrolysis of DIPSbH alone is the reductive coupling reaction, producing C₃H₈. Additional products are C₃H₆ and C₆H₁₄ produced by disproportionation and recombination reactions, respectively, of C₃H₇ radicals produced during the second stage of DIPSbH pyrolysis. The mixture of DIPSbH with TMIn produces a nonvolatile adduct on the quartz walls immediately after mixing in the reactor even at room temperature. No products were evolved at room temperature. However, for reactor temperatures between 100 and 200°C, an alkane elimination reaction occurs, producing CH₄. The remaining solid product is postulated to be [(CH₃)₂InSb(C₃H₇)₂]_n (n=2or3). For temperatures greater than 200°C, the DIPSbH begins to pyrolyze independently. This temperature for the onset of DIPSbH pyrolysis is considerably above the temperature (125°C) at which pyrolysis begins for DIPSbH alone. This suggests that during co-pyrolysis formation of the adduct retards pyrolysis of DIPSbH. Apparently, dissociation of the adduct is necessary before the DIPSbH can pyrolyze independently. Co-pyrolysis of DIPSbH and TMIn between 250 and 375°C produces (C₃H₇)Sb(CH₃)₂ and (CH₃)₃Sb. Neither is found for the pyrolysis of DIPSbH alone. Considerably larger amounts of C₂H₆ are also detected at low temperatures (¾= 300°C). The ethane may come from the [(CH₃)₂InSb(C₃H₇)₂]_n via an intramolecular alkane elimination reaction. The high carbon contamination levels reported for InSb samples grown by OMVPE using these precursors at 300 and 325°C are postulated to be caused by the formation of (C₃H₇)Sb(CH₃)₂ during the co-pyrolysis of DIPSbH and TMIn, but not during the pyrolysis of each precursor alone.

Structural Characterization of Bulk GaN Crystals Grown Under High Hydrostatic Pressure
ZUZANNA LILIENTAL-WEBER,¹ C. KISIELOWSKI,¹ S. RUVIMOV,¹ Y. CHEN,¹ J. WASHBURN,¹ I. GRZEGORY,² M. BOCKOWSKI,² J. JUN,² and S. POROWSKI²
1--Center for Advanced Materials, Materials Science Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720. 2--High Pressure Research Center "Unipress," Polish Academy of Sciences, Warsaw, Poland.

KEY WORDS Bulk GaN, crystal polarity, excess Ga, microstructure, planar defects, surface roughness

This paper describes TEM characterization of bulk GaN crystals grown at 1500-1800K in the form of plates from a solution of atomic nitrogen in liquid gallium under high nitrogen pressure (up to 20 kbars). The x-ray rocking curves for these crystals were in the range of 20-30 arc-sec. The plate thickness along the c-axis was about 100 times smaller than the nonpolar growth directions. A substantial difference in material quality was observed on the opposite sides of the plates normal to the c direction. On one side the surface was atomically flat, while on the other side the surface was rough, with pyramidal features up to 100 nm high. The polarity of the crystals was determined using convergent-beam electron diffraction. The results showed that, regarding the long bond between Ga and N along the c-axis, Ga atoms were found to be closer to the flat side of the crystal, while N atoms were found to be closer to the rough side. Near the rough side, within 1/10 to 1/4 of the plate thickness, there was a high density of planar defects (stacking faults and dislocation loops decorated by Ga/void precipitates). A model explaining the defect formation is proposed.

Optimization of the Spacer Layer Thickness in AlInAs/InGaAs/InP MODFETs
MATTHEW L. SEAFORD,^1,2 GLENN MARTIN,¹ L.F. EASTMAN¹ DAVE HARTZELL,³ and SCOTT MASSIE³
1--Cornell University, Department of Electrical Engineering, Ithaca, NY, 14853. 2--Geocenters, Eatontown, NJ, 07724. 3--Quantum Epitaxial Designs, Inc., Bethlehem, PA, 18015.

KEY WORDS Channel conductance, InGaAs, InAlAs, InP, interface roughness, molecular beam epitaxy (MBE)

The effects of spacer layer thickness variations on single atomic planar doped (APD) AlInAs/InGaAs modulation doped field effect transistors grown by solid source molecular beam epitaxy have been studied and characterized. The thickness of the AlInAs spacer layer was varied between 0 and 100Å. Room temperature Hall measurements found the mobility exhibited an exponential relationship ranging from 6500 to 10800 cm²/Vs. The sheet charge varied linearly from 3.46 x 10¹² cm^-2 to 2.24 x 10¹² cm^-2. An optimum spacer layer thickness based on maximum channel conductance was found to be 40Å with a mobility of 9600 cm²/Vs and a sheet charge of 3.0 x 10¹² cm^-2. The loss of mobility due to remote ion scattering was examined. This loss was related to the distribution of the Si atoms in the atomic planar doped layer in order to obtain the standard deviation of the interface. This relationship will allow various growth parameters, such as substrate temperature, growth rate, and V/III ratio to be altered to determine the optimum conditions independent of the growth chamber used to create the structures.

Mass Spectroscopy Study of GaN Metalorganic Chemical Vapor Deposition
YONGJO PARK and DIMITRIS PAVLIDIS
Solid State Electronics Laboratory, Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI 48109-2122.

KEY WORDS GaAs, GaN, mass spectroscopy analysis, metalorganic chemical vapor deposition (MOCVD)

Metalorganic chemical vapor phase deposition of GaN on (100) GaAs has been studied using mass spectroscopy. With increasing substrate temperature, the amount of decomposed trimethylgallium (TMGa) was observed to increase exponentially with a characteristic energy of 1.5 eV. The presence of NH₃ was found to suppress the production of CH₃ in the gas phase. This implies that CH₃ of TMGa reacts with the hydrogen atom of NH₃, forming CH4 as a main gas product. Studies of nitrogen evaporation from the growth surface when TMGa flow was off lead to the conclusion that increased growth rate could result in decreased background electron concentration due to nitrogen vacancy. The presence of NH₃ significantly promotes the decomposition of TMGa. Desorption of excess Ga atoms from the growth surface at low NH₃ flow rates takes place as suggested by the increased ratio of peak intensity of Ga (m/e=69) to that of DMGa ((CH₃)₂Ga, m/e=99) with decreasing NH₃ flow rate.

A Model of the Interdiffused Multilayer Process
SPYROS A. SVORONOS,¹ WILBUR W. WOO,¹ STUART J.C. IRVINE,² HALUK O. SANKUR,³ and JAGMOHAN BAJAJ³
1--University of Florida, Department of Chemical Engineering, Gainesville, FL 32611. 2--Multidisciplinary Research and Innovation Centre, North East Wales Institute, Plas Coch, Mold Road, Wrexham, Clwyd LL11 2AW, United Kingdom. 3--Rockwell International Corporation, Science Center, 1049 Camino Dos Rios, Thousand Oaks, CA 91360.

KEY WORDS HgCdTe, in situ monitoring, interdiffused multilayer process, laser reflectance, MOCVD, modeling

The interdiffused multilayer process (IMP) is a novel approach to growing Hg_1-xCd_xTe. In this process, alternating thin films of HgTe and CdTe are grown and allowed to interdiffuse resulting in a bulk material of constant composition. A model of the IMP must include the effects of both the deposition of new material and the interdiffusion of the material. It must also be able handle the flush phases of the IMP where the growth rate decays to zero. Existing approaches to modeling epitaxial growth of Hg_1-xCd_xTe treat growth and interdiffusion as separate, sequential steps resulting in numerical stability problems, pseudodiffusion effects, or flush phase modeling problems. The model presented here, however, is based on an incremental balance where growth and diffusion occur simultaneously, resulting in a model exhibiting none of the difficulties mentioned above. The IMP growth model is integrated with a model for calculating reflectance from a laser directed at near normal incidence angle. The predicted reflectance is compared to experimental measurements and showed a good preliminary fit when the model employed default parameters. The agreement is greatly improved after parameter fitting.