Novas rotas para obtenção de silício grau solar a partir de silício grau metalúrgico estão em desenvolvimento e a solidificação direcional é uma etapa presente em todos os processos propostos. O refino de silício por solidificação direcional baseia-se no fenômeno de macrossegregação das impurezas. Experimentos de solidificação direcional transiente foram realizados em condições estáticas utilizando um equipamento projetado no âmbito desse trabalho. A partir de um experimento de referência, foram avaliadas as influências da alteração do material da base do cadinho, altura do lingote e condição de resfriamento do forno. Para estudo das condições de solidificação e dos mecanismos envolvidos no fenômeno de macrossegregação de solutos foi proposto e implementado um modelo matemático. Esse modelo considera as equações gerais de transporte no caso unidirecional e o transporte de espécies químicas por difusão macroscópica e convecção. A convecção foi tratada a luz da teoria da camada estagnada a frente da interface sólido-líquido. Variáveis como velocidade da interface sólido-líquido, gradiente de temperatura, perfis de concentração de soluto e de fração de sólido foram obtidos com o modelo matemático utilizando as temperaturas medidas no silício durante os experimentos de solidificação como condições de contorno do modelo. Os resultados experimentais mostraram que sob algumas condições foram obtidos lingotes com macroestrutura típica de solidificação unidirecional e, além disso, as microestruturas mostraram evidências de macrossegregação de solutos. Os resultados do modelo matemático mostraram que a solidificação ocorreu em diferentes condições de velocidade da interface sólido-líquido e gradiente de temperatura nos experimentos. Os resultados obtidos com a utilização do modelo matemático mostraram que a convecção teve papel fundamental no fenômeno de macrossegregação de solutos. / New process routes are under development to obtain solar grade silicon from metallurgical grade one, and the directional solidification is an essential step in all proposed process routes. The silicon refining by directional solidification is based on the impurities macrosegregation phenomena. Transient solidification experiments were conducted under a static condition in a furnace projected for this work. From a reference experiment it was analyzed the effects of the changing the material of the crucible base, the ingot height and the cooling condition of the furnace. A mathematical model was proposed and implemented in order to study the solidification conditions and the main mechanisms regarding the macrosegregation phenomena. The mathematical model considers the conservation equations in one direction and the transport of chemical species occurs by diffusion and convection. The convection was treated using the diffusion layer theory. The velocity of solid-liquid interface, temperature gradient and profiles of solute concentration and solid fraction were obtained using the temperatures on silicon during the solidification experiments as boundaries conditions of the model. The experimental results showed that under some conditions it was obtained ingots with typical unidirectional macrostructure and, besides, showed in the microstructure evidences of macrosegregation. The mathematical model results showed that the solidification took place under different conditions of solid-liquid interface velocities and temperature gradient in the experiments. The results from the mathematical model showed that convection plays an essential role in the macrosegregation phenomena.

Depuis la fin du 20ème siècle, notre société a pris conscience que la protection de l’environnement et le développement durable sont nécessaires à notre épanouissement futur. Des directives ont contraint les pouvoirs publics, les collectivités, ainsi que les industriels à réagir notamment en s’impliquant dans la recherche. Les cimentiers sont aussi concernés. La fabrication du ciment nécessite énormément de matières premières.L’objectif principal de cette thèse est donc de valoriser les sous-produits industriels en matériaux cimentaires. Le contexte et la problématique des déchets sont abordés, ainsi que la possibilité de valorisation par les ciments. La méthodologie mise en place et son originalité sont expliqués. Dans un premier temps, le protocole de caractérisation des déchets minéraux a été développé et les résultats en laboratoire sont prometteurs pour le remplacement du calcaire et de l’argile dans le cru d’un ciment. Les cuissons en laboratoire ont permis d’obtenir un ciment comportant les 4 phases d’un Portland. La synthèse en plus grande quantité d’une formulation en laboratoire a mis en évidence de bonnes performances mécaniques sur mortiers au jeune âge. Dans un deuxième temps, les meilleures formulations ont été testées à l’échelle industrielle, dans le four de ARF. Le cru, malaxé par trémies agitatrices, a été une réussite. Enfin, l’immobilisation des polluants dans une matrice cimentaire des résidus de combustion, provenant des combustibles secondaires de ARF a permis de déclasser le déchet stabilisé/solidifié par réduction granulométrique du matériau avant tabilisation/solidification. / Since 20th century’s end, our society started to take conscience of environmental protection and sustainable development are required to our future blooming. Directives have obliged public authorities, communities, such as industry to be reacted particularly while being implied in research. The cement-manufacturers are also concerned. The process cement-manufacturer requires large quantities of raw materials. The main objective of the thesis is to value industrial by-products by cementitious materials. Context and problem of wastes were discussed, as well as the possibility of valorization such as the cements. The methodology set up for the thesis and its originality are explained. First, the protocol of characterization of mineral wastes was developed. The results in laboratory are promising for the replacement of limestone and clay in the cement’s raw meal and different formulations. The cookings in laboratory allowed obtaining a cement containing 4 phases of Portland. The synthesis in bigger quantity of a formulation in laboratory highlighted good mechanical performances on mortars at the young age. In a second stage, the best formulations were tested on an industrial scale in the ARF’s kiln. The raw meal, mixed by hoppers agitators, was a success. Finally, immobilization of pollutants in a cement matrix of the residues after combustion, resulting from secondary fuels of ARF, was studied to be able to downgrade the waste stabilized/solidified by grading reduction of the material before stabilization/solidification.

Cette thèse s’inscrit dans le projet PRINCIPIA (PRocédés INdustriels de Coulée Innovants Pour l'Industrie Aéronautique) de l’ANR MATETPRO (Matériaux et Procédés pour des Produits Performants). L'objectif de ce projet est la promotion de nouveaux alliages aluminium-cuivre-lithium à destination de l'industrie aéronautique afin d'apporter une alternative aux composites. Cependant, ces alliages sont sujet à une importante porosité pour deux raisons : une forte solubilité à l'hydrogène et une facilité d'oxydation. Dans ce projet, le but de la thèse était d'établir un modèle de prédiction de la porosité à l'échelle du produit. La porosité se forme lors de la solidification de l’alliage à cause d'une plus faible solubilité de l'hydrogène dans le solide. La teneur en hydrogène dans la phase liquide va augmenter par ségrégation et provoquer la nucléation des pores. Il est donc nécessaire de prendre en compte la solidification dans la modélisation de la porosité. De plus, la composition locales modifie la cinétique de croissance des pores et la microstructure exerce une contrainte mécanique sur les pores qui modifie leur équilibre chimique. Après une première partie consacrée à améliorer les connaissances sur les phénomène de transport dans la coulée semi-continue d'aluminium, nous avons modélisé la formation de porosité en se basant sur les modèles disponibles. Le modèle a reproduit l'inhomogénéité de la porosité observée expérimentalement sur une plaque d'alliage aluminium-magnésium. L'analyse nous a montré que la limitation de la croissance par le temps de diffusion de l'hydrogène était responsable de ce profil particulier. La densité volumique des pores est critique pour la limitation de la croissance par la diffusion de l’hydrogène. En fonction de la densité, la croissance passe d'une croissance limitée à une croissance non limitée / This thesis is part of the project PRINCIPIA (PRocédés INdustriels de Coulée Innovants Pour l'Industrie Aéronautique) of the ANR MATEPRO (MATériaux Et PROcédés pour des produits performants). The goal of this project is the promotion of new aluminum-copper-lithium alloys for the aeronautic industry in order to propose an alternative to composite materials. Unfortunately, these alloys are highly sensitive to the appearance of porosity during the alloy creation process. It is due to a high hydrogen solubility and oxidation. Inside this project, my work was to establish a porosity model at the scale of the ingot. Porosity starts to develop during the solidification process due to a lower solubility of hydrogen in the solid phase. Hydrogen content in liquid phase increases by segregation and leads to pores' nucleation. Thus, it is necessary to take into account solidification for porosity-modeling purposes. It is even more important because the alloys' local composition alters the pores' growth and the microstructure modifies the chemical equilibrium by pinching effect.After a first part dedicated to general improvement of knowledge about transport phenomena in DC casting, the porosity formation model was developed based on model found in literature. The model was able to reproduce the inhomogeneity experimentally observed in an aluminum-magnesium ingot. This profile is explained by the hydrogen diffusion time which limits the pore growth. The pore density is critical for the growth limitation by hydrogen diffusion. Depending on the density, the growth switch from a non limited to a limited growth.

Le présent mémoire constitue une étude numérique sur la méthode globale de collocation par fonctions radiales pour la résolution de l'équation de diffusion. Ce type de méthode ne comporte pas de maillage contrairement aux méthodes par différences finies ou éléments finis. Elle peut s'appliquer à des ensembles de points dispersés au hasard. La résolution de l'équation de diffusion avec des conditions de Neumann de flux massique nul à la surface a été expérimentée en une dimension et deux dimensions. Divers paramètres ont été étudiés tel que le nombre de points, leur disposition spatiale et autres configurations pertinentes afin d'optimiser la méthode. D'excellents résultats ont été obtenus en une dimension avec un faible nombre de points comparativement aux méthodes classiques. Par contre, une résolution adéquate en deux dimensions s'est avérée impossible pour cause de diverses instabilités numériques.

This dissertation mainly focuses on the development of new numerical models to simulate transport phenomena and predict the occurrence of macrosegregation defects known as freckles in directional solidification processes. Macrosegregation models that include double diffusive convection are very complex and require the simultaneous solution of the conservation equations of mass, momentum, energy and solute concentration. The penalty method and Galerkin Least Squares (GLS) method are the most commonly employed methods for predicting the interdendritic flow of the liquid melt during the solidification processes. The solidification models employing these methods are computationally inefficient since they are based on the formulations that require the coupled solution to velocity components in the momentum equation Motivated by the inefficiency of the previous solidification models, this work presents three different numerical algorithms for the solution of the volume averaged conservation equations. First, a semi explicit formulation of the projection method that allows the decoupled solution of the velocity components while maintaining the coupling between body force and pressure gradient is presented. This method has been implemented with a standard Galerkin finite element formulation based on bi-linear elements in two dimensions and tri-linear elements in three dimensions. This formulation is shown to be robust and very efficient in terms of both the memory and the computational time required for the macrosegregation computations. The second area addressed in this work is the use of adaptive meshing with linear triangular elements together with the Galerkin finite element method and the projection formulation. An unstructured triangular mesh generator is integrated with the solidification model to produce the solution adapted meshes. Strategies to tackle the different length scales involved in macrosegregation modeling are presented. Meshless element free Galerkin method has been investigated to simulate the solidification processes to alleviate the difficulties associated with the dependence on the mesh. This method is combined with the fractional step method to predict macrosegregation. The performance of these three numerical algorithms has been analyzed and two and three dimensional simulations showing the directional solidification of binary Pb-Sn and multicomponent Ni base alloys are presented.

Twin roll casting (TRC) offers a promising route for the economic production of Mg sheet, but unfortunately, it produces strip with coarse and non-uniform microstructures and severe centre line segregation. Recently, a novel magnesium strip casting process termed melt conditioned twin roll casting (MC-TRC) was developed that, compared with the conventional TRC process, emphasizes solidification control at the casting stage rather than hot rolling. This was achieved by melt conditioning under intensive forced convection prior to twin roll casting resulting in enhanced heterogeneous nucleation followed by equiaxed growth. In this study the development of TRC and MC-TRC processes and a microstructural comparison of the MC-TRC Mg-alloy strip with that of conventional TRC strip, have been investigated. Emphasis has been focused on the solidification behaviour of the intensively sheared liquid metal, and on the mechanisms for microstructural refinement and compositional uniformity in the MCTRC process. The results of the process development indicate that the MC-TRC process reduces considerably or eliminates defects such as the centre line segregation, voids and cracks at or near the strip surface that are always present in conventional TRC strip. The newly-designed homogenization treatment investigated for TRC and MC-TRC magnesium alloy strips was based on microstructural evolution obtained during heat treatment. The results of the MC-TRC strips showed a much faster recrystallization rate with finer recrystallized grains, which are due to more homogeneous and a finer grain size of the as-cast MC-TRC strips compared with the as-cast TRC strips. During down-stream processing, the effects of MC-TRC process on microstructural evolution of hot-rolled magnesium strips have been understood thoroughly by accurate control of the hot-rolling procedure during each step of strip thickness reduction. This study indicates that the MC-TRC strip requires fewer rolling steps when compared to TRC strip, thus offering reduced processing cost and carbon footprint. Mechanical properties at room temperature of MC-TRC as-cast and rolled sheets are much improved when compared with the conventional TRC as-cast and rolled sheets which can result in a higher quality of final components. The mechanical properties at elevated temperature shows for the first time that the higher elongation and lower yield strength of MC-TRC as-cast strips at a temperature close to its optimised hot-rolling temperature results in better ability for rolling and higher ductility of MC-TRC Mg strip compared with the TRC Mg strip.

Diffusion controlled phase transformations are studied bymeans of the phase-field method. Morphological evolution ofdendrites, grains and Widmanst\"atten plates is modeled andsimulated. Growth of dendrites into highly supersaturated liquids ismodeled for binary alloy solidification. Phase-field equationsthat involve both temperature and solute redistribution areformulated. It is demonstrated that while at low undercoolingheat diffusion does not affect the growth of dendrites, i.e.solidification is nearly isothermal, at high cooling rates thesupersaturation is replaced by the thermal undercooling as thedriving force for growth. In experiments many crystals with different orientationsnucleate. The growth of randomly oriented dendrites, theirsubsequent impingement ant formation of grain boundaries arestudied in two dimensions using the FEM on adaptive grids. The structure of dendrites is determined by growthconditions and physical parameters of the solidifying material.Effects of the undercooling and anisotropic surface energy onthe crystal morphology are investigated. Transition betweenseaweeds, doublons and dendrites solidifying out of puresubstance is studied and compared to experimental data. Two-and three-dimensional simulations are performed in parallel onadaptive and uniform meshes. A phase-field method based on the Gibbs energy functional isformulated for ferrite to austenite phase transformation inFe-C. In combination with the solute drag model, transitionbetween diffusion controlled and massive transformations as afunction of C concentration and temperature is established byperforming a large number of one dimensional calculations withreal physical parameters. In two dimensions, growth ofWidmanstaetten plates is governed by the highly anisotropicsurface energy. It is found that the plate tip can beapproximated as sharp, in agreement with experiments. Keywords:heat and solute diffusion, solidification,solid-solid phase transformation, microstructure, crystalgrowth, dendrite, grain boundary, Widmanstaetten plate,phase-field, adaptive mesh generation, FEM. / <p>NR 20140805</p>

Alloy castings are usually solidified with a coarse columnar grain structure under normal casting conditions unless the mode of the solidification is carefully controlled. It is desirable for the grain structure to be fine and equiaxed to improve their mechanical performance as finished castings. It is possible to develop a fine and equiaxed grain structure either by increasing the number of nucleation sites or by grain multiplication. Immiscible alloys with a microstructure in which a soft phase is dispersed homogeneously in a hard matrix have significant potential applications in advanced bearing systems, especially for the automotive industry. Despite considerable efforts made worldwide, including extensive space experiments, no casting techniques so far can produce the desired immiscible microstructure of alloys. Experimental results on Al-Sn-Cu immiscible alloys have confirmed that intensive shearing using melt conditioning by an advanced shearing technology (MCAST) unit, is an effective way to achieve a fine and uniform dispersion of the soft phase without macro-demixing, and that such a dispersed microstructure can be further refined in alloys with precipitation of the primary Al phase prior to the demixing reaction. In addition, it was found that melt shearing at 200 rpm for 60 s will be adequate to produce a fine and uniform dispersion of the Sn phase, and that a higher shearing speed and prolonged shearing time can only achieve further minor refinement. A study of Al-Si hypoeutectic and hypereutectic alloys presents the effects of the processing temperature and intensive shearing on the microstructural and mechanical properties which have been investigated systematically. Attempts have been made to explain the solidification mechanism with intensive melt shearing. The sheared melt was cast into tensile test samples by high pressure die caster (HPDC) to examine the microstructures and mechanical properties. The experimental results reveal that significant grain refinement and uniformity of grains was achieved by the intensive shearing and also a considerable increase in mechanical properties with pouring temperature by changing intermetallic particles morphology, the position of defect band and reduced microscopic defects.

As energias renováveis têm ganhado destaque como possíveis soluções para a questão da matriz energética mundial. Dentre as diferentes alternativas está a energia solar. As células fotovoltaicas são produzidas majoritariamente com silício metálico de alta pureza que é produzido tradicionalmente por um processo químico de alta complexidade e custo. Alternativas vêm sendo desenvolvidas para produção deste material, sendo a rota metalúrgica uma delas. Nesta rota parte-se de um silício de menor pureza e removem-se as impurezas através de operações metalúrgicas. Um dos elementos a serem removidos é o carbono, que está majoritariamente presente precipitado na forma de carboneto de silício (SiC). Visando o estudo da formação da macrossegregação de partículas de SiC em Si, ensaios de solidificação em um forno do tipo Bridgman com diferentes velocidades de extração e diferentes diâmetros de molde foram realizados. Ensaios para verificar o efeito de decantação destas partículas na macrossegregação também foram conduzidos. Nos ensaios de solidificação, os resultados mostram que um aumento da velocidade de extração do molde para fora da região quente do forno resulta em uma estrutura de grãos menos alinhada na direção axial dos lingotes cilíndricos. Adicionalmente, o aumento do diâmetro do lingote também possui um efeito na macroestrutura, dependo da velocidade de extração. A macrossegregação de carbono para o topo do lingote é mais severa quanto menor for a velocidade de extração e maior for a direcionalidade dos grãos. Esta macrossegregação é causada pelo empurramento das partículas de SiC pela interface sólido/líquido. A decantação de partículas de SiC causou macrossegregação de C para a base dos lingotes de Si, porém menos acentuada que aquela observada nos ensaios de solidificação direcional. A alteração do período de tempo reservado no experimento para ocorrência da decantação de uma para seis horas não influenciou significativamente a macrossegregação. Para um mesmo tempo de processo, a solidificação direcional a uma velocidade de extração de 5 µm/s resultou em uma maior macrossegregação de carbono do que aquela observada após decantação, apresentando um maior potencial para a remoção do carbono contido no silício. / Renewable energies have been considered as possible solutions to the world energy matrix issue. Solar energy is one of the alternatives. Photovoltaic cells are mainly made of high purity silicon, which is produced by a complex and costly chemical process. New alternative processes are under development and one of them is the metallurgical route. In this route, a less pure silicon is refined and impurities are removed by metallurgical operations. One of the elements to be removed is carbon, which is mainly present as silicon carbide (SiC). Aiming at an investigation about the macrosegregation of carbon and SiC particles in Si, solidification experiments were conducted in a Bridgman type furnace using different extraction velocities and different mold diameters. Experiments to investigate the effects of settling of SiC particles were also carried out. The results show that the mold extraction velocity plays an important role in the grain macrostructure formation: higher velocities give rise to a less directional grain structure. Additionally, the change in the ingot diameter also has an effect on the macrostructure depending on the extraction velocity. Carbon macrosegregation to the ingot top is more severe for lower mold extraction velocities. This macrosegregation is a result of SiC particles being pushed by the solid/liquid interface to the ingot top. The settling of SiC particles causes carbon macrosegregation to the ingot bottom, but less intense than that observed in the directional solidification experiments at the ingot top. The change of settling times from one to six hours, have no significant effect on the degree of macrosegregation. For equal processing times, directional solidification at a mold extraction rate of 5 µm/s causes more carbon macrosegregation than settling, representing a higher potential alternative for carbon removal from silicon.

Em processo de fabricação de polióis, são gerados dois resíduos de catalisadores (RNi e RAl) considerados perigosos ao meio ambiente, devido aos seus respectivos altos teores de níquel e alumínio. A presente Tese trata do estudo da solidificação/estabilização simultânea desses dois resíduos catalíticos com cimento Portland tipo II, com o intuito de minimizar os impactos ambientais e verificar a possibilidade de uso do produto solidificado como elemento estrutural. Os ensaios realizados consistiram em: análise térmica diferencial não convencional (NCDTA), análises termogravimétricas (TG/DTG), análise de difração de raios X (DRX), ensaios de fluorescência de raios X (FRX), ensaio de lixiviação e resistência mecânica. A análise das primeiras etapas de hidratação do cimento assim como de amostras hidratadas em diferentes idades nos primeiros 28 dias, possibilitou avaliar os efeitos da presença dos rejeitos no processo, identificar as etapas onde ocorrem e permitiu quantificar as principais fases do cimento hidratado. Pastas com relação água/cimento igual a 0,5 foram utilizadas, às quais diferentes quantidades de cada resíduo foram adicionadas. Foi verificado que nos estágios iniciais de hidratação do cimento ocorrem efeitos de retardamento e aceleração, respectivamente, devido à presença de RNi e RAl. A utilização simultânea dos dois resíduos de catalisadores no processo de solidificação/estabilização em cimento, indica a ocorrência de um efeito sinérgico, permitindo melhores condições de solidificação do que quando cada resíduo é tratado separadamente. Os ensaios de lixiviação feitos para pastas e argamassas após 28 dias de solidificação, prazo padrão para avaliação de processos de solidificação, apresentaram valores abaixo dos limites permitidos para a concentração final de Ni e Al nos extratos de lixiviação, indicando que o processo de solidificação simultânea dos rejeitos atende à legislação ambiental e elimina o seu impacto ambiental original. Pastas e argamassas analisadas também após 28 dias por testes de resistência à compressão, apresentaram resultados aceitáveis para possível uso na construção civil. / In polyols production two catalyst wastes (RNi and RAI) are obtained, which are considered hazardous, due to their respective high nickel and aluminum contents. This Thesis presents the study of the simultaneous solidification/stabilization of both wastes with type II Portland cement (CPII), in order to avoid environmental impacts and to check the possibility of the use of the solidified products as structural elements. The experimental research performed for this study consisted of non-conventional differential thermal analysis, thermogravimetric analysis, X ray diffraction, X ray fluorescence, leaching and compressive strength tests. The analysis of the first stages of the cement hydration, as well as of samples hydrated at different ages during the first 28 days of hydration, allowed evaluate the effects of the presence of the wastes on the process, identify the steps where the changes occur and have a quantitative information about the main cement hydrated phases. Pastes with water/cement ratio equal to 0.5 were used, into which different amounts of each waste were added. In the early stages of cement hydration retarding and accelerating effects occur, respectively due to RNi and RAl presence. During the simultaneous use of the two waste catalysts for their solidification/stabilization in cement, there is a synergic effect, which allows better operating conditions than when each waste is solidified separately. The leaching tests done for solidified pastes and mortars, after the standard evaluation period of 28 days of solidification, presented values for the final concentration of Ni and Al below accepted limits, indicating that the simultaneous solidification process attends environmental legislation, as well as eliminates the original environmental impact of the wastes. Pastes and mortars analyzed also after 28 days by compressive strength tests, presented acceptable results for the possible use of the solidified products in construction industry.