<p>An investigation of liquid film migration (LFM) in the aluminum-copper system is presented in this thesis. Experimental and numerical modelling results are discussed for the migration of liquid lenses, formed on melting of grain-boundary [straight theta] precipitates in Al-3.4 wt% Cu alloys which had been previously equilibrated at 400°C. Liquid films of comparable thickness, formed upon the melting of intragranular Widmanstätten precipitates were found not to migrate. Volume fraction studies indicate that after the initial liquation of the [straight theta] precipitates, the liquid phase undergoes an isothermal solidification. Three-dimensional reconstructions of LFM events were created using a combined SEM/microindenting/polishing method. The results indicate the difficulty in obtaining reasonable measurements of migration distances from two-dimensional cross-sections; a method was developed to obtain appropriate two-dimensional measurements based upon these results. The modelling studies are based on the hypothesis that the leading and trailing solid-liquid interfaces are each characterized by a constrained local equilibrium, as determined by chemical, coherency and capillary terms, and that the process is driven by the liquid concentration gradient resulting from different equilibrium concentrations at the two interfaces. The model is utilized to explore the sensitivity of the process to variations in physico-chemical parameters, and ultimately as an aid in the understanding of LFM and its subsequent cessation. It is determined that the capillary term has a strong influence on the migration process, even for liquid pool dimensions of several microns. From model simulations and an analysis of coherency loss in the leading solid it is concluded that the migration of liquid films in the Al-Cu system is driven by coherency strain energy in the early stages of the process. As migration continues, the increasing energy due to curvature is sufficient to decrease the overall driving force to a point where the film. slows down and coherency is lost. The liquid films become immobile (no reversal of migration) once loss of coherency occurs.</p> / Doctor of Philosophy (PhD)

<p>The relationships between the mechanical properties, fracture and forming characteristics and the microstructural features has been investigated in commercially available HSLA steels. In addition, attention has been focussed on the behaviour of these materials subjected to a reverse straining, as a very important characteristic for pipeline applications.</p> / Master of Engineering (ME)

<p>Transmission electron microscopy (TEM), atomic force microscopy (AFM) and scanning electron microscopy (SEM) have been used to study the strain relaxation mechanisms in In1-x Gax Asy P 1-y films grown on (100) InP substrates using gas-source molecular beam epitaxy (MBE). Highly anisotropic strain relief behavior was found in 2% tensile strained In0.25 Ga0.75 As and In0.72 Ga0.28 P films. In the first stages of film growth, the strain in [01¯1] cross-section was relieved by twinning, while it occurred by cracking in the orthogonal [011] cross-section. In the In0.25 Ga0.75 As film cracking was a transitory phenomenon. Crack healing was observed in the 500 nm thick film. Cracks were observed to penetrate into the substrate and deviate from an (01¯1) to (11¯1) or (1¯1¯1) planes. A critical stress intensity argument was developed to explain substrate cracking. A dislocation analogue for a surface crack was developed to successfully account for the experimental value of the ratio of crack opening displacement to normal surface displacement associated with cracks in In0.72 Ga0.28 P films. The 90° partial dislocations were found to form prior to the formation of either cracks and or 60° dislocations in In0.25 Ga0.75 As films, which is consistent with the critical thickness and nucleation calculations. Elastic strain energy computations show that the 90° partial dislocations also provide the most effective relief of elastic strain energy for films with smaller thickness, while cracks are the most effective strain relaxation mechanism for thicker films. It has also been shown both experimentally and theoretically that the twin thickness increases with an increase in the film thickness. The free surface plays an important role in determining the equilibrium position of misfit dislocations in thin epitaxial films. The computations based on a force argument show that the core of the dislocation lies close to the interface when the film is softer than the substrate. On the other hand, when the film is elastically stiffer than the substrate, the core of the dislocation is predicted to lie at some distance from the interface in the softer substrate. This prediction agrees with the experimental observations that the 90° partial dislocations bounded by stacking faults are frequently observed to locate inside the InP substrate over a range of a few hundred angstroms. The composition modulation in In1-x Gax As y P1-y films was found to be associated with the chemical spinodal in this alloy. Films with compositions lying within the chemical spinodal at growth temperature show fine scale composition modulation contrast. The composition modulation scales with the size of the strain-induced surface facets.</p> / Doctor of Philosophy (PhD)

<p>Pure and doped CaF₂ single crystals were grown from melts by Czochralski technique. The doped crystals contained two levels of Na⁺, K⁺, V³⁺ and Y³⁺ impurities. The crystals were subsequently annealed and analysed for impurities. Samples were mechanically tested in compression at 450 and 600°C, and the 0.1% proof stress noted. An attempt was made to explain the observed hardening by Fleischer's theory of solid solution hardening in the case of V³⁺ and Y³⁺ -doped CaF₂ and by precipitation hardening theory in the case of Na⁺ and K⁺ -doped crystals.</p> / Master of Engineering (ME)

<p>A numerical Solution to the one dimensional unsteady state heat transfer problem including the latent heat evolution of the pearlite reaction has been developed to predict thermal conditions in a Jominy bar. This has been combined with a model for pearlite nucleation and growth developed by Kirkaldy and used to predict the TTT curves for the austenite → pearlite transformation and the pearlite volume fraction as a function of time and Jominy number. The conversion of the TTT to OCT curves, which is a necessary step in the latter calculation, has been Investigated using the additivity rule and the approximation to it due to Grange and Kieffer. It was concluded that the latter approximation will often not be justified. Inclusion of the latent heat evolution in the heat transfer calculations was proven to be significant and as such should be included in any accurate algorithm for predicting Jominy curves.</p> / Master of Engineering (ME)

<p>The crystal structure of Cu₂V₂O₆ has been studied by X-ray diffraction techniques. CuV₂O₆ has been found to have a structure closely related to the mineral brannerite as do the majority of the metavanadates, but with lower symmetry. The refinement of the structure showed a bonding geometry consistent with the bond strength-bond length correlations for oxides which have been applied to other vanadate structures.</p> <p>The symmetry and unit cell parameters of Ca₂V₂O₇ have been determined. The attempted solutions are described and suggestions are made for future work on this structure.</p> / Master of Science (MS)

<p>A production basic oxygen furnace installation at Dominion Foundries and Steel Ltd., Hamilton, Ontario was used to develop and evaluate a method of measuring the rates of materials ejection during steelmaking. These rates are determined by periodic sampling at the mouth of the BOF with a tubular sampler, during the oxygen blow.</p> <p>The materials ejected are classified into two categories, slopping and metal ejection. From the measured rates and analyses of ejected materials, combined with the composition of the slag and metal bath as obtained by direct sampling, certain insights are made about the mechanisms of materials ejection. Among the process variables examined, the metal bath level, oxygen flow and course of slag development showed the largest influence on ejection rates.</p> / Master of Engineering (ME)

<p>Intrinsic degradation, which leads to the long-term decrease in the electroluminescence efficiency, has been a major limitation facing the new technology of organic light emitting devices (OLED). Traditionally, degradation has been speculated to be caused by morphological instability of the organic layers, especially the less stable hole transport layer (HTL), or by the formation of deep traps at the hole-injecting contact. These speculations were based on experimental observations showing that doping the organic layers or introducing a buffer layer at the hole-injecting contact can dramatically improve device stability. However, the real causes of OLED degradation remained uncertain. In this study, the cause of the long-term degradation of OLEDs based on tris(8-hydroxyquinolato) aluminum (A1Q3), the most widely used electroluminescent molecule, is investigated. OLEDs with various structures are studied. Results reveal that the injection of holes into the AlQ3 layer is the dominant factor responsible for device degradation. Cationic AlQ3 species are found to be unstable and their degradation products are fluorescence quenchers that lower the electroluminescence efficiency of OLEDs. In view of these findings, the effectiveness of stabilizing agents, such as, doping the HTL, introducing a buffer layer at the hole-injecting contact, or using mixed layers of hole and electron transporting molecules, is explained in terms of their role in slowing down the injection of holes into the AlQ 3 , which results in a higher electron density and thus a more rapid electron-hole recombination. Therefore, the lifetime of the unstable cationic AlQ3 species is reduced leading to a significant decrease in AlQ 3 degradation and consequently increases device stability. Other earlier observations pertaining to OLED degradation are also addressed. The degradation mechanism is further demonstrated on OLEDs with dual-layer HTL made of materials with different ionization potentials. The important features of a theoretical framework to model OLED degradation are also discussed.</p> / Doctor of Philosophy (PhD)

<p>Steelmaking dust is generated in high temperature refining step at a rate of about 2% of the steel produced. Electric Arc Furnace (EAF) dust is classified as a hazardous material due to its content of leachable heavy metals, i.e., Pb, Cd, and Cr. Therefore, it must be treated before disposal to meet certain environmental regulations. Most processes for the treatment of EAF dust involve recovery of valuable metals, such as zinc. In most cases alkali halides and lead in the dust are problematic in both processing and product quality. In the present work a pre-treatment is proposed to separate "more volatile species" which include the alkali halides, lead compounds and cadmium oxide, from "less volatile species" which include iron, zinc and calcium oxides. Evaporation of volatile species in this process takes place at about 900° C in a virtually closed system under reduced pressure, and condensation of the vapors occurs at a lower temperature. The thermodynamic aspects of the process are considered. With the use of three different experimental apparatuses the kinetics of the system are studied. The rate of individual kinetic steps; evaporation, condensation, heat and mass transfer, are calculated and compared with the observed overall rate of reaction. A numerical model for the heat transfer inside the dust bed is developed. Applying the principles of mass and heat transfer to the system under investigation, and using the results of the experiments and the numerical model of heat transfer, it is shown that heat transfer across the porous dust bed is most likely the rate controlling step. A rotary reaction chamber is designed to facilitate the heat transfer to dust particles and eliminate the slow kinetic step of heat transfer across the stationary dust bed. With the use of the rotary reaction chamber the duration of treatment is significantly reduced under otherwise identical conditions. The benefit of the rotational movement may be appreciated by direct comparison of the treatment of twenty five grains EAF dust. At a furnace temperature of 1100°C for 95% removal of the volatile species the time required (from the introduction of the apparatus at 25°C to its withdrawal from the furnace) is 12 minutes for the stationary and 8 minutes for the rotational chamber. At a lower furnace temperature of 950°C , the degrees of removal of lead and potassium are about 51% and 27% for 12 minutes in the stationary reaction chamber and about 83% and 78% for 10 minutes in the rotary reaction chamber. A secondary incinerator dust is also investigated in the present work. In 6 minutes at a furnace temperature of 950°C about 99% of the volatile species including NaCl , KCl , and lead compounds are removed from twenty five grams dust. The residue has an enrichment of zinc from 27%wt to 78%wt . This environmentally friendly and energy efficient process may be applicable for the separation of "more volatile species" in dust generated from most high temperature processes such as steelmaking, incineration, nonferrous processes and cement manufacturing.</p> / Doctor of Philosophy (PhD)

<p>The ionic stability of oxide particles in polar non-aqueous media is studied. Surface chemistry and interparticle forces are manipulated by controlling the acidity and ionic strength of the suspensions without dispersants. The acidity of ethanolic solution is determined using ion transfer functions, wherein the relationships between acidity, oxide particle-surface-charge, zeta-potential, stability and suspension rheological behaviour are established. The ionic stability of oxide particles in ethanol can be controlled by combination of potential determining ions and indifferent electrolyte to optimize the values of repulsive potential and repulsive force. It is shown oxide particles can be charge-stabilized, as in aqueous suspensions. The viscosity and flow curves for oxide/ethanol suspensions are acidity dependent. The flow curves of the suspensions follow the Casson model and the Casson yield value is used to evaluate their stability. Positive isoelectric point shifts were observed for alumina and magnesia in ethanol on increasing the solid concentration. However both dilute and concentrated aqueous suspensions of alumina give the same isoelectric point. Silica/ethanol suspensions are stable near the IEP. This result suggests the colloidal stability of silica in ethanol can not be explained exclusively by the ionic stability mechanism of DLVO theory. The discrepancy is believed due to a steric barrier consisting of a silicic acid gel network. The surface chemistry and rheological properties of alumina suspensions in EtOH and DMSO are strongly influenced by the ionic strength of the suspensions. Rheological measurements show the viscosity of the suspensions decreases with increasing salt concentration due to suppression of the second electroviscous effect. Solvent is found to have a marked influence on suspension rheology. The heterocoagulation behaviour of oxide-mixture/ethanol suspension systems is examined, elucidating the general principles underlying structure formation in mixed dispersions. It is demonstrated that the architecture of composites can be controlled by manipulation of the relative colloidal stability of the constituent primary particles.</p> / Doctor of Philosophy (PhD)