Low-Temperature Metal-Organic Chemical Vapor Deposition of Silicon Nitride

Amorphous silicon nitride films have been deposited on single-crystal silicon from the gas mixture of methylsilazane and ammonia at 873 to 1073 K. The films have been characterized by ellipsometry, Fourier transform infrared spectroscopy, and Auger electron spectroscopy. The Si-C, Si-H, and C-H bonds in methylsilazane can be effectively cleaved and the associated C and H species removed. The structure and composition of the films do not show any apparent dependence on the deposition temperature.     

Coal Slag Protection of Silicon Carbide with Chemically Vapor Deposited Mullite Coatings

The corrosion characteristics of bulk alumina, SiC, mullite, and CVD mullite coatings on SiC in contact with coal slag were investigated. Uncoated SiC corroded in the presence of coal slag, forming mixed FeSi phases and carbon. Bulk Al₂O₃ and slag formed a diffusional phase believed to be the spinel hercynite (Fe,Mg)O·(Al,Fe)₂O₃. After exposure to coal slag, a compositional difference was observed at bulk mullite/coal slag interfaces, yet this diffusional phase did not appreciably degrade the mullite samples and no cracking was observed. CVD mullite coatings offered protection to SiC in a simulated coal gasification atmosphere with corrosion protection dependent on the uniformity of the coating. Microprobe analysis of the CVD mullite coating/slag interface showed the formation of a Fe(Mg)Al.     

Chemical Vapor Deposition for Silicon Cladding on Advanced Ceramics

Polycrystalline Si was used to clad several advanced ceramic materials such as SiC, Si₃N₄, sapphire, Al₂O₃, pyrolytic BN, and Si by a chemical vapor deposition (CVD) process. The thickness of Si cladding ranged from 0.025 to 3.0 mm. CVD Si adhered quite well to all the above materials except Al₂O₃, where the Si cladding was highly stressed and cracked or delaminated. A detailed material characterization of Si-clad SiC samples showed that Si adherence to SiC does not depend much on the substrate surface preparation; that the thermal cycling and polishing of the samples do not cause delamination; and that, in four-point bend tests, the Si–SiC bond remains intact, with the failure occurring in the Si.     

Thermodynamic Calculations on the Chemical Vapor Deposition of Silicon Nitride and Silicon from Silane and Chlorinated Silanes

After a discussion of the thermochemical values of the Si–H–Cl–N system which occur in the literature, CVD phase diagrams are presented which include contours of constant deposition efficiency. The temperature range considered is from 800 to 2600 K. A number of chlorinated silanes as well as silane can be used as a silicon source, while ammonia is used as the nitrogen source. The effects of pressure variation and dilution by nitrogen and hydrogen are also included. Some initial calculations concerning silicon diimide are made. The CVD phase diagrams are used to describe several mechanisms occurring during the formation of silicon nitride from the gas phase.     

Surface Pretreatments of Silicon Nitride for CVD Diamond Deposition

The efficiency of different surface pretreatments (four standard chemical etchings and four diamond powder abrasive procedures) on silicon nitride (Si₃N₄) substrates for chemical vapor deposition (CVD) of diamond has been systematically investigated. Blank Si₃N₄ samples were polished with colloidal silica (∼0.25 μm). Diamond nucleation and growth runs were conducted in a microwave plasma chemical vapor deposition apparatus for 10 min and 6 h, respectively. Superior results concerning nucleation density ( N _d∼ 10¹⁰ cm⁻² after 10 min), film uniformity, and grain size (below 2 μm after 6 h) were obtained for the mechanically microflawed samples, revealing that chemical etchings (hot and cold strong acids, molten base or CF₄ plasma) are not crucial for good CVD diamond quality on Si₃N₄.     

Simultaneous Chemical Vapor Deposition of Boron Nitride and Aluminum Nitride

Two chemically different phases, hexagonal BN and AIN, were simultaneously produced by chemical vapor deposition (CVD) using an impinging jet reactor and the BCl₃─AlCl₃─NH₃─Ar reagent system. The microstructure of the BN + AIN composite coatings was strongly dependent on temperature, pressure, and BCl₃ and AlCl₃ concentrations. The growth characteristics of BN and AIN in the codeposition system were similar to those expected from the single-phase deposition processes (i.e., BN-CVD and AIN-CVD), except the growth of AIN whiskers was accentuated, and competition between BN and AIN deposition in the composites was suspected to be the cause of less-crystalline deposits. In both BN + AIN-CVD and AIN-CVD, the growth of AIN whiskers became more apparent with increasing pressure or temperature. The codeposition behavior observed experimentally was compared with thermodynamic predictions.     

Oxidation of Chemically-Vapor-Deposited Silicon Nitride and Slnglem-Crystal Silicon

Oxidation of {111} single-crystal silicon and dense, chemically-vapor-deposited silicon nitride was done in clean silica tubes at temperatures of 1000° to woo°C. The oxidation rates of silicon nitride under various atmospheres (dry O₂, wet O₂, wet inert gas, and steam) were several orders of magnitude slower than those of silicon under the identical conditions. The activation energy for the oxidation of silicon nitride decreased from 330 to 259 kJ/mol in going from dry O₂ to steam while that for Si decreased from 120 to 94 kJ/mol. The parabolic rate constant for Si increased linearly as the water vapor pressure increased. However, the parabolic rate constant for silicon nitride showed nonlinear dependency on the water vapor pressure in the presence of oxygen. The oxidation kinetics of silicon nitride is explained by the formation of nitrogen compounds (NO and NH₃) at the reaction interface and the counterpermeation of these reaction products.     

Kinetic Analysis of Chemical Vapor Deposition of Boron Nitride

BN was deposited on Al₂O₃ substrates from the BCl₃─NH₃─Ar reagent system. An impinging jet reactor configuration was used to obtain kinetic data in the temperature range of 800° to 1000°C and in the pressure range of 4 to 20 kPa. The BN deposition could be described by a simple kinetic rate expression with an activation energy of about 39 kcal/mol and a first-order dependency on BCl₃ concentration. The BN deposition rate increased with temperature, pressure, and BCl₃ concentration. The microstructure of the BN coatings was not strongly influenced by the process parameters.     

Thermodynamic and Experimental Study of High-Purity Aluminum Nitride Formation from Aluminum Chloride by Chemical Vapor Deposition

Thermodynamic calculations in the systems Al-Cl, Al-Cl-N, and H-Al-Cl-N were used to assess the capabilities of AlCl₃ or mixtures of AlCl₃ with Al to produce AlN by chemical vapor deposition (CVD) techniques. Direct nitridation (N₂ as reaction agent) is possible only at high temperatures (≥1500 K), using AlCl₃–Al mixtures. Reaction with NH₃ at equilibrium gives low yields but the suppression of NH₃ dissociation yields near 100%, which makes the method suitable for powder production, coating, and single-crystal growth. AlN with less than 1 wt% oxygen was obtained from technical grade AlCl₃ by this process. The formation of both amorphous AlN powder and crystalline AlN coatings was observed. It is assumed that the formation of AlCl₃· x NH₃ adducts by mixing of Al-Cl vapor and NH₃ at temperatures ≤1273 K prevents NH₃ dissociation and favors the production of amorphous AlN.     

Growth Mechanism of Silicon Carbide Films by Chemical Vapor Deposition below 1273

SiC films were prepared from the reaction of Si₂H₆ with C₂H₄ or C₂H₂ at relatively low temperatures ranging from 873 K to 1273 K by low-pressure chemical vapor deposition. The deposition rate profiles determined by gravimetry and the alloy composition (carbon content, x, for Si_1-xC_x) profiles determined by X-ray photoemission spectroscopy in the reactor were mainly investigated. The results revealed that the carbon content, x , was a function of the temperature, ratio of partial pressures of source gases, and source gas species (C₂H₄, C₂H₂). From these results we deduced that SiC formation occurred by two major competing reaction processes: (1) the silicon deposition processes, SiH₂ Si (wall) and Si₂H₆ Si (wall), and (2) the carbon deposition process, C₂H₄+ SiH₂ vinylsilane or C₂H₂+ SiH₂ ethynylsilane.     

Modeling of Laser-Induced Chemical Vapor Deposition of Silicon Carbide Rods from Tetramethylsilane

Finite-difference fluid-dynamics modeling has been used to predict deposition rates, fractional amounts of phases, and deposition morphology for the codeposition of silicon carbide and pyrolitic carbon from tetramethylsilane via laser-induced chemical vapor deposition (LCVD). Calculated results agree fairly well with rod deposition experiments. The morphologic features of rods that have been grown using LCVD are examined and explained using the results of the finite-difference calculations.     

CVD Mullite Coatings in High-Temperature, High-Pressure Air–H2O

Crystalline mullite was deposited by chemical vapor deposition (CVD) onto SiC/SiC composites overlaid with CVD SiC. Specimens were exposed to isothermal oxidation tests in high-pressure air + H₂O at 1200°C. Unprotected CVD SiC formed silica scales with a dense amorphous inner layer and a thick, porous, outer layer of cristobalite. Thin coatings (∼2 μm) of dense CVD mullite effectively suppressed the rapid oxidation of CVD SiC. No microstructural evidence of mullite volatility was observed under these temperature, pressure, and low-flow-rate conditions. Results of this preliminary study indicate that dense, crystalline, high-purity CVD mullite is stable and protective in low-velocity, high-pressure, moisture-containing environments.     

Effects of Thin Mullite Coating on the Environmental Stability of Sintered Si3N4

We investigated the effects of a chemically-vapor-deposited mullite coating (∼100 nm) on the oxidation resistance of sintered Si₃N₄ in air and steam environments. The coating was sacrificially incorporated into the thermally grown oxide (TGO) on Si₃N₄ during isothermal oxidation in air at 1400°C, leading to significantly reduced TGO growth as well as markedly improved TGO morphology. This improvement can be attributed to the refractory and viscous nature of the SiO₂-Al₂O₃ system, compared with SiO₂, when under the influence of alkali and/or alkaline-earth fluxing elements. However, the mullite coating had little effect on the stability of the ceramic in the steam environment at 1200°C, due likely to high activity of SiO₂ in mullite.     

Low-Temperature Chemical Vapor Deposition Tungsten Carbide Coatings for Wear/Erosion Resistance

A new chemical vapor deposition (CVD) process has been developed to deposit hard coatings, containing tungsten carbide, at temperatures below 500°C. These coatings, which have been applied to both ferrous and nonferrous alloys, exhibit excellent resistance to wear and erosion. The coatings comprise a mixture of tungsten and the tungsten carbide, the latter being present as W₂C, W₂C + W₃C, or W₃C. The coatings' composition and properties can be controlled by varying the CVD process parameters. The unique lamellar, fine-grained microstructures of these coatings contribute to their good tribological properties.     

Formation and Characterization of Amorphous Aluminum Nitride Powder and Transparent Aluminum Nitride Film by Chemical Vapor Deposition

Based on thermodynamic predictions that, in the gas-phase system AlCl₃/NH₃, AlN can be produced up to the theoretical yield, the performance of a chemical vapor deposition (CVD) reactor suitable for continuous powder production is investigated. Under the applied reaction conditions and reactor configuration, fine spherical AlN powders and transparent AlN films could be preferentially obtained. The generated amorphous CVD AlN powder is characterized by a BET surface area of 23.5 m²/g. The powder deposition rate was determined to be 3.5 g/h.     

Kinetic and Thermodynamic Analyses of Chemical Vapor Deposition of Aluminum Nitride

AIN coatings were prepared by chemical vapor deposition from the AlCl₃—NH₃—Ar reagent system using an impinging jet reactor in the temperature range of 700° to 1100°C. A mass transfer model and thermodynamic calculations were used to analyze the deposition data. The AIN-CVD process could be approximated by calculating mass transfer—thermodynamic limits at low AlCl₃ concentrations. The AIN deposition rate decreased drastically with increasing temperature above 1000°C in agreement with thermodynamic predictions. At high AlCl₃ concentrations, a surface kinetic mechanism involving AlCl₃ adsorbed on the deposition surface appeared to be the rate-limiting step. The AIN deposition rate decreased on increasing the AlCl₃ concentration or total pressure. The crystalline structure of AIN was strongly influenced by the processing parameters. The AIN coatings became highly crystalline and preferentially oriented with an increase in the AlCl₃ concentration or pressure.     

等离子体增强化学气相沉积法制备氢化硅膜的自晶化倾向性

为了探寻生长过程中硅膜的自晶化沉积,采用等离子体增强化学气相沉积(PECVD)法沉积了氢化硅薄膜,系统研究了不同沉积阶段所得硅膜微观结构的迁变规律。结果表明,硅膜的显微结构依赖于沉积时间,当沉积时间仅为30min时,所得硅膜的结构为非晶;而当沉积时间延至60min时,硅膜形成微晶颗粒;此后随着沉积时间的增加,晶化程度提高,且非晶区域面积相应减小。另外,硅膜的沉积速率也随沉积时间的增加而增加。在硅膜沉积过程中,随时间不断变化的界面状态可能为其自晶化的主要原因。     

Preparation of Dispersed Phase Ceramic Boron Nitride and Aluminum Nitride Composite Coatings by Chemical Vapor Deposition

Dispersed phase ceramic composite coatings containing BN and AIN were prepared by chemical vapor deposition. The BN + AIN coatings were deposited on Al₂O₃ from the BCI₃–AICI₃-NH₃ reagent system in the temperature range of 700° to 1200°C. Also, single-phase BN and AIN coatings were prepared for comparison purposes. The composite coatings consisted of very small BN and AIN regions which were either crystalline or amorphous, depending on deposition temperature and reagent concentrations. For example, a composite containing AlN whiskers of less than 100-nm diameter in a matrix of turbostratic BN of 2-nm grain size was deposited at 1100°C and 20 kPa.     

Thermodynamic Analysis of Chemical Vapor Deposition of BN + AIN Composite Coatings

Thermodynamic calculations were performed using a modified solgasmix-pv computer program in order to study the feasibility of codepositing boron nitride (BN) plus aluminum nitride (AIN) by chemical vapor deposition. Reactants considered were AICl₃, BCl₃ or B₂H₆, NH₃, and H₂. Deposition diagrams were generated for the BCl₃-AICl₃-NH₃ system over a range of processing conditions such as temperature, total system pressure, and reagent concentrations. Codeposition of BN + AIN was predicted by the calculations for temperatures in the range of 900 to 1700 K and pressures of 10.13 to 101.3 kPa. The predicted deposition efficiency at equilibrium was much higher for BN than for AlN at most reagent compositions. The AlN deposition efficiency increased with decreasing temperature and decreasing BCl₃ content, with increasing NH₃ content, or with the addition of H₂. Aluminum chlorides were found to be the dominant gaseous species.     

Fracture Strength of Plate and Tubular Forms of Monolithic Silicon Carbide Produced by Chemical Vapor Deposition

The fracture strength of silicon carbide (SiC) plate deposits produced by chemical vapor deposition (CVD) was determined from room temperature to 1500°C using a standard 4-point flexural test method (ASTM C1161). CVD SiC materials produced by two different manufacturers are shown to have only slightly different flexural strength values, which appear to result from differences in microstructure. Although CVD deposition of SiC results in a textured grain structure, the flexural strength was shown to be independent of the CVD growth direction. The orientation of machining marks was shown to have the most significant influence on flexural strength, as expected. The fracture strength of tubular forms of SiC produced by CVD deposition directly onto a mandrel was comparable to flexural bars machined from a plate deposit. The tubular (O-ring) specimens were much smaller in volume than the flexural bars, and higher strength values are predicted based on Weibull statistical theory for the O-ring specimens. Differences in microstructure between the plate deposits and deposits made on a mandrel result in different flaw distributions and comparable strength values for the flexural bar and O-ring specimens. These results indicate that compression testing of O-rings provides a more accurate strength measurement for tubular product forms of SiC due to more representative flaw distributions.