Orientation imaging microscopy (OIM) was performed on hot twisted polycrystalline 304 stainless steel specimens in order to study the characteristics of dynamic recrystallization (DRX) in low stacking fault energy FCC metals. Microtexture analyses were carried out and the misorientation statistics of the recrystallized areas were derived. The results revealed that multiple twinning controls the formation of new orientations. DRX macrotextures were also measured in torsion and compression. These textures are highly sensitive to the deformation (Zener-Hollomon) parameter Z. High Z values lead to overall textures controlled by nucleation, whereas, at low Z values, selective growth is more predominant. / A deformation and recrystallization model based on the OIM measurements and the steady state properties is proposed. The observed strong brass-type deformation textures were simulated by introducing self hardening coefficients. The transition from single to multiple-peak flow during DRX was investigated, together with the evolution of the grain size distribution. In particular, the effect of twinning on the evolution of the latter was simulated. Finally, it is shown that the model simulates texture development during DRX very well.

Solid wettability plays an important role in many industrial processes. Two examples of processes dependent on solid wettability are: Bubble generation from porous media (project one) and the bitumen froth treatment process in the recovery of oil from oil sands (project two). / Project one. Bubble size has a profound effect on flotation efficiency controlling particle collection and froth stability. Models of bubble generation at a rigid sparger usually include a wettability effect (i.e., contact angle). The role of sparger wettability on bubble formation was examined using three rigid spargers exhibiting water contact angles of 0, 64 and >90°. The wettability was varied by heating the sparger, and the contact angle was determined by the Washburn and modified Washburn methods. By measuring permeability, it was determined that heating had no effect on sparger pore structure. The results showed no detectable wettability effect on bubble formation over the practical operating range of column flotation. The lack of wettability effect may be attributed to the highly irregular morphology of the sparger surface. The bubble size, it was shown, can be predicted by using the concept of sparger equivalent pore diameter and active pore number, which are estimated by a back-calculation routine. / Project two. Production of oil from oil sand deposits in northern Alberta involves open pit mining, mixing the ore with water, extraction of bitumen from the slurry by a flotation-related process (Hot Water Extraction Process), removal of water and solids from the froth formed (froth treatment process), and upgrading the heavy bitumen to liquid hydrocarbons. The froth treatment process to remove fine solids and water from the bitumen froth depends on the wettability of the solids. Fine solids were extracted from samples of bitumen froth using heptane. A mixture of heptane and toluene (diluent) was used to study fine solid wettability. The contact angle (sessile drop method) and partition of the fine solids among the aqueous, diluent and interphase regions were determined. The effect of diluent composition, sample drying, and surface washing was examined. The partition of the particles correlated well with their wettability, and the results helped interpret observations from plant practice.

Tensile testing was carried out on five microalloyed steels in order to investigate the hot ductility behaviour between 800 and 1300$ sp circ$C, and between 1400$ sp circ$C and the solidus temperature. The influence of an increase of the carbon content up to the peritectic composition ($ approx$0.14 wt%) in Nb-Ti-Al bearing steels, and the effect of microalloying additions (Al, Al-Nb-Ti and, Al-Nb-Ti-V-Cr) were examined. / Above 1400$ sp circ$C, intergranular embrittlement occurs by incipient melting at the grain boundary at temperatures between 0 and 30$ sp circ$C below the solidus. The increase in the carbon content from 0.7 to 0.14 wt% leads to a decrease of the grain boundary incipient melting temperature, although the bulk solidus temperatures for all the steels remain more or less constant at about 1500$ sp circ$C. This is explained by the grain boundary segregation behaviour of carbon. / Above 1400$ sp circ$C, fractography reveals that the grain size in the steel with the peritectic composition (0.14 wt% C) is coarser than that of the low carbon grades (0.07-0.1 wt% C). This effect is attributed to: (i) the higher $ sb{ rm Y}/ delta$ transformation temperature in the steel with the peritectic composition, thus allowing more grain coarsening, and (ii) the lower amount of $ delta$ ferrite, which reduces the pinning effect of this phase on austenite grain growth. In the Cr containing steel the relatively low degree of segregation also produces a coarser structure than that of the Nb/Ti/Al steel. / In the temperature range 800-1300$ sp circ$C, the ductility is strongly dependent on prior heat treatment. For the Nb-Ti-Al bearing steels, an increase of the carbon content up to 0.14 wt% slightly affects hot ductility after direct heating and annealing at 1330$ sp circ$C. However, after sensitizing at 1480$ sp circ$C the width of the ductility trough is increased in the 0.14 wt% C steel. / For the C-Mn-Al steel, the ductility is the lowest following the direct heat treatment because of the presence of a greater volume percent of AlN precipitates at the grain boundaries compared to the Nb-Ti-Al steels. In contrast, after annealing and sensitizing, the C-Mn-Al grade displays the highest ductility because of the sluggish rate of precipitation of AlN in austenite. / Finally, Ti additions to Nb containing steels improve the ductility because some of the Nb is combined with TiN particles both as complex TiNb(C,N) and by precipitating as NbC on existing particles. The addition of V to the Nb-Ti-Al steels (which contains also Ni and Cr) increases the ductility because of the retardation of Nb(C,N) precipitation. However, after the sensitizing treatment, at low test temperatures, the V containing grade displays the lowest ductility because of the coarser grains. (Abstract shortened by UMI.)

Continuous temperature measurement of molten steel in new as well as existing steelmaking processes is not always feasible with current technology. However, it is a well known fact that molten steel temperature, especially during the continuous casting operation, can have a profound effect on the quality of the product. A knowledge of the instantaneous steel temperature in the tundish can serve as a valuable tool in the quest for better quality. With the implementation of statistical process control techniques such occurrences as breakouts, tundish freeze up or the nucleation of an extensive area of columnar grains in the cast section can be reduced. / However, conventional thermocouples cannot withstand the severe corrosion around the slag line for a sustained period of time. Even though the sensing wires that comprise the hot junction can, if well protected, be used for long periods of time without serious deterioration, a suitable refractory cover or sleeve must be applied to the thermocouple prior to using it. / For the present work, two methods for continuously measuring melt temperature were developed. One of these relies on deducing melt temperature while the other is based on providing a cooled sleeve for the thermocouple to minimize slag line corrosion. The first technique involves the use of multiple thermocouples embedded in a refractory section at various displacements. When the refractory is contacted by the melt, transient heat transfer is initiated through the section. By analyzing this transient behaviour with a suitable heat transfer model, it is possible to infer the temperature of the melt. A mathematical model that adequately describes this process has been developed and tested. / An alternative approach that was developed uses heat pipe technology to prevent the corrosion of the thermocouple probe through solidification of slag onto the heat pipe which can serve as a protector of the probe. Heat pipes are devices capable of transferring large quantities of heat with very small temperature differences. Advantages of the heat pipe when used as a heat transmission device are constructional simplicity, flexibility, high heat transport capability, and no need for external pumping device. The heat pipe consists of a closed evacuated tube, porous wicking material, and working fluid. Heating one part of the external surface leads to evaporation of the working fluid and the establishment of a pressure gradient within the heat pipe. The resulting pressure difference drives the vapor from the evaporator to the condenser where the pressure and temperature are slightly lower. The effective conductance of the heat pipe can be several orders of magnitude higher than that of an equivalent solid copper bar. In the second part of this thesis, the feasibility of incorporating heat pipe technology to solidify and maintain a thin layer of slag on the heat pipe while it is in operation is presented.

The crystallization behaviour of iron intermetallics during solidification and the dissolution behaviour of iron intermetallics during heat treatment and their subsequent effect on mechanical properties have been studied. / Addition of iron decreases DAS and affects eutectic silicon particle size as nucleation and growth of silicon crystals takes place on the $ beta$-AlFeSi platelets, thus minimizing isothermal dendrite arm coarsening at the eutectic temperature. Increase of iron also increases the size and volume percent of the iron bearing intermetallics. / In the absence of manganese, the iron intermetallics crystallize in the $ beta$-phase, at all cooling rates ranging from 0.1 to 20$ sp circ$C/s when cast from a normal casting temperature of 750$ sp circ$C. In the presence of manganese, the iron intermetallic crystallizes in $ alpha$-phase at low cooling rates and in both the $ alpha$- and $ beta$-phases at high cooling rates. This reverse crystallization behaviour is explained based on the segregation effect displayed by the phase diagram. / When the melt is superheated to a high temperature (about 200 to 300 degrees above the liquidus temperature), the iron intermetallic crystallizes in the $ alpha$-phase at high cooling rates. This behaviour is attributed to the fact that $ gamma$-alumina which forms at low melt temperatures ($ le$750$ sp circ$C) acts as a nucleus for crystallization of $ beta$-phase. When the melt is superheated to a high temperature ($ ge$850$ sp circ$C), the $ gamma$-alumina transforms to $ alpha$-alumina. The $ alpha$-alumina is found to be a poor nucleus for the $ beta$-phase crystallization, and as a result the $ alpha$-phase forms. The importance of nucleation and growth undercooling for the crystallization of iron intermetallics is highlighted. / Investigation of the dissolution behaviour of the iron intermetallics on non-equilibrium heat treatment indicates that the $ beta$-phase platelets dissolve slowly through concurrent fragmentation and then dissolution at the plate tips. Addition of manganese hinders the dissolution of iron intermetallics. The amount of liquid phase formed during non-equilibrium heat treatment increases dramatically once a critical temperature is exceeded. This critical temperature is estimated to be 520 and 515$ sp circ$C for samples initially solidified at 10 and 15$ sp circ$C/s respectively. / The 0.15% Fe alloy exhibits the highest tensile strength and percent elongation compared to 1.0% Fe and 1.0% Fe + 0.5% Mn alloys under as-cast and equilibrium heat treated conditions. However, under non-equilibrium heat treatment conditions, ie., 30 degrees above the T4 equilibrium solution temperature, the strength properties of 1.0% Fe alloy exceed, or at least equal, that of the equilibrium heat treated 0.15% Fe alloy. / An attempt is made to correlate the iron intermetallics present in the microstructure and mechanical properties with the associated fracture mode in this alloy.

In this work, an artificial standard material was developed and used to examine the robustness of various different stereological correction procedures. The four correction procedures that were examined were: large-sections correction, Hill's fast approximation, Barbery's correction and PARGEN correction. / There were three steps to this work: (1) refinement of the standard material--certain modifications were made to make the standard material (developed as an M.Eng. project) more versatile and easier to use. (2) sectioning and correction of computer-generated spheres--different liberation distributions of single-capped spheres were computer-generated and sectioned. The four correction procedures were used to correct the stereological bias in the sectioning data. The corrected liberation distributions were compared with the true liberation distributions. (3) sectioning and correction of standard material particles--the standard material was used to re-create the same distributions that were computer-generated. The particles were mounted, sectioned and the sectioning data measured with an electron microscope and image-analyzer. The data were corrected using the correction procedures and the corrected and true distributions were compared. / A two-phase (glass/lead borate) standard material was successfully developed. This standard material can be made to exhibit granular, layered or simple locking. / The correction of the sectioning of the computer-generated spheres and standard material particles yielded similar observations about the different correction procedures. / The large-sections correction provided a simple, uncorrupted correction. It performed better in the sphere cases than in the standard material cases. / Hill's fast approximation performed well except in the cases of narrow liberation distributions. The sectioning data of the standard material cases appeared to support the assumption in this correction that the locked section and locked particle distributions are identical. / Barbery's correction performed better in the standard material cases than in the sphere cases. The correction had problems in situations where an incomplete beta function could not be fitted to the true liberation distribution. / The PARGEN correction was able to provide a good estimate of the true amount of free material, but it had difficulty estimating the locked particle distribution.

Heat dissipation from silicon chips is becoming critical as a result of increasing circuit density and power applied to these chips. In this regard, aluminum nitride (AlN) is the most promising candidate to be used as a substrate since it possesses high a thermal conductivity and thermal expansion coefficient close to that of silicon. In addition, metal-ceramic interfaces determine the heat dissipation through a circuit. In order to minimize the effect of interfaces, direct bonding (DB) of AlN to Cu was studied. DB process has been reported to provide a low thermal barrier and an interface which is free of thermal fatigue. / In this work process parameters of DB were optimized based on time, temperature and thickness of the Cu-foil for Cu-Al$ sb2$O$ sb3$ system in a N$ sb2$ atmosphere containing 500 ppm O$ sb2$ in a temperature range of 1065 to 1075$ sp circ$C. These conditions were then applied to the Cu-AlN system. Wettability of AlN by Cu was studied and improved through oxidation of AlN and modification of Cu by adding 1 at.% O$ sb2$. The activation energy for oxidation of AlN was found to be 94 kJ/mol. It was then shown that direct bonding of Cu to AlN can be performed without any intermediate layer. The average peel strength of AlN-Cu, A$ sb2$O$ sb3$-Cu and AlN-Al$ sb2$O$ sb3$-Cu systems were measured to be 42, 49 and 14.7 MPa, respectively.

The recrystallization behavior of three Nb-bearing HSLA steels was investigated during multipass deformation under continuous cooling conditions. The niobium concentrations of these steels varied from 0.05 to 0.09 wt%. The specimens were tested on a computerized torsion machine using a simulation schedule of 17 passes. Deformation temperatures of 1180$ sp circ$C to 700$ sp circ$C were employed, together with pass strains of 0.1 to 0.7, strain rates of 0.2 to 10 s$ sp{-1}$, and interpass times of 2 to 200 seconds. By means of mean flow stress vs. 1000/T diagrams, the effect of reheating temperature, initial grain size, and chemical composition on the T$ sb{ rm nr}$ (temperature at which recrystallization is no longer complete) was determined. Reheating temperature and microalloying elements have significant influence on the T$ sb{ rm nr}$, while initial grain size has no effect on the T$ sb{ rm nr}$ during multipass deformation. / Furthermore, the effect of deformation parameters such as the pass strain, strain rate, and interpass time on the T$ sb{ rm nr}$ during multipass deformation was investigated in this way. The T$ sb{ rm nr}$ decreases with increasing strain and also decreases slightly with increasing strain rate. There is a T$ sb{ rm nr}$ minimum at times of about 12 $ sim$ 15 seconds and both increases and decreases from this value raise this characteristic temperature. When the interpass times are short, solute atoms control the rate of recrystallization, the extent of which decreases as the time is decreased. When the interpass times are long, precipitation takes place and retards recrystallization, so that the extent of softening decreases as the time is extended instead. / The evolution of Nb(C,N) precipitation during simulated rolling was studied with the aid of carbon extraction replication and electron microscopy. Finally, by applying the additivity rule to the isothermal model of Dutta and Sellars, continuous cooling T$ sb{ rm nr}$'s were predicted from recrystallization-precipitation-temperature-time (RPTT) diagrams; these are shown to be in good agreement with the experimental observations.

The work of this thesis centres around two related themes, the first being the characterization of the through-thickness inhomogeneity found in rolled sheet at various stages in its production. The second is involved with attempting to understand how and to what extent the existing inhomogeneity affects the correlations with the anisotropies of some physical or mechanical properties. For this, the inhomogeneity of six steels were measured. / Three were specimens removed from the sheet during the early processing of conventional grain-oriented ferrosilicon steel, one sample taken after the first cold rolling stage, another taken after the intermediate anneal, while the last was taken after the following stage, i.e. after the second cold rolling. / To contrast the extensive inhomogeneity expected in the ferrosilicon steels, the remaining three steels consisted of two continuous-annealed, interstitial-free, extra-low-carbon steels and one aluminum-killed, batch-annealed, low-carbon steel, all three of which are commercial final product, deep drawing steels.

In hot strip mills, in the steel industry, reheat furnaces are utilized to reheat slabs, billets, blooms, etc., to uniform rolling temperatures prior to hot working. / During this reheating operation, hot steel surfaces react with the in-furnace oxidizing atmosphere resulting in the formation of an iron oxides layer (scale). The yield loss due to this phenomenon depends on furnace operating conditions, i.e. steel temperature, excess combustion air, steel residence time in the furnace, etc., and ranges between 1.5 to 3% of reheated steel. / In this research, the oxidation of a mild steel during reheating cycles as function of furnace atmosphere, steel residence time in the furnace and steel temperature was investigated. / It has been found that scale formation was most sensitive to oxygen levels in the furnace, steel residence time and temperature. A saving of up to 35% of steel lost to scale was achieved by reducing the excess air in the furnace from 70% to 20%. Also, longer residence time and higher temperatures of the steel in the furnace resulted in the formation of excessive amounts of scale.