Texture in films develops during deposition processes and annealing of patterned wafers. Recent studies show that texture influences the performance and reliability of both aluminum and copper interconnects. To improve the current understanding of this matter, the influence of texture on reliability was studied. / The influence of texture on electromigration and stress-induced failures in aluminum interconnects was studied since these are the most often responsible for failures observed in aluminum interconnects. Results obtained showed that a strong {111} texture in aluminum interconnects improves their median time-to-failure. The grain boundary character distribution and percentage of coincidence site lattice (CSL) boundaries, was quantified using orientation imaging microscopy. It was found that the median time-to-failure of specimens increased as the number of low angle and CSL boundaries increased. These boundaries are known to have low diffusivity. It was also demonstrated that while the investigated specimens had grains of comparable size, the grains of similar orientations were clustered in the specimens having the stronger {111} texture. This phenomenon contributed to the longer median time-to-failure of the interconnects by reducing the frequency of high angle grain boundaries. The experimental data obtained shows that the residual stress in films decreases as the intensity of the {111} texture increases. A model based on Monte-Carlo simulation of texture formation during the deposition of aluminum film was proposed to suggest the optimum conditions for a growth of a strong {111} texture component. A low deposition rate and a high mobility of atoms on the surface, which corresponds to a high substrate temperature, can strengthen {111} texture. / Copper has been recently used as an interconnecting material because of its good electromigration resistance and low electrical resistivity. One of the major problems of copper as an interconnecting material is that it easily oxidizes at relatively low temperatures. The formation of oxide degrades the electrical and mechanical properties of copper interconnects. The influence of substrate texture on the oxidation kinetics was studied to suggest methods to reduce copper oxidation. Copper single crystals having (100), (110), (123), (314), (111) and (311) orientations were oxidized at 200ºC in air. Only the Cu2O phase was formed during oxidation. The oxidation of the (100) single crystal substrate was much faster than that of the others. This is attributed to a large number of fine oxide grains on the (100) crystal in the initial stages of oxidation. It is recommended that the {100} texture in copper interconnects should be avoided in order to reduce oxidation rate. A quantitative model was proposed to predict the oxidation kinetics of copper from the texture of the specimens. Reasonable agreement was obtained comparing the model predictions and the experimental results obtained from the test of oxidation of polycrystalline copper specimens. However, further improvement of the model can be done if more data from single crystal experiments are obtained.

A newly developed creep method is described for following the kinetics of manganese sulfide precipitation in four electrical steels. The technique was also applied to a Ti steel, in which the precipitation kinetics were previously determined using a stress relaxation method. The resulting creep rate is sensitive to the occurrence of precipitation; thus the slope of the true strain-log(time) curve decreases immediately after the initiation and increases on the completion of precipitation. The precipitation-time-temperature diagrams determined in this way are of classical C shape on all of the tested materials. Because higher dislocation densities and internal stress levels are maintained, the present technique is more effective for monitoring the precipitation events occurring in the ferritic phases and at high temperatures than the previously developed stress relaxation method. / The nucleation sites for precipitation during creep testing were investigated with the aid of electron microscopy. It was concluded that the progress of MnS precipitation can be divided into three stages: nucleation, growth and coarsening. There exists a strong tendency for the nuclei to be formed on dislocations at lower temperatures. However, nucleation at grain boundaries is preferred as the temperature is increased. The classical diffusion-controlled growth theory is not satisfactory for strain-induced MrS precipitation. Thus, a modification was made in such a dynamic case by taking the effects of deformation vacancies into account. The kinetics during the coarsening stage are controlled by both bulk and grain boundary diffusion. / The microstructural examinations revealed that both grain deformation and grain boundary sliding take place in the electrical steels at hot working temperatures. Quantitative evaluations showed that the latter process contributes about one-sixth to one-tenth the total strain under the current experimental conditions. The strong retardation of grain deformation after the occurrence of precipitation is explained in terms of the interfacial pinning of mobile dislocations by these incoherent particles. It is also demonstrated that the finely-dispersed precipitates slow down the boundary sliding rate. (Abstract shortened by UMI.)

Thermomechanical processing, which is a combination of deformation and heat treatment, is an optimum method to control the microstructural evolution and, accordingly, to generate the desired mechanical properties of materials. TRIP (Transformation-Induced-Plasticity) behavior is a powerful mechanism with which to improve mechanical properties. The basis of TRIP behavior is the retention of austenite with optimum characteristics (volume fraction, stability, size, morphology and composition) at room temperature. The transformation of retained austenite to martensite during deformation can lead to TRIP-enhanced properties. This work deals with the effects of thermomechanical processing parameters on the microstructural characteristics of TRIP steels, primarily from the point of view of the retained austenite condition in Si-Mn and Si-Mn-Nb bearing TRIP steels. Initially, a new test technique based on continuous cooling compression (CCC) testing was developed to find the critical temperatures of thermomechanical processing (TMP). A major finding from the CCC test is the ability to determine the $Ae sb3$ (equilibrium austenite-to-ferrite transformation) temperature. In the CCC test, the $Ae sb3$ appeared to be associated with an increase in the rate of increase in flow stress with decreasing temperature. In order to further evaluate this result, neutron diffractometry at high temperatures was used to monitor any crystallographic changes associated with the metastable region. The results revealed an increase in the rate of contraction of the austenite lattice as the temperature decreases through the metastable state (below $Ae sb3),$ compared with that observed as the temperature decreases through the stable austenite region. Having defined the critical TMP temperatures, the effects of processing parameters on the state of the retained austenite were examined by changing the thermomechanical processing conditions. These effects were more fundamentally considered by tak

The synthesis of silicon nitride (Si$ sb3$N$ sb4$) by carbothermal reduction of silica was sensitive to the processing steps used for the starting materials (SiO$ sb2$, C) and to the reaction conditions. A good degree of intimacy between the starting materials was required in order to obtain fine ceramic powders. / The formation of Si$ sb3$N$ sb4$ was known to occur over a narrow temperature range (1450-1550$ sp circ$C). The morphology of the silicon nitride powder produced was also shown to vary widely depending on the reaction temperature. / The amount and morphology of the silicon carbide formed was sensitive to impurities present in the precursors prior to the reaction. The nitrogen flow rate during reaction as well as the position of pellets in the reactor played a key role in the final SiC content of the powder. / Results showed that the strength values of two carbothermal powders were excellent (507 and 577 MPa) considering their lower sintered densities which was the result of residual carbon present in the powder. (Abstract shortened by UMI.)

In the process of extracting silver from lead bullion, zinc is added to the lead bullion to form a zinc crust. The crust is subsequently distilled and metallic zinc is condensed for reuse. In the present work, four aspects of the process were studied: the thermodynamic properties of zinc in zinc crust (the Pb-Ag-Zn ternary system); the heat and mass transfer of zinc vapour in a vertical tube condenser in the presence of a non-condensable gas; the cooler-condenser design for the vapour and gas mixture; and the zinc vapour condensation behaviour.

The rejection of iron sulphides (pyrite and pyrrhotite) using nitrogen was evaluated on three Canadian complex sulphide ores. Laboratory scale work (at McGill and on-site) was conducted and, in addition, continuous minicell testwork was conducted on-site. / The promotion of pyrite flotation using nitrogen was confirmed. It was evaluated at Brunswick Mining ahead of zinc flotation (using the minicells), and at Kidd Creek Mines in a laboratory scale investigation of reverse flotation. This latter also showed enhanced chalcopyrite flotation with nitrogen. / The explanation of pyrite flotation using nitrogen was based on weakening of galvanic coupling with other sulphides. / Nitrogen proved more effective than air in depression of pyrrhotite in processing a pentlandite/chalcopyrite ore. Under some conditions, chalcopyrite was also depressed by nitrogen (apparently when the pulp potential was less than O V vs. S.H.E.). / The depressant action of nitrogen on pyrrhotite was interpreted as a reduction in the density of surface OH$ sp-$ sites necessary for electrostatic interaction with xanthate ion in the Hodgson/Agar model of pyrrhotite floatability.

The aqueous behaviour of rhodium (III) (Rh) in chloride solutions was investigated with the objective of evaluating its response to solvent extraction with 8-hydroxyquinoline derivatives. A theoretical analysis of the abundance of the various chloro-complexes in solution revealed that monoaquo-pentachlororhodate and hexachlororhodate are the most common forms of Rh(III) in chloride solutions ($>$0.5M Cl$ sp{-}$) and that the relative abundance of hexachlororhodate may be lower than was previously reported. / The extraction degree of Rh(III) with 8-hydroxyquinoline derivatives from solutions ranging in acidity from 0.7 to 8.0M and in chloride concentration from 0.7 to 4M was found to follow the calculated abundance of hexachlororhodate (about 5 to 40% in the region investigated). Also, the extraction mechanism was deduced to be ion-pair formation. / The extraction degree of Rh(III) from similar solutions but previously treated with SnCl$ sb2$, was also investigated, and in this case the extraction was quantitative throughout the acidity and chloride concentration ranges tested. Virtual quantitative stripping can be achieved with four five-minute contacts with 1.7M H$ sb2$SO$ sb4$ + 1M Na$ sb2$SO$ sb4$. A preliminary flowsheet for the industrial application of this solvent extraction process for Rh is presented.

The amount of foam volume generated in relation to the gas flowrate has long been used to define the foaminess of a solution and to characterize surfactants. The vessel geometry and gas flowrate range can affect the foam volume results. A fixed area column was tested alongside a variable area conical vessel with three surfactants (commercial frothers). It was found that the fixed area vessel would cause foam volume to expand or contract, which masks the effect of the surfactant. The conical vessel showed two regimes of foam volume production dependant on gas flowrate. The low gas flowrate regime was tentatively associated with a quiescent bubbly zone while the high gas flowrate regime associated to a turbulent bubbly zone. Foaming properties of a surfactant are best determined with a variable area vessel at low gas flowrate. / L'étude du volume de mousse généré en corrélation avec le flux de gaz est utilisée depuis longtemps afin de déterminer la moussabilité d'une solution et d'en caractériser les agents de surface. La géométrie de la cuve ainsi que l'amplitude du flux de gaz sont connus pour avoir un effet sur les résultats de volume de mousse généré. Une colonne avec une aire de surface prédéfinie a été testée en parallèle avec une cuve conique de surface variable ; trois agents de surface (des moussants commerciaux) ont servi de base d'étude. Les résultats ont montré que la cuve possédant une aire de surface fixe causerait l'expansion ou la contraction de la mousse, masquant l'effet de l'agent de surface. Avec la cuve conique, deux régimes de production du volume de mousse, chacun dépendants du flux de gaz, ont été déterminés. Le régime correspondant à un flux de gaz faible est le produit d'une zone quiescente d'évolution des bulles alors que le régime correspondant à un flux de gaz élevé montre des propriétés typiquement cinématiques dues à une zone turbulente d'évolution des bulles. Les propriétés de moussage d'un agent de surface donné sont déterminées avec une meilleure exactitude en utilisant une cuve de surface variable avec un flux de gaz faible.

This research has focused on the behavior of gold (III) chloride in acidic solutions containing dissolved iron (III) sulphate and arsenic (V) that have been subjected to neutralization or hydrothermal treatments. In particular, the scope of the research was the study of the co-precipitation behavior of gold (III) chloride with various iron (III) – arsenic (V) precipitates that form either in pressure oxidation and leaching autoclaves, or during the neutralization of pressure oxidation/leaching discharge solutions.Gold (III) chloride (5x10-4 M in 0.05 M HCl) was found to co-precipitate with ferrihydrite formed during neutralization of ferric sulphate solutions, of variable concentration, at pH 4. The amount of gold chloride that co-precipitated from solution increased with increasing temperature (22 to 90 °C) and initial iron concentration (0.075 to 0.3 M) or equivalently the surface area of the precipitate. In the presence of arsenic (V) the degree of gold chloride co-precipitation decreased due to apparent competition for adsorption on the surface of ferrihydrite. It was found that gold chloride does not adsorb onto ferric arsenate itself, but only onto co-precipitating ferrihydrite, the latter being favored at elevated temperatures and at Fe/As ratios over 1. During the heterogeneous growth of scorodite, no gold (III) chloride co-precipitated but remained in solution. Under autoclave conditions (>200 °C) gold (III) chloride was found to co-precipitate completely during the formation of hematite, scorodite, ferric arsenate sub-hydrate (FAsH) and basic ferric arsenate sulphate (BFAS). Part of the gold chloride co-precipitated via adsorption or substitution, and the balance via reduction to a metallic state. The latter, based on thermodynamic analyses performed with the OLI software package - may be attributed to inadequate concentration of excess chloride. However, during the precipitation of natrojarosite at 200 °C there was no co-precipitation of gold chloride via adsorption or reduction. / Cette étude porte sur le comportement du chlorure d'or (III) en solution acide contenant du sulfate de fer (III) et de l'arsenic (V) sous forme de solutés soumis à des traitements de neutralisation ou bien de type hydrothermique. En particulier, le champ d'application de la recherche liée à cette étude s'étend au comportement de co-précipitation du chlorure d'or (III) avec des précipités de fer (III) – arsenic (V) qui se forment soit par oxidation sous pression et/ou dans des autoclaves de lixiviation, soit lors de la neutralisation de solutions récupérées après oxidation sous pression/lixiviation. Les résultats de cette étude indiquent que le chlorure d'or (III) (5x10-4 M dans 0.05 M HCl) se co-précipite avec la ferrihydrite formée pendant la neutralisation des solutions de sulfate de fer à concentration variable et à pH 4. La quantité de chlorure d'or co-précipitée à partir de la solution augmente avec la température (22 à 90 °C) et la concentration initiale en fer de la solution (0.075 à 0.3 M) mais également avec l'aire de surface du précipité. En présence d'arsenic (V), le degré de précipitation du chlorure d'or diminue, apparemment dû au mécanisme compétitif d'adsorption de l'arsenic (V) à la surface de la ferrihydrite. L'étude indique également que le chlorure d'or ne s'adsorbe pas à l'arséniate de fer mais seulement à la ferrihydrite co-précipitée, cette dernière étant favorisée par des conditions de températures élevées et de taux de Fe/As supérieurs à 1. Lors de la croissance hétérogène de la scorodite, aucun chlorure d'or n'est co-précipité mais reste en solution.L'étude réalisée en autoclave (>200 °C) à montré que le chlorure d'or se co-précipite dans son intégralité durant la formation d'hématite, de scorodite, d'arséniate ferrique sous-hydratée (FAsH) et d'arséniate et de sulfate de fer à caractère basique (BFAS). Une partie du chlorure d'or se co-précipite par adsorption ou substitution et le restant par réduction en or métallique. D'après des analyses thermodynamiques réalisées à l'aide du logiciel OLI, la réduction de l'or à l'état métallique pourrait s'expliquer par la présence inadéquate d'un excès d'ions chlorures. Cependant, lors de la précipitation de la natrojarosite à 200 °C, la co-précipitation du chlorure d'or par adsorption ou par réduction n'a pas été observée.

The cold gas dynamic spraying of commercially pure titanium coatings was investigated. Specifically, the relationship between several key cold spray parameters on the quality of the resulting coatings was studied in order to gain a more thorough understanding of the cold spray process. To achieve this goal, three distinct investigations were performed.The first part of the investigation focussed on the effect of propelling gas, particularly helium and nitrogen, during the cold spraying of titanium coatings. Coatings were characterised by SEM and were evaluated for their deposition efficiency (DE), microhardness, and porosity. In selected conditions, three particle velocities were investigated such that for each condition, the propelling gasses temperature and pressure were attuned to attain similar particle velocities for each gas. In addition, a thick and fully dense cold sprayed titanium coating was achieved with optimised spray parameters and nozzle using helium. The corresponding average particle velocity was 1173 m/s. The second part of the investigation studied the effect of particle morphology (spherical, sponge, and irregular) and size distributions (mean particle sizes of 20, 29, and 36 µm) of commercially pure titanium on the mechanical properties of the resulting cold sprayed coatings. Numerous powder and coating characterisations were performed. From these data, semi-empirical flow (stress-strain) curves were generated based on the Johnson-Cook plasticity model which could be used as a measure of cold sprayability. Cold sprayability can be defined as the ease with which a powder can be cold sprayed. It was found that the sponge and irregular commercially pure titanium powders had higher oxygen content, poorer powder flowability, higher compression ratio, lower powder packing factor, and higher average particle impact velocities compared to the spherical powders. XRD results showed no new phases present when comparing the various feedstock powders to their corresponding coatings. For all feedstock powder morphologies, it was observed that the larger the particle size, the higher the temperature generated on impact. For the spherical powders, the higher the temperature generated on impact, the lower the stress needed to deform the particle. In addition, as the kinetic energy of the impacting particle increased, the flow peak stress decreased while the final strain increased. Furthermore, higher final flow strains were associated with higher coating ΔHV10 (between the coatings and the feedstock powders). Similar relationships are expected to exist for the sponge and irregular feedstock powders. Based on porosity, the spherical medium powder was found to have the best cold sprayability.The final part of the investigation focussed on the effect of substrate surface roughness and coating thickness on the adhesion strength of commercially pure titanium cold sprayed coatings onto Steel 1020, Al 6061, and Ti substrates. Adhesion strength was measured by tensile/pull tests according to ASTM C-633-01 standard. Through-thickness residual stresses of selected coatings were measured using the modified layer removal method (MLRM). In addition, mean coating residual stresses were calculated from MLRM results. It was found that adhesion strength increases with increasing substrate surface roughness and decreases with increasing coating thickness. Furthermore, mean coating residual stresses were correlated with adhesion strength and it was suggested that higher adhesion strengths are associated with higher mean compressive stresses and a higher probability for adiabatic shear instability to occur due to the higher particle impact velocities. In general, it was found that under similar cold spray conditions and substrate surface preparation method, adhesion strength was strongest for commercially pure titanium coatings deposited onto Al 6061, followed by Ti, then Steel 1020. / Des revêtements de titane pur produits par projection à froid ont été étudiés. La relation entre les paramètres de projection à froid et la qualité des revêtements a été étudiée afin d'acquérir une compréhension approfondie du processus de fabrication. Pour atteindre cet objectif, trois études ont été effectuées.La première étude portait sur l'effet du gaz de propulsion (hélium et azote) lors de la projection à froid sur les propriétés des revêtements de titane pur. Pour chaque gaz, les conditions de pression et de température ont été variées afin de déterminer leur effet sur les vitesses moyennes des particules déterminées par une technique de diagnostic optique par temps de vol. Trois conditions de vitesse des particules identique pour les deux gaz ont étés choisis. Les résultats ont clairement démontré que pour la même vitesse des particules, les propriétés des revêtements diffèrent selon la nature du gaz employé. Cette étude a permis de conclure que la température du substrat, la température des particules ainsi que la pression exercée par le flux de gaz supersonique sur le substrat influences notablement les propriétés des revêtements de titane déposés par projection à froid.La deuxième étude portait sur l'effet de la morphologie des particules de titane pur (sphérique, éponge, et irrégulière) et la distribution de taille (taille moyenne: 20, 29 et 36 µm) sur les propriétés mécaniques des revêtements projetés à froid. Pour chaque poudre, plusieurs conditions de projection ont été testées et les vitesses de particules résultantes de même que l'efficacité de déposition ont été mesurées et les températures des particules ont été calculées. Des courbes semi-empiriques de contrainte-déformation ont été générées en utilisant le modèle de plasticité de Johnson-Cook. Ces courbes peuvent être utilisées comme une indication de la facilité de déposition. Pour toutes les poudres, les résultats ont montré que plus la taille des particules est grande, plus la température générée lors de l'impact est élevée. Pour les poudres sphériques, plus la température est élevée lors de l'impact, plus la contrainte requise pour la déformation des particules est basse. En outre, les observations ont montré que plus l'énergie cinétique de la particule sphérique augmente, plus la contrainte maximale est basse et plus la déformation maximale est élevée. Par ailleurs, plus la déformation maximale de la particule sphérique est élevée, plus la différence en microdureté entre le revêtement et la poudre avant la projection est élevée (ΔHV10). Des relations similaires sont attendues pour les poudres éponge et irrégulière. Basé sur la porosité des revêtements, la poudre sphérique avec la taille moyenne de 29 µm s'avère avoir la meilleure facilité de déposition lors de la projection à froid. La dernière étude porte sur l'effet de la rugosité de la surface du substrat et l'épaisseur du revêtement sur la force d'adhérence des revêtements de titane pur projetés à froid sur acier 1020, Al 6061, et Ti. Les contraintes résiduelles en fonction de l'épaisseur des revêtements et des différents substrats ont été mesurées en utilisant la méthode MLRM (Modified Layer Removal Method). Les résultats ont montré que la force d'adhérence augmente avec la rugosité de la surface du substrat initial et diminue avec l'épaisseur du revêtement. De plus, il a été suggéré que les forces d'adhérence élevées sont associées à des contraintes moyennes élevées en compression et à une probabilité plus élevée d'obtenir une instabilité de cisaillement adiabatique de se produire en raison d'une vitesse plus élevée des particules lors de l'impact. En général, les résultats ont montré que pour une condition de projection à froid et pour une préparation de surface quelconque, la force d'adhérence a été plus élevée pour les revêtements de titane pur déposés sur Al 6061, suivie par Ti, puis en acier 1020.