The manufacture of ceramic composites has traditionally been a complex and often expensive process. A new processing method, the DIMOX$ sp{ rm TM}$ process, takes advantage of the high temperature oxidation behaviour of aluminum alloys to produce a ceramic-metal composite. Although this process is fairly simple to implement, there has not been any attempt to link the starting materials and manufacturing conditions to the properties of the final composite. This work attempts to identify some critical parameters in the manufacturing process and how they affect the end product. The reaction temperature, alloy composition, and powder bed condition were optimized with respect to the growth process, final composition, and resulting microstructure. These characteristics were then related to the elasticity, strength, fracture toughness, and fracture mode of the final composite. / Aluminum-magnesium-silicon alloys were oxidized into an alumina bed of either Alcan C-70 UNG power or Struers' 400 grit. The process conditions were optimized in air at 1120$ sp circ$C with a 10% silicon, 2% magnesium alloy. The growth rate was dependent on the powder bed. The material was composed of alumina, silicon, aluminum, and trace amounts of magnesium aluminate spinel. The fracture mode was dependent on the composition of the material and the alumina bed.

In this work, the carbothermal reduction of Al$ sb2$O$ sb3$ to AlN was studied. Several kinds of aluminum oxides including $ alpha$-Al$ sb2$O$ sb3$, $ gamma$-Al$ sb2$O$ sb3$, $ theta$-Al$ sb2$O$ sb3$ and boehmite (AlOOH) were examined in order to observe the differences in reaction behaviour and powder characteristics obtained from each type of precursor. Cane sugar (sucrose) and carbon black were used as carbon sources. Reaction conditions studied were carbon to alumina ratio, temperature and reaction time. Sucrose resulted in a close-to-stoichiometric ratio of Al$ sb2$O$ sb3$:C (1:3.2) achieving full conversion to AlN and produced a regular powder morphology, whilst carbon black required higher ratio ($>$1:4) to reach full conversion with agglomeration of the AlN powder. The optimal reaction temperature was 1600$ sp circ$C with the reaction time being dependent on the Al$ sb2$O$ sb3$ source. The results of the thermodynamic study for the Al-N-O-C system suggest a solid-state reaction process which is consistent with the experimental observations. Moreover, flowing N$ sb2$ gas flushes out the product CO gas and thus forces the equilibrium in favour of AlN formation. Reaction mechanisms are proposed for the two forms of carbon precursor.

Vacuum levitation experiments have been conducted to study the decarburization kinetics of levitated steel droplets in order to determine the factors and relationships which control the rate of decarburization especially at C levels below 30 ppm. It was found from the experiments that (1) vacuum chamber pressure had a significant effect on the rate of decarburization when the carbon content was below 35 ppm; (2) sulfur did not show any significant effect on the rate of decarburization due to the strong stirring inside the droplet caused by magnetic levitation field; (3) the rate of decarburization of levitated droplets was 3 ppm/sec at (C) = 30 ppm which was 40 times higher than the overall rate of decarburization in the RH process at (C) 30 ppm; (4) high initial oxygen contents improved the rate of decarburization at high carbon contents. / The following suggestions are made: (1) increase the amount of liquid steel droplets without increasing the size of the droplets; (2) increase the fraction of the amount of decarburization reaction inside the molten steel by gas and powder injection; (3) further reduce the partial pressure of CO and CO$ sb2$ gas in the gas phase. (Abstract shortened by UMI.)

The addition of frother in flotation has two main functions, to help reduce bubble size and help produce a stable froth. A role of frother on bubble behavior in pulp zone is usually not considered. A previous study showed that as frother type was changed the same gas holdup was given by different size bubbles. This implies that bubble rise velocity depends on the nature of the surfactant (frother type). A study using bubble swarms appears to support the frother type effect but bubble interactions are a possible confounding factor. This study resolved the question by measuring terminal rise velocity profile of single bubbles (ca. 1 to 2 mm) as a function of frother type. It is shown that at the concentrations of interest in flotation, 1-pentanol hardly alters the velocity compared to water alone while F150 (a polyglycol) reduces the velocity by up to 50%. The results become in 1-pentanol bubble did not reach terminal velocity. For high concentration of 1-pentanol (>130ppm) the rise velocity is reduced comparable to F150. To investigate, experiments were performed using aliphatic alcohols from 1- butanol (C4) to 1-octanol (C8). It was found there is a minimum concentration for the frother to give terminal velocity close to the Clift et al. contaminated eater result. The concentration decreases as molecular weight (chain length) of alcohol increases. Larger bubbles (1.8 vs. 1.5mm) require higher minimum concentration. To study the influence of molecular structure, three 6-C alcohols, 1-hexanol, MIBC and 2-hexanol, were used. The results show that molecular structure influences rise velocity through the position of OH group, and whether the alcohol is straight chained or branched. The observation can make a useful link frother to chemistry for understanding frother influence on bubble rise and possibly its function in flotation. The influence of three industrial frother, MIBC, F150 and DF250, was studied and / Deux raisons principales commandent à l´ajout de la mousse dans le processusde flottation, à savoir, la réduction de la bulle et la production d´une moussestable. Léffet de la mousse sur le comportement de la bulle en zone pulpairenést pas pris en compte. Un travail antérieur a démontré que pour la mêmefraction d´air transporté par les bulles, le type de mousse a de l´influence sur lataille des bulles. Cela implique que la vélocité de la bulle dépend de la nature dusurfactant (type de mousse). Une étude basée sur l´usage de plusieurs bullessemble s´accorder avec l´hypothèse relative au type de mousse, néanmoins lesinteractions des bulles rendent le problème complexe.La présente étude a résolu cette question en mesurant la vélocité d´une bulledont les dimensions sont presque égales à 1 ou 2 mm, suivant le type demousse. Il en est résulté que dans l´intervalle des concentrations d´importanceen flottation, le pentanol peine à influencer la vélocité, alors que léau toute seuleen serait capable. En revanche, le F150 (un polyglycol) réduit la vélocité de 50%.Pour les concentrations élevées de pentanol (>100 ppm), la vélocité décroit etdevient comparable à celle engendrée par la présence du F150. Il va de soi quel´observation antérieure est confirmée dans l´intervalle pratique desconcentrations.En guise d´investigations, des expériences au cours desquelles les alcoolsaliphatiques allant du butanol (C4) à l´octanol (C8), ont été réalisées. Il a étédémontré que léffet du type de mousse sur la vélocité dépend de laconcentration. Il existe une concentration minimale de la mousse en dessous delaquelle, la bulle monte comme dans le modèle de léau contaminée de Clift etal. La concentration minimale critique dépend de la dimension de la bulle et dudéplacement mesuré à partir du point terminal du tube capillaire. L´on rapporteque les bulles les

The addition of Nb significantly retards recovery and recrystallization through solute drag and precipitation pinning effects. It is important to describe the precipitation behaviour precisely because finely distributed precipitates retard the restoration process abruptly. The isothermal precipitation behaviour has been studied frequently. However, during industrial hot deformation processing, the temperature decreases continuously, so that isothermal data cannot be applied directly to predict the precipitation under these conditions. The focus of this study is therefore on the CCP (Continuous-Cooling-Precipitation) behaviour of Nb carbonitride in austenite. / To calculate the CCP behaviour, isothermal PTT (Precipitation-Time-Temperature) data in terms of the P$ sb{ rm s}$ (precipitation start) and P$ sb{ rm f}$ (precipitation finish) times are first required. In 1987, Dutta and Sellars developed a thermodynamic model to predict P$ sb{ rm s}$ times for the precipitation of niobium carbonitride, and in 1989, Liu and Jonas developed a model for titanium carbonitride. In the present work, the L-J (Liu-Jonas) model is used to calculate the P$ sb{ rm s}$ time at a given temperature from experimental data. Since the prediction models are only limited to the P$ sb{ rm s}$ times, a new calculation method for the P$ sb{ rm f}$ time, based on reaction kinetics and classical nucleation and growth theory, is proposed in this study. Two models are developed to follow the precipitation process, and the time exponent and rate constant for the kinetics are formulated for each model. The additivity rule, which was developed by Scheil in 1935, is then used to calculate the extent of precipitation during continuous cooling. / Isothermal precipitation rates for 0.040% Nb steels are measured experimentally by the stress relaxation method. The CCP behaviour is then calculated and is then examined by continuous cooling testing, using a hot deformation dilatometer. Precipitates are observed by transmission electron microscopy of specimens quenched after a period of cooling at various cooling rates. The P$ sb{ rm s}$ and P$ sb{ rm f}$ times estimated from the particle size data show good agreement with the calculated CCP behaviour.

The subject of this work is gallium extraction from sulphate solutions--an additional source of this metal from hydrometallurgical zinc production--with organophosphorus acid reagents: di-2-ethyl hexyl phosphoric acid (D2EHPA) and OPAP, a mixed extractant consisting of mono- and di-octyl phenyl phosphoric acids. Extraction proceeds via cation-exchange and Ga$ sp{3+}$ is the reacting species. Gallium is extracted with D2EHPA mostly as GaR$ sb3$ - HR. The results for OPAP suggest existence of four reactions, which form GaM$ sb3,$ GaM$ sb2$D, GaMD$ sb2,$ and GaD$ sb3;$ this explains and allows prediction of behaviour over a wide range of OPAP compositions. Sulphate complexation causes decrease in concentration of the reacting species, and thus lower $D sb{ rm Ga}$ values and extraction rates. Prior knowledge on gallium aqueous complexes is used, and an algorithm developed, to allow quantitative prediction of complexation effects on extraction. The model of mass-transfer with chemical reaction, verified with several known criteria for reaction site determination, describes well the kinetic data for the Ga-D2EHPA system. The model is further developed to account for the stronger acidity and the monomer/dimer equilibria typical for the kind of extractants used. A detailed reaction mechanism is proposed and the first organic ligand addition is found as rate-limiting. The model parameters, estimated from extraction kinetic data, are reasonable, when compared with those obtained for other metals elsewhere. The model's predictions agree with the results from stripping kinetics; the equilibrium conditions (zero rate) can also be satisfactorily predicted, as found by comparisons with the equilibrium data. Ga-D2EHPA and Ga-OPAP systems are compared with an emphasis given to the potential for metal separation; the importance of the ligand exchange rate constant is illustrated with the example of Ga and Al extraction/stripping and their separation based on different rates with

For TRIP steels to be included as a future structural steel in automobile industries, the drawbacks associated with high silicon (more than 1%) must be solved. To this end, TRIP steels, with aluminum as a substitute for the silicon, have been under investigation by many groups. / In this work, the approach is to investigate the effect of hot deformation on the microstructure and the mechanical properties of the aluminum containing TRIP steels. The problem with these steels is that the strength is reduced compared to classical Si TRIP steels. In this work, a number of mechanisms could contribute to the strengthening, including refining the austenite grain and work hardening of ferrite. These may also affect the retained austenite characteristics. / Two different C-Mn-Al TRIP aided steels with aluminum contents of 1.43 wt % and 2.0 wt % were investigated and compared with a typical TRIP steel containing 1.5 wt % Si. To determine the critical hot working temperatures, the FactSage software was used to predict the phase diagram and the critical temperatures were confirmed by heat treatment and metallography of the quenched specimens. Using these data, hot deformation experiments were performed to investigate the influence of the deformation temperature and the strain within the intercritical region by using a Gleeble simulator. (Abstract shortened by UMI.)

Steel media wear has been well studied and quantified; one method of quantification is the Comminution Dynamics Lab's Total Media Wear model. It combines simulations, abrasion and corrosion experimental trials to determine wear for the components of milling and grinding ores and minerals for the mining industry. Breakage is an important measure of process efficiency, the greater the number of breakage events per unit energy consumed, the greater the throughput and production. The classic breakage parameter, the Bond Work Index generates an energy term used by mill operators to determine this performance criterion. It has been suggested that the abrasion test, used in the Total Media Wear Model, would be a suitable alternative to the labour and time-consuming Bond Locked-cycle Test. Extensive tests were performed in order to ascertain the possibility of obtaining these two desired results from a single two-minute test. Different steel media samples were tested at different energy levels, and finally by testing under wet and dry conditions. Dry and wet testing did not generate the same wear results. Wear and breakage rates were higher under wet conditions. In general, the size distribution of the abrasive feed evolved into a product with a finer size distribution. Statistical analysis of the data obtained suggests that there is indeed a linear relationship between the energy input into the system and the resultant Work Index value. These results support the suggestion that this test will be able to recreate the Bond Work Indices for minerals; however, more work is required in order to build a working database and determine appropriate correlation factors. / L'usure des composants ferriques dans les procédés minéralurgiques et un sujet bien étudié. Le laboratoire « Comminution Dynamics » a McGill à crée un modèle nommé « Total Media Wear Model » pour prédire le temps de l'avance nécessaire avant lequel il faudra remplacer les blindages et boulettes. Cette modèle comprend des simulations informatiques et des tests de laboratoire. Une autre mesure utilisée par les contremaitres et operateurs c'est le « Bond Work Index ». L'index donne une valeur approximative pour prédire la consommation d'énergie des moulins. Un test du style « Bond » exige beaucoup d'effort, et temps. Récemment, il a été suggérer qu'on pouvait obtenir des valeurs pour le « Bond Work Index » en utilisant un procédé modifié pour un des tests du « Total Media Wear Model ». Différentes nuances d'acier on été testé avec des niveaux d'énergie variées et sous des conditions sec et humide. Les résultats on démontré que les tests sec/humides n'était pas nécessairement équivalent, de plus, un changement dans la distribution granulométrique s'est produit entre la charge et le produit. Et finalement, l'analyse des données suggère qu'il y a une relation entre la valeur calculée de « Work Index » et l'énergie introduit dans la système. Ces résultats confirment les croyances qu'on pourra déterminer la valeur de « Bond Work Index » en même temps que les données du « Total Media Wear Model ». Cependant, il faudra des tests supplémentaires pour plusieurs types de minéraux. Cela pourra crée une base de données avec lequel des facteurs de corrélations pourront être déterminer.

Gravity-Recoverable-Gold (GRG) is defined as gold present in a particle in sufficient quantities as to be selectively recoverable from gangue via gravity methods. The McGill standard GRG test is an ore characterization test using three stages of sequential liberation and recovery with a Knelson KC-MD3 centrifuge to determine the size distribution of GRG. This thesis describes the development and testing of two simplified versions of the GRG test, using two and one stages of recovery respectively. Both tests use a feed mass of 20 kg, as opposed to the 40 to 100 kg normally used for the standard test. Eighteen differing ore samples were processed with the simplified GRG tests. For non-abrasive ores the one-stage simplified test returns a similar GRG content and size distribution, making the two-stage test superfluous. For abrasive ores, the one-stage test returns a GRG content that can be as much as 33% relative lower than that of the standard test, with a much finer size distribution. The two-stage test exhibited similar poor performance, though to a slightly lesser degree due to and additional stage of recovery attempted prior to grinding the abrasive material. The GRG lost typically reports to size fractions coarser than 25 &mgr;m, strongly suggesting smearing onto gangue particles. Because of the lower feed mass used, both simple tests are susceptible to the nugget effect; feed representativity also becomes challenging for ore samples of a head grade of 1 g/t or less.

The INCO finishing procedure represents the near final stages of copper production at the Copper Cliff Smelter in Sudbury, Ontario, whereby semi-blister copper, at 4--6 wt % Ni and 2--3 wt % S, is processed into blister copper, at 100 ppm S and 0.4 wt % Ni in a modified Pierce-Smith converter combining top-blown oxygen injection and bottom-blown nitrogen stirring of the melt to accelerate semi-blister desulfurization. The procedure is performed in two stages of equal duration: in the first, melt saturation with oxygen and sulfur elimination from 2--3 wt % S to 300 ppm is achieved by 45 minutes of oxygen top-blowing and nitrogen bottom stirring; in the second stage, oxygen injection is terminated and sulfur removal from 300 ppm to 30--100 ppm. proceeds by the ppm to 30--100 ppm proceeds by the reaction of dissolved S and O under a further 45 minutes of nitrogen stirring. / Cu-S and Cu-Ni-S alloys were also exposed to the open atmosphere to examine desulfication of stagnant melts under air. (Abstract shortened by UMI.)