This thesis aims at a deeper understanding of Supercritical Fluid Chromatography (SFC). Although preparative SFC has started to replace Liquid Chromatography (LC) in the pharmaceutical industry - because of its advantages in speed and its less environmental impact - fundamental understanding is still lacking. Therefore there is no rigid framework to characterize adsorption or to understand the impact of changes in operational conditions.   In Paper I we demonstrated, after careful system verification, that most methods applied to determine adsorption isotherms in LC could not be applied directly in SFC. This was mainly due to operational differences and to the fact that the fluid is compressible which means that everything considered constant in LC varies in SFC.   In Paper II we showed that the most accurate methods for adsorption isotherm determination in LC, the so called plateau methods, do not work properly for SFC. Instead, methods based on overloaded profiles should be preferred.   In Paper III a Design of Experiments approach was successfully used to quantitatively describe the retention behavior of several solutes and the productivity of a two component separation system. This approach can be used to optimize SFC separations or to provide information about the separation system.   In Paper IV severe peak distortion effects, suspected to arise from injection solvent and mobile phase fluid mismatches, were carefully investigated using experiments and simulations. By this approach it was possible to examine the underlying reasons for the distortions, which is vital for method development.   Finally, in Paper V, the acquired knowledge from Paper I-IV was used to perform reliable scale-up in an industrial setting for the first time. This was done by carefully matching the conditions inside the analytical and preparative column with each other. The results could therefore provide the industry with key knowledge for further implementation of SFC. / This thesis aims at a deeper understanding of Supercritical Fluid Chromatography (SFC). Although preparative SFC has started to replace Liquid Chromatography (LC) in the pharmaceutical industry - because of its advantages in speed and its less environmental impact - fundamental understanding is still lacking. Therefore there is no rigid framework to characterize adsorption or to understand the impact of changes in operational conditions.   In Paper I-II it was demonstrated why most methods applied to determine adsorption isotherms in LC could not be applied directly for SFC. Methods based on extracting data from overloaded profiles should be preferred.   In Paper III a Design of Experiments approach was successfully used to quantitatively describe the behavior of several solutes in a separation system. This approach can be used to optimize SFC separations or to provide information about the separation system.   In Paper IV severe peak distortion effects often observed in SFC were carefully investigated and explained using experiments and simulations.   Finally, in Paper V, the prerequisites for performing reliable and predictable scale-up of SFC were investigated by small and large scale experiments. / <p>Paper 4 ("Evaluation of scale-up from analytical to preparative...") ingick som manuskript med samma titel i avhandlingen. Nu publicerad. </p>

Increase in demand for uniform ceramic coatings on larger industrial components have led to a need for a PP-MOCVD coating system scale up. The objective of this thesis is to develop a fully functional coating system operating in the PP-MOCVD regime that is able to deposit thin film ceramic coatings on commercial or industrial components with complex 3D geometries. This can be achieved by applying engineering and vacuum science theories, coupled with the established fundamentals of PP-MOCVD. A larger system was designed and assembled around the boundaries set by the dimensions and geometry of a stainless steel water pump impellor acting as the base substrate. Most of the components were sourced off the shelf from vacuum and fluid specialists. Components which were unavailable for various reasons were designed, and machined in-house by the departmental workshop. Initial test depositions were conducted using small stainless steel disk substrates, heated using a resistive heater similar to the one utilised on the research scale system. The test depositions were performed with the heater and substrate combination placed in strategic locations. This is to test the overall uniformity of precursor flux in the chamber volume. The resulting coating uniformity on the disk surfaces were fair but problems such as the large collection of unreacted precursor on the chamber viewport and valve timing issues had to be addressed. Before making any improvements to the system, each of the process areas leading to a successful deposition needed to be understood. Five process areas were developed: ‘Liquid Delivery’, ‘Atomization’, ‘Evaporation’, ‘Transport and Reactor Geometry’, and ‘Droplet Management’. Each of the process areas were analysed individually and changes were made to push for a maximum evaporation efficiency. xviii The improved system provided opportunities to perform depositions that were once not possible for PP-MOCVD. Two sets of deposition tests were designed and conducted. Firstly, the improvements were justified with a series of depositions using flat stainless steel plates with dimensions 65x65x5mm. The other set of 3D case study depositions involve observing the effects of the operational parameters of PP-MOCVD on the uniformity and penetration depths of the coatings into different sized macro blind trenches. Five geometric setup conditions were used to justify the improvements made to the system. These are: ‘Substrate positioned in the direct line of spray’, ‘Use of an unheated receptor’, ‘Use of a heated receptor’, ‘Use of an unheated receptor with a non-axial substrate setup’, and “Choked Flow’. As expected, the uniformity of the coatings on both sides of the plate varied significantly when the substrate is placed over the line of sight of the precursor spray. Similarly, the coating produced under the induced choked flow condition resulted in low conformality. The introduction of an unheated receptor plate resulted in an increase in uniformity on both sides of the plate. Further prove that PP-MOCVD is geometry independent is provided by the deposition made with the non-axial substrate placement resulting in a coating of similar result to the unheated receptor. The use of a heated receptor provided a source for a secondary evaporation of the larger precursor droplets collected resulting in an increase in coating thickness while maintaining good conformality. The effects of temperature, pressure, injection volume, and concentration were explored in the final case study. With maximum depths of 50mm, the macro blind trenches has an aspect ratio of 1:1 and cross-sectional areas of 3x3mm, 9x9mm, and 15x15mm. The final results show that as the temperature rises, the depth penetrated into the trench decreases. This could be due to the change in rate limiting steps as homogeneous reactions begin to increase at higher temperatures. Similar trends were observed with increasing pressure. As the pressure difference between the volume of the trenches and the rest of the chamber decreases, the push needed to xix force the precursor down the trench also decreases, resulting in less depth penetration. The effects of injection volume and concentration observed, can be explained by how much precursor molecules are present during one pulse cycle. The more that is available at any given time, the more likely a reaction will occur and deeper the penetration will get. Of course a ceiling or a limit exists where the molecules in the chamber will get evacuated without being reacted. The future work made possible as a result of the scaled up system are proposed. These include a scale up of the operational parameters to suit any given substrate geometry, improvements to the heating source to achieve greater thermal uniformity, further improvements to the overall system accessibility, and performing other depositions using different substrate materials and precursor types.

Les coûts de revient des alliages de titane limitent leurs applications à des domaines de forte valeur ajoutée. La diminution des coûts de mise en forme des pièces tout en maitrisant les microstructures et les propriétés mécaniques est donc un enjeu important. Les procédés de mise en forme issus de la métallurgie des poudres comme le Spark Plasma Sintering (SPS) peuvent permettre d'atteindre ces objectifs. Les premiers instants du frittage ont été étudiés selon la compacité initiale du corps granulaire. Lorsque cette dernière est faible, les surfaces des cous entre les granules, mises en évidences par microtomographie, sont plus importantes. Ces observations peuvent être expliquées par le fait que les densités de courant au niveau des points de contacts sont plus importantes quand les surfaces de contacts sont plus faibles. De plus, les densités finales très proches malgré les compacités initiales différentes laissent penser que les mécanismes de frittage en action sont différents. Une modélisation Electro-Thermique et Mécanique par éléments finis a permis de décrire de façon très réaliste la densification du TA6V. La loi de densification du matériau a été paramétrée par le biais de campagnes d'essais en fluage menés dans le SPS ainsi que d'essais de compaction du corps granulaire. Des voies d'amélioration du modèle ont également été proposées. Dans la dernière partie de l'ouvrage, un panel des microstructures obtenues par frittage SPS du TA6V ainsi qu'un transfert d'échelle des échantillons modèles cylindriques de petites tailles vers des formes plus complexes et massives ont été étudiés. De plus, les propriétés mécaniques en traction sur pièces près des cotes ont été caractérisées. Les domaines (température et pressions) d'obtention de microstructures denses, homogènes et répétables ont été identifiés. Le transfert d'échelle révèle que la répétabilité du SPS est validée tant en termes de microstructures que de propriétés mécaniques des pièces densifiées. De plus, les propriétés en traction sont au niveau des matériaux forgés, notamment avec des allongements à rupture de près de 13%. / The cost prices of the titanium alloys limit their applications to strong added value fields. The decrease of the pieces shaping costs while mastering microstructures and mechanical properties is an important issue. Shaping processes from powder metallurgy as Spark Plasma Sintering (SPS) can allow to reach these goals. The first moments of the sintering were studied according to the initial compactness of the granular body. When the latter is low, the area of necks between granules, highlighted by microtomography, are bigger. These observations can be explained by the fact that current densities at contact points are higher when the contact areas are lower. Furthermore, the very close final densities in spite of the different initial compactness let think that the mechanisms in action during sintering are different. An electro-thermal and mechanical modelling by finished element allowed to describe in a very realistic way the densification of the TA6V. The densification law of the material was configured by means of creep tests programmes performed dine the SPS chamber as well as of compaction tests of the granular body. Ways of improvement of the model were also proposed. In the last part of the word, a panel of microstructures obtained by SPS of TA6V as well as a scale transfer from the small cylindrical samples towards more massive and complex shapes was studied. Furthermore, the tensile properties of near-net-shape specimens were characterised. Domains (temperature and pressure) of dense, homogeneous and repeatable microstructures obtaining were identified. The scale transfer reveals that the SPS repeatability is both validated in terms of microstructures and mechanical properties of the densified specimens. Furthermore, tensile properties are at the same level as the forged materials, in particular an elongation at break of about 13%.

A conversão de biomassa vegetal em produtos químicos e energia é essencial a fim de sustentar o nosso modo de vida atual. O bagaço de cana-de-açúcar, matériaprima disponível em abundância no Brasil, poderá tanto ajudar a suprir a crescente demanda pelo etanol combustível como ser empregado para obtenção de produtos de valor agregado, tais como xilitol, além de trazer vantagens econômicas para o setor sucroalcooleiro. O xilitol, um poliol com poder adoçante semelhante ao da sacarose e com propriedades peculiares, como metabolismo independente de insulina, anticariogenicidade e aplicações na área clínica, no tratamento de osteoporose e de doenças respiratórias, é obtido em escala comercial por catálise química de materiais lignocelulósicos. A produção biotecnológica de xilitol como alternativa ao processo químico vem sendo pesquisada e os resultados revelam que a presença de compostos tóxicos nos hidrolisados hemicelulósicos resultantes do processo de hidrólise ácida contribui para sua baixa fermentabilidade. Isto se deve à inibição do metabolismo microbiano causada principalmente por compostos tais como ácidos orgânicos, fenólicos e íons metálicos. No presente trabalho foi avaliado o efeito de diferentes fontes de carbono (xilose, glicose e mistura de xilose e glicose) empregadas no preparo do inóculo de Candida guilliermondii FTI 20037 sobre a bioconversão de xilose em xilitol a partir de fermentações em frascos Erlenmeyer de hidrolisados hemicelulósicos submetidos a procedimentos de destoxificação. A condição de favorecimento deste bioprocesso foi empregada para a avaliação da ampliação de escala em fermentadores de 2,4L para 16L, utilizando como critério de ampliação o KLa (igual a 15h-1). De acordo com os resultados, os máximos valores dos parâmetros fermentativos como fator de conversão de xilose em xilitol e produtividade em xilitol foram alcançados com a utilização de inóculo obtido em xilose durante fermentação do hidrolisado destoxificado por resinas (YP/S = 0,81 g g-1 e QP = 0,60 g L-1 h-1, respectivamente), embora o emprego de carvão ativado tenha gerado valores de rendimento próximos para as diferentes fontes de carbono (YP/S variando de 0,78 a 0,80 g g-1). Considerando o valor de fator de conversão e que o procedimento de destoxificação com carvão ativado é o de menor custo e de mais fácil manipulação em comparação ao processo com resinas, os experimentos de ampliação de escala da produção de xilitol por C. guilliermondii foram realizados nesta condição de destoxificação e empregando-se xilose como fonte de carbono para o inóculo. Nesta etapa ficou evidente a viabilidade de ampliação de escala de produção de xilitol de fermentador de 2,4L para 16L, já que os valores dos parâmetros fermentativos avaliados foram semelhantes entre os fermentadores (valores médios: YP/S &#8776; 0,68 g g-1 e QP &#8776; 0,28 g L-1 h-1). No entanto, tais valores foram inferiores aos obtidos em frascos Erlenmeyer, possivelmente devido às condições de disponibilidade de oxigênio diferirem nos fermentadores de bancada, uma vez que o oxigênio é o parâmetro mais crítico neste bioprocesso. / The conversion of vegetable biomass into chemicals and energy is essential to sustain our current style of life. Sugarcane bagasse, a raw material abundantly available in Brazil, greatly contributes to the supply of the evergrowing demand for ethanol. Furthermore, biomass can be employed for obtaining value-added products, such as xylitol, as well as bring economical advantages for the sugar-ethanol sector. Xylitol, a polyol with sweetener power similar to that of saccharose and peculiar properties such as insulin-independent metabolism, anticariogenic power, and applications in the clinical area, in the treatment of osteoporosis and respiratory diseases, is obtained on a commercial scale by chemical catalysis of lignocellulosic materials. The biotechnological production of xylitol as an alternative to the chemical process has been researched and the results reveal that the presence of toxic compounds in hemicelllosics hydrolysates resulting from acid hydrolysis process contributes to its low fermentability. Such toxicity could be due to the inhibition of microbial metabolism promoted mainly by compounds such as organic acids, phenols and metallic ions. In the present work, the effect of different carbon sources (xylose, glucose and a mixture of xylose and glucose) used in the inoculum preparation of Candida guilliermondii FTI 20037 for the xylose-to-xylitol bioconversion by fermentation of hemicellulosics hydrolysates submitted to detoxification procedures in Erlenmeyer flasks was evaluated. The best condition for this bioprocess was employed to evaluate the scale up from the 2.4L to 16L fermentors, using KLa (equal to 15h-1) as scale-up criteria. According to the results the highest values of fermentative parameters such as xylitol yield and productivity were achieved with the use of inoculum cultivated on xylose during the fermentation of hydrolysate detoxified with resins (YP/S = 0.81 g g-1 and QP = 0.60 g L-1 h-1, respectively), although with the use of charcoal the yield value was similar (YP/S ranging for 0.78 to 0.80 g g-1), regardless of the carbon source employed. Considering the value of xylitol yield and that detoxification with activated charcoal is less expensive and more easily manipulated when compared to detoxification procedure with resins, the experiments for scale up xylitol production by C. guilliermondii were performed in such detoxification condition with xylose as the carbon source for the inoculum. At this stage it was evident the scale up xylitol production from a fermenter of 2.4L to 16L was feasible, since the values of fermentative parameters evaluated were similar to those of the fermentors (medium values YP/S &#8776; 0.68 g g-1 e QP &#8776; 0.28 g L-1 h-1). However, these values were lower than those obtained in Erlenmeyer flasks, maybe due to conditions of oxygen availability for they differ from those in fermentors, since oxygen is the most critical parameter in this bioprocess.

A literatura corrente possui informações limitadas sobre o projeto da área de troca térmica de tanques com jaqueta, serpentina helicoidal, serpentina espiral e chicana tubular vertical, em operação com fluidos não-Newtonianos. A presente tese teve por objetivo principal analisar a transferência de calor, potência consumida e ampliação de escala em tanques com impulsores mecânicos na agitação de fluidos não-Newtonianos com duas superfícies de transmissão de calor, chicana tubular vertical e serpentina em espiral. O trabalho também visou fornecer métodos de ampliação de escala de tanques com agitação para fluidos não-Newtonianos que sigam o modelo reológico da lei das potências. A unidade experimental contemplou dois tanques de acrílico, com volume de 10 litros e 50 litros, respectivamente, chicanas tubulares verticais e serpentina em espiral. Os impulsores mecânicos utilizados foram o axial com 4 pás inclinadas a 45° e o radial turbina com 6 pás planas. Como fluidos utilizaram-se soluções aquosas de carboximetilcelulose (0,5%, 1,0% e 1,5%), solução aquosa de carbopol 940 (1,5%), solução aquosa de sacarose (50%) e água. Todos os experimentos foram conduzidos em batelada. Com os dados obtidos, empregou-se o uso de regressões para a obtenção da Equação de Nusselt, as quais forneceram valores de coeficiente de determinação ajustados entre 0,83 e 0,89 com Reynolds no intervalo de 20 a 405000, Prandtl na faixa de 4 a 6400 e índice reológico do modelo da lei das potências entre 0,45 e 1,00. Observou-se que no aquecimento realizado com a chicana tubular vertical, o impulsor radial forneceu coeficientes de convecção 20% acima quando comparado com o impulsor axial, entretanto o consumo de potência foi cerca de 66% maior em relação ao impulsor axial. No caso da serpentina espiral, o impulsor axial promoveu coeficientes de convecção por volta de 15% superiores em relação ao impulsor radial com um consumo de potência 65% menor. Desse modo, em processos em que não é necessária uma elevada turbulência, recomenda-se o uso do impulsor axial com a serpentina espiral, porém, se o processo demandar uma turbulência significativa, deve-se usar o impulsor radial com a chicana tubular vertical. Em uma última análise, os modelos não-lineares obtidos para ampliação de escala forneceram erros entre 11% e 20% na predição da rotação no tanque industrial, os quais são válidos para Reynolds modificados de Metzner e Otto (1957) na faixa de 20 a 4000 e para fluidos não-Newtonianos pseudoplásticos com índices reológicos entre 0,45 e 1,00. / Current literature has limited information on the design of the thermal exchange area of tanks with jacket, helical coil, spiral coil and vertical tuber baffle, in operation with non-Newtonian fluids. The main purpose of this thesis was to analyze heat transfer, power consumption and scale-up in tanks with mechanical impellers in the agitation of non-Newtonian fluids with two heat transfer surfaces, vertical tube baffle and spiral coil. The work also aimed to provide methods of scale-up tank scale with agitation for non-Newtonian fluids that follow the rheology model of the law of powers. The experimental unit included two acrylic tanks, with a volume of 10 liters and 50 liters, respectively, vertical tube baffles and spiral coil. The mechanical impellers used were the 45° pitched blade turbine (PBT) and the Rushton turbine (RT). Aqueous solutions of carboxymethylcellulose (0.5%, 1.0% and 1.5%), aqueous solution of carbopol 940 (1.5%), aqueous solution of sucrose (50%) and water were used as fluids. All the experiments were conducted in batch. With the obtained data, we used the regressions to obtain the Nusselt Equation, which provided coefficient of determination values adjusted between 0.83 and 0.89 with Reynolds in the range of 20 to 405000, Prandtl in the range of 4 to 6400 and rheological index of the power law model between 0.45 and 1.00. It was observed that in the heating performed with the vertical tube baffle, the RT provided convection coefficients 20% higher when compared to the axial impeller, however the power consumption was about 66% higher in relation to the PBT. In the case of the spiral coil, the PBT promoted convection coefficients around 15% higher than the RT with 65% lower power consumption. Thus, in processes where high turbulence is not required, it is recommended to use the PBT with the spiral coil, but if the process requires significant turbulence, the RT must be used with the vertical tubular chassis. In a final analysis, the nonlinear models obtained for scaling provided errors between 11% and 20% in the prediction of rotation in the industrial tank, which are valid for Metzner and Otto (1957) modified Reynolds in the range of 20 to 4000 and for non-Newtonian pseudoplastic fluids with rheological indexes between 0.45 and 1.00.

O trabalho visa o desenvolvimento do sistema para medidas de distribuição de corrente e ampliação de escala (50 cm²) buscando aperfeiçoar as condições de preparação do conjunto eletrodo membrana (MEA) quanto às condições de operação da célula e avaliar a melhor geometria. Foram realizados estudos de síntese de catalisadores de Pt-M e avaliação do desempenho desses materias e das rotas de síntese utilizadas com objetivo de aplicar estes materias em sistemas de maior escala. A insuficiência do desempenho e estabilidade dos catalisadores são fatores que ainda inviabilizam o uso em larga escala das células a combustível de eletrólito polimérico sólido, destacando-se as perdas associadas ao desempenho do cátodo. Os catalisadores preparados foram nanopartículas bimetálicas PtM/C (M = Fe, Co e Ni) suportadas em carbono de elevada área superficial, por duas rotas sintéticas. Foram utilizadas as rotas: ácido fórmico e etilenoglicol modificado (EG). Em ambas as rotas se buscou catalisadores com alto grau de incorporação do segundo metal, tamanho de partícula pequeno e bom desempenho catalítico do cátodo. Observou-se que pela rota do ácido fórmico com modificações no processo de síntese é possível obter a incorporação nominal do segundo metal no catalisador, porém há desvantagem de o tamanho de partícula ser elevado. Pela rota do EG obteve-se catalisadores com pequeno tamanho de partícula, porém a incorporação do segundo metal mostrou-se ineficiente. Os estudos de ampliação de escala foram realizados em células de 50 cm2 variando-se as condições de operação; i) diferentes placas de distribuição de gás, e ii) diferentes valores de fluxo dos gases reagentes. Foi observado que a baixos fluxos de gases a quantidade de reagente é insuficiente para ser difundida por todo eletrodo, o que ocasiona reação apenas na região de entrada de gases no sistema, ocasionando uma rápida limitação em obter-se densidades de corrente alta. Pode-se observar que a diferença de desempenho entre as placas é pequena, porém a placa serpentina 6 apresentou melhor desempenho. O desempenho dos cátodos preparados com catalisadores comerciais e os sintetizados no laboratório nas células de 50 cm² mostrou sofrer bastante influência das condições de operação comparada com as células de 4,6 cm². / The work aims at the development of the system for current distribution measurements and scale (50 cm²) seeking to improve the conditions of preparation of the membrane electrode assembly (MEA) as the cell operating conditions and evaluate the best geometry. We were performed catalyst synthesis studies of Pt-M and performance evaluation of these materials and the synthetic routes used in order to apply these materials in larger-scale systems. The failure of the performance and stability of the catalysts are factors that make yet infeasible the scale up use for solid polymer electrolyte fuel cells, highlighting the losses associated with the performance of the cathode. The nanoparticles of bimetallic catalysts Pt-M/C (M = Fe, Co and Ni) supported on high surface area carbon were prepared by two different synthetic routes. The synthesis route utilized were formic acid and modified ethylene glycol (EG). In both route search catalysts with a high degree of incorporation of the second metal, small particle size and good catalytic performance. It was observed that the route of formic acid with modifications in the synthesis process it is possible to obtain the average incorporation of the second metal in the catalyst, but there is a disadvantage of the particle size is high. The route of EG modified was obtained catalysts with small particle size, but incorporation of the second metal was inefficient. The scale-up studies were performed in 50 cm2 cell varying operating conditions: i) different gas distribution plates and ii) different values of flow of reactant gases. It was observed that at low gas flows the amount of reagent is insufficient to be spread throughout the electrode, causing reaction only in the region of entry of gases in the system, causing a fast limitation in obtain high current densities. It can be observed that the performance difference between the plates is small, but the plate serpentine 6 showed the best performance. The performance of the cathodes prepared with commercial catalysts and synthesized in the laboratory in 50 cm² cells showed considerable influence the operating conditions compared to cells 4,6 cm².

Un nouveau procédé intensifié de vinification continue avec un mélange de levures S.cerevisiae et Sch.pombe immobilisées dans des billes d’alginate est proposé. A l’échelle laboratoire, l’étude de la teneur en billes et de la configuration du réacteur conduit à l’obtention d’un réacteur de type lit fixe permettant une production de vin en 35 heures. Des validations du procédé aux échelles pilote (170 L) puis industrielle (120 hL) montrent que, en cave, du vin de qualité semblable au témoin est produit en 2 à 3 jours. Une analyse du comportement du réacteur a identifié des raisons de pertes de performances liées à l’hydrodynamique lors du changement d’échelle ainsi que des axes améliorations possibles. Ce procédé permet d’obtenir un vin de qualité maitrisée et un gain de temps de plusieurs semaines / From a batch to another, produced wines are usually different because of the different alcoholic and malolactic fermentation courses. To blend wines quality and continue wine production industrialization, a new continuous process, using Ca-alginate immobilized yeast cells, was developed for red wine-making. Working with a blending of S.cerevisiae and Sch.pombe allowed the regrouping of the alcoholic and malolactic fermentations in a unique step. After testing different reactor set-ups at lab scale, the selected process, a vertical bed reactor, was used in real wine-making conditions, firstly in a pilot reactor (170 L) and then in an industrial one (120 hL). The results showed that continuous wine-making was possible in 2 to 3 days. The wine presented nearly the same sensory profile compared to a classical one. Thanks to the analysis of the reactor behaviour, we were able to explain the efficiency losses linked to the hydrodynamic, observed during the scale-up. This new intensified process enables to obtain a wine with a controlled quality and to save several weeks of production time

Les objectifs de cette thèse sont, d’une part, de développer un système d’agitation performant en termes de transfert d‘oxygène et de forces de cisaillement afin qu’il soit adapté aux cultures de levures, de bactéries et de champignons filamenteux. D’autre part, de mettre au point un procédé de production des spores de champignons en milieu liquide pour remplacer la culture sur milieu solide (boîte de roux) utilisée au démarrage de la thèse par la société DSM. Ces travaux ont permis de sélectionner le mobile type « oreilles d’éléphant » comme système d’agitation. Son efficacité de transfert d’oxygène a été caractérisée dans une large gamme de débits d’air et de vitesses d’agitation. Les données obtenues ont été modélisées pour développer un outil de prédiction du KLa et d’extrapolation de procédé d’agitation. L’approche utilisée pour l’extrapolation est basée sur la similitude non géométrique et le maintien du coefficient de transfert d’oxygène (KLa) constant entre les échelles. L’efficacité du nouveau mobile d’agitation a été validée à l’échelle pilote en présence des cultures de bactéries, de levures et de champignons. Parallèlement au choix du mobile d’agitation, un procédé de production de spores de champignons filamenteux en milieu liquide a été développé. Ainsi, le milieu de culture et les conditions opératoires de chaque étape de la fermentation (germination, croissance végétative et sporulation) ont été optimisés. Ce nouveau procédé permet d’obtenir jusqu’à 200.106 cellules par millilitre de culture aussi bien à l’échelle du laboratoire que pilote (900L). Il a été aussi validé en présence de cultures de levures et de bactéries. Sur le même principe, les données obtenues à l’échelle pilote ont permis de développer un outil d’extrapolation à l’échelle industrielle de 10 m3 / The objectives of this thesis are, on one hand, to develop an efficient system of agitation in terms of oxygen transfer and shear stress, suited for the fermentation of yeasts, bacteria and filamentous fungi, on the other hand, to develop a process of spore production in a liquid medium. This new process will be used instead of the solid culture in Roux flasks used by DSM at the beginning of the PhD program. In a first step, we have selected “Elephant Ears” impellers as stirrers; its effectiveness of oxygen transfer was characterized in a broad range of stirring speeds and airflow rates. The data obtained were modeled and a predictive tool for KLa and process extrapolation was developed. The extrapolation is based on the non-geometrical similarity and the coefficient of oxygen transfer (KLa) was kept constant between scales. The effectiveness of the new stirrer was validated in a pilot-scale in the presence of the cultures of bacteria, yeasts and filamentous fungi. Besides the studies on the stirrer, a process in a liquid medium of filamentous fungi was developed. Thus, the fermentation medium and the operating conditions of each steps of the fermentation (germination, vegetative growth, and sporulation) were optimized. This new process makes possible to obtain up to 200.106 cells per milliliter of culture on both laboratory and pilot scale (900L). Finally, the data gathered on the pilot scale was used to extrapolate this process to the industrial-scale of 10m3

Anaerobic digestion, which is the process by which bacteria breakdown organic matter to produce biogas (renewable energy source) and digestate (biofertiliser) in the absence of oxygen, proves to be the ideal concept not only for sustainable energy provision but also for effective organic waste management. However, the production amount of biogas to keep up with the global demand is limited by the underperformance in the system implementing the AD process. This underperformance is due to the difficulty in obtaining and maintaining the optimal operating parameters/states for anaerobic bacteria to thrive with regards to attaining a specific critical population number, which results in maximising the biogas production. This problem continues to exist as a result of insufficient knowledge of the interactions between the operating parameters and bacterial community. In addition, the lack of sufficient knowledge of the composition of bacterial groups that varies with changes in the operating parameters such as temperature, substrate and retention time. Without sufficient knowledge of the overall impact of the physico-environmental operating parameters on anaerobic bacterial growth and composition, significant improvement of biogas production may be difficult to attain. In order to mitigate this problem, this study has presented a nonlinear multi-parameter system modelling of mesophilic AD. It utilised raw data sets generated from laboratory experimentation of the influence of four operating parameters, temperature, pH, mixing speed and pressure on biogas and methane production, signifying that this is a multiple input single output (MISO) system. Due to the nonlinear characteristics of the data, the nonlinear black-box modelling technique is applied. The modelling is performed in MATLAB through System Identification approach. Two nonlinear model structures, autoregressive with exogenous input (NARX) and Hammerstein-Wiener (NLHW) with different nonlinearity estimators and model orders are chosen by trial and error and utilised to estimate the models. The performance of the models is determined by comparing the simulated outputs of the estimated models and the output in the validation data. The approach is used to validate the estimated models by checking how well the simulated output of the models fits the measured output. The best models for biogas and methane production are chosen by comparing the outputs of the best NARX and NLHW models (each for biogas and methane production), and the validation data, as well as utilising the Akaike information criterion to measure the quality of each model relative to each of the other models. The NLHW models mhw2 and mhws2 are chosen for biogas and methane production, respectively. The identified NLHW models mhw2 and mhws2 represent the behaviour of the production of biogas and methane, respectively, from mesophilic AD. Among all the candidate models studied, the nonlinear models provide a superior reproduction of the experimental data over the whole analysed period. Furthermore, the models constructed in this study cannot be used for scale-up purpose because they are not able to satisfy the rules and criteria for applying dimensional analysis to scale-up.

Répondre aux besoins énergétiques croissants de nos sociétés et limiter leur impact sur l’environnement est un enjeu actuel majeur. De nouvelles technologies alternatives comptent tirer profit de sources d’énergie négligées. Le potentiel énergétique des eaux usées peut être exploité par de nouvelles technologies telles que les piles à combustible microbiennes (PACM). Ces piles, pouvant produire de l’énergie électrique à partir d’eaux usées, montrent une diminution de leur rendement énergétique lorsque leur taille augmente, ce qui ne permet pas encore leur application industrielle. Ces travaux de thèse visent à identifier certains verrous de ce changement d’échelle et à proposer de nouvelles directions pour leur optimisation. Une première partie s’est intéressée à l’influence des collecteurs de courant anodiques sur les performances électriques et sur le développement du biofilm électro-actif. Nous avons émis l’hypothèse qu’à grande échelle, les collecteurs de courant peuvent être un élément limitant à la production d’électricité. Pour vérifier cette hypothèse, quatre PACM avec une anode de 490 cm² connectée de différentes manières ont été étudiées. L’augmentation du nombre de collecteurs a permis une hausse de la puissance produite par les PACM. La disposition des collecteurs affecte la répartition du potentiel sur la surface d’anode et peut engendrer dans certains cas, des gradients de potentiel qui influencent la structure microbiologique du biofilm, en particulier Geobacter. Par ailleurs, des mesures d’impédance ont montré que multiplier les collecteurs augmente la capacité de double couche de l’anode et engendre un courant capacitif dont l’importance pour les performances de fonctionnement en cycles de charge/décharge est non négligeable. La suite du travail s’est attachée à prendre en compte différents aspects physiques, notamment l’aspect hydrodynamique, afin de modéliser leur fonctionnement. Pour cela, trois PACM de volumes différents ont été mises en œuvre et testées à différents débits. Les données de configuration, d’opération et de performances de ces piles ont permis de construire des modèles statistiques de régression linéaire multiple prédisant la valeur de puissance maximale. Ces deux modèles ont montré que la puissance maximale produite était principalement corrélée à la vitesse de l’électrolyte circulant dans la pile et à la contrainte de cisaillement appliquée à l’anode par le mouvement du fluide. Ces deux parties ont également montré que l’abondance dans le biofilm de Geobacter, une bactérie électro-active très répandue dans les PACM, n’était pas corrélée avec la puissance maximale. Tout en étant très abondante, son seul nombre n’explique donc pas entièrement les performances électriques d’une PACM. / Facing increasing energy needs and limiting their impact on the environment are current and major issues for society. Renewable energy development is needed and new alternative technologies could benefit from exploiting neglected energy sources, such as microbial fuel cells (MFC), for energy production. MFCs can be operated with wastewater and produce a reasonable quantity of energy at the small laboratory-scale. Unfortunately, when their size is increased, their efficiency dramatically decreases, which prevents their industrial use. This thesis aims at identifying some obstacles to scale-up of MFC and proposing new directions for its optimization. The first part of the study was focused on the influence of anodic current collectors on electrical performance and on electroactive biofilm development. Our hypothesis was that they could be a limiting factor for electricity production at large scales. To test this hypothesis, four MFCs were operated with a 490 cm² anode connected to the external circuit in a different ways. Increasing the number of collectors improved the power. Collector’s layout influenced electrical potential on the anode surface and created an electrical potential gradient on the anode and this gradient shaped the microbiological structure of the biofilm. This effect especially concerns Geobacter, whose clade G. metallireducens is favored at strongly negative potentials. In addition, impedance measurements showed that multiplying collectors increased the double layer capacitance and, thus, generated a capacitive current that was important for MFC functioning in cycles of charge/discharge and that would improve its performance. Then, MFCs were considered as bioreactors and their different aspects, notably hydrodynamics, were taken into account to model their power output. Three MFCs of different volumes were operated under continuous-flow conditions and tested at four different flow rates. Configuration, operation and performance data were used to build two multiple linear regression statistical models: the first with variables selection through LASSO, the second with dimensionless numbers created with the Vaschy-Buckingham theorem. These two data-driven models showed that the maximal power was mostly correlated to electrolyte transfer rates inside MFC chamber and to shear stress at the anode generated by fluid movement. These two major experimental projects also showed that the abundance of Geobacter, an electroactive bacteria, inside the biofilm was widespread in MFCs, but it was not correlated to maximal power. Despite its large abundance, its quantity alone does not entirely explain the performance of a MFC. In order to succeed at MFC scale-up, fundamental research on electroactive biofilms, process engineering and modeling need to be associated and generalized as empirical results and their explanation.