Impact damage behaviour of CVD-coated silicon nitride for gas turbines

Impact tests were conducted on the silicon nitride substrates coated with Si₃N₄ and SiC by chemical vapour deposition (CVD). For both 100- and 200-m-thick Si₃N₄-coated silicon nitride, Hertzian crack extension was reduced by debonding at the interface. Although Hertzian crack extension was not reduced for 100-m-thick SiC-coated silicon nitride, it was reduced for 200-m-thick SiC-coated silicon nitride. Theoretical calculations suggest that debonding at the interface consumed the fracture energy of Hertzian crack extension in the case of Si₃N₄ coatings, but it was observed that Hertzian cracks were not arrested at the interface.     

Pulsed laser deposition of silicon nitride thin films by laser ablation of a Si target in low pressure ammonia

We report the deposition of Si-N films by multipulse excimer laser ( = 308 nm, _FWHM = 30 ns) ablation of Si wafers placed in a slow flow of NH₃ in the pressure range (1 bar-1 mbar). The films are deposited on to a Si collector placed parallel to the Si target. We succeeded in depositing pure amorphous Si₃N₄ films at a pressure of 1 mbar of NH₃. The deposition rate reached a maximum value of 0.2–0.3 nm per pulse. At lower pressures, the deposited films consist of a fine mixture of three amorphous phases (amorphous stoichiometric silicon nitride, amorphous non-stoichiometric silicon nitride and amorphous silicon). The amorphous silicon is prevalent in films deposited at a pressure of several to several tens of bars. Droplets of polycrystalline -Si are sometimes visible on the film surface. The experimental evidence, is analysed with a view to elucidating the participation in the chemical synthesis of the three main stages of the process: the substance expulsion from the target by laser ablation, the transition through the gas of the expulsed substance and it's final impact on the collector. We conclude that silicon nitride is mostly synthesized during the impact on the collector of the flow of the ablated substance.     

Silicon nitride crystal structure and observations of lattice defects

In view of the considerable progress that has been made over the last 40 years on the microstructural design of silicon nitride and related materials of tailored properties for specific applications, a clear review of the current understanding of the crystal structure and crystal chemistry of silicon nitride is timely. The crystal structures, crystal chemistry, and lattice defect nature of silicon nitride are critically reviewed and discussed, with emphasis placed firstly on the structural nature of -silicon nitride (whether it is a pure silicon nitride, or should better be regarded as an oxynitride); and secondly on the space group of -silicon nitride (whether it is P6₃/m or P6₃). In conjunction with recent observations of vacancy clusters in -silicon nitride, a comprehensive view compatible with all the experimental facts with respect to the structural nature of -silicon nitride is tentatively presented.     

α-β phase transformation of silicon nitride — computer simulation

The simulation model of - silicon nitride phase transformation was developed for the case when the crystallization of the phase is the rate controlling step. The results gained on the base of the present model indicate that the temperature and total free surface area of silicon nitride phase present in the firing body limit the rate of transformation and total amount of transformed silicon nitride phase. These results are acceptable from the point of view of experimental experience and support the applicability of the presented model.     

Synthesis and characterization of silicon nitride whiskers

Silicon nitride whiskers were synthesized by the carbothermal reduction of silica under nitrogen gas flow. The formation of silicon nitride whiskers occurs through a gas-phase reaction, 3SiO(g)+3CO(g)+2N₂(g)=Si₃N₄()+3CO₂(g), and the VS mechanism. The generation of SiO gas was enhanced by the application of a halide bath. Various nitrogen flow rates resulted in different whisker yields and morphologies. A suitable gas composition range of N₂, SiO and O₂ is necessary to make silicon nitride stable and grow in a whisker form. The oxygen partial pressure of the gas phase was measured by an oxygen sensor and the gas phase was analysed for CO/CO₂ by gas chromatography. Silicon nitride was first formed as a granule, typically a polycrystalline, and then grown as a single crystal whisker from the {1 0 0} plane of the granule along the 210 direction. The whiskers were identified as-sialon with Z value ranging from 0.8 to 1.1, determined by lattice parameter measurements.     

The development of strength in reaction sintered silicon nitride

The development of strength in reaction sintered silicon nitride has been investigated by determining the elastic moduli, fracture mechanics parameters, strengths and critical defect sizes of silicon compacts reacted to various degrees of conversion using static or flowing nitrogen. The relationship between each property and the nitrided density is shown to be independent of the green silicon compact density but is influenced by the nitriding conditions employed. Young's moduli, rigidity moduli and strengths vary linearly with the nitrided density. After an initial period when increases may occur, the critical defect sizes in both static and flow materials decrease continuously with increasing nitrided density, although at any particular density they are larger in material produced under flow conditions. A model is suggested for the development of the structure of reaction sintered silicon nitride involving the development of a continuous silicon nitride network within the pore space of the original silicon compact. The experimental data are discussed in terms of the proportion of silicon nitride which contributes effectively to the continuous network.     

Some factors influencing the formation of reaction-bonded silicon nitride

Several salient factors influencing the formation of reaction-bonded silicon nitride (RBSN) compacts have been studied. These include the effects of mullite and alumina furnace tubes typically employed during high-purity nitridation studies, pre-sintering of green silicon compacts, free powder versus compact nitridation, and the influence of metal/metal oxide additions. The latter studies have provided experimental evidence for enhancement due to dissociated nitrogen, and suggest that (1) -Si₃N₄ formation does not necessarily require a liquid phase, (2) atomic nitrogen stimulates -phase formation, and (3) the liquid phase provides an efficient source for volatile silicon, promoting -Si₃N₄. These conclusions are consistent with accepted mechanisms for the formation of the two phases.     

Thermal conductivity of unidirectionally oriented Si3N4w/Si3N4 composites

-silicon nitride whiskers were aligned unidirectionally in silicon nitride sintered with 2 wt% Al₂O₃ and 6 wt% Y₂O₃. It was be densified by the Gas Pressure Sintering (GPS) method. Thermal conductivity of the sintered body with different amount of - silicon nitride whiskers was measured by the direct contact method from 298 K to 373 K. This unidirectionally oriented -silicon nitride whiskers grew into the large elongated grains, and improved also the thermal conductivity. The amount of -silicon nitride whiskers changed the microstrcuture, which changed the thermal conductivity.     

Factors influencing the crystallization of ultrafine plasma-synthesized silicon nitride as a single powder and in composite SiC-Si3N4 powder

A study of various parameters affecting the crystallization process of amorphous silicon nitride produced by plasma gas-phase reaction was undertaken to determine the conditions under which whiskers are formed. This process is influenced by the ammonium chloride content of the starting powder and the presence of nitrogen in the furnace atmosphere. This last parameter is also influential on the / phase ratio, along with other factors like the silica content, temperature and duration of the thermal treatment. Heat treatment at 1500°C for 30 min under argon produced well-defined -Si₃N₄ crystals with a hexagonal cross section, a mean length around 0.8 m, and no sign of agglomeration. Under the same conditions, crystallization of silicon nitride in SiC-Si₃N₄ composite did not give crystals, but Si₃N₄ whiskers. Therefore silicon carbide plays a major role in their formation.     

Sepiolite-PAN intercalation used as Si3N4 forming precursor

Silicon nitride (Si₃N₄) formation was investigated using sepiolite and polyacrylonitrile as silicon and carbon source, respectively. It was found that purified sepiolite could readily adsorb a sufficient amount of acrylonitrile (AN) reagent without pre-treatment. Polymerisation of sepiolite-AN mixtures and subsequent cyclization of the polymerised complex yielded a precursor, which was found ideal as a starting material in carbothermal reduction-nitridation (CRN) for the formation of silicon nitride powders. The quantity of silicon nitride, grain size and morphology were found to be dependent on the reaction conditions and starting reagent. Fine grain size, high surface area (up to approx. 12.4 m²/g) powders of mainly -Si₃N₄ were obtained via pyrolysis of sepiolite-PAN complex after 4 h heating at 1400°C in 1000 ml/min nitrogen flow with a heating rate of 300°C/h. Mg retained in the molecular structure of the mineral must have promoted the formation of -grains in CRN process.     

Effects of promoters on the product quality of nanophase SiC/α-Si3N4 composite powders synthesized through carbothermal reduction nitridation

The effect of additives is investigated for the carbothermal reduction synthesis of nanophase silicon carbide/silicon nitride composite powders. Mixtures of silica, carbon, seed silicon nitride, and additive are reacted in a thermogravimetric analyzer. The mass loss information combined with compositional and spectroscopic analysis allows product quality (morphology, surface area, -Si₃N₄ and -SiC contents, oxygen content, etc.) information to be obtained. It was observed that all of the additives used in this study increased the reaction rate. Lithium carbonate produced a silicon nitride/silicon carbide composite that was not significantly different from experiments without promoter. However, the product quality was severely affected in other instances.     

Structural and optical properties of microcrystalline silicon films deposited by plasma enhanced chemical vapour deposition

Hydrogenated microcrystalline (c) silicon films were prepared by plasma enhanced chemical vapour deposition using an Ar-diluted SiH₄ gas at various deposition conditions. The substrate temperature and RF power were varied from 150 to 400 C and from 10 to 120 W, respectively. Structure and microstructure were examined by X-ray diffraction, Raman spectroscopy, and scanning electron microscopy. Hydrogen bonding and optical properties were investigated by FTIR spectra and UV transmission spectra. The crystal fraction of the films increased as the deposition temperature decreased and RF power increased. More definite columnar morphology was developed with increasing crystal fraction. The existence of c-Si above a critical RF power (>30 W) suggests that SiH₂ radical in plasma plays an important role for the formation of columnar morphology and c-Si. IR absorption analysis showed that the SiH₂/SiH bonding ratio in the silicon films increased as the crystal fraction increased. The UV absorption coefficient of the films became smaller as the deposition temperature and RF power increased.     

Low-temperature plasma-enhanced chemical vapor deposition of tungsten and tungsten nitride

Tungsten and tungsten nitride layers have been deposited by plasma-enhanced chemical vapor deposition (PECVD). Tungsten layers deposited at low deposition temperatures T150 °C using this method showed good uniformity over dielectric and silicon substrate areas. As the deposition temperature decreased, the silicon consumed during the deposition reaction decreased, at T150 °C no silicon consumption was measurable. PECVD tungsten nitride layers were deposited directly on oxidized silicon substrates with no requirement for a nucleation layer. As the NH₃ flow rate was increased, whilst maintaining all other parameters constant, deposited layers were found to change from metal tungsten to tungsten-rich amorphous layer to W₂N. The resistivity of the layers was found to be high compared to published literature for higher-temperature deposited layers. The high resistivity is attributed to the incorporation of fluorine into the layer at low deposition temperatures. A deposition process was established for smooth amorphous tungsten-rich W _x N layers at 150 °C.     

Reaction sintered silicon nitride

The strength, composition and structure of reaction sintered silicon nitride formed in the presence of oxygen or water vapour have been investigated. It is shown that chemical contamination of the nitriding gas by oxygen and water vapour to relatively high concentrations is not the cause of abnormally low strength silicon nitride. The coefficient of variation in strength and apparent crystallite size for a particular batch of samples are related to the water vapour concentration in the nitriding gas. The formation of - and -silicon nitride is discussed and it is suggested that compositional variations in -silicon nitride must be small.     

Deposition characteristics and properties of iron nitride films by CVD using organometallic compound

Iron nitride films were prepared by chemical vapour deposition from the gas mixture of Fe(C₅H₅)₂-NH₃-H₂-CO₂. The effects of deposition parameters on the deposition characteristics were investigated. Iron nitride films were deposited above 500 ° C and the films of -Fe₄N single phase were deposited above 700 ° C. At 700 ° C and under the total gas flow rate from 1 to 8 l min^–1, the deposition rate of the film may be controlled by the transport of Fe(C₅H₅)₂ molecules to the surface of the deposits. At 700 °C and under the total gas flow rate of 4 l min^–1, the phases and nitrogen contents of the films were determined bypNH₃/pH₂ ^3/2, the controlling factor of the nitrogen contents of the films. Decreasing of the total gas flow rate and increasingpCO₂ increased the nitrogen contents of the films and phases with higher nitrogen were deposited. On the other hand, increasingpFe(C₅H₅)₂ and the absence ofpCO₂ increases the carbon contents of the films, and the phase with a greater solubility in carbon, i.e. -Fe₂N, was codeposited with -Fe₄N. The saturation magnetization of the films deposited at 700 ° C was in good agreement with that reported for the bulk iron nitride, which depended not on the deposition conditions but on the nitrogen contents of the films.     

The α- to β-Si3N4 transformation in the presence of liquid silicon

The compacts consisted of , -Si₃N₄ and free silicon are heat treated in the range 1650° C to 1750° C in an argon atmosphere in order to observe the following behaviours; the to phase transformation and variations of the microstructure during heat treatment in silicon nitride. For the microstructural observation of the heat treated specimens, the same grains in the polished surface were investigated before and after eliminating the retained silicon by etching. The to phase transformation, in this case, occurs via silicon melts irrespective of added -Si₃N₄. Both and phases are soluted and precipitated into molten silicon and their morphology are changed from an equiaxed shape to prismatic one. Although elongated grains are precipitated at low temperature or in the early stage of heat treatment, fine precipitated grains are mainly observed with increasing heat treating temperature.     

Synthesis of silicon nitride powder through nitrogen gas atomization

The feasibility of synthesizing silicon nitride powder utilizing reactive atomization processing was analysed. The range of times required for the flight time of particles, the cooling rate of the silicon melt, the reaction time of silicon and nitrogen, and the diffusion of nitrogen through silicon nitride layers were obtained and compared. The results of this study indicated that the production of silicon nitride powder through the reactive atomization process would be limited by diffusion of nitrogen through the nitride (ash) layer, assuming the nitride layer was coherent and the unreacted core model was a valid representation of the liquid silicon-silicon nitride system.Nomenclature k(T) reaction rate constant at temperature, T(s^–1) - k ₀ Arrhenius coefficient - E activation energy (kJ mol^–1) - R gas constant - T temperature (K) - fraction of normalized conversion of -phase in time t - fraction of normalized conversion of -phase in time t - k reaction rate constant for -phase (s^–1) - k reaction rate constant for -phase (s^–1) - k ⁱ intrinsic first-order rate constant for -phase (s^–1) - x conversion fraction of -phase in time t - x conversion fraction of -phase in time t - n reaction order for -phase = 1 - n reaction order for -phase = 0.5 - J diffusion flux (mol m^–2 s^–1) - D diffusivity, or diffusion coefficient (m² s^–1 or cm² s^–1) - dC change in concentration (mol m^–3) - dl change in distance, l (m) - A(_g) gaseous reactant A - B reactant B (may be solid or liquid) - P solid product P - b stoichiometric coefficient of reactant B - p stoichiometric coefficient of product P - t time of reaction passed (s) - time for complete reaction of a particle (s) - X _B conversion fraction - r _c core radius (m) - R _p particle radius (m) - _B molar density of reactant B (mol m^–3) - k _g mass transfer coefficient between fluid and particle (m s^–1) - C _Ag concentration of gaseous reactant A (mol m^–3) - D _e effective diffusion coefficient of gaseous reactant in ash layer (m² s^–1)     

Effect of selected impurities on the high temperature mechanical properties of hot-pressed silicon nitride

The effect of impurities on the high temperature mechanical properties of hot-pressed silicon nitride has been determined. Selected impurity additions were made to both relatively pure -phase and -phase silicon nitride starting powders. These powders were hot-pressed to full density using 5 wt % MgO as the pressive additive. The silicon nitride hot-pressed from the -phase powder exhibited higher strength at both 25 and 1400 C than that fabricated from the -phase powder. The impurity additions had no effect on the room temperature mechanical properties. The CaO additions had the most significant effect on the high temperature mechanical properties. In both the material hot-pressed from the -phase and -phase powders, increasing CaO additions severely reduced the high temperature strength and increased the amount of non-elastic deformation observed prior to fracture. Although alkali additions (Na₂CO₃, Li₂CO₃, K₂CO₃) also tended to have the same effects as the CaO, the high volatility of these compounds resulted in a much reduced concentration in the hot-pressed material, thus minimizing somewhat their tendency to enhance the high temperature strength degradation. The Fe₂O₃ and Al₂O₃ had no apparent effect on the high temperature mechanical properties.     

Analysis of the α to β phase transformation in Y-Si-Al-O-N ceramics

The silicon nitride- sialon phase transformation in the Y-Si-Al-O-N system was investigated by transmission and scanning electron microscopy and X-ray diffractometry using samples which contained up to 40% liquid-forming components at temperatures between 1500 and 1600 °C. Completely dense samples suitable for TEM analysis, in which the to transformation could be examined, were prepared using hold times as short as 5 min under a nominal uniaxial pressure followed by rapid cooling. Spheroidal, partially dissolved, silicon nitride grains, together with acicular grains of sialon, were observed in a glass phase containing a very low nitrogen content (undetectable by electron microprobe analysis). This absence of nitrogen build-up in the liquid phase between the dissolution and precipitation sites during the to transformation indicates that the diffusion of nitrogen through the liquid phase is extremely rapid. Nucleation of the sialon was almost entirely homogeneous and the unconstrained nature of the liquid environment resulted in growth of defect-free sialon grains with curved growth fronts perpendicular to the c-axis. The technique described allows direct observation of the effect of various additives on the to phase transformation.     

Development,preliminary testing and future applications of a rational correlation for the grain densities of vapour-deposited materials

It is conjectured and found in this work that the grain densities (suitably normalized) of vapour-deposited solid materials depend principally on competition between the successful arrival rate of their reagent molecules and the surface diffusion rate of admolecules on their growing surfaces. The ratio of these two rates defines an important dimensionless Damköhler number, called here the burial parameter, . Available grain density data for seven vapour deposited materials [silicon (Si), gallium arsenide (GaAs), silicon carbide (SiC), silicon nitride (Si₃N₄), titanium oxide (TiO₂), boron nitride (BN) and graphite (C)] are used to establish and test the universality of the proposed normalized grain density versus burial parameter correlation. As anticipated, these data show that the normalized grain densities of these materials increase with their corresponding burial parameters. Moreover, for estimated burial parameters much less than unity, the deposits formed are indeed reported to be amorphous, while the deposits are observed to be crystalline under conditions for which 1 is estimated. As the burial parameter decreases, the reported grain densities of turbostratic, layered, materials are found to decrease more gradually than for materials with no turbostratic structure. While the present implementation of this basic hypothesis cannot be regarded as complete, it is argued that a rationally-based, reasonably universal vapour deposit density correlation of this general form can be quite useful in making rational predictions of deposit quality. Moreover, it appears that this path to such mechanistically plausible correlations, which, using available experimental data, can be implemented/tested even in the absence of a complete theory, can be broadened to include other important deposit characteristics via the introduced of additional characteristic time ratios.