Supersonic cluster beam techniques have recently produced some fascinating new information as to the surface chemistry and physics of small (2–100 atom) clusters of various semiconductors. For example, it appears that silicon clusters exhibit a remarkably pronounced alternation in reactivity as a function of cluster size. For ammonia chemisorption certain clusters such as Si33+, Si39+, and Si45+ have been found to be almost completely inert, while neighboring clusters such as Si36 and Si45+ chemisorb ammonia readily. Such sharply patterned reactivity results may provide significant clues as to the detailed nature of semiconductor surface restructuring and the consequent effects of this restructuring upon the surface chemistry.

This article summarizes recent studies of the chemistry of aluminum and silicon cluster ions containing 3–27 atoms.

Past syntheses of polymer composites have largely evolved from chemical reduction or thermal decomposition of organometallic or inorganic precursor molecules in polymers, or plasma and thermal co-deposition of metal vapors and carbonaceous free radicals. Our approach involves the site-specific capture of metal atoms deposited in vacuum to give isolated, high energy mononuclear organometallic centers within a polymer film. These centers can be converted at ambient or sub-ambient temperatures (ie, below the polymer glass transition temperature) to, for example, metal oxide microclusters.

We describe the results of our studies of a prototypical system involving chromium atoms and their conversion to corundum-type oxide microclusters in arene-functionalized polymer films. Thus Cr was deposited into 150 K liquid tetrahydrofuran solutions of polystyrene or poly(styrene-isoprene-styrene) triblock, spun in vacuo as thin films on the surface of a rotating glass cryostat. Evidence from epr spectrscopy shows that the resulting polymer-anchored (inter/intra-chain) bis(arene)Cr sandwich complex is locally mobile in the macroscopically rigid film at room temperature. The Cr atom is discharged from the rings by subsequent reaction with oxygen diffused into the film. Although α Cr2 O3 is a classic twosublevel antiferromagnet that is not epr active above 308 K, we observe an intense signal even at 77 K in these films. Cr2O3 microclusters are indicated, and these are confirmed by in situ measurements of the oxidation and aggregation process.

The metal atom methodology has also been used to synthesize silver microsphere/polymer composites. With quadratic electrooptic phase modulation, these composites were found to show a third order susceptibility enhanced by coupling the dipolar surface plasmon mode of the particles with incident light.

The precursor method is being used increasingly in the preparation of solid state inorganic materials. Use of this general method allows the isolation of otherwise inaccessible phases and the preparation of known phases under much milder conditions. One approach which holds promise is the use of organometallic precursors for the preparation of both thin films and bulk samples of inorganic materials. In this paper I describe our syntheses of several metal chalcogenides from organometallic reagents.

The reactions of (Me3Si)3As with group III halides have been utilized to prepare AlAs, GaAs and InAs. The adduct (Me3Si)3AsAlCl3 has been isolated as an intermediate in the formation of AlAs from (Me3Si)3As and AICI3. The crystal structure of its toluene solvate has been determined.

The goals of the research are the design and synthesis of a new class of precursor compounds for III/V compound semiconductor materials, growth of films with these precursors and developoment of an understanding of the relationships between precursor structure, film growth reactions and film properties. Conventional OMCVD of III/V compound materials has a number of inherent safety and processing problems associated with the group III alkyl and group V hydride sources. Our approach to these problems is the synthesis of a single precursor with a fixed III:V stoichiometry and a direct two center, two electron sigma III V bond., These, compounds have the general formula,[R2M(R 2 E)] 2 and R2M(R 2 E) 2M R2 (MM = Al, Ga, In; E=P, As; R, R = alkyl, aryl). The III V bond in these compounds is stronger than the other bonds and the minium deposition temperature can be controlled by employing subsituents that undergo facile hydrocarbon elimination.

A typical example is the use of [Me 2Ga(µ t Bu 2 As)] 2 as the single source for GaAs films. The organometallic precursor is a solid crystalline powder which is maintained at 130°C to generate enough vapor for OMCVD. Typical film growth conditions involve the use of H2 or He as the carrier gas, substrate temperatures of 500 to 700°C, and a total system pressure of 0.0002 Torr. GaAs(100), Si(100) (As doped 30 off toward (011) and quartz have been used as substrates. Film composition has been established with XPS. The Ga 3d, As 3d, and C ls signals at 18.8, 40.9, and 284.6 eV, respectively, reveal the films to be 1:1 Ga:As and void of carbon. The carbon levels are less than 1000 ppm. X ray diffraction and SEM results suggest polycrystalline GaAs on quartz and epitaxial GaAs on GaAs(100) and Si(100). (2 K) photoluminescence measurements on GaAs, grown on semi insulating GaAs(100) and Si doped GaAs(0 100) at 570 C. produce PL signals indicating that crystalline domains are present, the measurements indicate degeneratively n doped material and show that good Ga:As ratios and low levels (ca. 1 ppm) of impurities are present. Growth rates:∼ 1.0 mm/hour.

Organometallic molecules containing covalently linked gallium and arsenic or indium and phosphorus have been synthesized and characterized spectroscopically and by X-ray diffraction. These precursors can be transformed into the corresponding III-V materials in a chemical reaction proceeding at ambient temperature. The compound semiconductors prepared in this way are obtained as amorphous powders. During the reaction, quantum size effects may be observed by UV-VIS spectroscopy as the particles grow.

The growth of ZnSe on GaAs substrates by metalorganic chemical vapor deposition was investigated in a specially designed vertical downflow reactor. Dimethylzinc was used as the Zn source while different Se source compounds (hydrogen selenide (H2Se), diethylselenide and methylallylselenide) were employed to determine the effect of different source combinations on morphology, thickness uniformity, growth rate, electrical properties and photoluminescence (PL) characteristics of the grown ZnSe films. The H2Se was produced in situ by reaction of H2 and Se followed by distillation to control the amount of H2Se entering the reaction zone. H2Se produced very high mobility films with good PL spectra but poor surface morphology. Diethylselenide led to layers of good morphology and PL characteristics but the films were highly resistive. Unusual surface features were observed for methylallylselenide.

Organometallic precursors with low decomposition temperatures are essential in the fabrication of high performance mercury cadmium telluride (Hg1-xCdxTe) infrared detectors by pyrolytic and photolytic metal-organic chemical vapor deposition (MOCVD). Film growth temperature is governed by the relative stability and/or reactivity of the organotellurium precursor, which is determined by the strength of the Te-C bonds. Since the rate-determining step in the pyrolysis of organometallic compounds involves bond breaking and free radical formation, we have concentrated on the synthesis of a variety of organotellurium precursors with substituents that possess low activation energies for the formation of hydrocarbon free radicals. The synthesis, characterization, and properties of methylallyltelluride, ethylallyltelluride, isopropylallyltelluride, tertiarybutylallyl-telluride, methylbenzyltelluride, and methylpentadienyltelluride are reported. These unsymmetrical tellurides were characterized by 1H, 13 C, and 125Te NMR spectroscopy. The potential applicability of these organotellurium precursors to lower film-growth temperatures in MOCVD is discussed.

In order to reduce the growth temperature of (Hg,Cd)Te by MOVPE below 350–400°C, alternative organometallic precursors will be required which either decompose at a lower temperature than existing precursors or which absorb strongly at a suitable wavelength in the ultraviolet. The features required for a programme of assessment of organometallics are discussed. UV absorption spectra for dimethyltelluride, dimethylditelluride, diethyltelluride, di-iso-propyltelluride and diallyltelluride are presented and their usefulness as photolytic MOVPE precursors discussed.

InP has been prepared by the reaction of InX3 (X = Cl, Br, I) with P(SiMe3)3. The intermediates and product have been characterized by XPS, elemental analysis and X-ray powder diffraction.

The thermal decomposition of AsH3 and TMG is measured insitu under different experimental conditions. Simultaneously the production of H2, CH4 and C2H6 is observed. The data indicate a situation where AsH3 is only partially decomposed at the GaAs surface. The hydrogen released removes additional CH3 groups from the trimethyl-gallium (TMG) molecule, enhances the decomposition of TMG, and thereby forms methane.

A study of the room temperature gas-phase interactions between gallium and indium alkyls was undertaken using a mass spectrometer sampling system, mounted on a low pressure organometallic vapor phase epitaxial reactor. Mixtures of triethylindium with triethylgallium or trimethylgallium were investigated. Both combinations formed addition compounds; moreover, the triethylindium-trimethylgallium mixture underwent alkyl exchange. Both admixtures showed reduced reactivity towards arsine. A structure for the addition compound is proposed.

The growth of GaAs from triethylgallium (TEG) and trimethylgallium (TMG) with tertiarybutylarsine (tBAs), triethylarsenic (TEAs), and trimethylarsenic (TMAs), has been investigated by using a reactor equipped with a recording microbalance for in situ rate measurements. Rate data show that the growth with these precursors is dominated by the formation of adduct compounds in the gas lines, by adduct related parasitic gas phase reactions in the heated zone, and by the surface reactions. A model is proposed for the competition between deposition reactions and the parasitic gas phase reactions. Model predictions are in very good agreement with experimental data for all combinations of precursors except for TEG/TMAs where extensive gallium droplet formation is observed at low temperatures. Growth of reasonable quality GaAs with Hall mobilities of 7600 cm2/Vs at 77 K using TEG and tBAs is reported for the first time.

A general MOVPE model has been used to assess the applicability of simplified representations for surface kinetics. With the general model, predictions for GaAs deposition on (111 )Ga using trimethylgallium and arsine show excellent agreement with observed growth rates. However, if Langmuir-Hinshelwood kinetics is assumed, the model only matches the deposition rates over a narrow range of operating conditions, even when several rate-limiting steps are included. This limitation arises because combinations of equilibrium constants and local partial pressures often do not give reasonable approximations for the surface concentrations of reactive intermediates. The form of the Langmuir-Hinshelwood relation(s) and the parameter values can be fitted empirically to experimental data, but this could lead to erroneous conclusions concerning process behavior and the model would have limited predictive capabilities. An alternative approach is to use surface reaction probabilities, but they can only be applied in an empirical fashion and their magnitudes depend on gas flow rate, inlet composition, and reactor pressure as well as surface temperature.

A kinetic model for metalorganic chemical vapor deposition (MOCVD) of GaAs from trimethylgallium and arsine is presented. The proposed mechanism includes 15 gas-phase species, 17 gas-phase reactions, 9 surface species and 29 surface reactions. The surface reactions take into account different crystallographic orientations of the GaAs substrate. Sensitivity analysis and existing experimental observations have been used to develop the reduced mechanism from the large number of reactions that might in principle occur. Rate constants are estimated by using thermochemical methods and reported experimental data. The kinetic mechanism is combined with a two-dimensional transport model of a hot-wall tubular reactor used in experimental studies. Model predictions of gas-phase composition and GaAs growth rates show good agreement with published experimental studies. In addition, the model predicts reported trends in carbon incorporation.

The thermodynamic equilibrium state of the Ga-Al-As-C-H system was determined theoretically by means of an iterative equilibrium constant method. This method of calculation is presented and discussed. With very little operator input, the program is capable of computing the partial pressures of the gas-phase species present in the equilibrated system.

In these calculations the system was considered to be saturated with solid-phase A1GaAs and included 58 plausible gas-phase intermediates which evolved from the initially present gas species; trimethylgallium, trimethylaluminum, arsine, and hydrogen. Temperature and total system pressure ranges investigated were 750–1100 K and 0.1 atm-1.0 atm, respectively. The effects of temperature and pressure variations, in addition to effects caused by changes in the appropriate atom ratios, have been delineated. The properties of this equilibrated system are compared with those from recent thermodynamic research efforts on AlGaAs systems consisting of only gaseous constituents.

Alkylated arsenic compounds have shown some promise as alternatives to arsine as the group-V source gas for GaAs MOCVD. However, little is known about the fundamental chemical interactions of these compounds with the GaAs surface. We have investigated the adsorption and reactivity of trimethylarsenic (TMAs) on GaAs(100) using temperature programmed desorption (TPD), Auger electron spectroscopy, and LEED. For the exposures and temperatures studied, TMAs did not pyrolytically decompose on the GaAs(100). TPD results indicate that TMAs chemisorbs, apparently non-dissociatively, and desorbs ≅330 K. Multilayers of TMAs desorb ≅140–160 K. Exposure of adsorbed TMAs to 70 eV electrons results in irreversible decomposition of the molecule. After electron irradiation, TPD shows that methyl radicals desorb at 660 K, which corresponds to a desorption activation energy of ≅40 kcal/mol. At higher temperatures, As2, H2, C2H2, and a smaller amount of methyl radicals desorb, and a small coverage of carbon remains on the surface.

Despite their importance, the detailed surface reactions and rearrangements which occur during chemical vapor deposition remain largely undetermined because of the lack of suitable experimental probes. In principle, x-ray scattering and spectroscopy techniques are well suited to studying these near atmospheric pressure processes but advances in this area have been limited both by the lack of suitable x-ray sources and by the difficulty of integrating the growth and measurement experiments. We have developed equipment and techniques to perform in situ x-ray scattering studies of the structure of surfaces during organometallic vapor phase epitaxial (OMVPE) growth using the extremely bright undulator radiation from the PEP electron storage ring. In this paper, we describe our initial experimental results studying cleaning and subsequent reconstruction of GaAs (001) surfaces in a flowing H2 ambient. These results demonstrate the excellent surface sensitivity, low background and high signal levels necessary to study the dynamic processes associated with semiconductor growth using OMVPE.

A secondary ion mass spectroscopy study of Ga diffusion in CdTe grown on (001) GaAs was carried out. The layers were grown by pulsed laser evaporation and epitaxy. Two characteristic regions with increased Ga concentration were found. The first was the CdTe/GaAs interface where the concentration of Ga decreases rapidly to the detection limit of −8 × 1014 cm−3. This region was usually less than 300 nm wide. The second was a surface region with a Ga accumulation of up to −1017 cm−3. Ion imaging revealed that in the (001) CdTe layers, Ga accumulates near the surface at localized spots, up to about 8 μm in diameter. This feature is less apparent in the (111) CdTe layers. Annealing at 500 °C for 1 h increased the Ga concentration in the whole layer to above 1016 cm−3. We also observe thermal annealing induced precipitation of Ga at the surface of bulk CdTe samples which were originally uniformly doped with Ga.