Cogliati, Belén. (2011). Modelos viscosos em mecânica dos solos: análise de uma equação visco-hipoplástica. Dissertação de Mestrado, Escola Politécnica da Universidade de São Paulo, São Paulo. Esta dissertação estuda o comportamento de um modelo visco-hipoplástico proposto por Niemunis (2002), com as funções constitutivas da equação hipoplástica de Nader (2003). Para entender o papel da viscosidade no comportamento do solo são discutidos o adensamento secundário, a influência da velocidade de deformação na resistência não-drenada e a variação do coeficiente de empuxo com o tempo. Como etapa preliminar, são apresentados os modelos reológicos simples em uma dimensão, formados por um só elemento (modelos de Hooke, Newton e Saint-Venant) e modelos compostos pela combinação desses elementos (modelos de Maxwell, Bingham, Kelvin- Voigt, sólido linear padrão e visco-plástico com endurecimento). São deduzidas as equações de fluência e relaxação para todos esses modelos. Em três dimensões, são apresentadas as formulações do modelo visco-hipoplástico de Niemunis (2002) com as funções constitutivas de Nader (2003). São deduzidas as expressões simplificadas desse modelo para ensaios triaxiais. Em seguida, as equações são aplicadas à simulação de ensaios de compressão isotrópica e compressão não-drenada, com o objetivo de investigar a relaxação e a fluência bem como para analisar a influência dos parâmetros na resposta do modelo. / This thesis studies the behavior of the visco-hypoplastic model proposed by Niemunis, using Nader\'s hypoplastic constitutive equations. To understand the importance of viscosity in soil behavior the following topics are first examined: secondary consolidation, strain rate effects on undrained strength and the time variation of the coefficient of lateral pressure at rest. As a preliminary step, the present work discusses one-dimensional rheological models formed by a single element (Hooke\'s, Newton\'s and Saint-Venant\'s models) or by the combination of these elements (Maxwell\'s, Bingham\'s, Kelvin-Voigt\'s models; the standard linear solid model and the visco-plastic hardening model). For all the rheological models creep and relaxation are investigated. Niemunis\' visco-hypoplastic model with Nader\'s constitutive equations is presented next. First, simplified expressions of this model for triaxial test are deduced. Then the equations are applied to the simulation of isotropic compression and undrainded compression tests, with the aim of investigating relaxation and creep as well as of analyzing the influence of each parameter on the model response.

O chocolate é um dos principais produtos oriundos do cacau, sendo comercializado em sua maior parte na forma de bombons maciços ou recheados. Bombons com recheio de fruta podem conter a fruta ou pedaços da fruta distribuídos no recheio. Eles podem ser produzidos em base fondant, gordura, geléia ou com o próprio chocolate (trufas), entre outros, adicionando-se aromas e corantes ou não. O objetivo deste trabalho foi produzir preparados de frutas e recheios para bombons em base fondant, gordura hidrogenada e chocolate branco (trufa), utilizando estes preparados como substituto parcial ou total de aromas e corantes na formulação final dos recheios. Os preparados foram elaborados utilizando-se açúcar invertido, xarope de glicose e água como matriz, adicionados de morango e maracujá liofilizados e casca de laranja, para conferir o aroma da fruta in natura, reduzindo de maneira significativa o uso de aromas e corantes artificiais. Avaliou-se pH nos preparados de frutas e nos recheios; teor de sólidos solúveis nos preparados de fruta; atividade de água nos recheios base fondant, gordura hidrogenada e \"trufa\". Os valores de pH apresentaram-se menores que 4,0 para os preparados e menores que 6,0 para os recheios. A atividade de água não apresentou variação durante o período estudado. O teor de sólidos solúveis encontrado foi de 68 ºBrix, 69 ºBrix e 73 ºBrix, respectivamente para os preparados de morango, maracujá e laranja. A composição centesimal foi feita para caracterizar os bombons quanto ao teor de gordura, proteínas, umidade, cinzas e valor energético. Os bombons recheados com açúcar fondant apresentam os menores teores de lipídeos, foram encontrados valores de proteínas maiores para as trufas. O valor energético destes produtos variou de 500 a 600 kcal por 100 g de produto. O estudo da textura dos bombons recheados foi feito a 25 ºC e caracterizou os bombons quanto a sua dureza, sendo que as trufas se mostraram os bombons de menor dureza. Os dados de atividade de água e dureza foram analisados através da ANOVA e Teste de Tukey. Na análise sensorial dos bombons recheados foi aplicado teste de aceitação global aos consumidores. A análise dos dados foi feita através da ANOVA, Teste de Tukey, Mapa de Preferência e Análise de Cluster. O estudo da reologia foi feito para os preparados de fruta, para os recheios base fondant e gordura hidrogenada a 25 ºC. Para as trufas foi feito o estudo reológico através do teste rotacional à 0 ºC e 70 ºC. Os ensaios oscilatórios utilizaram varredura de freqüência. A análise sensorial apresentou resultados significativos quanto à aceitação dos produtos, sendo que as trufas foram as que obtiveram as maiores notas, seguidas dos bombons com recheio base gordura hidrogenada. Os ensaios oscilatórios demonstraram que os preparados de fruta e os recheios nos sabores morango, maracujá e laranja são estáveis quando sujeitos a uma baixa freqüência. No teste rotacional, a curva de fluxo determinou o comportamento dos preparados. As curvas de fluxo obtidas através do teste rotacional, foram avaliadas quanto ao melhor modelo reológico aplicado. A utilização de um preparado de fruta, dando aos recheios características da fruta in natura, conferindo sabor e aroma sem a necessidade de nova adição de aroma e corante, em combinação com açúcar fondant, com gordura hidrogenada e nas trufas, mostrou-se uma maneira interessante de produzir um recheio de baixo custo, com longa vida de prateleira se comparado aos produtos artesanais presentes no mercado, conferindo aos produtos atributos bem aceitos pelo consumidor. / Chocolate is one of the most important products made from cacao, being major commercialized as chocolate filled or plain. Chocolate filled with fruit can contain the fruit or parts of it mixed in the filling. The fillings could be produced using fondant, fat, jelly or the white chocolate, as the truffles, with the addition of flavor and colorants or not. The objective of this work was to produce fruit mixes and filling for chocolate, using fondant, hydrogenated fat and White chocolate (truffles), using those mixes as replacers of flavor and colorants in the final formulations. The mixes were elaborated using inverted sugar, glucose syrup and water, mainly, freeze dried strawberry and passion fruit, and orange, was added to gave flavor of in nature fruit, reducing the use of artificial flavor and colorants. The pH was evaluated for the mixes and fillings; soluble solids content in the mixes; water activity in the fillings. The pH values showed to be lower than 4,0 to the mixes , and lower than 6,0 for the fillings. Water activity didn\' t show variability during the evaluated period. The soluble solids content was 68 ºBrix, 69 ºBrix and 73 ºBrix, respectively for strawberry, passion fruit and orange mix. The centesimal composition was done to characterize the chocolates filled showing the content of fat, protein, humidity, ashes and energetic value. The chocolates filled with fondant sugar presented the lower values of fat, protein was greater for the truffles. The energetic value varied between 500 to 600 kcal per 100 g of product. The texture study of chocolates filled was done at 25 ºC and characterized the hardness of them. The truffles presented the lower values for hardness. The data of water activity and hardness were analyzed trough ANOVA and Tukey test. Sensory evaluation of chocolates filled was done applying global evaluation test to consumers. The data analysis was done trough ANOVA, Tukey test, Internal Preference Mapping and Cluster Analysis. The rheological study was done at 25 ºC for the mixes, fondant and hydrogenated fat fillings, for the truffles the rotational test was done at 0 ºC and 70 ºC. The oscillatory test was done using a frequency sweep. The sensory evaluation presented significant results, regarding product acceptance, the truffles gotten the greater grades, followed by the hydrogenated fat chocolate filled. The oscillatory study showed stability for all the mixes and fillings, when those are subjected to lower frequencies. In the rotational study, the flow curve determined the mixes behavior. The flow curves obtained trough rotational test, were evaluated for the best rheological model which fitted them. The use of a fruit mix, giving characteristics of in nature fruit, with flavor and color without the need of adding artificial ones, combined with fondant, hydrogenated fat and for the truffles, showed to be an efficient way to produce fillings with low cost, and a greater shelf life, if compared with artisanal products of the market, being well accepted by consumers.

A enzima lipoxigenase através da farinha integral de soja é amplamente utilizada na panificação com objetivos de branqueamento e melhora da reologia. Com base nesse pressuposto, o objetivo inicial deste estudo foi de observar a ação oxidante dessa enzima na presença de dois outros oxidantes (ácido ascórbico e peróxido de benzoíla) nas propriedades reológicas e sensoriais de pães. Sete misturas com diferentes combinações dos três compostos foram elaboradas utilizando-se um delineamento experimental do tipo \"centroide-simplex\", sendo aplicadas à farinha de trigo para o preparo de pães. A enzima apresentou sinergia com o ácido ascórbico na elasticidade da massa, sugerindo uma inter-relação bioquímica entre os dois compostos no fortalecimento da matriz protéica do glúten. A mesma sinergia entre a lipoxigenase da soja e o acido ascórbico foi observada na redução da tonalidade amarela dos pães medida instrumentalmente. A partir desses resultados, buscou-se avaliar a ação da enzima frente a variações tanto na força da farinha como no tempo de fermentação, reproduzindo as condições reais de processamento. Desta vez, um planejamento experimental do tipo Box-Behnken com três fatores (atividade da enzima, força da farinha e tempo de fermentação) em três níveis de variação, foi aplicado no preparo das amostras. Os resultados mostraram que a lipoxigenase da soja (fração 1) é uma enzima oxidante de ação rápida e seu efeito no branqueamento dos carotenóides da farinha de trigo exibiu interação positiva tanto com a força da farinha como o tempo de fermentação. Respostas reológicas e sensoriais decorrentes da ação conjunta de diferentes oxidantes e condições de processamento em pães foram pela primeira vez descritas através de modelos polinomiais com capacidade preditiva. / Soy lipoxygenase enzyme through soy flour is widely used in breadmaking for bleaching and rheological improvement. Based on this fact, the initial objective of this study was to observe the enzymes\' action in the presence of two other oxidants (ascorbic acid and benzoyl peroxide) in the rheological and sensory properties of breads. Seven mixtures with different combinations of the three compounds were prepared using a \"centroid-simplex\" design and applied to wheat flour for breadmaking. The enzyme showed synergism with ascorbic acid in the dough elasticity suggesting a biochemical interrelation between the two compounds in strengthening the gluten protein matrix. The same synergy between soy lipoxygenase and ascorbic acid was observed with the yellow hue reduction of breads measured instrumentally. From these results, the enzyme action was evaluated with wheat strength and proofing time variations reproducing real processing conditions. This time, a Box-Behnken experimental design with three factors (enzyme activity, wheat strength and proofing time), and three variation levels were applied in the sample preparation. The results showed that soy lipoxygenase (type 1) is a fast-acting oxidizing enzyme and its bleaching effect on wheat flour carotenoid exhibited a positive interaction with wheat strength and proofing time. Rheological and sensory results attributed to the combined action of the different oxidants and breadmaking conditions were for the first time described through polynomial models with predictive capacity.

Mudanças econômicas e hábitos alimentares incorretos estão contribuindo para maior ingestão calórica nas populações dos centros urbanos. Intensificam-se as pesquisas para a obtenção de ingredientes com menor valor calórico. Neste trabalho, foram avaliados os efeitos da Polidextrose (Litesse®II), Salatrim (Benefat®) e Sucralose (Splenda®) em substituição à gordura hidrogenada e sacarose em formulação de pão para hambúrguer. Avaliou-se a dureza, volume, umidade, cor e valor calórico pela Metodologia da Superfície de Resposta (MSR). Os resultados mostraram que os ingredientes utilizados sugerem grande potencial para o desenvolvimento de produtos panificados com valor energético reduzido, e a sobreposição das superfícies permitiu obtenção de formulações otimizadas. / Economic changes and incorrect nutrition habits are causing more calories ingestion in the populations of the urban centers. Substitution of fat and sugar on diet has being object of intense research in the attempt to producing foods without these ingredients or with reduction of amount normally used. In this work the effect of the Polidextrose (Litesse® II), Salatrim (Benefat®) and Sucralose (Splenda®) had been tested in substitution to the hydrogenated fat and sucrose in hamburger buns formulas. It was evaluated firmness, volume, moisture, color and caloric value using Response Surface Methodology (RSM). The results showed that the used ingredients suggest great potential for the development of bakery products with reduced energy value. Surfaces intersections allowed to obtain optimized bread formulas.

A tese consistiu no estudo por técnicas reológicas da dinâmica de géis e dispersões da gliadina do glúten do trigo em meios de dimetilsulfóxido (DMSO), dimetilformamida (DMF) e formamida puros e em suas misturas com a água. Investigou-se o efeito da temperatura e das concentrações das espécies componentes (proteína e constituintes dos sistemas solventes) sobre o comportamento de fluxo e de deformação dos sistemas. Foram realizados testes em regime estacionário, ensaios transientes e ensaios dinâmicos (oscilatórios). A caracterização da proteína nos diversos meios foi realizada através de técnicas de espalhamento de luz dinâmico e estático e de espalhamento de raios X em baixos ângulos. Esses recursos permitiram descrever a distribuição de populações da proteína nos géis e dispersões, em suas formas livres ou associadas, assim como estimar aspectos de tamanho e mobilidade a partir da determinação dos raios hidrodinâmicos e coeficientes de difusão. As composições de aminoácidos da gliadina, na fração bruta e em uma fração isolada em meio etanólico, foram determinadas por técnicas de CLAE Cromatografia Líquida de Alta Eficiência. Um isolamento preliminar das sub-frações de gliadina foi obtido por outra técnica cromatogrática (FPLC Fast Polymer Liquid Chromatography). Foram ainda desenvolvidos estudos relativos ao efeito de gliadina sobre a bicamada lipídica de vesículas gigantes de fosfolipídio, através da técnica de aspiração por micropipetas, com monitoramento por vídeo-microscopia. O objetivo foi o de se verificar o eventual efeito da proteína sobre a constante de elasticidade, kc, da bicamada. Os resultados obtidos podem ser resumidos como se segue: a) o comportamento viscoelástico apresentado pela gliadina mostrou dependência com a natureza e a composição do meio de dispersão. Foi verificada a ocorrência de comportamento tixotrópico ou reopéxico como função da temperatura e um comportamento cíclico para a viscosidade em função do tempo. b) foi possível correlacionar alguns aspectos da estrutura molecular e polaridade do solvente com a maior ou menor capacidade de formação de gel de gliadina: solventes com regiões apolares mais expressivas, como DMSO e DMF (ε= 47,24 e 38,25, respectivamente) induzem a formação de géis rígidos, enquanto um composto como formamida, destituído de um domínio apolar minimamente significativo e ε maior que o da água (ε = 84), tem capacidade significativamente menor de promover a formação de géis com as mesmas características de rigidez. c) os géis são formados por unidades com raios de giro da ordem de 15Å, além de agregados de tamanhos bem maiores. Considerando-se a massa molar média de 32 350 g/mol para a gliadina em nossa amostra, é de se supor que a proteína encontre-se muito compactada no gel (considere-se que lisozima, com massa molar muito menor, igual a 14300 g/mol, possui raio de giro nesse mesmo valor). Um modelo de enovelamento interpenetrante, proposto em nosso grupo através de métodos de dinâmica molecular para a conformação da gliadina, seria assim compatível com essas observações experimentais. d) gliadina ω, seletivamente extraída por tratamento com etanol, mostrou-se capaz de reduzir a elasticidade de membranas de microvesículas de fosfolipídios, aumentando o valor de sua constante de curvatura kc. Um efeito de saturação foi observado, para razões de massa gliadina/DOPC ~2-3%. e) coeficientes de difusão para dispersões de gliadina são da ordem de 10-10 cm2.s-1, com raios hidrodinâmicos na faixa nm-µm. Duas populações dinamicamente distintas foram identificadas nos géis de gliadina. f) o grau de polidispersão e o tamanho dos agregados mostraram dependência com a concentração da proteína nos géis: são gerados agregados menores (diminuição da massa molar média) em tamanhos mais uniformes (menor polidispersão), com o aumento da concentração de gliadina. Essa mesma tendência de variação é observada como função do tempo, indicando a ocorrência de rearranjos estruturais na fase de envelhecimento dos géis. / The thesis consisted in a rheological study of the dynamics of gels and dispersions of gliadin from wheat gluten in pure dimethylsulfoxide, dimethylformamide or formamide and in their mixtures with water. The effect of temperature and that of the concentration of the component species (protein and solvent constituents) on flow behavior and deformation were evaluated. Steady-state, transient and dynamical (oscillatory) tests were performed. Protein characterization in the various media was performed through dynamic and static light scattering and small angle X ray scattering techniques. The application of techniques allowed to describe the distribution of protein populations in the gels and dispersions, in their free and associated forms, as well as to estimate their size and mobility, through the determination of the hydrodynamical radii and diffusion coefficients. The amino acid compositions of gliadin, both in the crude protein and in an isolated fraction in ethanol, were determined by HPLC technique. A preliminary isolation procedure of gliadin sub-fractions was achieved by another chromatographic technique (FPLC). The effect of gliadin on the phospholipid bilayer of giant vesicles was studied through the technique of micropipet aspiration coupled to videomicroscopy. The aim was to verify the possible effect of protein on the elasticity constant, kC of the bilayer. Results obtained can be summarized as follows: a) the viscoelastic behavior presented by gliadin was found to be dependent on the nature and composition of the dispersing media. Temperature-dependent tixotropic or rheopexic behavior were observed, as well as a cyclic behaviour for viscosity as a function of time. b) it was possible to correlate some aspects of solvent molecular structure and polarity with its greater or lesser capacity of gel formation: solvents with more expressive apolar regions, such as DMSO and DMF (ε = 47,24 and 38,25, respectively) induce formation of rigid gels, whereas a compound such as formamide, devoid of a minimally significant apolar domain and ε greater than that of water (ε = 84), is significantly less able to promote gel formation with the same characteristics of rigidity. c) gels are formed by units with gyration radius around 15Å and also by aggregates of greater size. Considering that the average molar mass in our sample for gliadin is 32350, one can assume that the protein is very densely packed in the gel (consider for instance that lysozyme, with a much lower molar mass, 14300 g/mol, has a gyration radius of this same size). A compact, interpenetrating folding model for gliadin, developed in our group through molecular dynamics, is compatible with such experimental observations. d) ω gliadin, selectively extracted in ethanol, was found to be able to reduce membrane elasticity of phospholipid microvesicles, by increasing its elastic curvature constant kC. A saturation effect was observed, for a mass ratio as low as gliadin/DOPC ~2-3%. e) diffusion coefficients for gliadin dispersions are circa 10-10 cm2.s-1, with hydrodynamical radii in the nm-µm range. Two dynamically distinct populations are identified in the gliadin gels. f) gel polydispersity and aggregate size depend on protein concentration: smaller aggregates (lower average molar mass) of more uniform sizes are produced for increasing gliadin concentration. Variation with time followed the same pattern, indicating that structural rearrangements take place during gel aging.

L'objectif des travaux présentés dans ce mémoire est de caractériser la texture et la rhéologie des milieux granulaires en fonction de la polydispersité de forme et de taille des particules mais aussi en fonction de la cohésion interne des particules lorsqu'elles sont fragmentables, en vue d'une meilleure compréhension du procédé de fabrication des compacts. Pour ces études, nous avons utilisé la méthode de Dynamique des Contacts avec un modèle de rupture des particules. Nos analyses montrent que la polydispersité de forme peut jouer un rôle important lorsque la polydispersité de taille est faible. Par exemple, la résistance au cisaillement est presque indépendante de la polydispersité de taille mais elle diminue lorsque les formes deviennent plus irrégulières. La fragmentation des particules est fortement hétérogène en raison de la redistribution des contraintes au sein du matériau. Celui-ci garde la mémoire de la distribution granulométrique initiale, mais une classe intermédiaire de tailles se développe avec une distribution en loi de puissance des tailles et un rapport d'aspect moyen proche du nombre d'argent. Au cours du cisaillement, les déformations se localisent dans des bandes avec une compacité supérieure au reste du matériau en raison de la fragmentation des particules. La fragmentation tend également à annuler la dilatance et le pic de contrainte. / The general objective of the work presented in this dissertation is to investigate the microstructure and rheology of granular materials as a function of size and shape polydispersity of the particles, and to analyze the role of particle fragmentation as a function of the internal cohesion of particles in view of a better understanding of the manufacture process of powder compacts. For this work, we used numerical simulations by means of the Contact Dynamics method with a model of particle fracture. Our results suggest that shape polydispersity may play an important role when size polydispersity is low. For example, the shear strength is nearly independent of size distribution but declines when the particles becomes increasingly more irregular in shape. The process of particle fragmentation is found to be highly inhomogenious as a result of stress redistribution. The memory of the initial size distribution is mainly conserved in the class of larger particles while a class of intermediate sizes develops with a power-law size distribution and a mean aspect ratio close to the silver number independently of the initial size distribution. During shear, the strain is localized in shear bands of large solid fraction as a consequence of particle fragmentation and enhanced size polydispersity. Particle fragmentation tends to reduce dilatancy and the peak shear strength.

La convection dans le manteau terrestre est la principale force motrice du mouvement des plaques tectoniques. Alors que les parties inférieures du manteau supérieur se déforment de manière ductile, les plaques tectoniques sont rhéologiquement plus rigides que l'asthénosphère sous-jacente. Pour comprendre le couplage entre la convection profonde et les plaques tectoniques à la surface de la Terre, il est essentiel de comprendre les mécanismes de déformation visqueuse et frictionnelle du manteau lithosphérique. Mais à ce jour, la rhéologie du manteau supérieur juste au-dessous de la discontinuité de Mohorovicic est encore mal comprise. De plus, les premiers stades de la déformation viscoplastique à des températures intermédiaires (600-1000 ° C) pertinentes pour le manteau lithosphérique, ne sont ni bien documentés ni quantifiés. Dans le passé, la plupart des expériences de déformation étaient effectuées à des températures très élevées (> 1200 ° C). Pour fournir des valeurs mécaniques précises pour le manteau lithosphérique, nous avons besoin de données mécaniques mais aussi de la caractérisation de la microstructure associée pour comprendre la physique des mécanismes en jeu lors de la déformation permanente des roches riches en olivine. Dans cette thèse, nous avons réalisé des expériences de déformation en compression axiale à l'aide d'une presse Paterson (Géosciences Montpellier, Université de Montpellier, France) à haute pression et température (300 MPa, 1000-12000 ° C) et en torsion (‘rotary shear frictional testing machine’ au laboratoire de mécanique des roches, université de Durham, Royaume-Uni) à pression et température ambiantes. Les échantillons ont été caractérisés par microscopie électronique à balayage, diffraction d’ d'électrons rétrodiffusés et microscopie électronique en transmission. Après un chapitre d'introduction où l'état de l'art est détaillé et un chapitre consacré aux méthodes expérimentales et analytiques utilisées dans les projets scientifiques, la thèse s'organise en trois chapitres, chacun correspondant à trois articles scientifiques: le premier est publié (1) Évolution de la contrainte et des microstructures associées au fluage transitoire de l'olivine à 1000-1200 °C (Phys. Earth Planet. Int., doi: 10.1016/ j.pepi.2018.03.002. (https: //hal.archives- ouvertes.fr/hal-01746122) et les deux autres sont en préparation, (2) Densité de disclinaisons dans l'olivine polycristalline déformée expérimentalement à 1000 ° C et 1200 ° C (3) Déformation par cisaillement de l'olivine nano- et micro-cristalline. Le premier projet du chapitre III a montré que le durcissement mécanique observé ne peut pas provenir d'une simple augmentation de la densité de dislocations (e.g., la forêt) et que d'autres mécanismes doivent être mis en œuvre pour compenser les limites de glissements des dislocations. Dans le chapitre IV, les densités de dislocation géométriquement nécessaires (GND, défauts de translation) et les disclinaisons (défauts de rotation) sont quantifiées sur une série de roches déformées à différentes températures, déformations finies et niveaux de contrainte, mais aucune corrélation n'a été identifiée entre la densité de disclinaisons, et la contrainte, la déformation finie, ou la densité de GND. Le rôle des disclinaisons serait donc limité à la migration aux joints de grains, ce qui peut être suffisant pour débloquer les dislocations dans l'agrégat d'olivine polycristalline. Au chapitre V, les expériences de torsion ont confirmé l'effet négligeable de la taille du grain (olivine de 0,7 à 70 µm) sur la diminution drastique du coefficient de frottement, mais la caractérisation des échantillons n’a pas permis d'élucider le mécanisme principal de déformation. Cette thèse a permis de mieux caractériser la transition fragile-ductile d'une roche de type dunite à grains fins soumise à une déformation permanente aux températures du manteau sommitale. / Convection in Earth’s mantle is the major driving force behind the movement of tectonic plates. While the lower parts of the upper mantle deform in a ductile way, the plates themselves are rheologically more rigid than the asthenosphere beneath. To understand how convection yields tectonic plates, it is vital to quantify the viscous and frictional strength of the lithospheric mantle. Yet to date, the rheology of the uppermost mantle just below the Mohorovicic discontinuity is still poorly understood. Furthermore, the early stages of visco-plastic deformation at intermediate temperatures (600 – 1000 °C) relevant of the lithospheric mantle are not well documented or quantified. In the past, most deformation experiments were performed at high temperatures (> 1200 °C). To provide accurate mechanical values for the lithospheric mantle, we need mechanical data but also a characterization of the associated microstructure to understand the deformation mechanisms at play during permanent deformation of olivine-rich rocks. In this thesis, I have performed deformation experiments in axial compression using a Paterson press (at Géosciences Montpellier, University of Montpellier, France) at high pressure and temperature (300 MPa, 1000 -12000 °C) and in torsion using a low to high velocity rotary shear frictional testing machine (Rock Mechanics Laboratory, Durham University, UK) at room pressure and temperatures. The recovered samples were characterized using scanning electron microscopy, electron backscatter diffraction and transmission electron microscopy. After an introduction chapter where the state-of-the-art is detailed, and a chapter focusing on experimental and analytical methods used during scientific projects, the thesis is organized as three subsequent chapters, each of them corresponding to three scientific articles: one is published (1) Stress evolution and associated microstructure during transient creep of olivine at 1000-1200 °C (Phys. Earth Planet. Int., doi: 10.1016/j.pepi.2018.03.002.); and the two others are in preparation, (2) Disclination density in polycrystalline olivine experimentally deformed at 1000 °C and 1200 °C; and (3) Shear deformation of nano- and micro-crystalline olivine at seismic slip rates. Chapter III has shown that the observed mechanical hardening can not come from a simple increase in dislocation density (e.g., entanglement) and that other mechanisms must be at play to compensate for the limitations of dislocation slip. For the first time, in chapter IV the densities of geometrically necessary dislocations (GND, translational defects) and disclinations (rotational defects) are quantified on a series of rocks deformed at different temperatures, finite strains and stress levels. No correlation has been identified between disclination density and stress, strain or GND. The role of the disclinations will therefore be limited to migration at grain boundaries, which may be sufficient to unblock dislocations in the polycrystalline olivine aggregate. In chapter V, torsion experiments confirmed the negligible effect of grain size (olivine from 0.07 to 70 μm) on the drastic decrease of the coefficient of friction, but the characterization of the samples did permit to shed light on the main mechanism of deformation. Thanks to an experimental approach and up-to-date material characterization, this thesis permitted better characterization of the brittle-ductile transition of a fine-grained dunite-type rock subjected to permanent deformation at uppermost mantle temperatures.

Le caoutchouc naturel (CN) est un élastomère produit à partir du latex d'Hevea brasiliensis. Ce matériau est constitué en moyenne de 94% (m/m) de poly(cis-1,4-isoprène) et de 6% (m/m) de composés non-isoprènes avec notamment des lipides (1,5-3,0%), des protéines (~ 2,0%) et des éléments minéraux (0,2%). Ces composés non-isoprènes seraient responsables de la structure associative ainsi que des propriétés remarquables du CN. Ce matériau présente une grande variabilité de ses propriétés de mise en œuvre, liée aux conditions agronomiques (saison, clones, système d'exploitation) et aux différents procédés de fabrication existants. Le projet CANAOPT, dans lequel s'inscrit cette thèse, vise à identifier et à quantifier les sources de variabilité, ainsi qu'à développer de nouveaux grades de CN à variabilité réduite et contrôlée pour les applications antivibratoires. Ces travaux de thèse visent à mieux comprendre l'origine du gel du CN et son influence sur les propriétés rhéologiques du matériau. Cette phase gel (ou gel total) est constituée d'une fraction insoluble en solvant organique appelée le macrogel, et d'une fraction de microagrégats dispersés dans la fraction soluble et appelée le microgel.Le premier axe de recherche de cette thèse vise à identifier l'effet du gel sur les propriétés rhéologiques du CN. Dans ce but, des méthodes ont été développées afin de quantifier les différents taux de gels du CN, de caractériser la structure des chaînes de poly(cis-1,4-isoprène) et des microagrégats (SEC-MALS en mode TBABr), mais également de mesurer les propriétés rhéologiques en double cisaillement (DMTA). Une méthode permettant de caractériser rapidement le taux de composés non-isoprènes des échantillons par spectroscopie FT-IR a également été mise au point. Des corrélations ont d'abord été mises en évidence entre les masses molaires moyennes des macromolécules et les propriétés rhéologiques d'une sélection de 25 échantillons différents. Enfin, une corrélation a également été mise en évidence entre le taux de gel total et le paramètre K' décrit par le modèle de Perez. Ce paramètre est déterminé graphiquement par la représentation Cole-Cole. En parallèle, ces travaux se sont portés sur une caractérisation physico-chimique du microgel et du macrogel du CN. Une méthodologie d'extraction du microgel a été mise en place, et a permis la caractérisation des microagrégats du CN. La composition, le degré de réticulation, et les propriétés rhéologiques des macrogels issus d'une sélection de 11 échantillons de CN ont également été étudiés. Il a été montré que le taux de protéines contenus dans le macrogel (5-51% m/m) est directement lié à son degré de réticulation et impacte ses propriétés rhéologiques, très supérieures à celles du CN brut et de la fraction soluble correspondants. / Natural Rubber (NR) is an elastomer made from Hevea brasiliensis latex. It contains about 94% (w/w) of cis-1,4-polyisoprene and 6% (w/w) of non-isoprene compounds, such as lipids (1.5-3.0%), proteins (~ 2.0%) and minerals (0.2%). These non-isoprene compounds are assumed to be responsible for the associative structure and the outstanding properties of NR. However, substantial variability in NR properties is caused by diverse agronomic conditions (season, clones and tapping system) and by the different manufacturing processes. This PhD thesis is part of the French CANAOPT project, which aims to create new NR grades with reduced and controlled variability for antivibratory applications. This PhD work aims for a better understanding of the origin of the NR gel phase and its influence on rheological properties. The gel phase of NR (or total gel) is composed of a fraction that is insoluble in organic solvent (macrogel), and of microaggregates dispersed in the soluble fraction (microgel). The first component of this PhD research set out to identify the impact of the gel phase on NR properties. To that end, characterization methods were developed in order to investigate the different gel contents, the structures of cis-1,4-polyisoprene chains and microaggregates (SEC-MALS in TBABr mode), and also rheological properties in double-shear solicitation (DMTA) of NR samples. The non-isoprene contents (lipids and proteins) were also investigated using a new characterization method in FT-IR spectroscopy, developed during this PhD work. Correlations between the average molar masses of the macromolecules and the rheological properties of 25 NR samples were highlighted. Finally, a correlation was found between the total gel content and a K' parameter, introduced by the Perez model and graphically determined by the Cole-Cole representation.At the same time, physicochemical characterization of the microgel and the macrogel of NR samples was undertaken. First, a new method was developed in order to extract and study NR microaggregates. The composition, the degree of crosslinking and the rheological properties of macrogels extracted from 11 NR samples were also investigated. It was found that the protein content (5-51% w/w) of the macrogel was clearly correlated to its degree of crosslinking. The macrogel protein content also influenced its rheological properties, which were much greater than those of the corresponding NR and soluble fraction.

L’incorporation de nanoplaquettes d’argile dans une matrice ou un mélange de thermoplastiques s’avère prometteur pour en améliorer les propriétés d’usage. Récemment, un procédé de synthèse hydrothermale a permis d’obtenir un hydrogel chargé de nanoplaquettes de talc, ouvrant des perspectives d’élaboration de nanomatériaux à base de talc, potentiellement concurrents de ceux à base d’argile modifié. La grande quantité d’eau, introduite dans les appareils de mise en œuvre à l’état fondu, conduit à une perte de matière minérale importante et à l’agrégation partielle du talc. Ces contraintes rendent obligatoire l’utilisation de la mini-extrudeuse, de faible capacité, pour élaborer ces systèmes multiphasiques innovants, comme les nanocomposites polyamide/talc synthétique ou les mélanges polypropylène/polyamide (PP/PA) chargés de talc synthétique. Dans une phase polaire, le talc synthétique présente une structure constituée majoritairement de nanoparticules, mais aussi d’agrégats formés lors du malaxage. Il a été montré qu’une polarité accrue de la phase polyamide facilitait la nanodispersion du talc synthétique. Dans le cas des mélanges chargés, le talc synthétique est préférentiellement localisé dans les nodules de PA. Pour un polyamide peu polaire, l’effet émulsifiant est amplifié, ce qui s’explique par la rupture des nodules au niveau de zones peu cohésives PA/talc synthétique. D’autre part, l’augmentation de la viscosité de la matrice PP entraîne une localisation partielle du talc synthétique à l’interface matrice/nodules.Les relations structure-rhéologie des systèmes binaires et ternaires ont été discutées. / Addition of clay nanoplatelets within a thermoplastic matrix or immiscible thermoplastic blend usually improves final properties of materials. Recently, a hydrogel containing talc nanoplatelets has been obtained from hydrothermal synthesis, opening possibilities of development of talc based nanomaterials, which could rival the nanocomposites based on organically modified clay.The presence of large amount of water in the mixing chamber leads to significant losses of inorganic matter and to the presence of a few synthetic talc aggregates. These observations require the use of a twin screw mini-extruder which is appropriate to the elaboration of innovative multiphase materials, such as the synthetic talc/polyamide nanocomposites or the synthetic talc/polyamide/polypropylene blends.The structure of nanocomposites based on synthetic talc is mainly composed of nanometric entities, but also of a few micrometric aggregates. It has been shown that an increase of the polarity of the polyamide favours the dispersion of synthetic talc at nanometric scale.In the case of filled blends, the synthetic talc particles are preferentially located within PA nodules. The reduction of nodule size is explained by the breakup of nodules, due to numerous cohesion defects between synthetic talc nanoparticles and polyamide dispersed phase. Moreover, increasing polypropylene matrix viscosity leads to a selective localization of the synthetic talc at the nodule/matrix interface.Relationships between structure and rheology of binary and ternary systems have been discussed.

[pt] Polpas concentradas de caulim (70 por cento em peso de sólidos) são utilizadas na Indústria de Papel como cobertura. Durante a aplicação da polpa são alcançadas altas taxas de cisalhamento que, associadas às altas viscosidades de polpas concentradas, provocam dificuldades no movimento da lâmina acarretando riscos, bolhas ou até mesmo rasgando o papel. Para minimizar esse efeito, são adicionados agentes dispersantes que causam a redução nos valores da viscosidade. O presente trabalho teve por finalidade caracterizar fisicamente as amostras beneficiadas do caulim da região do Prado/BA e, utilizando-se da Teoria DLVO Clássica, contribuir para uma melhor compreensão sobre os aspectos fundamentais que norteiam a reologia de polpas concentradas de caulim a partir dos fundamentos de química coloidal. A caracterização tecnológica das amostras foi efetuada pelas técnicas de difratometria de raios-X, medição de alvura, análise granulométrica e análise morfológica (MEV e MET). As amostras analisadas revelaram ser, predominantemente, de caulinita; de alvura adequada a Indústria Papeleira para as amostras Coat 90 e Coat 87; de distribuição bimodal para as amostras Coat 90 e Coat 87 e de formato lamelar de perfil irregular, tendo sido encontradas partículas de perfil hexagonal. Posteriormente, foram apreciadas as propriedades eletrocinéticas das amostras. Foi medido o potencial zeta das amostras na ausência e na presença do dispersante hexametafosfato de sódio em função do pH e da concentração do dispersante em força iônica constante. A análise dos resultados revelou: que as amostras Coat 90 e Coat 87 possuem ponto isoelétrico de 3,5, enquanto o caulim Extra Fino possui ponto isoelétrico de 3,2; que o potencial zeta assume valores mais negativos com o aumento do pH e com o aumento da concentração de dispersante e que o ânion proveniente da dissociação do hexametafosfato de sódio adsorve quimicamente à superfície das amostras. De posse dos dados de potencial zeta, utilizou-se a Teoria DLVO para calcular a energia de interação total existente entre as partículas da amostra de caulim Coat 90. Após a análise dos resultados, ficou evidente que, para valores de pH maiores ou iguais a 5, existe uma barreira energética que aumenta de intensidade com o aumento do pH e da concentração de dispersante onde, com o aumento do pH, houve uma tendência à estabilização e, no caso da concentração, foi obtido um valor máximo em 3kg/t de dispersante (após este valor não houve aumento da barreira energética). Nessas curvas também foi observada a existência de um mínimo secundário que, embora pequeno (~2kT), revelou a possibilidade de adesão reversível. A análise reológica da amostra de caulim Coat 90 foi realizada a partir da comparação das viscosidades aparentes encontradas para taxas de cisalhamento constantes (100s-1 ou 1000s-1) e após a mesma quantidade de tempo (280s) já que as polpas eram tixotrópicas. Esses dados revelaram que, com o aumento do pH, as polpas atingem valores mínimos de viscosidade, onde para cada concentração de dispersante foi atingido um patamar diferente. O menor valor de viscosidade foi obtido em 3kg/t de dispersante, onde a partir desse ponto a viscosidade passou a aumentar levemente. Finalmente, os dados reológicos foram confrontados com os calculados pela Teoria DLVO Clássica, onde foi revelada a existência de um valor de energia de interação total (~400kT) comum a todas as concentrações de dispersante que indica um grau máximo de estabilidade da polpa, isto é, as partículas encontram-se tão afastadas umas das outras que o movimento relativo entre elas durante o cisalhamento não é mais afetado. Concluiu-se que, apesar das interações interparticulares serem de natureza microscópica e a viscosidade uma medida macroscópica, a Teoria DLVO foi capaz de explicar qualitativamente o compor / [en] Concentrated kaolin pulps (70 percent solid weight) are utilized as coatings in the paper industry. During the pulp application, high shear rates are obtained. Shear rates associated to high viscosities incite difficulties on the lamina movement generating risks, bubbles or even tearing the paper. To minimize this effect dispersant agents are added. The aim of the present work was to characterize physically the Prado region (Bahia State) kaolin particles and, using the DLVO theory, contribute for a better understanding about the fundamental aspects of kaolin concentrated pulps rheology via colloidal chemistry. Sample technological characterization was carried out with X-Ray difractometry technique, brightness measurements, size measurements, particle distribution and morphological analysis (MEV and MET). Samples were, predominantly, of kaolinite; samples Coat 90 and Coat 87 were of adequate brightness for Paper Industry and presented two peaks and all samples were platelike with irregular edges where hexagonal particles were found. Sample electrokinetic properties were appreciated afterwards. Zeta potential measurements were taken in the absence and in the presence of sodium hexametaphosphate dispersing agent as a function of pH and dispersant concentration using constant ionic strength. Result analysis revealed: Coat 90 and Coat 87 samples had 3,5 of isoelectric point, while Extra Fino sample had 3,2 of isoelectric point; zeta potential values became more negative with pH and dispersant concentration rising and the anion of sodium hexametaphosphate dissolution chemically adsorbs to sample surfaces. Zeta potential values were used to obtain total interparticular interaction energy by means of DLVO theory. These procedures were done only for Coat 90 sample. It became clear that, for pH values equals or higher than 5, exists an energy barrier whose values rise as pH and dispersant concentration increases. It has been noted that for pH increase, the tendency of energy barrier was to stabilize. In the case of dispersant concentration increase, was obtained maximum value (3kg/t). By data observation, a secondary minimum was noted. However the secondary minimum is low (~2kT), it revealed the possibility of reversible adhesion. The rheological analysis of Coat 90 kaolin sample was realized by comparison of apparent viscosities found on constant shear rates (100s-1 and 1000s- 1) and after the same time (280s) because pulps were thixotropic. The investigation showed that with pH increase, pulps had their viscosity reduced. Each dispersant concentration value generates a different viscosity, where the smallest viscosities were achieved in 3kg/t. Within the increase of dispersant concentration, viscosity had its values improved. Finally, rheological data were coordinated with those calculated from DLVO theory. It was observed that there is a total interaction energy value (~400kT) common to all dispersant concentrations. This point means that pulp stability is maximum, in other words, particles are so far from each other that relative movement among them is not affected anymore. The conclusion of this work is that, however interparticular forces are of microscopic nature and viscosity is a macroscopic measure, DLVO theory was able to explain qualitatively the rheological behavior of kaolin concentrated pulps.