A three-dimensional mathematical model for the simulation of vertical direct chill (DC) slab casting of aluminum alloys has been developed. The model is based on the solution of the 3D time-averaged turbulent momentum (Navier-Stokes) and energy equations. The momentum equations are modified with a Darcy-type source term to simulate motion of the melt in the mushy region. The buoyancy force term is implemented in the model through the Boussinesq approximation. The low Reynolds number k-ε turbulence model of Launder and Sharma is used to calculate the Reynolds stresses and the turbulent heat fluxes. A variable heat transfer coefficient is used along the ingot surface to account for the different cooling regions. The mathematical model is qualitatively and quantitatively verified by comparing the computed results with a physical water model and a real casting experiment, respectively of independent researchers. Each of the comparisons showed a good agreement. The quantitative verification of the solidification front depths is improved when the thermal buoyancy force effect is included in the model. / A parametric study has been carried on two casters of variable aspect ratio each using a different type of inlet melt distribution system. In the case of the small aspect ratio caster, the physical properties of aluminum Al-3104 are used. For this caster, the studied parameters are the casting speed, the primary cooling rate, the melt superheat and the combo-bag dimensions. Also, the effect of complete blockage of the bottom windows of the distribution bag is studied. An in-depth understanding of some behaviors of the melt flow and solidification profile in the steady state operational phase of the DC casting process is gained. For example, the roles played by the angle flow and the upward component of the vertical recirculation at the wide symmetry plane in controlling both the solidification front depth and the mushy layer thickness at the slab center are ascertained. This study has revealed the influence of the melt stream issued from the bottom window of the bag on the depth and uniformity of the solidification front. The model has successfully identified a faulty design of the short combo bag. This industrially favorite design causes what is called the reverse flow, that is, the melt from the surrounding sump enters the combo bag through the bottom window. (Abstract shortened by UMI.)

Different types of zinc bicrystals were tested under creep conditions. Observations of grain boundary sliding along boundaries with different misorientations were made. It was established that Sigma = 9 coincidence boundary slides faster than boundaries of general type. Grain boundary sliding along coincidence boundary was activated at higher stress and only in the presence of intragranular slip. Sliding along coincidence and near-coincidence boundaries is accompanied by regular boundary migration whereas sliding along general boundary is not. The sliding-migration behavior of coincidence boundaries was explained in terms of DSC-dislocations associated with steps. Sliding along general boundaries was described in terms of non-DSC grain boundary dislocations having infinitesimal Burgers vector. / It is established in bicrystals with individual boundaries that crystallographic slip can increase significantly the rate of grain boundary sliding. Two mechanisms of stimulation of sliding by slip are proposed. Intragranular slip can make a direct contribution to grain boundary sliding in incompatible conditions of deformation. Identical deformation of grains can increase the rate of sliding only by its accommodation at boundary irregularities. The close relationship between sliding and slip at incompatible deformation makes intragranular slip strongly dependent on boundary structure and geometry of slip and independent of the Schmid factor. / It is demonstrated that deformation of textured fine-grained Zn-1.1%Al alloy in the direction favorable for intragranular slip decreases the contribution of grain boundary sliding to the total strain. Deformation in the direction unfavorable for slip increases its contribution. This shows that grain boundary sliding and intragranular slip are concurrent and independent processes during superplastic deformation.

Visualization and quantitative oxidation kinetic experiments on 100 g samples of Pb-As at 600°C; thermal analysis and phase-equilibrium measurements of Pb-PbO-As2O3 samples under argon over the temperature range of 420°C to 875°C; computational thermodynamic solution modeling; and phase diagram and equilibrium calculations using FACTSage(TM) were performed to elaborate the poorly documented thermodynamics of the slags in the lead softening stage in the pyrometallurgical refining of lead. In the softening stage, the minor element impurities: arsenic, antimony and tin are removed from lead bullion by oxidation and are transferred to a skimmable oxide slag phase. / It was found that optimizing an ionic molten oxide solution model that was conceptualized to contain Pb2+ and O2- with AsO3-4 and AsO3-3 ions, or with SbO3-4 and SbO3-3 ions in the respective PbO rich regions of the Pb-As-O and the Pb-Sb-O systems, was able to accurately reproduce the measured and published thermodynamic data. It was also found that the subsystems in the PbO-As2O 3-As2O5 and PbO-Sb2O3-Sb 2O5 systems showed small deviation from the ideal ionic solution model and small magnitude excess Gibbs energy parameters were sufficient to fit the predicted liquidus curves to the experimental measurements. / Arsenic in the +3 and +5 oxidation states was measured in the PbO rich region of the Pb-As-O liquid solution in the temperature range of 420°C to 875°C. The variability in the ratio of trivalent arsenic to the total arsenic content, as well as the complex variation of arsenic distribution between metal and oxide phases found strong interaction between the lead, arsenic and oxygen atoms at the 3PbO to 1AS2O3 molar ratio thus suggesting a short range ordering corresponding to the formation of AsO3-3 groupings, and indicating that the Pb3(AsO3) 2(l) species was likely to be present in the PbO rich region of the Pb-As-O system and contributing to an understanding of the Pb-As-O liquid oxide structure. Also, two new compounds (Pb3(AsO3) 2(s), Pb2AsO4(s)) were identified in the Pb-PbO-As 2O3 quenched samples via wavelength-dispersive spectrometry using the electron microprobe. The present work has application in commercial oxygen partial lead softening (OPLS), as uniquely practiced at Teck Cominco Ltd., British Columbia. There, pure oxygen gas is injected into the bath of impure bullion through a number of submerged lances in order to oxidize only part of the arsenic, antimony and tin into a slag phase. For such an operating practice, it was concluded from the visualization and quantitative oxidation experiments that the formation of solid oxides as the product of oxidation produced a physical barrier to the progress of oxidation and resulted in the commercially observed, highly-problematic, process initiation issues. When the product was liquid, there was much less of a barrier to rapid oxygen mass transfer to the minor element impurities and the softening reactions were easy to initiate. Such a change in the physical state of the products of oxidation was correlated to the optimized ternary Pb-As-O and Pb-Sb-O phase diagrams. / A current point of interest in partial lead softening is to increase the arsenic content of the slag phase. Arsenic distribution between lead bullion and slag calculated by the optimized solution model of the Pb-As-O system suggests that this can be achieved in a counter-current contacting of the slag and bullion.

Magnesium additions to aluminum enhance strength and hardness, lowering density, and improve corrosion resistance. However, during melting and casting processes, significant amounts of magnesium are lost due to selective oxidation. Preventing these losses would reduce the production cost and improve the quality of the final product. / The effects of various levels of strontium addition on the oxidation behaviour of aluminum-magnesium alloy melts were investigated by monitoring sample weight gains with time using a thermogravimetric balance at 750°C. Sample cross-sections have been examined in detail using Hitachi S-4700 Field Emission Gun Scanning Electron Microscope, and phases formed on the oxide layer and in the alloy morphology were identified by EDS, WDS and low angle X-ray diffraction techniques. It was observed that in the absence of Sr, the Al-Mg samples gained substantial amounts of weight by formation of spinel (MgAl2O4) phase at the oxide-metal interface. Samples containing Sr had significantly lower weight gains. The drop in total weight gain by Sr additions was about 98% in low Mg-containing Al-Mg alloys. This change in oxidation behavior was linked to the presence of a Sr-enriched liquid phase underneath the first formed MgO layer, suppressing the formation of spinel crystals. In addition, spherical equilibrium shapes are found in the different as-cast and oxidized alloys of Al-Mg-Sr.

Welding based coating deposition techniques allow high rates of material deposition and form a permanent metallurgical bond between the coating and the substrate material. Welding based methods can also provide an economic alternative over other industrial coating deposition processes where high initial capital investment and running costs can be restrictive. As with all technological sectors, the need for new and improved machinery and processes to meet industrial needs provides a drive for continued research. The controlled short-circuit MIG (CSC-MIG) welding system is a newly developed welding apparatus built to overcome several shortcomings associated with traditional MIG welding. It allows for greater control of many welding parameters and has reduced heat input during deposition when compared with conventional MIG welding systems. This project was conducted to understand the CSC-MIG welding system as a process and as a hardfacing deposition technique through examination of the microstructural features and transformations of Ni/WC coatings. Several coatings deposited with a Ni/WC electrode wire, with heat input ranging between 10 J/mm and 110 J/mm, were examined. It was found that the detrimental decarburization reactions acting on the WC particles, as seen in thermal spray systems, do not occur when welding with the CSC-MIG. Although the energy input during welding with the CSC-MIG system is significantly lower than for traditional MIG, dissolution of the reinforcing phase is an issue to be contended with and must be minimized through proper selection of welding parameters. Precipitation of a reaction layer around the WC/W2C reinforcing phase was identified as WC; the average thickness of which increased from 3.8 um to 7.2 um for the low and high heat input condition, respectively. Precipitation of newly formed WC particles was also observed; their size distribution increased from D50 = 2.4 um in the low heat input weldment to D50 = 6.75 um in the high heat input weldment. The hardness of the deposited coatings decreased from 587 HV10 to 410 HV10 when the energy input was increased from 10.1 J/mm to 108.7 J/mm. Using a pre-placed powder method, as in submerged arc welding, several coatings were embedded with either conventional, ultrafine or nanostructured WC powder. In the analysis of these tests, it was found that the method of embedding the WC particles into the coating had an effect on the overall dissolution of the reinforcing phase. Although the loss of the nanostructure was observed in coatings embedded with the nanostructured WC feedstock, the precipitation of ultrafine WC single crystals is likely to increase the wear resistance compared with conventional sized WC additions. / Les techniques de revêtement à base de soudage permettent de hauts niveaux de déposition et forment une liaison permanente entre revêtement et le substrat. Les techniques basées sur le soudage peuvent aussi fournir une alternative économique aux autres méthodes de revêtements industriels où les coûts sont souvent plus élevés. Comme dans tous les secteurs technologiques, le besoin pour des solutions industrielles nouvelles et améliorées contribue au dynamisme pour la recherche continuelle. Le système de soudage par MIG aux courts-circuits contrôlés (CSC-MIG) est une machine récemment développé pour surmonter quelques points faibles du soudage MIG classique. Celui-ci permet un meilleur contrôle des paramètres de soudage et un débit de chaleur moins élevé en comparaison avec les systèmes de soudage MIG traditionnels. Ce projet a été réalisé afin de comprendre le système CSC-MIG en terme de processus et de technique de déposition de revêtement en examinant les caractéristiques et transformations des couches déposées avec un fil-électrode à base Ni/WC. Plusieurs couches avec des apports énergétiques entre 10 J/mm et 110 J/mm furent examinées. Il a été démontré que la décarburation des particules de WC, qui caractérisent les revêtements obtenus par projection thermique, n'affecte pas les couches produites avec le CSC-MIG. Bien que les apports énergétiques avec le CSC-MIG soient réduits en comparaison avec le MIG traditionnel, la dissolution des particules de WC est un phénomène auquel on doit faire face tout en minimisant sa gravité en optant pour une sélection de paramètres de soudage appropriée. La précipitation d'une couche réactive a l'entour de la phase renforcent WC/W2C a été identifiée comme étant du WC; l'épaisseur de cette phase a crût de 3.8 um à 7.2 um pour les conditions de soudage à basse et haute intensité respectivement. La précipitation de nouvelles particules de WC a également été observée; leur distribution de taille a augmenté de D50 = 2.4 um pour la condition de soudage à basse intensité à D50 = 6.75 um pour la condition de soudage à haute intensité. La dureté des revêtements a diminué de 587 HV10 à 410 HV10 lorsque l'apport énergétique a été accrue de 10.1 J/mm à 108.7 J/mm. En utilisant une méthode de poudre pré-placé comme pour le soudage sous flux, il a été découvert que la méthode d'incorporage des particules de WC dans le revêtement a eu un effet sur la dissolution de la phase de renforcement. Malgré que la nanostructure des revêtements faites avec la méthode de poudre pré-placée et ayant des particules nanostructurées ait été perdue, la précipitation de particules ultra-fines de monocrystal risque tout de même d'augmenter le niveau de résistance à l'usure en comparaison avec des revêtements provenant de particules conventionnelles.

In this thesis, the effects of iron(II) on arsenic(V) removal from acidic sulphate solutions in lime neutralization systems were investigated. The role of Fe(II) was analyzed via different types of experiments. Firstly, 2-stage continuous coprecipitation (CCPTN) circuit experiments were run, involving variable Fe(II)/Fe(III) fractions whilst maintaining an Fe(tot)/As(V) molar ratio of 4, and the resultant products were subjected to stability testing. It was found that CCPTN results were reproducible; that increasing the Fe(II) content resulted in somewhat lower initial arsenic(V) removal, but still better results than those obtained from equimolar Fe(III)-As(V) solutions in the absence of ferrous pointing to the latter's beneficial effect on arsenic(V) precipitation and retention. Coprecipitates aged at constant pH 8, drifting pH and at various temperatures reached pseudoequilibrium after several months. Notably, long term stability testing of the coprecipitates showed that up to an Fe(II)/Fe(III) ratio of 1 at 20 °C, As(V) release was maintained below 1 mg/L after 463 days with "drifting pH" stabilized at 5 increasing only to 1.9 mg/L upon pH adjustment to 8. Secondly, the behaviour of Fe(II) was studied in batch reactor set-ups as part of hydrolysis and oxidation experiments with and without As(V). It was shown in the absence of As(V), Fe(II) precipitated out of solution completely between pH 7.5 and 8.5. Subsequent oxidation of the ferrous hydroxide slurry was found to proceed via a series of transformations starting from ferrous hydroxide to green rust, to magnetite and finally goethite. The oxidation kinetics were governed by oxygen mass transfer. In the presence of As(V) both Fe(II) and As(V) precipitated from solution starting at pH 3 with the latter ultimately dropping below 1 mg/L between pH 6.5 to 9 via the proposed precipitation of a ferrous arsenate compound (symplesite). Subsequent oxidation of the Fe(II)-As(V) slurry at constant pH 8 led to destabilization of the ferrous arsenate phase and concomitant partial release of As(V). The bulk control of As(V) in the latter case switched from ferrous arsenate to ferric arsenate or arsenate adsorption on freshly formed iron(III) hydroxide. / Cette thèse à pour objet d'étudier les effets des ions ferreux (Fe(II)) sur la stabilisation, par neutralisation à base de chaux, de l'arsenic (As(V)) contenu dans des solutions acides sulfatées. Le rôle des ions ferreux a été analysé à l'aide de différents types d'expériences. Premièrement, des essais de co-précipitation en circuit continu (CCPTN) comprenant deux étapes ont été réalisés pour différentes fractions de Fe(II)/Fe(III), tout en conservant un rapport molaire Fe(tot.)/As(V) égal à 4; les produits obtenus ont par la suite été soumis à des tests de stabilité à long terme. Les résultats de ces tests ont montré de façon reproductible qu'une augmentation de la teneur en ions ferreux réduisait l'effet de stabilisation de l'As(V) initialement présent; ces résultats étaient cependant meilleurs que dans le cas de tests de stabilisation de l'As(V) présent dans des solutions équimolaires de Fe(III)-As(V), en l'absence d'ions ferreux, validant l'effet positif des ces derniers sur la précipitation et la rétention de l'As(V). Après plusieurs mois de vieillissement dans des conditions variées de pH constant (ajusté à pH 8), de pH non-ajusté et de températures, les produits de co-précipitation ont fini par atteindre un état de pseudo-équilibre. Notamment, les tests de stabilité à long terme ont montré que pour une fraction molaire Fe(II)/Fe(III) allant jusqu'à 1 et une température de 20 °C, la libération d'As(V) en solution après 463 jours était maintenue en-dessous de 1 mg/L, respectivement 1.9 mg/L, dans le cas d'une solution au pH non-ajusté (se stabilisant à pH 5), respectivement d'une solution au pH constamment ajusté à pH 8. Deuxièmement, le comportement des ions ferreux a été étudié à l'aide d'un réacteur discontinu, dans le cadre de tests d'hydrolyse et d'oxydation, en présence ou non d'As(V). Les résultats de cette partie de l'étude ont montré qu'en l'absence d'As(V), les ions ferreux précipitent intégralement entre pH 7.5 et 8.5. Par la suite, l'oxydation de la suspension d'hydroxyde de fer (II) procède selon une série de transformations allant de la rouille verte, à la magnétite et finalement à la goethite. Les résultats ont également montré que la cinétique d'oxydation était gouvernée par le transfert de masse d'oxygène. En présence d'As(V), la précipitation du Fe(II) et de l'As(V) à été observée à partir de pH 3, sous la forme suggérée d'un composé d'arséniate de fer (II) (symplésite), la concentration finale d'As(V) non-précipité atteignant moins d'1 mg/L entre pH 6.5 et 9. Par la suite, l'oxydation de la suspension de Fe(II)-As(V) à pH 8 constant a entrainé la déstabilisation de la phase d'arséniate de fer et la remise en solution partielle d'As(V). En effet, dans ce cas particulier, le control de l'As(V) à entrainé la conversion de la majorité de la phase d'arséniate de fer (II) en arséniate de fer (III) ou possiblement son adsorption à la surface d'hydroxyde de fer (III) fraichement oxydé.

Detection of particulate matter thinly dispersed in a fluid medium with the aid of the difference in electrical conductivity between the pure fluid and the particles has been practiced at least since the last 50 to 60 years. The first such instruments were employed to measure cell counts in samples of biological fluid. Following a detailed study of the physics and principles operating within the device, called the Electric Sensing Zone (ESZ) principle, a new device called the Liquid Metal Cleanliness Analyzer (LiMCA) was invented which could measure and count particles of inclusions in molten metal. It provided a fast and fairly accurate tool to make online measurement of the quality of steel during refining and casting operations. On similar lines of development as the LiMCA, a water analogue of the device called, the Aqueous Particle Sensor (APS) was developed for physical modeling experiments of metal refining operations involving water models. The APS can detect and measure simulated particles of inclusions added to the working fluid (water). The present study involves the designing, building and final application of a new and improved APS in water modeling experiments to study inclusion behavior in a tundish operation. The custom built instrument shows superior performance and applicability in experiments involving physical modeling of metal refining operations, compared to its commercial counterparts. In addition to higher accuracy and range of operating parameters, its capability to take real-time experimental data for extended periods of time helps to reduce the total number of experiments required to reach a result, and makes it suitable for analyzing temporal changes occurring in unsteady systems. With the modern impetus on the quality of the final product of metallurgical operations, the new APS can prove to be an indispensable research tool to study and put forward innovative design and parametric changes in industrially practised metallurgical operat / La détection de particules finement dispersées dans un fluide à l'aide de la différence de conductivité entre le liquide pur et les particules est pratiquée depuis environ 50 à 60 ans. Le premier instrument utilisant cette technique fut employé pour mesurer le nombre de cellules dans des échantillons de fluide biologique. Suivant une étude détaillée de la physique et du principe d'opération de l'appareil appelé zone de détection électrique (Electric Sensing Zone -- ESZ), un nouvel appareil, baptisé analyseur liquide de propreté en métal (Liquid Metal Cleanliness Analyzer -- LiMCA), pouvant mesurer le nombre de particules d'inclusion dans les métaux en fusion fut inventé. Cet instrument constituait un outil rapide et précis pour faire des mesures en ligne de la qualité de l'acier au cours du raffinage et de l'opération de coulée. Dans la même lignée de développement que le LiMCA, un outil analogue, spécifique à l'eau, appelé senseur de particules aqueuses (Aqueous Particle Sensor -- APS) fut développé pour la modélisation physique des expériences reliées aux opérations de raffinages des métaux impliquant des modèles aqueux. L'APS peut détecter et mesurer les particules d'inclusions simulées ajoutées au fluide (l'eau). Cette recherche comprend la conception, la réalisation ainsi que l'application finale d'un APS (dans l'eau) nouveau et amélioré appliqué aux expériences de modélisation afin d'étudier le comportement de l'inclusion dans une opération de répartiteur. L'instrument bâti sur mesure démontre une performance supérieure, comparé à ses équivalents commerciaux, applicable dans des expérimentations comprenant de la modélisation physique des opérations de raffinage de métaux. En plus d'une précision accrue et d'un plus grand éventail de paramètres d'opérations, sa capacité de traiter des données expérimentales en temps réel pour de longues périodes rend l'outil en question plus appropriée p

Physical and mathematical modelling of fluid flow phenomena play an important role in studies of steel quality associated with the various steps in steelmaking and steel processing operations. / In the continuous casting process, the tundish not only serves as an intermediate buffer, but with increasing demands for high quality steel, it has also evolved into a useful reactor for liquid steel refining. It is now a proven fact that a modern tundish is designed to carry out different metallurgical operations such as inclusion separation and floatation, alloy trimming, superheat control, calcium doped inclusion modification, thermal and particulate homogenization. / To carry out those operations effectively, certain parameters like Residence Time Distribution (RTD), Residual Ratio of Inclusions (RRI), Slag Entrainment and Vortex formation phenomena are generally studied and optimized using various tundish designs, and flow modification devices. These are usually specific to a given plant practice and to a targeted grade of steel. / A full scale physical model of a delta shaped, four strand tundish using water as the simulating fluid for liquid steel was used to study the fluid flow behavior of a 12 t tundish producing aluminium-killed steel billets., The above mentioned parameters were studied to predict the performance of the tundish using two different types of flow modification systems.

The focus of this research is the interaction of gold(III)chloride with activated carbon in acidic sulphate solutions. More specifically in this research the in-situ adsorption of gold (III) chloride on activated carbon within an autoclave in the presence of iron sulphide oxidation products, is investigated. Two types of experiments are presented in this research. In the first one, gold(III) chloride (1.5x10-4 M Au(III) plus 0.05 M extra chloride) was allowed to adsorb onto activated carbon (10, 20 and 30 g/L) at pH 1.5 and at different temperatures (22, 60 and 95 oC). Complete gold removal was achieved within a few minutes of contact. The kinetics were found to increase with increasing carbon loading as well as temperature and exhibit "first order" diffusion-controlled behaviour. However an increase in chloride concentration from 0.05 M to 1 M was found to slow down gold adsorption. Via X-ray photoelectron spectroscopic analysis (XPS), it was determined that gold undergoes a reduction upon adsorption on carbon and is present partly as Au(I)Cl and partly as metallic gold. The relative amount of metallic gold increased with temperature but decreased with high chloride ion concentration.The second type (carbon-in-autoclave) of experiment involved pressure leaching (at 155 and 200 oC) of pyrite particulates (acting as model sulphide mineral) spiked with metallic gold (precipitated via cementation) in the presence of low chloride ion concentration and activated carbon. Under these conditions gold was found to lixiviate forming gold chlorocomplexes that reported (in part) on the available activated carbon. The loaded gold, analyzed by XPS was found to be ~2/3 metallic and ~1/3 Au(I)Cl. Because of a "preg-robbing" effect exhibited by hematite forming in-situ, excess sodium was added to the system favouring natrojarosite as the latter has no affinity for gold. In this way gold loading on carbon was maximized. However, significant occurrence of gold deposition on metallic surfaces was observed. By using a glass autoclave the gold plating occurrence was eliminated allowing up to 97% recovery of gold on the activated carbon. Similar results were obtained when copper concentrate material was substituted for pyrite. This work shows that even in small chloride ion concentrations (<0.2 M) gold is easily leached under autoclave oxidation conditions allowing conceptually for its in-situ recovery on activated carbon. / L'objectif de cette recherche est d'analyser le comportement du complexe Au(III)Cl4- en contact avec du charbon actif dans un environnement hydrométallurgique acide contenant des ions de sulfate. Cette recherche détaillera plus précisément l'adsorption du complexe Au(III)Cl4- sur le charbon actif dans un autoclave à température et pression partielle d'oxygène élevée en présence des produits d'oxydation de sulfure de fer. Cette recherche présentera deux types d'expériences. En premier lieu, le chlorure d'or(III) (1.5x10-4 M Au(III) et 0.05 M chlorure) a été adsorbé sur le charbon actif (10, 20, 30 g/L) à un pH de 1.5 et à différentes températures (20, 60 et 95 oC). Le complexe aurifère est complètement adsorbé par le charbon actif après quelques minutes. La cinétique augmente proportionnellement avec une augmentation de la teneur d'or sur le charbon actif de même qu'avec l'augmentation de la température tout en présentant un comportement de diffusion du premier ordre. Toutefois, une augmentation de la concentration de chlorure de 0.05 M à 1 M ralentit l'adsorption d'or sur le charbon actif. Suite à une analyse par rayons-X spectroscopiques photoélectriques (XPS), il a été possible de déterminer que le complexe aurifère (Au(III)Cl4-) subit une réduction sur le charbon actif et est présent en partie comme or natif (métallique) et or complexé Au(I)Cl. La quantité relative d'or métallique adsorbée sur le charbon actif augmente avec la température, mais diminue avec une concentration de chlorure élevée.Le deuxième type (Carbon-in-autoclave) d'expérience consiste à oxyder sous pression (à 155 oC et 200 oC) des particules de pyrite (agissant comme minerais de sulfure idéal) dopé avec de l'or métallique (précipité via cémentation) en présence d'une faible concentration d'ions de chlorure et de charbon actif. Dans cet environnement, l'or est lixivié formant des chlorocomplexes d'or qui ont été adsorbés (en partie) sur le charbon actif. L'or chargé sur le charbon actif a été analysé par XPS et les résultats démontrent la presence de 2/3 métallique et 1/3 Au(I)Cl. Suite au phénomène de « preg-robbing » dû a l'hématite formée durant l`oxydation de la pyrite, un surplus de sodium a été ajouté afin de favoriser la formation du natrojarosite qui n'a aucune affinité avec le chlorure d'or(III). Ainsi, l'adsorption de l'or sur le charbon actif est maximisée. Toutefois, une cémentation importante d'or a été observée sur les surfaces métalliques internes de l'autoclave. En utilisant un autoclave en verre, la quantité d'or plaqué a été éliminé permettant ainsi une récupération de 97% d'or sur le charbon actif. Des résultats similaires ont été obtenus lorsqu'un concentré aurifère de cuivre a été oxydé à la place de la pyrite. Ce travail démontre donc que même de faibles concentrations de chlorure (<0.2M) peuvent facilement lixivié l'or métallique en autoclave permettant théoriquement sa récupération sur du charbon actif durant le procédé d'oxydation.

The research aims to develop the use of ozone in nickel oxidation-precipitation from dilute solutions. The process presents three main steps: ozone transfer from gas to liquid, ozone decomposition by hydroxyl ions, and nickel oxidation-precipitation. Each step was studied using a semi-batch bubble reactor. For ozone gas-liquid transfer, a methodology to calculate the volumetric mass transfer coefficient ("kLa") was developed. Parameters evaluated included: diffuser porosity, volume of liquid and superficial gas velocity. To calculate the mass transfer coefficient ("kL") interfacial area ("a") is estimated from gas holdup and bubble size measurements. Ozone decomposition by OH" was examined as a function of pH and two ranges were identified, above and below pH 7.2. A kinetic equation for pH > 7.2 is proposed. Decomposition rate below pH 7.2 is low and considered negligible. Ozone oxidation-precipitation of nickel (II) in aqueous sulfate solution was determined as a function of pH, ozone and nickel concentration. Xray diffraction (XRD) analysis determined that the products comprised Ni(OH)2, NiOOH, Ni302(OH)4 and Ni202(OH)4. From the pattern of ozone consumption and the XRD results the proposed reaction sequence is hydrolysis - precipitation - oxidation rather than oxidation - hydrolysis - precipitation suggested in the literature.