La gestion des déchets radioactifs se fait par confinement des déchets radioactifs dans une matrice vitreuse. Il est réalisé par fusion inductive à 1150°C dans des pots constitués de superalliages riches en chrome. Ces pots sont soumis à une corrosion par le verre entrainant à terme leur remplacement. La protection contre la corrosion par le verre est assurée par l’établissement d’une couche d’oxyde protectrice Cr2O3 à la surface de l’alliage. Un alliage binaire chromino-formeur (Ni-30Cr) a été choisi comme alliage d’étude. Trois compositions de verre sodo- et borosilicatées sont choisies dans le but de discerner l’influence de la basicité et de la viscosité des verres sur les cinétiques de corrosion. La décorrélation des phénomènes de formation et de dissolution a permis la modélisation des cinétiques de croissance globale de la couche d’oxyde en milieu liquide silicaté. Dans ce cadre, la cinétique de formation de l’oxyde en milieu verre est assimilée à la cinétique d’oxydation de l’alliage Ni-30Cr en milieu gazeux sous des pressions partielles d’oxygène représentatives du pouvoir oxydant d’un verre. Les études des cinétiques d’oxydation ont montré que la cinétique d’oxydation de l’alliage Ni-30Cr est indépendante de la pression d’oxygène entre 10-13 et 10-3 atm O2 à 1150°C. Les travaux réalisés ont montré que la cinétique de dissolution de la couche d’oxyde Cr2O3 est limitée par la diffusion du CrIII dans le verre. Cette cinétique de dissolution a été ainsi évaluée à partir de la limite de solubilité et du coefficient de diffusion du CrIII dans les verres. Enfin, la cinétique de la croissance expérimentale de la couche d’oxyde Cr2O3 dans le verre a été modélisée avec succès pour chacun des verres, à partir de (i) la limite de solubilité du CrIII, (ii) son coefficient de diffusion et (iii) la cinétique d’oxydation de l’alliage. Le modèle établi permet également de quantifier l’influence de ces différents paramètres sur la cinétique de croissance de l’oxyde protecteur / Nuclear wastes management consists in the confinement of the radioactive wastes in a glass matrix. This is made by inductive melting in a hot crucible at an operating temperature around 1150°C. These crucibles are constituted of nickel based superalloys with high chromium content. They are submitted to a harsh corrosion by the molten glass, eventually leading to their replacement. The protection of the crucible against corrosion is best provided by the establishment of a protective chromium oxide layer at the surface of the alloy. A binary chromia-forming alloy (Ni-30Cr) is studied in this work. Three different binary and ternary glass compositions are chosen in order to understand the influence of the glass basicity and glass viscosity on the corrosion kinetics. Besides, the decorrelation of the formation and dissolution kinetics of the oxide layer allows the modelling of the overall oxide growth in the molten glass. For that purpose, the oxide formation kinetics in molten glass media is assimilated to the oxidation kinetics of the alloy in gaseous media with oxygen partial pressure that are representative of the redox properties of the glasses. Studies of the oxidation kinetics and of the diffusion mechanisms have shown that the oxidation kinetics is independent on the oxygen pressure in the range of 10-13 up to 10-3 atm O2 at 1150°C. The present work has shown that the dissolution kinetics of the oxide layer is governed by the diffusion of CrIII in the glass melt. This dissolution kinetics has been evaluated from the diffusion coefficient and the solubility limit of CrIII in the glass. Finally, the overall growth kinetics of the Cr2O3 layer in the glass has been successfully modelled for each glass, thanks to the knowledge of (i) the solubility limit of CrIII, (ii) its diffusion coefficient in the glasses and (iii) the oxidation kinetics of the alloy. The presented model also allows quantifying the influence of each of these parameters on the protective oxide layer growth kinetics

The present thesis gives a summary of results obtained through theoretical and experimental studies of systems with relevance to copper smelting and converting. Many chemical elements are involved in the copper production pro­cesses and a detailed experimental study would be very time- consuming and expensive. A complicating fact is also the corrosivity of the liquid phases towards container material. A powerful alternative is equilibrium calculations, in which models for the liquid phases as well as reliable basic thermodynamic data are needed. In the present thesis, a generalized structure based model for liquid silicates was developed and used in assessments of the sys­tems PbO-SiO2, Fe-O-SiO2, CuO0.5-SÌO2 and Cu-Fe-O-SiO2. In the model, the non-ideal silicate melt is treated as an ideal solu­tion but containing a few complexes. The PbO-Si02 melt could be described by introducing the complexes Pb3Si207, Pb4Si4010 and Pb13Si12O37 in addition to the components PbO and Pb2Si04. The species considered in the Fe-O-SiO2 melt were FeO, FeO1.5, Fe2Si04, Fe3Si207, Fe3Si6O15 and in the CUO0.5-SiO2 melt CuO0.5 ana CU4SiO4. Trie calculated phase diagrams, the activities of me­tal oxides and the oxygen partial pressures were all in good agreement with the published data. Two of the papers in this thesis concern the determination of Gibbs free energies for Cu2S(s,l) and Ca2Fe2O5(s) through emf measurements utilizing a solid electrolyte. Activities and termi­nal solubilities in the solid solution [Fet,Ca]0 were also deter­mined. The results obtained from the quantitative equilibrium calcula­tions for conventional copper smelting and converting were used to outline the overall reactions taking place and the outcome of changes in process parameters. Comparison with observed values, however, showed that the copper and magnetite contents in slag were calculated too low. These discrepancies could be completely explained by using a non-equilibrium approach in which the con­verter was assumed to consist of several segments with concentra­tion gradients between the segments. / digitalisering@umu.se