Cette thèse porte sur l’étude structurale de mélanges de fluorures fondus utilisés pour la production d’aluminium par électrolyse. Ce procédé est obtenu par dissolution de l'alumine (Al₂O₃) dans un bain de cryolithe fondue (NaF-AlF₃ plus quelques additifs) à 960°C. Afin d’obtenir des données d’entrée pour la modélisation électrocinétique des phénomènes de transport durant le procédé, une meilleure description de la spéciation dans les bains en fonction de leur température et de leur composition est nécessaire.Pour déterminer cette spéciation in situ, nous avons développé une démarche originale combinant la spectroscopie par Résonance Magnétique Nucléaire (RMN) à haute température avec des simulations de Dynamique Moléculaire (DM) couplés à des calculs de premiers principes. Ces derniers basés sur la théorie de la fonctionnelle de la densité (DFT) ont permis de calculer les paramètres d’interaction mis en jeu lors des expériences RMN et d’évaluer la qualité des modèles issus de la DM par comparaison avec les données expérimentales.Les mesures et simulations effectuées dans les systèmes binaires MF-AlF₃ (M= Na ou K) ont confirmé la présence d’ions F‾ et d’espèces anioniques [AlFₓ]³‾ˣ avec une durée de vie comprise entre 5 et 25 ps. La probabilité de former des dimères du type [Al₂Fm]⁶‾ᵐ est inférieure à 10 %. Quel que soit l’alcalin, l’ajout d’alumine dans ces systèmes affecte la proportion d’espèce AlF5²‾ pour former des oxyfluoroaluminates : [Al₂OF6] ²‾, [Al₂OF8]⁴‾, [Al₂O₂F4] ²‾, [Al₂O₂F6]⁴‾ et [Al₂O₃F₂]²‾. Ces derniers sont peu influencés par la présence de CaF₂. En se dissociant partiellement, celui-ci contribue à modifier les équilibres chimiques vers les milieux plus basiques et à augmenter la coordinence moyenne des ions aluminium. / This thesis deals with the structural study of molten fluorides mixtures used for aluminium production by electrolysis. This process is obtained by dissolving alumina (Al₂O₃) in a molten cryolite bath (NaF-AlF₃ plus some additives) at 960°C. In order to obtain input data for the electrokinetic modeling of transport phenomena during the process, a better description of speciation in baths as a function of their composition and temperature is required.To determine this in situ speciation, we developed an original approach combining high temperature nuclear magnetic resonance (NMR) spectroscopy with Molecular Dynamics (MD) simulations coupled with first principle calculations. Based on the density functional theory (DFT), it was possible to calculate the interaction parameters involved in NMR experiments and to evaluate the quality of the models derived from MD when compared to the experimental data.The measurements and simulations carried out in the MF-AlF₃ binary systems (M = Na or K) confirmed the presence of F- and anionic species [AlFₓ]³‾ˣ with a lifetime between 5 and 25 ps. The probability of forming dimers [Al2Fm]6-m is less than 10%. Regardless of alkali, the addition of alumina in these systems affects the proportion of AlF5²‾ species to form oxyfluoroaluminates: [Al₂OF6] ²‾, [Al₂OF8]⁴‾, [Al₂O₂F4] ²‾, [Al₂O₂F6]⁴‾ et [Al₂O₃F₂]²‾. The latter are little influenced by the presence of CaF₂. By partially dissociating, this contributes to modifying the chemical equilibria towards the more basic media and to increasing the average coordination of the aluminium ions.

O presente trabalho tem como objetivo principal um estudo sistemático de líquidos iônicos imidazólicos, ou seja, sais líquidos a temperaturas relativamente baixas derivados do anel imidazol, tendo como técnica principal a espectroscopia Raman. Foram estudados os cátions 1-alquil-3- metilimidazólio e 1-alquil-2,3-dimetilimidazólio, sendo o grupo alquil composto por 2, 4, 6, 8 ou 10 átomos de carbono, com os ânions brometo, hexafluorofosfato ou bis(trifluorometanosulfonil)imida (TFSI-), totalizando 30 sistemas distintos. O estudo foi dividido em três partes. Na primeira, foram estudados compostos derivados de imidazol, neutros e catiônicos, para entender as diferenças entre os cátions formadores (o ânion foi desconsiderado) de ILs e seus precursores. A espectroscopia Raman forneceu informações sobre as alterações nos modos vibracionais com as substituições, sendo possível constatar que os modos do anel se tornaram menos intensos com as substituições e a contribuição de grupos -CH se tornou mais importante. Cálculos de cargas de Mulliken foram realizados para estas espécies e os resultados obtidos reforçam as interpretações dos espectros vibracionais. A segunda etapa consistiu no estudo de ILs puros, analisando tanto cátion quanto ânion. A cadeia carbônica tem grande influência no espectro Raman, sendo observadas bandas atribuídas a confôrmeros diferentes, que aparecem em maior número e menor intensidade à medida que a cadeia se torna menor. Os ânions foram estudados nos ILs e em sais inorgânicos, sendo observados espectros muito semelhantes destes nos diversos ILs, porém diferentes nos sais inorgânicos. Os mesmos resultados foram encontrados nas medidas de XANES (espectroscopia de raio-X próxima à borda de absorção), que permitiu sondar a estrutura eletrônica ao redor de diversos átomos, utilizada como técnica auxiliar. Após analisado cátions e ânions, foi possível caracterizar o par iônico formado por estes íons, enfocando as bandas Raman das cadeias carbônicas, já que, como a interação de Coulomb é baixa, estes se ligam principalmente via ligação de hidrogênio. Para os ILs com -H no carbono 2, o ânion Br- forma um par iônico mais forte com os diversos cátions, enquanto PF6- e TFSI-</sup formam pares iônicos fracos, já que são menos coordenantes. A adição do metil no carbono 2 altera essa dinâmica, e mesmo o brometo não consegue formar um par iônico forte. Por fim, a terceira etapa visou o estudo destes sistemas em solução com um solvente molecular, a dimetilformamida (DMFA), usada como sonda, pois foi analisado o deslocamento da banda correspondente ao grupo C=O. Em um primeiro momento foram estudadas soluções equimolares de ILs e DMFA, em que foi possível observar a estruturação dos ILs na mistura, diferenciando efeito do cátion e do ânion. Posteriormente, selecionou-se alguns sistemas para um estudo em diversas frações molares de IL e DMFA. Em baixas concentrações de IL, o comportamento foi semelhante ao observado para solventes moleculares. Conforme se aumentou a concentração destes sais, observou-se um comportamento anômalo no deslocamento da banda C=O, que pode ser atribuída ao caráter iônico destes sistemas. / The present work aims to be a systematic study, by Raman spectroscopy, of imidazolic ionic liquids, which means low temperature molten salts derived from imidazole ring. The study involved the cations 1-alkil-3-methylimidazolium and 1-alkyl-2,3-dimethylimidazolim, being alkyl chain composed of 2, 4, 6, 8 or 10 carbon atoms and with the anions bromide, hexafluorophosphate and bis(trifluoromethanesulfonyl)imide (TFSI-), in a grand total of 30 species with distinct properties. The study has been separated in three steps. The first one aimed to study both neutral and cationic molecules derived from imidazolic ring to understand the difference between cations that can form ILs and their precursors. Raman spectroscopy gave information about the changes in the vibrational modes with the substitutions. The results show that the intensity of modes from the ring decreases with the substitutions, while the contribution of -CH modes increases. In accordance with this are the results obtained from theoretical calculation of Mulliken charge, used as an auxiliary technique. The second step consisted in the study of pure ILs, analyzing both cation and anion effects. The carbonic chain plays a major role in Raman spectra, exhibiting bands attributed to different conformers. These bands become more abundant and less intense when the chain length increases. The anions have been studied in both ILs and inorganic salts, being observed very similar spectra for all the ILs but different ones for the inorganic salts. These results have been confirmed by XANES (X-ray absorption near edge structure) spectroscopy, which makes possible to probe the electronic structure of different atoms. After the analysis of both cation and anion, it was possible to better understand the ion pair formation in ILs, focusing in hydrogen bonds in the carbonic chain, since the Coulombic interaction is too weak. For ILs with hydrogen in carbon 2, it could be observed that bromide anion forms a stronger ion pair when compared to PF6- and TFSI-, less coordinating anions. The addition of methyl group in carbon 2 modifies this behavior, and even bromide does not form the strong ion pair. At last, the third step focused in the better understanding of binary systems containing ILs and a molecular solvent, dimethylformamide (DMFA), used also as probe since the shift of the band attributed to C=O group was used in this study. In a first stage it has been studied equimolar solutions of ILs and DMFA, in which it has been observed the organization of ILs in the mixture, and the contribution of both cation and anion. The second stage was dedicated to the understanding of these mixtures with different molar fraction of ILs and DMFA, for which some ILs have been selected. For low concentration of ILs, their behavior was similar to molecular solvents, but when the concentration of IL was increased, an anomalous behavior in the shift of C=O band has been observed, which can be attributed to the ionic character of these systems.

Thermal barrier coatings (TBC) are used in the hot sections of gas turbine engine in order to insulate the substrate at high temperature. Molten salt infiltration retards the durability of TBCs. The current standard material, i.e. 8YSZ is susceptible to molten salt infiltration. Therefore, alternate TBC materials are desirable. In addition to material composition, the TBC microstructure plays an important role in mitigating molten salt infiltration. Therefore, in this work, three different TBC variations were investigated. The first variation was a columnar microstructured 48YSZ TBC processed by SPS (48YSZ-SPS). The second variation was a columnar microstructured 8YSZ TBC processed by SPS (8YSZ-SPS), and the third variation was a lamellar microstructured 8YSZ TBC deposited by APS (8YSZ-APS). The as-sprayed TBC specimens were characterized by SEM/EDS, porosity analysis and XRD measurements. Later, the TBC specimens were exposed to hot corrosion test and their interaction with the molten salts were investigated using SEM (EDS and XRD). It was shown that an increase in stabilizer content (yttria content) in zirconia (in the case of 48YSZ) leads to an improved hot corrosion resistance due to the adequate amount of yttria content, which restricts the molten salt infiltration by forming needle like YVO4 phase. In terms of microstructure comparison, the infiltration behavior was similar for columnar microstructured 8YSZ and lamellar microstructured 8YSZ-APS as the molten salts infiltrated the coatings completely compared to the 48YSZ TBC. Furthermore, it seems that the molten salt infiltrates the TBC through globular pores, delamination cracks and splat boundaries in the case of APS-TBCs whereas the column gaps favor easier infiltration of molten salts in the case of columnar microstructured SPS processed TBCs.

Le néodyme appartient à la série des lanthanides se trouvant dans le tableau périodique. Il est le composant clé des aimants permanents Nd2Fe14B implantés dans différents appareils électroniques mais aussi dans appareils dis « écologiques » comme les éoliennes, les voitures ou vélos électriques. De nos jours seulement 1 % du néodyme présent dans les déchets électroniques est recyclé. De fortes tensions géopolitiques ainsi que la croissance rapide de la demande pour ce type d’aimants conduisent à d’importantes spéculations et fluctuations sur le prix du néodyme aggravées par l’appauvrissement annoncé de la ressource.Depuis quelques décennies, plusieurs procédés de recyclage du néodyme ont été développés au niveau des laboratoires sans avoir été utilisés à l’échelle industrielle. Les procédés existants ont plusieurs désavantages comme un grand nombre d’étapes avant d’obtenir le produit final ou encore l’utilisation de grandes quantités de produits chimiques.Le but de ces recherches est le développement d’un procédé pyrochimique pouvant être une alternative à ceux déjà existants. Ce procédé permettrait de récupérer le néodyme sous sa forme métallique en seulement une étape. Pour atteindre cet objet, il a fallu élaborer un réacteur pouvant résister à de hautes températures ainsi qu’aux sels fondus pouvant être corrosifs selon leur composition. Par la suite, les propriétés électrochimiques du néodyme ont été étudiées dans différents sels fondus à base de chlorures et de fluorures afin trouver la configuration optimale pour l’extraction du néodyme contenu dans les aimants permanents.Nous avons ainsi prouvé qu’il était possible d’extraire le néodyme sous sa forme métallique contenu dans de vieux aimants permanents en une seule étape durant une électrolyse. / Neodymium belongs to the lanthanide’s serie of the period system and is the key component of the permanent magnets Nd2Fe14B which are implemented in electronic devices and “green” technologies like wind turbines or electric cars and bicycles. Nowadays, only one percent of the neodymium in electronic scraps is recycled. Due to the geopolitical considerations and a strong increase of the use of permanent magnets, there is an impoverishment of the raw material resources leading to an instable market.Since some decades, few recycling processes have been developed on a lab scale without any upscaling to the industrial scale. The existing processes have several drawbacks like multiple steps to obtain the final desired product. This means that these methods have a long process time or use a large amount of chemical productsThe aim of this research was the development of a pyrochemical process, which could be an alternative to the existing recycling processes in order to extract neodymium from electronic scraps with less steps, a smaller amount of chemical products and a higher recovery rate of the rare neodymium. For this purpose a special reaction chamber has been created which resists to the severe experimental conditions induced by the use of high temperatures and corrosive molten salts. Moreover the electrochemical behaviour of neodymium in different chloride- and fluoride-based molten salts was studied in order to find the most appropriated setup.We could prove that the pyrochemical method could be the solution to recover neodymium from the old permanent magnets under its metallic form in just one single reaction step during an electrolysis.

Aluminium die casting is a process used to transform molten aluminium material into automotive gearbox housings, wheels and electronic components, among many other uses. It is used because it is a very efficient method of achieving near net shape with the required mechanical properties. Life Cycle Assessment (LCA) is a technique used to determine the environmental impacts of a product or process. The Life Cycle Inventory (LCI) is the initial phase of an LCA and describes which emissions will occur and which raw materials are used during the life of a product or during a process. This study has improved the LCI technique by adding in manufacturing and other costs to the ISO standardised methods. Although this is not new, the novel application and allocation methods have been developed independently. The improved technique has then been applied to Aluminium High Pressure Die Casting. In applying the improved LCI to this process, the cost in monetary terms and environmental emissions have been determined for a particular component manufactured by this process. A model has been developed in association with an industry partner so this technique can be repeatedly applied and used in the prediction of costs and emissions. This has been tested with two different products. Following this, specialised LCA software modelling of the aluminium high pressure die casting process was conducted. The variations in the process have shown that each particular component will have different costs and emissions and it is not possible to generalise the process by modelling only one component. This study has concentrated on one process within die casting but the techniques developed can be used across any variations in the die casting process.

The increasing miniaturization of electric and mechanical components makes the synthesis and assembly of nanoscale structures an important step in modern technology. Functional materials, such as the ferroelectric perovskites, are vital to the integration and utility value of nanotechnology in the future. In the present work, chemical methods to synthesize one-dimensional (1D) nanostructures of ferroelectric perovskites have been studied. To successfully and controllably make 1D nanostructures by chemical methods it is very important to understand the growth mechanism of these nanostructures, in order to design the structures for use in various applications. For the integration of 1D nanostructures into devices it is also very important to be able to make arrays and large-area designed structures from the building blocks that single nanostructures constitute. As functional materials, it is of course also vital to study the properties of the nanostructures. The characterization of properties of single nanostructures is challenging, but essential to the use of such structures. The aim of this work has been to synthesize high quality single-crystalline 1D nanostructures of ferroelectric perovskites with emphasis on PbTiO3 , to make arrays or hierarchical nanostructures of 1D nanostructures on substrates, to understand the growth mechanisms of the 1D nanostructures, and to investigate the ferroelectric and piezoelectric properties of the 1D nanostructures. In Paper I, a molten salt synthesis route, previously reported to yield BaTiO3 , PbTiO3 and Na2Ti6O13 nanorods, was re-examined in order to elucidate the role of volatile chlorides. A precursor mixture containing barium (or lead) and titaniumwas annealed in the presence of NaCl at 760 °C or 820 °C. The main products were respectively isometric nanocrystalline BaTiO3 and PbTiO3. Nanorods were also detected, but electron diffraction revealed that the composition of the nanorods was respectively BaTi2O5/BaTi5O11 and Na2Ti6O13 for the two different systems, in contradiction to the previous studies. It was shown that NaCl reacted with BaO(PbO) resulting in loss of volatile BaCl2 (PbCl2 ) and formation and preferential growth of titanium oxide-rich nanorods instead of the target phase BaTiO3 (or PbTiO3 ). The molten salt synthesis route may therefore not necessarily yield nanorods of the target ternary oxide as reported previously. In addition, the importance of NaCl(g) for the growth of nanorods below the melting point of NaCl was demonstrated in a special experimental setup, where NaCl and the precursors were physically separated. In Paper II and III, a hydrothermal synthesis method to grow arrays and hierarchical nanostructures of PbTiO3 nanorods and platelets on substrates is presented. Hydrothermal treatment of an amorphous PbTiO3 precursor in the presence of a surfactant and PbTiO3 or SrTiO3 substrates resulted in the growth of PbTiO3 nanorods and platelets aligned in the crystallographic &lt;100&gt; orientations of the SrTiO3 substrates. PbTiO3 nanorods oriented perpendicular to the substrate surface could also be grown directly on the substrate by a modified synthesis method. The hydrothermal method described in Paper II and III was developed on the basis of the method described in Appendices I and II. In Paper IV, a template-assisted method to make PbTiO3 nanotubes is presented. An equimolar Pb-Ti sol was dropped onto porous alumina membranes and penetrated into the channels of the template. Single-phase PbTiO3 perovskite nanotubes were obtained by annealing at 700 °C for 6 h. The nanotubes haddiameters of 200 - 400 nm with a wall thickness of approximately 20 nm. Excess PbO or annealing in a Pb-containing atmosphere was not necessary in order to achieve single phase PbTiO3 nanotubes. The influence of the heating procedure and the sol concentration is discussed. In Paper V, a piezoresponse force microscopy study of single PbTiO3 nanorods is presented. The piezoelectric properties were studied in both vertical and lateral mode. Piezoelectric activity and polarization switching was observed in the vertical mode, demonstrating the ferroelectric nature of the nanorods. The nanorods decomposed after repeated cycling of the dc bias at one spot on the nanorod, which resulted in parts of the nanorod disappearing and/or accumulation of particles on the surface of the nanorod. In Paper VI, a method to contact single nanorods by electron beam induced deposition of platinum is presented. An organometallic compound, (trimethyl)-methylcyclopentadienylplatinum(IV), was used as precursor. A home-made apparatus was constructed for the purpose and was mounted onto a scanning electron microscope. Calculations based on apparatus geometry and molecular flow were used to estimate the deposition time and the height of the deposits. The location and height of the deposits were controlled so that single nanorods could be successfully contacted at the ends of the nanorods. Fabrication of a sample device for piezoresponse force microscopy studies of single nanorods using an axial dc bias setup is described in Appendix IV. A proposed experimental setup for such studies is also presented.

The recovery boiler in a pulp and paper mill plays a dual role of recovering pulping chemicals and generating steam for either chemical processes or producing electricity. The efficiency of producing steam in the recovery boiler is limited by the first melting temperature of ash deposits that accumulate on the superheater tubes. Above the first melting temperature, the molten salt reacts with the protective oxide film that develops and dissolves it. The most protective oxide is determined by evaluating how little it dissolves and how its solubility changes in the molten salt. Solubility tests were done on several protective oxides in a known salt composition from a recovery boiler that burns hardwood derived fuel. ICP-OES was used to measure concentration of dissolved metal in the exposure tests while EDS and XRD were used to verify chemical compositions in exposure tests. NiO was found to be the least soluble oxide while Cr₂O₃ and Al₂O₃ had similar solubility with Fe₂O₃ being less soluble than Cr₂O₃ but more soluble than NiO. Exposure tests with pure metals and selected alloys indicated that even though Fe₂O₃ has little solubility, it is not a protective oxide and causes severe corrosion in stainless steels. The change in performance of iron based alloys was due to the development of a negative solubility gradient for Fe₂O₃ where Fe₂O₃ precipitated out of solution and created a continuous leaching of oxide. Manganese was found to be beneficial in stainless steels but its role is still unknown. Nickel based alloys were found to be least corroded due to nickel's low solubility and because it did not form a negative solubility gradient.

High grade galvanized steel in large amounts is needed to match the increasing demand of automotive industry both in our country and in the world. Stainless steels, used in fabrication of zinc bath hardware of continuous galvanizing lines, lose their corrosion resistance due to various mechanisms in such mediums containing molten metals like zinc and aluminum. Consequently they corrode to the levels where they should be taken to maintenance or replaced. In this study, corrosion performance and the effect of typical galvanizing and age treating heat treatments to mechanical properties of 4 newly developed austenitic stainless steels and AISI 316L grade stainless steel were investigated and compared with each other. Experimental studies involved immersion corrosion tests for 168 and 504 hours followed by weight loss determinations and comparisons of corrosion performances of age treated and solution annealed stainless steels. Parallel with corrosion testing, delta ferrite content v determinations with 3 different methods, tensile tests and v-notch impact tests at 4 different heat exposure conditions were carried out and discussed. 2 new stainless steel compositions were selected to be used in fabrication of galvanizing hardware based on the comparisons of corrosion &amp / mechanical performances of candidate steels.

The Liquid-Salt-Cooled Very High-Temperature Reactor (LS-VHTR) was modeled using the neutronics analysis code SCALE6.0 and the thermal-hydraulics and kinetics modeling code RELAP5-3D with objective to devise, analyze, and evaluate the feasibility and stability of a start-up procedure for this reactor using natural circulation of the coolant and under the Loss Of Offsite Power (LOOP) conditions. This Generation IV reactor design has been studied by research facilities worldwide for almost a decade. While neutronics and thermal-hydraulics analyses have been previously performed to show the performance of the reactor during normal operation and for shutdown scenarios, no study has heretofore been published to examine the active or passive start-up of the reactor. The fuel temperature (Doppler) and coolant density coefficient of reactivity of the LS-VHTR were examined using the CSAS6 module of the SCALE6.0 code. Negative Doppler and coolant density feedback coefficients were calculated. Two initial RELAP5 simulations were run to obtain the steady-state conditions of the model and to predict the changes of the thermal-hydraulic parameters during the shutdown of the reactor. Next, a series of step reactivity additions to the core were simulated to determine how much reactivity can be inserted without jeopardizing safety and the stability of the core. Finally, a start-up procedure was developed, and the restart of the reactor with natural convection of the coolant was simulated. The results of the simulations demonstrated the potential of a passive start-up of the LS-VHTR.

No description available.