Silicon carbide fibre/silicon nitride matrix composites have been fabricated using the reaction bonded silicon nitride (RBSN) and sintered reaction bonded silicon nitride (SRBSN) processing routes. A filament winding and tape casting system was developed to produce sheets of parallel aligned fibres within a layer of green matrix ('prepreg') which were cut, stacked and hot pressed to form a plate. This was nitrided and (in the case of SRBSN matrix composites) hot pressed at 1700°C to density the matrix. The magnesia (MgO) and the yttria/alumina (Y2O3/AI2O3) additive SRBSN systems were investigated as matrices for ease of processing and compatibility with the matrix. The MgO additive Si3N4 matrix reacted with the outer carbon rich layer on the surface of the fibres, framing a reaction layer approx. 2pm in thickness. A reaction layer was also observed with the Y2O3/AI2O3 additive matrix, but was thinner (< 0.5um), and was identified as silicon carbide from the electron diffraction pattern. X-ray mapping in the SEM was used to investigate the spatial distribution of elements within the interface region to a resolution < lum, including light elements such as carbon. The 6wt%Y203/ 2wt%Al203 additive SRBSN system was chosen for more detailed investigation, and the majority of characterisation was performed using this composition. Oxidation of composite samples was carried out at temperatures between 1000°C and 1400°C for up to 1000 hours. Little damage was visible after 100 hours for all temperatures, corresponding to a relatively small drop in post oxidation bend strength. After 1000 hours at 1000°C both carbon rich outer layers and the central carbon core of the fibre were removed. Samples were severely oxidised after 1000 hours at 1400°C, having a glass layer on the outer surface and replacement of near surface fibre/matrix interfaces with glass. The post oxidation bend strengths for both conditions were approx.2/3 of the as fabricated strength. Less damage was observed after 1000 hours at 1200°C, and the post oxidation bend strength was higher than the 1000°C and 1400°C samples. Mechanical properties of the SRBSN matrix composite were investigated at room temperature and elevated temperatures (up to 1400°C). The average room temperature values for matrix cracking stress and ultimate strength (in bend) were 651.1 and 713.2 MPa respectively, with corresponding Weibull moduli of 5.7 and 8.7. The stresses are comparable to similar monolithic silicon nitrides. Room temperature tensile matrix cracking and ultimate strength were 232MPa and 413MPa, lower than the bend test results, which were attributed to bending stresses in the sample, lowering the apparent failure stresses. The samples failed in a composite like manner (i.e. controlled rather than catastrophic failure), with a substantially higher woric of fracture than monolithic materials. The average matrix cracking and ultimate bend strength at 1200°C were 516MPa and 554MPa, dropping to 178MPa and 486MPa at 1400°C (the matrix cracking stress was indistinct at 1400°C due to plasticity). The creep and stress rupture properties at 1300°C were investigated in four point bend, using dead-weight loading. The creep rate was KH/s at a stress of 200MPa, lower than a hot pressed silicon nitride with MgO additive, and higher than a hot isostatically pressed Y2O2/SÍO2 additive silicon nitride. A cavitation creep mechanism was deduced from the stress exponent, which was >1. Failure by stress rupture did not have a lower limit, which is also associated with cavitation of the amorphous grain boundary phase.

A range of composites have been prepared by dispersing various sources of SiC whiskers and platelets into a Si3N4 based matrix intended for high temperature applications. The main objective was to improve the fracture toughness of the Si3N4 monolith without any detrimental effect on other properties such as the excellent high temperature deformation behaviour. The matrix composition was tailored so that the intergranular phase formed desirable crystallisation products within the Si3N4 - Si2N20- Y2Si207 phase system. A slip casting technique was used to prepare the composites, in order to preferentially align the dispersoids, and conventional powder pressing methods were used to achieve a random distribution of dispersoids. Composites containing at least 30 weight % dispersoid content were fully densified using ABB Cerama's glass encapsulated HIP process. The resulting composites showed steadily increasing fracture toughness values with increasing SiC content, although the maximum increment was only approximately 30%. Platelet containing composites had slightly higher values than composites containing a similar volume fraction of whiskers, although there was a significantly lower fracture strength in the former. A microstructural survey of the materials indicated that the mechanical properties were influenced by the dispersoid /matrix interfacial characteristics. Fracture surfaces of the platelet containing composites showed there was significant debonding at the interface, whereas the whiskers appeared strongly bonded to the matrix. Whisker reinforced Si3N4 composites showed excellent creep resistance at 1400°C, which appeared slightly better than the monolith. Both materials had a similar stress exponent value close to 1, which suggests the deformation mechanisms in both these materials are similar and follow a grain boundary diffusional flow mechanism. There appeared to be a threshold stress level of 250 MPa at 1400°C for both materials where failure did not occur after prolonged deformation. This result and the absence of cavity formation wiihin the intergranular region suggests there is a sub- critical residual glass volume at this stress level for the nucleation of cavities and hence for slow crack growth. Both monolith and composites show good oxidation resistance at these elevated temperatures.

Melt-quenched glasses containing SiO2, CaO, Na2O and P2O5, and sol-gel derived glasses containing SiO2 and CaO are known to have bioactive properties. Foaming of binary sol-gel derived bioactive glasses containing SiO2 and CaO can be used to produce 3D porous scaffolds which mimic the structure of trabecular bone, increasing the potential for these glasses to be used as bioactive bone-regenerative materials. A range of experimental techniques have been used to investigate the atomic scale structure of these materials, and also to observe the reaction mechanisms which occur when these materials are immersed in a simulated physiological solution (simulated body fluid, SBF) and a standard cell culture medium (tris buffer solution, TBS). A robust structural model of the most bioactive of the melt-quenched glasses, namely Bioglass®, has been produced by combining high energy X-ray and neutron diffraction data, magic angle spinning nuclear magnetic resonance (MAS NMR) and reverse Monte Carlo (RMC) modelling. It has been shown that Ca clustering occurs in the glass, which is of direct relevance to the understanding of the facile nature of calcium within such glasses giving rise to its relatively rapid diffusion from the solid into solution. Hydroxyapatite has been confirmed as the calcium phosphate phase which grows on the surface of Bioglass® when immersed in the standard cell culture medium, TBS. A new method which can be used for in-situ time resolved high-energy X-ray diffraction studies of reaction mechanisms, such as those involved when a bioactive glass is immersed in a simulated physiological solution, is decribed in this thesis. Small-angle X-ray scattering has enabled the growth of mesopores to be observed during the foamed sol-gel stabilisation process. In-situ simultaneous small and wide angle X-ray scattering measurements of a foam in SBF have shown that the mesoporous network facilitates the rapid growth of relatively high-density HCA, which will therefore eventually replace the initial silicate glass as the material bounding the macropores. The data presented herein reveal the structure of highly important materials in the field of biomaterials and enable a link to be made between the atomic scale structure of the materials and their bioactive properties.

The thesis focuses on the analysis and determination of the structure of various metal doped phosphate glasses, which are of interest for their potential biomedical properties. The structures have been determined principally by X-ray and neutron diffraction but are also supported by complimentary X-ray absorption spectroscopy measurements and computational modelling. Such studies contribute to about half of the work presented in this thesis. Among the glasses of interest are silver-doped calcium sodium phosphates, which exhibit antimicrobial properties when the Ag+ ions are released over time in an aqueous environment. The advanced probe technique of neutron diffraction with isotropic substitution (NDIS) has been applied to elucidate the structural role of silver in these glasses. The results revealed that silver occupies a highly distorted octahedral environment analogous to that in crystalline Ag2SO4. Another glass study herein is associated with zinc titanium calcium sodium phosphate, which is biomedically interesting since the release of Zn2+ ions is shown to enhance cell attachment and proliferation. Structural analysis of multi-component glasses such as these tends to be difficult, but diffraction techniques and X-ray absorption spectroscopy have been used together to reveal the cation first neighbour coordination environments. The other significant element of the work presented here has been the development of data analysis techniques, with the emphasis on the creation of a program, which allows co-fitting of X-ray and neutron diffraction data of amorphous (and potentially crystalline) data. The code is written in MATLAB and makes use of the Nelder-Mead simplex method to minimise a set of “best guess” structural parameters supplied by the user. Extrema bound constraints are implemented by means of a sinusoidal parameter transform. Ultimately, the code is to be compiled and made available to users via the ISIS Pulsed Neutron Facility, UK.

ẞ +glass sialon materials are prepared with an Nd203 sintering additive. These materials exhibit similar property characteristics to equivalent ¥203-based compositions. With the absence of an Nd phase equivalent to YAG, phase relationships in the Nd-Si-Al-O-N system are examined to identify alternative grain boundary devitrification products. Materials in which the residual glass after sintering is recrystallised to give a mixture of Nd-N-O wollastonite and NdA103 are shown to exhibit high temperature property characteristics comparable to the conventional ẞ'+YAG sialon materials. The glass forming region in the Nd-sialon system is more extensive than for yttrium and new phases have been identified within the expanded volume. These also offer potential as fully crystalline matrix devitrification products. The most significant of these has the composition Nd3Si3Al30i2N2• B'«Nd3Si3Al30i2N2 materials are prepared and found to exhibit excellent properties up to * 1300°C. The preparation of pure a' and α +β’sialon materials by transient liquid phase sintering is particularly sensitive to starting composition. The fabrication of α +’β +glass materials allows greater flexibility but high temperature properties are still sensitive to the chemistry and concentration of the liquid sintering additives and to post sintering heat treatments. α +’β +glass materials are developed with a very minor amount of residual glass of a composition which allows full devitrification to form mainly YAG. Upon annealing, the a' species is particularly receptive to the non-stoichiometric elements which results in removal of the intergranular residual glass and subsequently increased solid/solid contact. To alleviate the interfacial energy anisotropy the YAG crystals are diffusively rearranged to an isolated equiaxed morphology. The limitations on high temperature use are mitigated by the discontinuous nature of this YAG phase. Component field trials and the potential of this new range of generic sialon materials are discussed.

Oxychloride glasses in the ternary system Sb2O3 - PbCl2 - ZnCl2 have been prepared. The thermal properties and the crystallisation behaviour of these glasses have been investigated by differential scanning calorimetry, viscometry, x-ray diffraction, scanning electron microscopy with energy dispersive x-ray analysis and infra-red spectroscopy. Thermal stability gaps, T1 - T4 up to 149°C are observed. The predominant crystal phases formed, depending on composition, are ZnCl2, Sb2O3 and a lead antimony oxychloride Sillan phase whose stoichiometry varies with glass composition from PbSb2O2.Cl2 to PbSb1O2.Cl. Relative crystal growth rates for the glasses have also been measured and values up to 0.16 μm sec−1 are observed. These values are found to be strongly dependent on the chlorine content of the glasses as are also the viscosity characteristics. The chemical durability of the glasses has also been studied under various conditions. Initial water dissolution rates between 10−2 to 10−4 gm cm day were observed. The corrosion rates were found to be chlorine content dependent and reaction mechanisms are suggested. The introduction of a heavy metal chloride such as BlCl2 or TlCl, as a fourth component, was found to increase the chlorine content without much affecting the thermal stability. However, an excessive amount of chlorine makes the glass less durable.

Ultrasonic velocity and attenuation measurements have been used to characterise a range of phosphate bonded, alumina filled, magnesia ceramics and other ceramic materials... Measurements were made over a range of frequency from 50kHz - 1 OM Hz, using a variety of commercial probes and equipment, and a variety of techniques. An ultrasonic double-probe method was used to monitor the setting process of the cementitious ceramics using commercial 2.25MHz and 2MHz transducers, for compressional and shear wave modes, respectively, in samples with alumina content in the range of 0 - 60 wt 0/0. The elastic properties of the material were determined from ultrasonic velocity measurements and were found to be dependent upon the filler volume fraction. The measured elastic moduli were found to Increase as porosity decreased, and this effect might possibly be used to estimate porosity. The composition dependence of the elastic moduli is compared with the Hashin and Shtrikman theoretical bounds for the elastic moduli of two-phase materials. All data lie between these bounds, suggesting that the alumina particles were well dispersed and well bonded to the matrix. However, the fact that the data are slightly above the lower bound suggested that the particles are not spherical, and this, together with other evidence obtained from an analysis of reaction rates, indicates the predominence of plate-like gram structures.

No description available.

Sliding wear of metal-on-ceramic, ceramic-on-metal, and ceramic-on-ceramic have been investigated using a tri-pin-on-disc machine. A technique has been developed for thin foil preparation for transmission electron microscopic examination perpendicular to the wear surface. The role of transformation toughening in the wear behaviour of zirconia ceramics has been investigated. In addition, the role of high strain deformation in a steel surface has been evaluated. The wear factor of 316L stainless steel pins worn against a zirconia disc was found to decrease as the load was increased, believed to be associated with metal oxide formation. TEM of the stainless steel revealed a worn surface which consisted of a mechanical mixture of metal oxide and heavily deformed metal. Deformation of the metal had occurred by shear banding with a microstructure similar to that observed in rolled specimens, although the texture formed was a wire texture rather than a rolling texture. The crystallite size was found to decrease towards the surface, demonstrating that the shear stress was a maximum at the surface. The shear bands at the surface had always been formed by the passage of the last asperity indicating that contact was plastic over the load range 6-60N/pin. The majority of wear occurred by transfer resulting from plastic overload, although a contribution to the material loss was made by metal extruded off the end of the pin as a result of the high strains. The depth of deformation correlated closely with the wear volume. The wear of the zirconia discs was found to be dominated by metal transfer. With Mg-PSZ, transformation occurred cooperatively in crystallographically determined bands. Microcrack coalescence led to preferential wear in these bands. However, with a Y-TZPdisc transformation appeared to have been responsible for widespread surface fracture. The wear of zirconia pins against a bearing steel disc gave limited metal transfer. Very little transformation of tetragonal to monoclinic was observed. However, milder forms of the transformation related wear mechanism did occur. Zirconia had formed a solid solution with the iron oxide, leading to the conclusion that the wear mechanism was tribochemically based. TZP worn against a ZTA disc showed evidence of very high temperature rises at the interface. The surface layer was amorphous and contained a mixture of alumina and zirconia suggesting that melting had occurred at the interface during sliding. At a depth of O.5pm. the surface consisted of heavily elongated tetragonal grains, with a low dislocation density, indicating a strain of at least 1.7. At a depth of 2-4pm a layer of monoclinic was found. There was evidence that the stresses imposed by friction extended to at least 8-10pm from the surface. TZP containing 20vol% SiC whiskers gave exceptionally low wear rates when worn against a ZTA disc. The greater wear resistance is believed to be a result of the improved load bearing capacity and of the higher thermal conductivity. It is clear that the poor thermal conductivity of zirconia dominates its tribological behaviour. Temperature generation was high enough to substantially reduce the driving force for transformation of the tetragonal to monoclinic, with a high enough temperature for plastic' deformation where a low thermal conductivity counterface was used. Where transformation occurred, its effect was to increase the wear rate.

No description available.