This thesis pertains to the processing of ultra-fine mineral particles and low rank coal using the hydrophobic--hydrophilic separation (HHS) method. Several explorative experimental tests have been carried out to study the effect of the various physical and chemical parameters on the HHS process. In this study, the HHS process has been employed to upgrade a chalcopyrite ore. A systematic experimental study on the effects of various physical and chemical parameters such as particle size, reagent dosage and reaction time on the separation efficiencies have been performed. For this, a copper rougher concentrate (assaying 15.9 %Cu) was wet ground and treated with a reagent to selectively hydrophobize the copper-bearing mineral (chalcopyrite), leaving the siliceous gangue minerals hydrophilic. The slurry was subjected to a high-shear agitation to selectively agglomerate the chalcopyrite and to leave the siliceous gangue dispersed in aqueous phase. The agglomerates were then separated from dispersed gangue minerals by screening and the agglomerates dispersed in a hydrophobic liquid (n-pentane) to liberate the water trapped in the agglomerates. The chalcopyrite dispersed in the hydrophobic liquid was separated from the medium to obtain a concentrate substantially free of gangue minerals and moisture. The copper recoveries were substantially higher than those obtained by flotation. The HHS process was also tested on ultrafine mono-sized silica beads. The results were superior to those obtained by flotation, particularly with ultrafine particles. The HHS process has also been tested successfully for upgrading subbituminous coals. Low-rank coals are not as hydrophobic as high-rank coals such as bituminous and anthracite coals. In the present work, a low-rank coal from Wyoming was hydrophobized with appropriate reagents and subjected to the HHS in a similar manner as described for processing copper. The results showed that the HHS process reduced the moisture substantially and increased the heating value up to 50% without heating the coal. Laboratory-scale tests conducted under different conditions, e.g., particle size, reagent type, reaction time, and pretreatments, showed promising results. Implementation for the HHS process for upgrading low-rank coals should help reduce CO2 emissions by improving combustion efficiencies. / Master of Science

Titre de l'écran-titre (visionné le 28 novembre 2023) / La chalcopyrite est un sulfure de cuivre et de fer trouvé dans une grande variété de gîtes minéraux. Sa composition en éléments traces est très variable. Parmi les facteurs les plus importants qui contrôlent sa concentration en éléments traces se trouvent la structure cristalline, les cœfficients de partage (chalcopyrite-liquide, chalcopyrite-minéraux qui co-cristallisent) et la composition du magma/fluide hydrothermal, lesquels reflètent les conditions physico-chimiques de l'environnement géologique de formation. Ces caractéristiques, ainsi que ses propriétés physiques, favorisent la chalcopyrite comme un potentiel minéral indicateur pour l'exploration. Avec l'objectif de déterminer les critères géochimiques permettant son utilisation comme un minéral indicateur, des échantillons appartenant à 8 types de dépôts à l'échelle mondiale, dont 2 magmatiques (sulfures à Ni-Cu et EGP lités) et 6 hydrothermaux (porphyres, sulfures massifs volcanogènes (SMV), iron-oxide-copper-gold (IOCG), skarn, or épithermal et or orogénique), ont été analysés par microsonde électronique (EPMA) et par ablation laser et spectrométrie de masse à plasma à couplage inductif (LA-ICP-MS). La composition en éléments traces de la chalcopyrite a été examinée à l'aide de méthodes statistiques pour caractériser sa variation dans les différents types de dépôts, lesquelles combinées avec l'apprentissage automatique, a permis le développement de modèles de discrimination. Cette étude montre que la différence la plus importante dans la composition de la chalcopyrite demeure entre les systèmes magmatiques et hydrothermaux. Le Ni détermine cette différence, enrichi dans la chalcopyrite magmatique par rapport à celle hydrothermale. Dans les systèmes magmatiques (chapitre 1), les processus ayant lieu pendant la cristallisation qui exercent une influence sur certains éléments ont été identifiés. La chalcopyrite des dépôts d'EGP lités est enrichie en Se et appauvrie en Te-Sn-Bi-In, comme une conséquence des rapports élevés magma/liquide sulfuré (facteur-R) pendant la cristallisation. La cristallisation fractionnée du liquide sulfuré dans les dépôts de sulfures à Ni-Cu impacte également sur la composition en éléments traces dans la chalcopyrite. Ainsi, la chalcopyrite formée à partir d'une solution solide intermédiaire (iss), cristallisant en dernier à partir d'un liquide résiduel riche en Cu, est enrichie en éléments incompatibles (Ag-Bi-Cd-Pb-Se-Sn-Tl), par rapport à celle formée d'une solution solide de monosulfure (mss) riche en Fe. Cela constitue la base des modèles de classification supervisée par l'analyse discriminante par les moindres carrés partiels (PLS-DA) développés dans cette étude, lesquels permettent de discriminer entre les deux types de dépôts magmatiques (sulfure à Ni-Cu vs. EGP lités) et le type de minerais. Dans les dépôts de SMV (chapitre 2), la concentration des éléments traces dans la chalcopyrite dépend de la composition (conditionnée par la roche encaissante), la température et de la nature du fluide hydrothermal. La PLS-DA indique que la composition en éléments traces de la chalcopyrite varie en fonction du sous-type lithostratigraphique. Ainsi, la chalcopyrite des SMV du sous-type ultramafique est enrichie en Ni-Co-Te, tandis que celle des SMV dans les environnements siliciclastique-felsique a des teneurs élevées en Sb-Bi-In. Dans les SMV en contexte mafique et siliciclastique-mafique, le Pb et Mn sont, respectivement, enrichis dans la chalcopyrite, tandis que dans celle des dépôts bimodal-felsique et mafique montrent une composition similaire. Des modèles de prédiction Random Forest (RF) ont été développés, un algorithme plus adapté et performant que la PLS-DA en raison du nombre élevé de classes (types de gîtes), afin de classifier la chalcopyrite parmi les 6 contextes lithostratigraphiques des SMV. Quatre algorithmes d'apprentissage automatique (Artificial Neural Network (ANN), RF, K-Nearest Neigbor (KNN) et Naïve Bayes (NB)) et une méthode statistique de classification supervisée (PLS-DA) sont comparés afin de classifier la chalcopyrite parmi les 8 types de dépôts (chapitre 3), dont le RF donne une performance plus élevée. Trois modèles RF ont été développés: le premier, permettant la classification entre chalcopyrite hydrothermale et magmatique; le deuxième, pour déterminer sa provenance selon le type de dépôt magmatique (sulfures à Ni-Cu vs. EGP lités); et le troisième, lequel discrimine la chalcopyrite parmi les 6 types de dépôts hydrothermaux. Ces modèles ont été testés sur des données compilées de la littérature et sur des grains analysés de chalcopyrite récupérés de sédiments de tills et eskers de la Province de Churchill (Québec, Canada), et démontre l'utilisation de la chimique de la chalcopyrite pour déterminer sa provenance aux fins d'exploration minérale. Une importante variation des éléments traces dans la chalcopyrite est présente dans chaque type de dépôt; néanmoins, sa composition a une signature particulière selon le contexte de formation. Cela met en évidence le fort potentiel de discrimination de la source de minéralisation, démontré par les modèles de discrimination. Cette étude démontre que la chalcopyrite peut être utilisée comme un minéral indicateur et offre l'opportunité d'appliquer ces méthodes dans les sédiments en exploration minérale. / Chalcopyrite is a Cu-Fe sulfide found in a wide diversity of mineral deposits. Its trace element composition is highly variable. The crystal structure, the partition coefficient (chalcopyrite-liquid, chalcopyrite-co-crystallizing minerals), and the composition of magma/fluid are major factors controlling the trace element content in chalcopyrite, which reflects the physicochemical conditions of geological environment of formation. These characteristics, as well as its physical properties, make chalcopyrite a potential indicator mineral for exploration. In order to determine geochemical criteria enabling the use of chalcopyrite as an indicator mineral, samples from 8 deposit types worldwide, including 2 magmatic (Ni-Cu sulfide and Reef-type PGE) and 6 hydrothermal (porphyry, volcanogenic massif sulfides (VMS), iron oxide-copper-gold (IOCG), skarn, epithermal gold and orogenic gold), were analysed by electron probe micro-analysis (EPMA) and LA-ICP-MS (laser ablation-inductively coupled plasma-mass spectrometry). The data were investigated with univariate, bivariate and multivariate statistical methods to characterise its variation in the different deposit types, which combined with the machine learning, enabled the development of discrimination models in order to determine the provenance of chalcopyrite. This study shows that the most important difference in trace element composition of chalcopyrite is observed between magmatic and hydrothermal systems. According to our results, Ni determines this difference, with values significatively higher in chalcopyrite formed in Ni-Cu sulfide and Reef-type PGE deposits than in hydrothermal deposits. In magmatic systems (chapter 1), the statistical analysis allowed the recognition of the processes occurring during crystallisation that affect the concentration of certain elements. In Reef-type PGE deposits, the high values of magma/sulfide liquid ratio (R-factor) during crystallisation led to Se enrichment (a strongly chalcophile element), whereas Te, Sn, Bi and In are depleted in chalcopyrite, compared to those formed in Ni-Cu sulfide deposits. Furthermore, the variation in trace elements in chalcopyrite is influenced by fractional crystallisation of sulfide liquid in Ni-Cu sulfide deposits. Thus, chalcopyrite from intermediate solid solution (iss), which crystallised from a Cu-rich residual liquid, is enriched in incompatible elements (Ag, Bi, Cd, Pb, Se, Sn and Tl) compared to chalcopyrite from an early-forming Fe-rich monosulfide solid solution (mss). These geochemical characteristics constitute the basis of the classification models with partial least squares-discriminant analysis (PLS-DA) to determine the deposit type (Ni-Cu sulfide vs. Reef-type PGE deposits) and ore type (Cu-rich vs. Fe-rich). In hydrothermal systems, the concentration of trace elements in chalcopyrite depends on hydrothermal fluid composition, which is a function of host rock composition, temperature, and nature of the fluid source. The PLS-DA results show that, in VMS deposits (chapter 2), trace elements in chalcopyrite vary in function of VMS setting, which is determined by the host rock composition. Chalcopyrite from ultramafic-hosted VMS is rich in Ni, Co and Te, whereas chalcopyrite from siliciclastic-felsic setting VMS has high values of Sb, Bi and In. Chalcopyrite from mafic and siliciclastic-mafic settings VMS is respectively high in Pb and Mn, whereas chalcopyrite from bimodal-mafic setting has similar trace element composition to those from bimodal-felsic subtype. Random Forest classifiers were developed and were shown to be more adapted and performance than PLS-DA because of high number of classes (deposit types), in order to classify chalcopyrite according to the 6 VMS lithostratigraphic settings. When 4 machine learning algorithms (Artificial Neural Network (ANN), Random Forest (RF), K-Nearest Neigbor (KNN) et Naïve Bayes (NB)) and a supervised classification multivariate method (PLS-DA) are compared to classify chalcopyrite according to 8 different deposit types (chapter 3), the RF model reported the highest performance. Three RF prediction models were developed. The first discriminates between magmatic and hydrothermal deposits; the second, discriminates between magmatic deposits (Ni-Cu sulfide and Reef-type PGE) and the third, discriminates between the 6 different hydrothermal deposits. These models were tested on literature data and chalcopyrite data from grains recovered from tills and eskers in Churchill Province (Québec, Canada), and demonstrate that chalcopyrite chemistry can be used to determine its provenance for mineral exploration. Although the trace element composition of chalcopyrite shows important variability within a deposit type, the results reveal that the chalcopyrite records a characteristic signature in its chemical composition according to the ore forming environment and highlights a strong potential to discriminate the source of mineralisation, as demonstrated by the RF models. This study demonstrates that chalcopyrite could be used as an indicator mineral and provides an opportunity to apply these methods in overburden sediments for mineral exploration.

Dans le domaine des cellules photovoltaïques à base de couches minces, l'alliage de Cu(In1-x,Gax)Se2 (CIGSe) constitue l'une des filières les plus avancées. Le passage à l'échelle industrielle soulève cependant de nouvelles problématiques. En effet, si le procédé standard de co-évaporation permet d'atteindre des rendements records supérieurs à 20 %, il reste relativement couteux à mettre en place. C'est ainsi que dans un contexte toujours plus compétitif, l'électrodépôt apparait comme une alternative de choix pour diminuer les coûts de production tout en garantissant des rendements compétitifs sur de grandes surfaces. Néamoins, de nombreuses propriétés spécifiques à ce mode de dépôt restent méconnues. En particulier les propriétés optiques à l'origine du photo-courant. Le premier objectif de cette thèse est donc de déterminer les constantes optiques de chaque couche du dispositif par ellipsométrie. Une attention particulière est donnée à la couche absorbante de CIGSe électrodéposée pour laquelle un protocole spécifique est employé. Une seconde partie de la thèse est dédiée à la réalisation d'un outil de caractérisation sans contact : la photoréflectance (PR). La PR va permettre de mesurer avec précision les énergies de transition interbandes d'un semi-conducteur, dont l'énergie de gap. Nous décrivons dans le détail le dispositif expérimental. Une implémentation originale utilisant une double modulation des sources a été développée et permet de réduire le bruit de mesure induit par la rugosité et la diffusion. L'étude de 14 échantillons de CIGS permet finalement de corréler des paramètres opto-électriques issus des caractéristiques courant-tension aux spectres PR. / Regarding, thin film photovoltaic market, Cu(In1-x,Gax)Se2 (CIGSe) based material is one of the most advanced technologies. Its high absorption coefficient allows it to absorb an important part of the solar spectrum with only two micron thickness. But while moving from fundamental research to the development of batch flow production, issues still remain. If the standard co-evaporation process lead to the best efficiency up to 20 %, high energy consumption is needed. In an increasingly competitive market, electroplating allows to reduce operating cost related to vacuum processes while guaranteeing competitive efficiencies on large scale modules. Nevertheless, due to the specificities of electroplating, new issues occur and some properties may differ from vacuum routes. In particular, optical properties which are responsible for photo-current generation. The first part of this thesis is thus devoted to obtain the optical constants for each layer of the device by spectroscopic ellipsometry. We pay special attention to the electrodeposited CIGSe absorber layer for which a specific method have been used in order to perform measurements on the back side. The second part of this thesis is dedicated to the development of an caracterisation tool : the photoreflectance (PR). The experimental setup is precisely described. An special implementation, using dual modulation technique, increases accuracy by removing luminescence and scattering perturbations. The study of 14 CIGS samples allows finally to correlate opto-electrical parameters from I(V) curves with PR spectra.

Semiconductor nanocrystals are interesting candidates as new light-absorbing materials for photovoltaic (PV) devices. They can be dispersed in solvents and cheaply deposited at low-temperature on various substrates. Also, the nanocrystals have unique optical properties depending on their size due to the quantum size effect and moreover it is easy to uniformly control their stoichiometry. CdTe/CdSe/CdTe heterostructure nanorods and I-III-VI₂ nanocrystals were selected to synthesize and investigate in order to utilize the benefits of colloidal nanocrystals described above. Colloidal nanorods with linear CdTe/CdSe/CdTe heterojunctions were synthesized by sequential reactant injection. After CdTe deposition at the ends of initially formed CdSe nanorods, continued heating in solution leads to Se-Te interdiffusion across the heterojunctions and coalescence to decreased aspect ratio. The Se-Te interdiffusion rates were measured by mapping the composition profile using nanobeam energy dispersive X-ray spectroscopy (EDS). The rate of nanorod coalescence was also measured and compared to model predictions using a continuum viscous flow model. The synthetic method of monodisperse chalcopyrite (tetragonal) CuInSe₂ nanocrystals was also developed. The nanocrystals have trigonal pyramidal shape with one polar and three non-polar surface facets. When drop-cast onto carbon substrates, the nanocrystals self-assemble into close-packed monolayers with triangular (honeycomb) lattice structure. Moreover, the effect of excess Cu precursor (CuCl) was studied for the formation of monodisperse trigonal pyramidal CuInSe₂ nanocrystals. The formation mechanism of monodisperse trigonal pyramidal CuInSe₂ nanocrystals was suggested with regard to excess amount of CuCl precursor, based on the nucleationgrowth model of colloidal nanocrystal formation. A new wurtzite phase of CuInS₂, CuInSe₂, and Cu(InxGa1-x)Se₂ (CIGS) was observed in nanocrystals synthesized by heating metal precursors and Se-(or S-)urea in alkylamine. X-ray diffraction (XRD) showed the predominant phase to be wurtzite (hexagonal) instead of chalcopyrite (tetragonal). High resolution transmission electron microscopy (TEM), however, revealed polytypism in the nanocrystals, with the wurtzite phase interfaced with significant chalcopyrite domains. / text

Solar photovoltaic modules provide clean electricity from sunlight but will not be able to compete on an open market until the cost of the electricity they produce is comparable to that produced by traditional methods. At present, modules based on crystalline silicon wafer solar cells account for nearly 90% of photovoltaic production capacity. However, it is anticipated that the ultimate cost reduction achievable for crystalline silicon solar cell production will be somewhat limited and that thin film solar cells may offer a cheaper alternative in the long term. The highest energy conversion efficiencies reported for thin film solar cells have been for devices based around chalcopyrite Cu(In, Ga)(Se, S)2 photovoltaic absorbers. The most efficient Cu(In, Ga)(Se, S)2 solar cells contain absorber layers deposited by vacuum co-evaporation of the elements. However, the cost of ownership of large area vacuum evaporation technology is high and may be a limiting factor in the cost reductions achievable for Cu(In, Ga)(Se, S)2 based solar cells. Therefore, many alternative deposition methods are under investigation. Despite almost thirty companies being in the process of commercialising these technologies there is no consensus as to which deposition method will lead to the most cost effective product. Non-vacuum deposition techniques involving powders and chemical solutions potentially offer significant reductions in the cost of Cu(In, Ga)(Se, S)2 absorber layer deposition as compared to their vacuum counterparts. A wide range of such approaches has been investigated for thirty years and the gap between the world record Cu(In, Ga)(Se, S)2 solar cell and the best devices containing non-vacuum deposited absorber layers has closed significantly in recent years. Nevertheless, no one technique has demonstrated its superiority and the best results are still achieved with some of the most complex approaches. The work presented here involved the development and investigation of a new process for performing one of the stages of non-vacuum deposition of Cu(In, Ga)(Se, S)2 absorber layers. The new process incorporates copper into an initial Group III-VI precursor layer, e.g. indium gallium selenide, through an ion exchange reaction performed in solution. The ion exchange reaction requires only very simple, low-cost equipment and proceeds at temperatures over 1000°C lower than required for the evaporation of Cu under vacuum. In the new process, indium (gallium) selenide initial precursor layers are immersed in solutions containing Cu ions. During immersion an exchange reaction occurs and Cu ions from the solution exchange places with Group III ions in the layer. This leads to the formation of an intimately bonded, laterally homogeneous copper selenide – indium (gallium) selenide modified precursor layer with the same morphology as the initial precursor. These modified precursor layers were converted to single phase chalcopyrite CuInSe2 and Cu(In, Ga)Se2 by annealing with Se in a tube furnace system. Investigation of the annealing treatment revealed that a series of phase transformations, beginning at low temperature, lead to chalcopyrite formation. Control of the timing of the Se supply was demonstrated to prevent reactions that were deemed detrimental to the morphology of the resulting chalcopyrite layers. When vacuum evaporated indium (gallium) selenide layers were used as initial precursors, solar cells produced from the absorber layers exhibited energy conversion efficiencies of up to 4%. While these results are considered promising, the devices were characterised by very low open circuit voltages and parallel resistances. Rapid thermal processing was applied to the modified precursor layers in an attempt to further improve their conversion into chalcopyrite material. Despite only a small number of solar cells being fabricated using rapid thermal processing, improvements in open circuit voltage of close to 150mV were achieved. However, due to increases in series resistance and reductions in current collection only small increases in solar cell efficiency were recorded. Rapid thermal processing was also used to demonstrate synthesis of single phase CuInS2 from modified precursor layers based on non-vacuum deposited indium sulphide. Non-vacuum deposition methods provide many opportunities for the incorporation of undesirable impurities into the deposited layers. Analysis of the precursor layers developed during this work revealed that alkali atoms from the complexant used in the ion exchange baths are incorporated into the precursor layers alongside the Cu. Alkali atoms exhibit pronounced electronic and structural effects on Cu(In, Ga)Se2 layers and are beneficial in low concentrations. However, excess alkali atoms are detrimental to Cu(In, Ga)Se2 solar cell performance and the problems encountered with cells produced here are consistent with the effects reported in the literature for excess alkali incorporation. It is therefore expected that further improvements in solar cell efficiency might be achieved following reformulation of the ion exchange bath chemistry.

In dieser Arbeit wurde das reaktive Anlassen von dünnen Kupfer-Indium-Schichten in elementarem Schwefel mittels energiedispersiver Röntgenbeugung untersucht. Durch die simultane Aufnahme der Röntgenspektren und der Messung der diffusen Reflexion von Laserlicht der Wellenlänge 635 nm an der Oberfläche der Probe während des Schichtwachstums von CuInS<SUB>2</SUB> konnte eine Methode zur Prozesskontrolle für ein Herstellungsverfahren von CuInS<SUB>2</SUB> etabliert werden. <br><br>Die Bildung von CuInS<SUB>2</SUB> aus Kupfer-Indium-Vorläuferschichten wird dominiert von Umwandlungen der intermetallischen Phasen. CuInS<SUB>2</SUB> wächst innerhalb der Aufheizperiode ab einer Temperatur von ca. 200°C aus der Phase Cu₁₁In₉. Jedoch zerfällt letztere metallische Phase in Cu₁₆In₉ und flüssiges Indium bei einer Temperatur von ca. 310°C. Das flüssige Indium reagiert im Falle von Kupferarmut mit dem Schwefel und führt zu einem zusätzlichen Reaktionspfad über InS zu CuIn₅S₈. Unter Präparationsbedingungen mit Kupferüberschuss wird das Indium in einer intermetallischen Phase gebunden.<br><br>Erstmals konnte die Phase Digenite bei Temperaturen über 240°C beobachtet werden. Beim Abkühlen auf Raumtemperatur wandelt sich diese Phase unter dem Verbrauch von Schwefel in Covellite um.<br><br>Für Proben mit Kupferüberschuss konnte eine Wachstumskinetik proportional zur Temperatur beobachtet werden. Dieses Verhalten wurde durch eine stress-induzierte Diffusion als dominierenden Reaktionsmechanismus interpretiert. Dabei werden während der Bildung von CuInS<SUB>2</SUB> durch unterschiedliche Ausdehnungen der metallischen und sulfidischen Schichten eine Spannung in der CuInS<SUB>2</SUB>-Schicht induziert, die nach Überschreiten einer Grenzspannung zu Rissen in der CuInS₂-Schicht führt. Entlang dieser Risse findet ein schneller Transport der Metalle zur Oberfläche, wo diese mit dem Schwefel reagieren können, statt. Die Risse heilen durch die Bildung neuen Sulfids wieder aus. / In this work the reactive annealing of thin copper and indium films in elemental sulphur was investigated by energy dispersive X-ray diffraction. Measuring simultanously laser light diffusively reflected from the growth surface, a simple method for process monitoring could be established. The process monitoring using 635 nm laser light can now independently be used in production.<br><br>The growth of CuInS<SUB>2</SUB> from copper-indium precursors is dominated by transitions between intermetallic phases. CuInS<SUB>2</SUB> growths in the heat up period above 200 °C from the phase Cu<SUB>11</SUB>In<SUB>9</SUB>. However the latter metallic phase decomposes into Cu<SUB>16</SUB>In<SUB>9</SUB> and liquid indium at a temperature of 310 °C. The liquid indium reacts in the case of copper deficiency with sulphur to InS. This leads to an additional reaction path via InS to CuIn<SUB>5</SUB>S<SUB>8</SUB>. Under preparation conditions with copper excess to the contrary, indium is bound in an intermetallic phase.<br><br>For the first time the phase Digenite could be observed in a growth experiment at temperatures above 240 °C. During cool down to room temperature this phase transforms to Covellite by consumption of sulphur.<br><br>For samples with copper excess a growth kinetic proportional to the temperature was observed. This behaviour is interpreted by a stress induced reaction mechanism: During the formation of CuInS<SUB>2</SUB>, strain in the CuInS<SUB>2</SUB> thin film is induced due to different expansion coefficients of the metallic and sulphurous phases. After transgression of a certain strain limit, cracks within CuInS<SUB>2</SUB> are formed. Along these cracks, fast transport of metallic species to the surface can occur. There these species can react with the sulphur. The cracks can heal up by the formation of new sulphides.

Ce mémoire est consacré à l'étude de deux composés à structure chalcopyrite que sont le thiogallate et le sélénogallate d'argent. La présentation de leur diagramme de phases et de leurs propriétés physiques a permis de mieux comprendre la complexité de leur cristallogénèse. L'objectif de cette étude a donc été d'améliorer chacune des différentes étapes intervenant dans leur élaboration afin d'optimiser leur qualité cristalline. Des mesures d'absorption résiduelle ainsi que des expériences de doublage laser à 10.6 µm ont permis de confirmer l'intérêt porté aux cristaux d'AgGaX2 (X = S ou Se), en vue de leur intégration dans des systèmes optiques non linéaires. Parallèlement à cette étude, une approche numérique de l'hydrodynamique de la phase fluide, a été effectuée afin de définir plus précisement les phénomènes inhérents à la cristallogénèse de ces composés.

This dissertation reports a study of the dissolution mechanism which governs the leaching of Cu from chalcopyrite (CuFeS2) in acidic media at atmospheric pressure and examines the differences between chemical (abiotic), leaching and bioleaching. An array of solution, solid surface and bulk speciation studies were used to make a comprehensive study of the CuFeS2 leaching process(es). / Thesis (PhD)--University of South Australia, 2008.

[pt] Os minérios de cobre-molibdênio são beneficiados juntos normalmente pelo processo de flotação com a ajuda de coletores sulfidrílicos de relativa seletividade. Os concentrados bulk obtidos devem ser processados com o intuito de remover aqueles coletores empregados, pois eles são prejudiciais na etapa de separação por flotação diferencial, ou seja, na obtenção de concentrados de molibdênio de alta qualidade com baixos conteúdos de cobre e ferro. O processo de flotação bulk é realizado em condições oxidantes na presença de coletores sulfidrílicos (xantatos, etc.). Na prática industrial na etapa de flotação diferencial, estes concentrados bulk são separados por modificação do potencial eletroquímico da polpa para valores considerados redutores -300 a -450 mV, com hidrosulfeto de sódio (NaSH) ou o reagente Nokes (Penta-sulfeto de fósforo dissolvido em hidróxido de sódio) além de cianeto de sódio em alguns dos casos, para melhorar a eficiência do processo. Neste quadro, o processo industrial de separação seletiva produz efluente e gases perigosos como o H2S e HCN que devem ser controlados adequadamente. Este trabalho visa avaliar o uso do reagente ácido pseudo glicol tiouréia (PGA) e ácido tioglicólico (TGA) como alternativas ao processo industrial voltado para a separação seletiva dos minerais de cobre e molibdênio. Para a avaliação do depressor foram empregadas medições de ângulo de contato, potencial zeta, testes de flotação em célula Partridge-Smith (PS), FTIR e testes em célula de bancada em um planejamento experimental. O reagente estudado apresentou boas propriedades depressoras para minerais de cobre. Nos resultados das medições de ângulo de contato, pelo método de bolha cativa, foram encontrados valores maiores para a calcopirita do que para a molibdenita (até 10 graus de diferença) e medidas entre 60 e 70 graus, porém no pH ácido esta diferença diminui e os dois minerais apresentam quase o mesmo ângulo de contato 60 mais ou menos 1 grau. Nas medidas efetuadas não foi empregado o coletor final (querosene) para a molibdenita devido a sua falta de homogeneidade na emulsão para realizar as medidas de ângulo de contato. Nos testes de flotabilidade na célula PS em misturas binarias de calcopirita-molibdenita foram encontradas na faixa ácida, valores de pH nos quais há melhores relações de concentração e recuperações, na ordem de 67 por cento para a molibdenita e apenas 10 por cento para a calcopirita no produto flutuado. Os resultados usando o reagente NaSH em condições similares apresentam recuperações similares de molibdenita, mas na faixa básica do pH. Baseados nestes resultados foram avaliadas diversas variáveis e gerado um planejamento experimental para passar à escala de bancada onde foram obtidas recuperações de 84 por cento no concentrado rougher com apenas 13 por cento de cobre corroborando as boas propriedades do depressor estudado. / [en] Copper-molybdenum porphyry deposits worldwide are exploited industrially by flotation of the ores of interest into a BULK concentrate with the aid of sulfhydryl collectors of high efficiency and low selectivity. The bulk concentrate obtained should be reprocessed in order to remove those collectors employed in the bulk stage since they impair the differential flotation during the separation step in order to obtain clean high quality molybdenite concentrates at low contents of copper and iron. The bulk flotation process is conducted under oxidizing conditions in presence of sulfhydryl collectors (xanthates, etc.). In industrial practice these bulk concentrates are then separated by modification of the pulp electrochemical potential to values considered reducing, between -300 to-450 mV, with sodium hidrosulphide (NaSH) or Nokes reagent (phosphorus pentasulfide dissolved in sodium hydroxide). Addition of sodium cyanide in some cases is used to improve process efficiency. Bulk concentrates produced are however a very small portion of the processed mineral (only 1/50 of the total ore milled), these industrial processes produce waste effluents and hazardous gases such as H2S and HCN if not adequately controlled. This dissertation seeks to assess the reagent pseudo glycol tiourea acid (PGA) and thioglycolic acid (TGA) as an alternative to industrial processing for the selective separation of minerals of copper and molybdenum. On the evaluation of this depressant were used contact angle measurements, zeta potential, flotation tests in Partridge-Smith cell (PS), FTIR and bench scale flotation testing cell with an experimental design. The reagent studied showed good depressant properties for copper minerals according to the results presented. Within the results of the contact angle measurements with the captive bubble technique were found greater contact angles for chalcopyrite than for molybdenite (up to 10 degrees difference) and values between 60 and 70 degrees, however in acidic pH this difference decreases and both minerals have almost the same contact angle as of 60 more or less 1 degree. Though on the contact angle measured values it was not employed the final collector (kerosene) for molybdenite due to lack of homogeneity on emulsions to make the contact angle measurements. On floatability tests (on PS cell) with binary mixtures of chalcopyrite-molybdenite were found pH values on the acidic range, at which good concentration ratios and recoveries in the order of 67 percent of molybdenite with only 10 percent of the chalcopyrite were obtained. The results using NaSH reagent under similar conditions gave similar recoveries for molybdenite, however it was found in the basic pH range. Based on the results, several variables were evaluated and an experimental design employed on bench scale flotation tests, the results showed recoveries of 84 percent were obtained in the rougher concentrate with 13 percent copper corroborating the good properties of this depressant.

Commodity projection studies anticipate an exponential increase in copper demand and decreasein mill grade of deposits currently exploited globally. Most of this demand is driven by theindispensable use of copper in electrification and currently, the growing demand in the contextof the green transition. To meet this challenge while supplying society with copper, producerssuch as Boliden AB have plans for exploration and expansion projects to increase their reservesand in turn increase production through efficient measures. As grinding is a bottleneck to mineralprocessing and the most energy intensive, more efficient comminution systems are beingconsidered for environmental sustainability. These are required to have lower energyconsumption while achieving the required liberation or better for subsequent processing byflotation. Consequently, the literature review shows there is a research query on whether more can beexpected from comminution to optimise for downstream processing. With most publishedresearch being on particles with a P80 less than 100 μm, there exists a gap in investigating theeffect of grinding on flotation for coarser size fractions which this degree project focuses on.Importantly, it includes evaluating chalcopyrite liberation and flotation performance when usinga Novel Comminution Device. Three grinding mills were considered, a vertical roller mill (VRM),a Novel Comminution Device (NCD), with a rod mill as the reference mill. The scope of this studywas an analysis of the mill products’ particle size distribution, chalcopyrite liberation andinvestigation of the flotation performance of the products from the three mills. This was done ona Cu-Au-Ag ore from the Boliden Aitik mine in northern Sweden. Particle size distribution (PSD) analysis revealed that at comparable P80 values, the VRMproduced particles had broader PSD curves than the rod mill, while the NCD provided a steeperproduct size distribution than its reference rod mill products. The chalcopyrite liberation analysisusing QEMSCAN automated mineralogy showed that for samples split into three size fractions, -45 μm, +45/-90 μm and +90 μm, the VRM and rod mill products had their highest proportion offree to liberated particles within the intermediate size fraction while for the NCD sample this waswithin the fine fraction. Overall, the NCD product had the highest free to liberated particles acrossall fractions. From shape analysis based on scanning electron microscopy (SEM) images, it wasobserved that the NCD produced particles with the higher elongation than the VRM and rod mill. The flotation performance was assessed with respect to the mass pull, kinetics, grade andrecovery in the concentrate and separation efficiency. The rougher batch flotation tests indicatedthat flotation of the VRM ground samples had higher mass pull, Cu recovery, faster kinetics andlower Cu grade than the samples from the reference rod mill. This was also a similar outcome forthe NCD with the exception of higher mass pull at an increased collector dosage in the rod millcase. For all the mills, the general trend showed that an increase in P80 was associated withreduced mass pull. Overall, at similar flotation conditions, despite having one of the coarsest P80s,the NCD product showed the fastest Cu recovery kinetics, highest total recovery and selectivityindex. The rod mill reference test samples gave the highest grade but the lowest recovery. The findings illustrate that there is a significant difference in PSD broadness between productsfrom different mills at comparable P80. The differences in flotation performance between thethree mill products were mainly attributed to differences in liberation, potential differences insurface activation between wet and dry grinding methods and to some extent, steepness of thePSD curve. Based on the high recovery yet low grade of the NCD ground product, this is best suitedfor the rougher flotation stage. Future studies should therefore include cleaner stage flotation andoptimisation tests for the NCD with respect to reagent dosages and hydrodynamic conditions.