<p>The photochemical etching behaviour of n-InP was examined by producing holes or vias through the samples. The rate of material removal and quality of the resulting vias were studied as a function of: electrolyte pH, temperature, and type of metal ion impurity, illumination level, frequency, and duty cycle, sample surface quality, area, and roughness, and a second light source. The project was undertaken to prepare thin samples for the transmission electron microscope (TEM). The quality of sample produced for the TEM was acceptable, as determined by comparing the thin sections to thin sections produced by other traditional sample preparation techniques. This new method of sample preparation has the important advantages of being able to thin at a precisely predetermined region and that the resulting thin regions are surrounded by a thick built-in stabilizing structure which improves handleability.</p> / Doctor of Philosophy (PhD)

<p>It has been appreciated for some time that low oxygen activity is required for effective steel desulphurization, and recently this has been recognized for hot metal desulphurization. In the present experiments, calcium carbide was injected into 70 kg melts of iron during which desulphurization and deoxidation were studied by continuously measuring oxygen activity and frequently sampling for sulphur content. In some experiments, carbon dioxide was used either as a carrier gas or as generated from calcium carbide/limestone mixtures. The experiments confirmed that the oxygen activity in hot metal was controlled by the silicon-silica equilibrium, except when aluminum was added. It was found that the rate of desulphurization was associated with the oxygen activity. The desulphurization rate increased as oxygen activities decreased in order from carbon dioxide-containing injections into iron-carbon-silicon melts, to calcium carbide/nitrogen injections into iron-carbon-silicon melts, to finally calcium carbide/nitrogen injected into iron-carbon-aluminum melts. The dependence of the incubation period with the melt oxygen activity was recognized. The incubation times coincide with periods of deoxidation before desulphurization proceeds. In order to interpret the experimental results, a kinetic model was developed that takes simultaneous desulphurization and deoxidation into account. This model demonstrates clearly that the rate of desulphurization is controlled by the melt oxygen activity and the powder feed rate. Thus, low oxygen is required for effective hot metal desulphurization. In a log h$\sb{\rm o}$-log h$\sb{\rm s}$ diagram, the refining trajectories determined from the kinetic model were found to be consistent with those obtained from the experiments. From the model calculations, it was found that the commercial calcium carbide which normally contains 30 per cent of lime can be employed for the melts pre-deoxidized with silicon or aluminum, without causing significant delay of desulphurization. Another aspect of this work is the particle-liquid contact. The contact was measured by monitoring the melt temperature during injection, and applying calorimetric principles to analyze the results. It was found that a significant fraction of particles does not come in contact with the melt during their rise through it. The particle-liquid contact fraction is approximately 30-50 per cent. The same conclusion was obtained by analyzing the experimental rate constants of desulphurization on the basis of mass transfer theory.</p> / Doctor of Philosophy (PhD)

<p>The primary purpose of this investigation was to examine the strengthening that arises when reinforcing an aluminum alloy with approximately equiaxed shaped SiC particles. The results indicate that the majority of strengthening that is derived from the particles develops at low strains. For a variety of test conditions explored, about 90% of the total strengthening derived is developed by a plastic strain of about.5%. This is primarily a result of the high strain hardening rate which is exhibit by the composite materials during the onset of plastic flow. An examination of the Bauschinger effect in the composites indicates that an internal stress, resulting from inhomogeneous plastic flow, develops rapidly in the low strain regime. Microstructural observations suggest that the nature of this inhomogeneous plastic flow is influenced by the distribution of the SiC particles. Initial plasticity appears to be associated with regions of the microstructure which are relatively particle free. A self-consistent model, based on Eshelby's equivalent inclusion method, has been developed which is capable of calculating the stress partitioning which occurs between the Al matrix and SiC particles during the elastic/plastic transition and beyond. Excellent quantitative agreement between the experimentally measured elastic modulus and that predicted by the model was achieved. When considering a uniform distribution of second phase particles, the model predicts an increase in the initial strain hardening rate. However, it underestimates the actual strain hardening rate which was measured experimentally. The influence of particle distribution was examined by representing the composite microstructure by a bimodal mixture consisting of a particle rich and particle free phase. The model predicts an increase in the initial strain hardening rate as a result of clustering the microstructure in this way. The magnitude of this increased hardening is comparable to that measured experimentally.</p> / Doctor of Philosophy (PhD)

<p>Pitting corrosion is one of the major causes of industrial failure of metallic parts. Unfortunately, such attack cannot always be detected before perforation occurs. Understanding the mechanism of pit development will provide us with a clue of how to prevent pitting corrosion in advance. From an engineering point of view, it is important to keep the pit, once it forms, repassivated rather than allow continuous growth, leading to the perforation. The key factors controlling pit development must therefore be determined. A comprehensive model of pitting processes is proposed which considers interfacial dissolution, salt film formation, multiple species in solution, reaction equilibria, migration and diffusion. The model is capable of determining local chemistry within pits and characterizing the conditions necessary for stabilization of pitting corrosion. The model is experimentally verified using microelectrodes developed to measure the local properties of pitting sites. The electrodes were characterized through studies on their potential stability, spatial resolution, and the interference of other ions involved in the corrosion of nickel. The study was performed on a nickel model pit in a saline environment. The in situ measurements obtained by microelectrodes provide the direct evidence of existence of a salt layer, comprised of corrosion products, at the bottom of a developing pit. The condition for formation of the salt film in active pits appears to require that the pit depth be sufficient to ensure the maintenance of a critical local solution chemistry. The experimental results are consistent with the theoretically predicted critical depth at which a metal chloride salt layer starts to form at the pit bottom. The measurements of the potential distribution within the salt film have provided data for the potential drop across the salt film, its thickness and therefore its conductivity. These help us to achieve a better understanding of the nature of the salt film. The anodic behaviour of nickel was studied in solutions which simulate the pit electrolyte, based on the in situ measurements of the local pH and chloride ion concentration within a pit at different stages of development. The factors controlling the pit growth were deduced.</p> / Doctor of Philosophy (PhD)

<p>A study consisting of theoretical and experimental work on kinetics of simultaneous interfacial reactions in multi-component metallic/ionic systems has been carried out in the present work. General rate expressions for simultaneous interfacial reactions are proposed based on the application of mass action law to the electrodic half-cell reactions and the constraint of no net electric current. The nature of coupling among interfacial reactions is discussed by defining the coupling factor which is a collective property of the system, contributed by and common to all interfacial reactions. Kinetic behavior of each element can be individually and simultaneously described. Experimental work is conducted in slag/metal system for the study of coupled interfacial reactions and diffusion. Considering interfacial reactions as the boundary conditions for diffusion in both phases, a mathematical model is developed. Computed results, based on thermochemical parameters mostly reported in the literature, are compared with diffusion profiles measured by electron probe microanalysis (EPMA) in silicate slag and Fe-Mn-Si alloy at 1763$\sp\circ$K. Values of reaction rate constants for transfer of iron, manganese and silicon are recommended through curve fitting with the experimental data. The conventional pseudo-binary approach in the formulation of rate equations for interfacial reactions in multi-component slag/metal systems may be deduced from the present theory with simplifying assumptions and as a limiting case.</p> / Doctor of Philosophy (PhD)

<p>Bubbly plume flow, which is encountered in many refining processes of the iron and steel making industry, has been studied with a water model ladle and simulated with a novel mathematical model. The main achievements of this project can be summarized as following: (1) A new experimental technique based on the Laser Doppler Anemometry and Electrical Probe has been developed to measure various flow parameters of the plume, such as: mean and turbulent liquid velocities, bubble velocities and distributions of void fraction and bubble frequency. (2) The turbulence feature and bubble behaviour in the gas/liquid two-phase zone has been experimentally studied with the newly developed LDA/EP technique in a water model ladle. It has been found that turbulence in the plume zone is close to an isotropic one and the relative velocity of bubbles is not affected by the void fraction in the plume. (3) A mathematical model based on the modified k-ε turbulence model has been developed to simulate the plume flow. The coefficients of the extra source terms of the modified k-ε model have been determined experimentally. The mathematical model yielded good predictions of both the mean and the turbulent liquid velocities. The behaviour of bubbles in the plume has also been dealt with to produce good predictions of the void fraction distributions.</p> / Doctor of Philosophy (PhD)

<p>A series of plane strain compression experiments and finite element method (FEM) calculations were performed on continuous fibre composites to examine the role of volume fraction and fibre arrangement on the mechanical behaviour when the fibre axis was perpendicular to the loading direction. The flow pattern in the matrix was quantified by measuring the shape change of a fine gold grid which had been fabricated on the surface of the sample prior to deformation. It was determined that the pattern of deformation in the matrix was controlled by the spacing and geometric arrangement of the fibres. The damage process (i.e. by fibre cracking or interfacial decohesion) can be understood by the local distribution of stresses (as calculated by FEM) in the fibres and at the fibre matrix interface. The development of the deformation texture in the matrix and the observations of the pattern of subsequent recrystallization events can be related to the flow pattern in the matrix. The combination of careful experiments with finite element method calculations represents a powerful approach to studying problems involving composite materials.</p> / Doctor of Philosophy (PhD)

<p>Thin films of ZnS and ZnO(1-x)Sx were prepared by electrodeposition and chemical vapour atomic layer deposition (CVALD), respectively. This represents the first effort to apply the technique of electrodeposition to ZnS for the purposes of electroluminescence. This is also the first time thin films of ZnO(1-x)Sx were grown by CVALD and represents the first analysis of how the films properties vary with a known stoichiometry. Characterization methods performed on the thin films include: Auger spectroscopy, Rutherford backscattering spectroscopy, scanning electron microscopy, x-ray powder diffraction, Hall measurements and uv/visible transmittance. The as electrodeposited films were nearly stoichiometric but conductive and oriented in a <200> direction. After annealing the films became discontinuous and lost their crystallinity, but their absorption characteristics more closely resembled that of evaporated films. Ion implantation with Mn was successful and the films exhibited yellow-orange cathodoluminescence. The value of x in the formula ZnO(1-x)Sx could be varied from 0 to 0.95 by changing the amount of hydrogen sulphide admitted to the reaction chamber. Films with values of x near 0 grow in the <200> direction (of ZnO) and have a fine grained structure. As the value of x increases the films become more and more amorphous but split into a two phase structure close to x = 0.55. At this point the films exhibit a minimum bandgap. As x increases further the films become one phase again and become oriented into the <111> direction of ZnS. The cathodoluminescent characteristics of these films also change with x. When x is zero the films appear to luminescence green or yellow depending on the deposition temperature. Even 5% S removes this luminescence however and the spectrum is dominated by a peak in the red. This peak steadily decreases then increases again as x increases.</p> / Doctor of Philosophy (PhD)

<p>The energies of one and two-dimensional dislocation arrays lying near<br />a free surface are evaluated directly from the stress fields of single dislocations<br />in a half-space. These results are used to obtain expressions giving the<br />equilibrium spacings of a number of different arrays relieving misfit in a<br />strained epitaxial system. Numerical calculations are performed for the case of<br />edge and 60º dislocations relieving strain in a silicon-germanium layer<br />deposited on a silicon substrate. This method is also used to calculate the<br />energies of various low angle grain boundaries in a half-space.</p> <p>Single-ended dislocation sources are observed using transmission<br />electron microscopy in two short-period Si-Ge superlattices grown on Si(100).<br />Their formation is linked to the development of non-planar layers during the<br />growth of the superlattices. The relaxation of these superlattices takes place<br />at significantly lower temperatures than equivalently strained homogeneous<br />epilayers.</p> <p>Si-Ge short period superlattices deposited on Si(100) are shown to<br />relax through twinning on {111} planes if the deposited layers become grossly<br />non-planar. Twinning is accompanied by the formation of a diamond<br />hexagonal phase. No 60· a/2(110) dislocations relieving misfit are present in the strained layer structure.</p> <p>The nature and origin of a new type of defect in Si₁_ᵪGeᵪ/Si strained<br />layer structures, the "pagoda" defect, is studied using transmission electron<br />microscopy. The defects are found to propagate in a direction determined by<br />the position of the Si source in unrotated substrates, and to have their origin<br />in the role played by SiC particles (left after cleaning the substrate) during<br />the growth process. Pits that form at the SiC particles are preserved during<br />MBE growth and perturb the strained layers, leading to the formation of<br />pagodas.</p> / Doctor of Philosophy (PhD)

<p>Steel-clad calcium wire was fed into 40 kg heats of AISI steel with various sulphur and aluminum contents under argon atmosphere. Samples were taken continuously during and after calcium injection for chemical and inclusion analysis from more than twelve heats of injection. Calcium dissolution was found to be enhanced by instantaneous reactions with oxygen and sulfur in the boundary layer. The unreacted calcium diffuses through the gas-liquid film into the melt for oxide and sulfide inclusion modification. The overall interaction between oxides and sulfides with the bulk drives towards the equilibrium of (CaO) + S = (CaS) + O. The system paths from the initial state to the final state were kinetically studied with the assistance of a_CaS-a_MnS diagrams, which were calculated by applying a regular solution model to the solid phase and an ideal model to the liquid phase of the CaS-MnS phase diagram at various steelmaking temperatures. A mathematical model for the multiphase kinetics was developed to interpret the calcium dissolution, deoxidation, desulphurization, oxide and sulfide inclusion modification during and after calcium injection. The implications for full-scale operation were discussed.</p> / Doctor of Philosophy (PhD)