<p>Metal-matrix composites (MMCs) have potential as load bearing components for particular functional requirements. However, MMCs are more difficult to form than conventional metals and alloys due to the tendency of suffering microstructural damage at low plastic strains. It is thus of importance to understand the forms of microstructural damage, its level and spatial distribution. In this work, three main areas of investigation related to damage accumulation in MMCs were considered: the hydrostatic extrusion of MMCs, a theoretical investigation of scale effects in MMCs and plane strain compression experiments on model composites to study load transfer and fibre distribution effects on the patterns of plastic flow in the matrix. In the experiments on the hydrostatic extrusion of particulate MMCs, it was found that damage occurred in the form of reinforcement cracking, that the amount of damage varied linearly with applied plastic strain and that the rate of damage accumulation depended on the morphology of the reinforcement. The extrusion process was examined in regard to its ability to limit or suppress damage through the action of high compressive hydrostatic pressures. A slip-line field analysis was adopted to relate the effect of process parameters such as the amount of reduction and die angle to the magnitude and distribution of hydrostatic pressure in the extrusion dies. The linear relationship between damage and applied plastic strain was explained using a micro-mechanical description of the response of MMCs to plastic flow and the effect of microstructural variables such as volume fraction, shape and size of the reinforcement on the rate of damage accumulation was discussed. In addition, the tensile properties of MMCs subsequent to hydrostatic extrusion were quantified and the effect of previous extrusion was described in terms of the strain hardening of the matrix materials and the damage accumulation in the reinforcement materials and their influence on the stability of the plastic flow process. In the work on plane strain compression of model composites, the effect of reinforcement distribution on the pattern of flow in Cu-W MMCs was described using maps of equivalent strain obtained from the analysis of deformed surface grids. The level and distribution of strain was found to be related to the spatial distribution of the reinforcements. The fiber distribution also influenced the efficiency of the fibers as obstacles to macroscopic flow. Comparison of the results with numerical analyses allowed a link to be established between the observed damage mechanism of decohesion and the state of stress and local strain level in the composite. An optical method was used to get information on the load transfer process in a Cu-sapphire MMC plastically deformed in plane strain compression. The fracture of the sapphire fiber was related to the measured stress state. In the theoretical analysis of scale effects in MMCs, scale dependent bounds on strength were developed and scale dependent composite design charts were constructed. The results highlighted the importance of the scale of the reinforcement in MMC design, as well as the importance of the processing route to fabricate MMCs on a fine scale. The conclusions of the various areas of investigation demonstrated how damage in MMCs links together issues of process design, microstructural optimization and resultant mechanical properties of MMCs.</p> / Doctor of Philosophy (PhD)

<p>The typical microstructure of silicon nitride consists of rigid Si3 N4 grains and intergranular amorphous films associated with the liquid phase sintering process involved during densification. The presence of the amorphous films may affect the creep behaviour of silicon nitride ceramics at elevated temperatures. The advent of high resolution transmission electron microscopy (HRTEM) coupled with fine-probe chemistry analysis enables us to investigate the structure and chemical composition of the nano-scale grain boundary amorphous films and the role they play in creep deformation of silicon nitride. The materials investigated consist of β-Si3 N4 grains with and without secondary crystalline phases. All grains were covered with a thin intergranular amorphous film at both homophase and heterophase boundaries. It was found that these amorphous films have a characteristic value of thickness, independent of grain misorientation, but dependent on the chemical composition of the film and the grains on either side of the film. The creep behaviour of the materials were evaluated by compressive and tensile testing. The grain-boundary film thickness distribution was measured before and after creep using both high-resolution lattice imaging technique and Fresnel fringe imaging technique. The results show a narrow range of film widths in the uncrept material but a bimodal distribution after creep. This provides, for the first time, direct evidence for the occurrence of viscous flow of intergranular amorphous films during creep deformation of silicon nitride. Finally, a model is developed to describe the viscous flow process in multi-phase Si3 N4 materials in contrast to prior models which are only applicable to "pure" Si3 N 4 materials. The creep response predicted by the model is consistent with the experiment.</p> / Doctor of Philosophy (PhD)

<p>Alumina composites with 10, 20 and 30 volume % SiC whiskers were fabricated using colloidal processing methods followed by uniaxial hot pressing. The tensile creep properties of these materials have been studied between 1200°C and 1400°C. The composite slurries showed the best stability at pH = 2, which led to uniform distribution of whiskers in the final products. However, at pH ≥ 6 flocculation occurred between whiskers, resulting in whisker agglomerates in the matrix. Distribution of whiskers was characterised using neutron diffraction methods, which indicated that the whisker orientation could not be altered significantly by adjusting pH. All the composites showed much superior tensile creep resistance compared to pure alumina and the effect of increasing whisker volume fraction was significant up to 30%. Relatively high stress exponents were found, which is most probably associated with much enhanced cavitational creep in tension. The activation energy varied with whisker volume fraction, temperature and applied stress in a complex manner. This combined with the temperature-dependent stress exponents makes the identification of creep mechanisms difficult. Nevertheless, it appears that at moderate stress level grain boundary diffusion and grain boundary sliding (GBS) become more significant as whisker volume fraction increases. The composites containing 20 and 30% whiskers showed significant anelastic strain recovery (∼0.001) following tensile creep, which is consistent with earlier reports that involved bending creep tests. The whisker bending effect was studied by measuring the peak width of (111) SiC planes (perpendicular to the whisker axis) at various conditions. The difference in the peak width at room temperature was found to be insignificant before and after creep. Moreover, during in-situ neutron diffraction measurement at 1400°C, no measurable variation in the peak width was recorded from the crept samples that were cooled under load. It may be that the neutron diffraction technique used in this study is not sufficiently sensitive to measure the small bending strains developed. However, These results along with other evidence in the literature suggest that Herztian contact deformation of networked whiskers, rather than bending deformation of whiskers, may be the dominant mechanism to explain the observed anelastic strain recovery. This mechanism predicts similar strain recovery in any composite that contains constrained 'hard' inclusions with sufficient contact numbers. Models based on this mechanism have been developed, which seem to predict the magnitude of the recoverable strain reasonably well.</p> / Doctor of Philosophy (PhD)

<p>BAZ microstructure stress and hydrogen level are the fundamental factors influencing the cold cracking susceptibility of the HAZ. Implant testing at various hydrogen levels over a range of steel compositions and heat inputs shows how the microstructure and hydrogen level influence the critical stress necessary for cold-cracking.</p> <p>Based on implant test data, a correlation formula predicting the critical stress necessary for cold cracking for given HAZ hardness martensite in the HAZ and hydrogen level, is proposed. Employing this correlation, together with prediction of martensite in the HAZ and HAZ hardness, based on heat transfer caculations, martensite-composition-cooling rate relations and hardness-composition-cooling rate relations, an algorithms which can predict the critical stress necessary for cracking for given implant composition, cooling rate and hydrogen level, is constructed.</p> <p>This method of predicting the critical stress necessary for cold cracking is an improvement over the existing regression formulas for estimating cold-cracking susceptibility.</p> <p>This formula has been successfully adapted to predict cold-cracking susceptibility data as obtained through other tests such as rigid restraint, and to recommend prehear levels necessary to avoid cold-cracking. Development and use of implant testing machine with an automatic welding and loading facility is also reported.</p> / Master of Engineering (ME)

<p>Five aspects of disordered solids and polymers have been studied, as follows. (1) The temperature and pressure modulation effects on structural relaxation have been formulated and simulated. The calculated heat capacity for the modulated conditions shows extra features over that for unmodulated conditions, which may cause misinterpretation of a disordered solid's characteristics. (2) A new mean field approximation in the lattice-hole model is developed for monodispersed polymer chains. Calculations of the configurational entropy, Sconf for polydispersed chains, has led to two predictions, (i) a maximum in the plot of configurational heat capacity against the extent of polymerization, and (ii) S conf remaining positive at 0 K. Both predictions have been verified by others. The lattice occupancy density contribution to S conf has been related to a liquid's viscosity and divergence of the viscosity-pressure plots explained. (3) From calorimetry and x-ray diffraction studies, a new phase of CuCN, which remains metastable on cooling to 77 K, has been discovered. It shows features characteristic of glasses. Sconf of this phase has been, calculated by using a flexible chain model. (4) A calorimetric method for determining the transition from mass-controlled to diffusion-controlled reaction kinetics during polymerization has been developed, and verified by experiments. In this transition range, the plot of the reaction rate against the reciprocal temperature at fixed value of the extent of polymerization deviates from the Arrhenius behavior. (5) Dielectric studies of linear chain polymerization of a melt in real time, and the polymers ultimately formed have shown that their properties depend upon the thermal history. This is attributed to different molecular level structures, e.g., chains and loops formed under different polymerization conditions. Altogether these theoretical and experimental studies have a broader consequence for our current understanding of the nature of disordered solids and of their formation from liquids, both molecular and polymeric.</p> / Doctor of Philosophy (PhD)

<p>The investigation concentrates on the morphology of the oxidation products produced in Ag-1.8, 7.5, 10.9, 12.9 and 18.7 wt % Zn alloys at 550°C under an oxygen pressure of 1 atm. Particular attention is given to the structure of the internal ZnO produced and the morphological development of the internal precipitation zone in the alloys containing Ag-1.8, 7.5 and 10.9 wt % Zn. Penetration depth of the internal ZnO was measured using a light microscope, and was observed to obey the parabolic law. A model based on volume diffusion control of oxygen has been proposed for the velocity of the internal ZnO precipitates.</p> <p>The Ag-12.9 wt % Zn alloy was observed to form an internal ZnO scale in combination with an external ZnO scale. An exclusive protective external scale of ZnO is formed on the Ag-18.7 wt % Zn alloy. The growth of the scale was found to obey parabolic kinetics. The morphological development of the external ZnO on the above alloys has also been studied.</p> <p>The theoretical basis for the transition from internal to external scale formation has been given according to the Wagner(²⁶) criteria for binary alloys.</p> <p>It was necessary to extend the Wagner binary treatment for the transition into a ternary thermodynamic and diffusion analysis. A portion of the Ag-Zn-O ternary isotherm has been proposed, thus making possible the placement of representative diffusion paths.</p> / Master of Engineering (ME)

<p>This thesis consists of two parts. The first describes work on swelling - the second, on creep damage accumulation.</p> <p>The first part develops a model for pressure-driven diffusional dedensification (or swelling) in a polycrystal containing isolated pores. The model is sufficiently general to be directly applicable also to pressure sintering. It is tested against experimental data for hot-pressed alumina (HPA) which swells during high-temperature, pressureless anneals in air.</p> <p>The model includes two independent, coupled, pore growth rate components: (1) diffusional (de)densification, driven by gas pressures in closed pores, and (2) coalescence of pores attached to grain boundaries, driven by grain growth against pore drag. Internal pore pressures may be due to inert trapped gas or to gas generated by chemical reactions.</p> <p>In a poly crystal with inert trapped gas, the trapped gas drives early swelling, but is self-dissipating. Pore coalescence, by progressively releasing capillarity constraint, becomes the dominant cause of swelling at longer times.</p> <p>The model predicts what magnitudes of chemically-generated pore pressure will increase swelling rates beyond those arising from inert trapped gas alone. And, for research on creep damage accumulation, it predicts what precreep anneals are required to dissipate an initial trapped-gas pore pressure, and the range of distinct, predamaged porous microstructures available.</p> <p>The second part of the thesis presents experimental data from flexural and uniaxial tensile creep tests. The data demonstrated that HPA's cavitation damage tolerance and failure strains are high enough that creep fracture tests should be done in uniaxial tension rather than in bending. A high temperature tensile test system, with optical extensometry, was designed, built and tested. The uniaxial tensile data suggest design improvements for testing ductile ceramics.</p> <p>The combined mechanical and microstructural data indicate that, at 1250 C, failures at high stress are controlled by slow crack growth from microstructural heterogeneities. An abrupt transistion at about 200 MPa seperates rapid failures from failures at strains on the order of 0.1. At about 50 MPa, a more gradual transition occurs, to fracture at higher strains, controlled by strain-driven damage. The internal microstructural data show internal creep damage at levels at which cavity coalescence generates macrocracks in HPA.</p> / Doctor of Philosophy (PhD)

<p>An industrial B.O.F., located at the #1 Melt Shop of Dominion Foundries and Steel Ltd., was used to develop and evaluate a method of measuring the rate of fume emissions during steelmaking.</p> <p>The fume rate was observed to decrease with increasing time into the blow. Approximately 60% of the total fume iron losses are emitted during the first one-third of the blowing time. The fume rate is influenced by such process variables as: metal carbon content, slag volume, metal temperature and lance practice.</p> <p>The mechanism of fume formation was assessed using the size, shape and chemical analyses of the fume material. The major mechanism is that of the explosive oxidation of metal droplets in the oxygen impact zone. Evidence of vaporization was also observed, but this represented less than 10% (by weight) of the total fume iron losses during the blowing period.</p> / Master of Engineering (ME)

<p>The mechanical deformation behaviour of the age-hardenable aluminum alloy 2036 was used to study the effects of differences in the microstructure due to (a) variations in commercial processing for one ageing condition (T4), and (b) variations in the ageing condition for one processing route. The effects of loading along different proportional straining paths have been considered with the aim of establishing correlations between the mechanical properties and formability, and for the case of (a) above, some rationalisation of selected press performance in terms of particular microstructural constituents has also been attempted. Changes in the fracture behaviour due to both (a) and (b) above are discussed. These changes emphasise the need to consider the role of fracture events in forming operations and their dependence on microstructure.</p>

<p>The role of damage on the mechanical response of a heterogeneous material was investigated through both mechanical testing and x-ray tomography. X-ray tomography was used to obtain quantitative information on the evolution of damage through processes occurring at both the local and global scale. The results indicate that heterogeneity in the spatial distribution of particles does influence the damage process. However, the influence of having one particle or more interacting with another is limited to reducing the tensile deformation required to initiate damage. The rate at which damage evolves is similar for both isolated and non-isolated particles and increasing the number of neighbors around a non-isolated particle was determined to have no additional influence on the evolution of damage. These results, coupled with mechanical testing measurements of both global and local properties, were used to develop models describing the flow response of composite materials as damage accumulates. Models were developed to predict the effects of both particle multiple cracking and micro-crack linkage on the composite flow response. The models predict that both damage processes reduce the load bearing capability of the material over that of an undamaged composite, however the loss in load bearing capability is much more severe when micro-crack linkage occurs. Micro-crack linkage rapidly leads to a loss in global stability so that the strain at which the composite fails is significantly less than previous models suggest. The experimental behavior of the composite materials investigated in the modelling work favors that predicted by the micro-crack linkage model. Ductility predictions resulting from the micro-crack linkage model were sensitive to both the volume fraction and the matrix work hardening exponent. By varying the matrix work hardening exponent the micro-crack linkage model captured the experimentally observed range of ductility values present in literature.</p> / Doctor of Philosophy (PhD)