La cristallisation des protéines est une étape cruciale dans l’élucidation de la structure tridimensionnelle des protéines. C’est un processus très délicat qui dépend de nombreuses variables environnementales dont le contrôle précis est difficile, voire impossible, dans les installations typiquement utilisées. Des approches basées sur la parallélisation massive des expériences et la réduction du volume d’échantillon par expérience, pour trouver les conditions initiales de cristallisation, montrent leur relative efficacité dans la recherche de ces conditions initiales mais aussi leurs limites quant à l’optimisation des cristaux obtenus.La méthode présentée dans cette thèse diffère de ce paradigme. La démarche proposée est une approche séquentielle plutôt que la parallélisation d’expériences et est basée sur la connaissance des diagrammes de phase. Cette thèse repose sur une suite de développements d’instruments mis en oeuvre pour maîtriser et rationaliser la cristallisation par la méthode de la dialyse, permettant ainsi l’exploration des diagrammes de phases sans consommer l’échantillon de protéine.Il en résulte un dispositif microfluidique permettant la cristallisation de protéines par la méthode de la dialyse, l’utilisation d’un flux continu d’agent de cristallisation et par conséquent l’échange continu de conditions de cristallisation ainsi que le contrôle de la température au cours de l’expérience. Il est compatible avec le rayonnement X pour la collecte de données de diffraction in situ sur les cristaux ayant poussés dans la puce microfluidique. Ce système microfluidique est basé sur la miniaturisation du banc de cristallisation qui a été amélioré d’un point de vue : de l’électronique pour l’automatisation, du transport de fluides pour le fonctionnement en flux continu, du développement logiciel pour le contrôle des paramètres de cristallisation, de la mécanique pour améliorer la cellule de dialyse et la thermorégulation, et enfin par l’intégration d’un système UV permettant de réaliser des mesures d’absorbance in situ qui offre pour l’avenir la possibilité de mesurer la solubilité de protéine au cours d’une expérience de cristallisation par la dialyse.Finalement les développements instrumentaux et méthodologiques ont été validés par la cristallisation de plusieurs protéines modèles dont les cristaux ont diffractés aux rayons X avec succès. En outre le transport d’espèces en solution par la dialyse a été étudié par une approche combinée expérimentale et théorique. / Protein crystallization is a key step in elucidating three-dimensional structure of proteins. This very sensitive process depends on many variables that are difficult to control precisely or simultaneously in the existing facilities. Instrumentation developments have concentrated on massive parallel experiments and sample volume reduction used by experiment. With this approach it is relatively easy to find initial crystallization conditions but their optimization to yield well diffracting crystals often proves to be more difficult.The method presented herein differs from the current paradigm, since we propose serial instead of parallel experiments based on the knowledge of phase diagrams. This project is based on a series of developments of instruments used to control and rationalize crystallisation using dialysis method, thus allowing phase diagrams exploration without consuming large quantity of protein sample.This results in a microfluidic device that allows crystallization of proteins by dialysis method, use of a continuous flow of crystallization agent and therefore continuous exchange of crystallization conditions as well as temperature control during experiment. It provides X-rays compatibility for in situ diffraction data collection of crystals grown in the microfluidic chip. This microfluidic system is based on the miniaturization of the crystallization bench which has been improved on electronics for automation, fluid transport to operate at a continuous flow, software development for the control of crystallization parameters, mechanics to improve both dialysis cell and thermoregulation, and finally by the integration of a UV system to perform in situ absorbance measurements that provide the future possibility to measure the solubility of proteins in a dialysis crystallization experiment.Finally both instrumental and methodological developments have been validated by the crystallization of several model proteins whose crystals diffracted succesfully X-rays. Furthermore understanding of the transport of species in solution by dialysis was investigated by combined experimental and theoretical approaches.

La neige est un réacteur photochimique multiphasique complexe capable d'échanger de nombreuses espèces réactives avec l'atmosphère. Les émissions du manteau neigeux peuvent donc influencer de manière considérable la composition et la réactivité des atmosphères polaires. Parmi les composés réactifs ainsi échangés se trouvent de nombreux composés carbonylés dont les aldéhydes qui font l'objet de cette étude. La photolyse des aldéhydes est une source importante de radicaux HOx dans l'atmosphère arctique et l'étude des échanges de ces gaz est donc indispensable à la compréhension de la capacité oxydante de l'atmosphère des régions polaires. Le formaldéhyde (HCHO) est l'aldéhyde le plus abondant dans l'atmosphère et a fait l'objet d'une part importante de notre étude. Une étude expérimentale de la solubilité et de la diffusion de HCHO dans la glace nous a permis de confirmer que ce composé formait une solution solide dans la glace. Les résultats que nous avons obtenus, associés à une analyse critique des données bibliographiques sur la solubilité de HCHO en phase aqueuse, ont été utilisés pour construire le diagramme de phase pression partielle – température du système H2O-HCHO. Afin de comprendre les mécanismes mis en jeu dans les échanges d'aldéhydes entre la neige et le manteau neigeux, nous avons suivi les concentrations en aldéhydes dans le manteau neigeux à Barrow, un site côtier à l'extrême Nord de l'Alaska, lors de la campagne polaire OASIS 2009. Nos mesures ont été associées à celles de la microphysique de la neige et à la mesure en phase gazeuse du formaldéhyde. En développant un modèle numérique de diffusion de HCHO dans les cristaux de neige, nous avons montré que l'évolution des concentrations dans la neige pouvait être reproduite de manière quantitative par l'équilibre thermodynamique et la cinétique de diffusion du formaldéhyde dans la glace. Ce travail ne se limite pas à la seule étude de HCHO. D'autres aldéhydes présents dans l'atmosphère peuvent potentiellement jouer un rôle important dans la chimie atmosphérique. Nous avons donc réalisé des améliorations à notre méthode analytique afin de permettre la mesure non seulement du formaldéhyde et de l'acétaldéhyde, mais aussi du glyoxal, du méthylglyoxal et de l'hydroxyacétaldéhyde. Cette méthode analytique déployée pendant la campagne OASIS nous a ainsi permis de mesurer simultanément et pour la première fois, l'ensemble de ces aldéhydes dans la neige polaire. Ces mesures ont révélé des concentrations importantes de glyoxal et de méthylglyoxal dans la neige, ces composés étant probablement présents dans les aérosols organiques piégés par la neige. Le rôle de cette matière organique particulaire dans la chimie de la neige nous semble mériter de plus amples études, afin de mieux caractériser sa structure et sa réactivité. / Snow is a complex multiphase chemical reactor that exchanges many reactive species with the atmosphere. One consequence of such emissions is that snow dramatically impacts the composition and the reactivity of polar the atmosphere. Carbonyl compounds, including aldehydes, are some of theses noteworthy species emitted by the snowpack. Here, we focus on aldehydes whose photolysis can yield significant amounts of HOx radicals. The knowledge of processes involved in air-snow exchanges is therefore required to understand how snow impacts the oxidative capacity of polar atmospheres. A major part of our work is focused on formaldehyde (HCHO), the most abundant aldehyde in the atmosphere. We first performed an experimental study to measure both the solubility and the diffusivity of HCHO in ice. Our results confirm that the formation of a solid-solution is the process of incorporation of HCHO into snow. We also performed a reanalysis of existing data on the solubility of HCHO in liquid water solutions. Our work made it possible to construct the partial pressure – temperature phase diagram for the H2O-HCHO system. To investigate the processes involved in air-snow exchanges of aldehydes, we monitored their concentration in snow during the OASIS 2009 field campaign which took place at Barrow, Alaska. Our measures were complemented by the monitoring of the snow physical properties and by the measurement of formaldehyde concentration in the gas phase. We developed a numerical code to model HCHO diffusion in and out of ice crystals and showed that it was possible to quantitatively reproduce snow concentrations by considering the equilibration of the H2O-HCHO by solid solution in ice. Our work also focused on other aldehydes that can potentially impact the atmospheric oxidative capacity. Improvements to our analytical method made it possible to measure not only formaldehyde and acetaldehyde, but also glyoxal, methylglyoxal and hydroxyacetaldehyde. This method deployed during the OASIS campaign provided the first measurements of all these aldehydes in polar snow. We found significant amounts of glyoxal and methylglyoxal in snow and we hypothesize that such compounds are located in the particulate organic matter scavenged by snow. The reactivity and structural composition of this organic matter must be further investigated to understand quantitatively the exchanges of these other aldehydes between the snow and the atmosphere.

No description available.

No description available.

Fundando-se em argumentos de equilíbrio e estabilidade termodinâmica, a solubilidade de soluto em um solvente é definida pelo limiar de existência da mistura como sistema homogêneo. De fato, entende-se tal grandeza sobre a coexistência de fases. A noção de solubilidade apresentada é desenvolvida através de modelos simples, que incluem resultados na rede bem como a extensão do modelo de van der Waals ao caso de duas componentes. Desta extrai-se uma condição genérica à observação de mínimos na solubilidade incidente sobre a topologia de seu diagrama de fases e de evidente correspondência física. / Concerning the equilibrium and stability in thermodynamics, the solubility of solute in a solvent is defined as the threshold of the mixture existing as a homogeneous system. Indeed, such quantity meets its meaning when understood on the coexistence surface. The idea of solubility presented is developed following the discussion of simple models, including results on the lattice as well as the van der Waals model extended to include two components. From these results we extract a general condition which connects the occurence of a minimum in solubility to the topological structure of the phase diagram.

La chaleur dégagée par des dispositifs à différentes échelles (microprocesseurs, automobiles, habitations, usines) constitue une source quasi inexploitée d’énergie. La conversion de cette énergie thermique en énergie électrique est possible via l’emploi de générateurs thermoélectriques. Cependant actuellement leur usage est peu répandu du fait de leur faible rendement, de la toxicité et du coût des matériaux. L’objectif de cette thèse qui fait suite à la thèse théorique de Kinga Niedziolka (Université Montpellier, 2014) était l’obtention de nouveau matériaux thermoélectriques performants à base de ZnSb sur la base des résultats obtenus par modélisation. Conformément aux calculs réalisés et en accord avec le partenaire industriel (Hutchinson), les dopants choisis pour ZnSb ont été l’aluminium (1% en substitution du zinc), le tellure (1% en substitution de l’antimoine, le titane (1% en substitution du zinc) et l’yttrium (1% en substitution du zinc). L’objectif ici est de synthétiser et de caractériser les matériaux qui ont été calculés de type n (conduction électrique assurée par les électrons). Les matériaux ont été synthétisés par fusion, trempe et recuit et ont été obtenus sous forme de lingot polycristallin. Ils ont été caractérisés par diffraction des rayons X, microscopie électronique à balayage et microsonde de Castaing. Leur mise en forme pour les caractérisations thermoélectriques a ensuite été mise au point. C’est en utilisant la technique du Spark Plasma Sintering (SPS) que nous avons atteint des densités très intéressantes de l’ordre de 98% pour tous les matériaux contenant un dopant. Les caractérisations thermoélectriques ont été conduites en fonction de la température jusqu’à 600 K au moyen d’un appareillage de type ZEM-3.Les travaux de dopage se sont concentrés sur les matériaux dopés au tellure présentant les résultats les plus prometteurs. Cependant les caractérisations thermoélectriques des matériaux synthétisés n’ont pas permis de mettre en évidence le passage au type n souhaité. Cependant, une avancée a été réalisée en ce qui concerne ZnSb non dopé qui est naturellement un matériau de type p (conduction électrique assurée par les trous): grâce à la mise en œuvre de la mécanosynthèse nous avons obtenu un matériau présentant de meilleurs propriétés que ceux existants. Des analyses de ce matériau au microscope électronique en transmission ainsi que des analyses par la méthode XPS ont conduit à l’identification de certains précipités responsables de l’amélioration de ses propriétés thermoélectriques. De plus, le matériau ZnSb obtenu, présente une grande stabilité en température ce qui renforce d’autant plus son intérêt d’utilisation par rapport à d’autres matériaux dans la même gamme de température. / The heat loss from different sources at different scales (microprocessors, cars, houses, factories) is an almost untapped source of energy. The conversion of this thermal energy into electrical energy is possible through the use of thermoelectric generators. However at the present time their use is not widespread due to their low efficiency, toxicity and because of the cost of the existing materials. The aim of this thesis that follows the DFT calculations previously made during Kinga Niedziolka’s thesis, (University of Montpellier, 2014) was to obtain new efficient thermoelectric materials based on ZnSb relying on the calculation results. Based on the calculations and in agreement with the industrial partner Hutchinson, aluminum (1 % at. in substitution of Zn), tellurium (1 % at. in substitution of Sb), titanium (1 % at. in substitution of Zn) and yttrium (1 % at. in substitution of Zn) were selected as dopants. The objective, was to synthesize and characterize the materials which were calculated as n type semiconductors (electron conduction). The materials were synthesized by melting, quenching and annealing. They were obtained in the form of polycrystalline ingots. They were characterized by X-ray diffraction, scanning electron microscopy and EDS and WDS microanalysis. Their shaping for thermoelectric characterization was then developed. The technique of Spark Plasma Sintering (SPS) was chosen to this aim and we achieved very interesting densities of around 98% for all doped materials. Thermoelectric characterizations were carried out up to 600 K by means of a ZEM-3 type apparatus. Only tellurium doped materials with the most promising results were thoroughly studied. However, the thermoelectric characterization of the synthesized doped materials didn’t show the expected transition to an n-type conduction. However, progress was made regarding ZnSb as a p-type material. The optimization of the mechanical alloying process led to a very promising ZnSb material. The analyses of this material by transmission electron microscopy and XPS analysis led to the identification of some precipitates which give rise to improved thermoelectric properties compared to existing p-type ZnSb. Furthermore the obtained material has a high thermal stability reinforcing its interest compared to other materials for applications in the same temperature range.

Fundando-se em argumentos de equilíbrio e estabilidade termodinâmica, a solubilidade de soluto em um solvente é definida pelo limiar de existência da mistura como sistema homogêneo. De fato, entende-se tal grandeza sobre a coexistência de fases. A noção de solubilidade apresentada é desenvolvida através de modelos simples, que incluem resultados na rede bem como a extensão do modelo de van der Waals ao caso de duas componentes. Desta extrai-se uma condição genérica à observação de mínimos na solubilidade incidente sobre a topologia de seu diagrama de fases e de evidente correspondência física. / Concerning the equilibrium and stability in thermodynamics, the solubility of solute in a solvent is defined as the threshold of the mixture existing as a homogeneous system. Indeed, such quantity meets its meaning when understood on the coexistence surface. The idea of solubility presented is developed following the discussion of simple models, including results on the lattice as well as the van der Waals model extended to include two components. From these results we extract a general condition which connects the occurence of a minimum in solubility to the topological structure of the phase diagram.

L'absorption d'hydrogène dans les sites interstitiels des métaux se situe au cœur de problématiques majeures comme la fragilisation des alliages ou le stockage de l'hydrogène à des fins énergétiques. En effet, ce phénomène modifie les propriétés physico-chimiques du métal hôte et peut conduire à la formation de composés ordonnés MHy appelés hydrures. Dans ce cadre, la méthode de modélisation Calphad (CALculation of PHAse Diagrams) constitue un outil pertinent pour comprendre et prédire le comportement des métaux et alliages en présence d'hydrogène. Toutefois, il n'existe pas de base de données Calphad centrée sur l'hydrogène, permettant de calculer des équilibres de phases dans des systèmes multi-constituants (ternaires, quaternaires…).Ainsi, la présente thèse est consacrée à l'étude de systèmes binaires métal-hydrogène (M-H), par une approche modélisatrice multi-échelle. Ces systèmes binaires représentent la première étape à la conception d'une telle base de données Calphad. Tout d’abord, des calculs DFT (Density Functional Theory) systématiques ont été réalisés pour 31 systèmes binaires M-H considérant 30 structures cristallographiques potentielles, soit 30 × 31 = 930 hydrures, stables ou métastables. Cette approche de criblage a permis notamment de déterminer les enthalpies de formation à 0 K, données d’entrée essentielles de la méthode Calphad. De nouveaux hydrures n'ayant jamais été observés expérimentalement ont été prédits à haute pression (TaH2, ZrH3…).Puis, des calculs de phonon dans l'approximation harmonique ont été réalisés sur les hydrures les plus stables. D’une part, ils permettent de corriger les enthalpies de formation issues de la DFT, en tenant compte de l'énergie et de l'entropie dues aux vibrations des atomes, non négligeables pour l'atome léger d'hydrogène. D’autre part, une étude à grande échelle a porté sur les modifications de l'enthalpie libre de formation résultant de la substitution de l'hydrogène par ses isotopes, dit « effet isotopique ». Des prédictions ont pu être réalisées sur la nature de cet effet en fonction de la température. De plus, pour étudier l'insertion aléatoire des atomes d'hydrogène en solution solide, nous avons utilisé des méthodes de thermodynamique statistique : CVM (Cluster Variation Method) et simulation de Monte-Carlo. Ces méthodes ont été implémentées dans des codes de calcul, appliqués aux métaux cubiques faces centrées (cfc) et cubiques centrés (cc). Les données d'entrée de ces codes sont les énergies associées aux interactions locales entre atomes plus proches voisins. Elles sont fournies par la CEM (Cluster Expansion Method) couplée à la DFT. Une étude comparée des systèmes Ni-H et Pd-H a mis en évidence les spécificités des comportements thermodynamiques des solutions solides interstitielles de chacun de ces systèmes. Par ailleurs, la pression en dihydrogène constitue un paramètre important car de nombreux hydrures ne se forment qu'à très haute pression. Afin d'améliorer la précision des modèles Calphad à pression élevée, le modèle de Lu et al. a été appliqué aux phases condensées des systèmes Ni-H, Rh-H et Mg-H. Ce modèle permet de déterminer la contribution à l'enthalpie libre du travail des forces de pression, en admettant en entrée aussi bien des calculs de phonon quasi-harmoniques que des données expérimentales. Enfin, la modélisation Calphad complète du système Ni-H a été réalisée en intégrant le modèle de Lu et al. L'enthalpie de formation calculée par DFT et l'enthalpie de mélange déterminée par CEM ont également été utilisées en données d'entrée, en complément des données expérimentales disponibles / Hydrogen absorption in the interstitial sites of metals is crucial for major issues such as alloy embrittlement or hydrogen storage for energy applications. This phenomenon modifies the physicochemical properties of the host metal and may lead to the formation of ordered MHy compounds called hydrides. Within this framework, the Calphad modeling method (CALculation of PHAse Diagrams) is a relevant tool for understanding and predicting the behavior of metals and alloys in the presence of hydrogen. However, there is no Calphad database centered on hydrogen for calculating phase equilibria in multi-constituent systems (ternary, quaternary…).The present thesis proposes to use a multi-scale modeling approach to study metal-hydrogen (M-H) binary systems, which are the first step in designing such a Calphad database. First, systematic DFT (Density Functional Theory) calculations were carried out for 31 binary M-H systems considering 30 potential crystal structures, resulting in 30 × 31 = 930 hydrides, stable or metastable. This high throughput approach allowed in particular to determine the enthalpies of formation at 0 K, which represent important input data for the Calphad method. New hydrides that have never been experimentally observed could be predicted at high pressure (TaH2, ZrH3 ...).Then, phonon calculations in the harmonic approximation were performed on the most stable hydrides. They allow, on the one hand, to correct the DFT calculated enthalpies of formation by considering the energy and entropy due to the atom vibrations, which are not negligible for the light hydrogen atom. On the other hand, a large-scale study focused on the modification of the free energy of formation due to hydrogen substitution by its isotopes, known as "isotopic effect". Predictions were made on the nature of this effect as function of temperature. Moreover, the random insertion of hydrogen atoms in solid solution was studied using statistical thermodynamic methods: CVM (Cluster Variation Method) and Monte-Carlo simulation. These methods have been implemented in calculation codes, applied to face centered cubic (fcc) and body-centered cubic (bcc) metals. The input data are the interactions energies between nearest neighbor atoms. They are provided by the CEM (Cluster Expansion Method) coupled with DFT calculations. A comparative study of the Ni-H and Pd-H systems revealed the specificities of the thermodynamic behaviors of both solid interstitial solutions. Furthermore, dihydrogen pressure is an important parameter because many hydrides form only at very high pressure. To improve the Calphad model accuracy at high pressure, the model of Lu et al. was applied to the condensed phases of the Ni-H, Rh-H and Mg-H systems. This model allows to determine the contribution to the free enthalpy due to the pressure force work. The input data may be both quasi-harmonic phonon calculation results and experimental data. Finally, a comprehensive Calphad model of the Ni-H system was carried out by integrating the model of Lu et al. The DFT enthalpy of formation and the mixing enthalpy determined by CEM were used as input data, to complement the available experimental data

A partir de medidas isotérmicas e isobáricas de susceptibilidade magnética diferencial, determinamos o diagrama de fases magnéticas do metamagneto Ni(NO3)2.2H2O, para pressões hidrostáticas de até cerca de 11 kbar. Determinamos o campo de transição à temperatura nula Hc(0,P) e à temperatura de Néel, ambos apresentando um crescimento aproximadamente linear com P. A partir de efeitos de desmagnetização, pudemos determinar o ponto de união das linhas de primeira e segunda ordem de cada diagrama, e para P=0 pudemos caracterizá-lo com um ponto tricrítico, com temperatura tricrítica Tt=3.70K. Observamos ainda a existência de uma transição metamagnética induzida por pressão para valores do intervalo de 0 a 8 kbar. Usando a Aproximação de Campo Médio num modelo elástico aproximado para metamagnetos com S=1, pudemos obter expressões para TN(P) e Hc(0,P), ambas apresentando uma dependência linear em P, em bom acordo com os dados experimentais. Usando este modelo obtivemos também uma expressão analítica para Tt(P). Uma vez que TN(P), Hc(0,P) e Tt(P) dependem dos parâmetros de \"exchange\", pudemos determinar a dependência de tais parâmetros com a pressão. Também através deste modelo simples, pudemos compreender que a transição metamagnética induzida por pressão está provavelmente associada a uma forte dependência do campo cristalino com a pressão / From isothermical and isobaric differential magnetic susceptibility measurements we have determined the magnetic phase diagram of the metamagnetic system Ni(NO3)2.2H2O, for several hydrostatic pressures up to 11 kbar. The zero temperature transition field Hc(0,P) and the Néel temperature TN(P) were determined, both increasing almost linearly with P. From demagnetizing effects we were able to determine the joint point of the first and second order lines of each diagram, which for P=0 is a tricritical point, with tricritical temperature Tt=3.70K. We also observed a pressure induced metamagnetic transition in the range from 0 to 8 kbar. Using Mean Field Theory in an approximated elastic S=1 metamagnetic model, we could obtain expressions for TN(P) and Hc(0,P), both showing linear dependence on P, in agreement with experimental data. We could also obtain, using this model, an analytical expression for Tt(P). Once TN(P), Hc(0,P) and Tt(P) are dependent on the exchange parameters, we cou1d obtain the pressure dependence of these parameters. From this simple model we could understand that the pressure induced metamagnetic transition is probably associated with a strong variation of the crystalline field with pressure

Orientadores: Everson Alves Miranda, Pedro de Alcântara Pessôa Filho / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-24T01:51:47Z (GMT). No. of bitstreams: 1 Hirata_GiseleAtsukoMedeiros_D.pdf: 4140620 bytes, checksum: 01025f5b18791e6b6698c34eea1dcda4 (MD5) Previous issue date: 2013 / Resumo: O segundo coeficiente virial osmótico, B22, tem sido considerado um preditor para o processo de cristalização. Uma faixa relativamente estreita de valores negativos de B22, -1x10-4 a -8x10-4 mL.mol/g2, é ideal à formação de cristais de acordo com George e Wilson (1994). Essa faixa de valores de B22 é denominada de "janela de cristalização", sendo utilizada para classificar condições adequadas de solvente à formação de cristais. Para valores maiores que -1x10-4 mL.mol/g2, a interação proteína-proteína não é suficientemente forte para a cristalização e nenhuma fase sólida é formada, enquanto para valores menores que -8x10-4 mL.mol/g2, as interações proteína-proteína são muito intensas e precipitados amorfos são formados. Dessa forma, os valores de B22 se tornaram um critério de seleção para a cristalização de proteínas, uma vez que esse coeficiente pode ser determinado por diversos métodos. Este trabalho teve como objetivo determinar experimentalmente diagramas de fases de proteínas (lisozima, insulina suína e bovina) e identificar nesses diagramas, através de análise dos valores do B22, as condições nas quais ocorre a formação de precipitado amorfo, cristalino ou outras fases (por exemplo, fase líquida). O "salting-out" foi o método escolhido para precipitar as proteínas, pois é considerado um dos mais simples e importantes métodos para induzir a cristalização. O sal volátil carbamato de amônio foi o agente de "salting-out" escolhido. As técnicas de espalhamento de luz estático (SLS) e cromatografia de auto-interação (SIC) foram usadas para determinar os valores de B22 para as proteínas em diferentes soluções aquosas de sal a 15 e 25 °C. O fenômeno de "salting-out" foi observado nos diagramas para as três proteínas estudadas. Valores negativos de B22 e altos valores da constante de "salting-out" - entre 1,07 a 3,77 kg/mol - confirmaram que o sal volátil carbamato de amônio empregado neste estudo é um bom agente precipitante. Os valores do B22 para a insulina suína (-250x10-4 a -18x10-4 mol.ml/g2 a 25 °C e -187x10-4 a -45,2x10-4 mol.ml/g2 a 15 °C) e insulina bovina (-999x10-4 a 6,7x10-4 mol.ml/g2 a 25 °C e -533x10-4 a -16,7x10-4 mol.ml/g2 a 15 °C) indicaram a precipitação, o que também foi confirmado pelos ensaios de cristalização. Já para a lisozima, obteve-se formação de cristais independente do valor de B22 encontrado (-20,4x10-4 a -3,6x10-4 mol.ml/g2 a 25 °C e -400x10-4 a -14,4x10-4 mol.ml/g2 a 15 °C). Além disso, os modelos teóricos disponíveis na literatura utilizados para a obtenção de uma estimativa do parâmetro B22 são adequados e válidos para as condições em que a medida experimental não é possível, podendo ser aplicados para o sistema proteína/sal volátil. Dessa forma, este trabalho mostrou que não existe uma "janela de cristalização universal" válida para todos os sistemas e o uso do sal volátil carbamato de amônio como agente de cristalização é uma alternativa ao uso de sais convencionais. / Abstract: The osmotic second virial coefficient, B22, has been used as a predictor of crystallization. A relatively narrow range of negative B22 values, -1x10-4 to -8x10-4 mL.mol/g2, is the ideal range for crystal formation according to George and Wilson (1994). This range, referred to as the "crystallization slot", has been used to classify suitable conditions under which proteins will assemble into crystals. For B22 values greater than -1x10-4 mL.mol/g2, the protein-protein interaction is very weak and no solid phase is formed, while for values less than -8x10-4 mL.mol/g2, the protein-protein interactions are very intense and amorphous precipitates are formed. Thus, the B22 value has become a selection criterion for protein crystallization, since this coefficient can be determined by various methods. This study aimed to determine the experimental phase diagrams for proteins (lysozyme and bovine and porcine insulin) and to identify those diagram conditions under which amorphous precipitate, crystals or other phases (for example, liquid phase) are formed, based on the values of B22. The salting-out method to precipitate proteins was chosen because it is considered one of the simplest and most important methods to induce crystallization. The volatile salt ammonium carbamate was chosen as the salting-out agent. Traditional static light scattering (SLS) and the novel self-interaction chromatography (SIC) technique were used to determine B22 values for the proteins in different aqueous salt solutions at 15 and 25 °C. The salting-out phenomenon was observed in the phase diagrams for the three proteins studied. Negative B22 values and high values of the salting-out constant - between 1.07 to 3.77 kg/mole (Cohn, 1925) - confirmed that ammonium carbamate was a good precipitant agent. The B22 values for porcine (-250x10-4 to -18x10-4 mol.ml/g2 at 25 °C and -187x10-4 to -45.2 x10-4 mol.ml/g2 at 15 °C) and bovine (-999x10-4 to 6.7x10-4 mol.ml/g2 at 25 °C and -533x10-4 to -16.7x10-4 mol.ml/g2 at 15 °C) insulin indicated precipitation that was confirmed experimentally. However, lysozyme was obtained as crystals, regardless of the B22 values found (-20.4x10-4 to -3.6 x10-4 mol.ml/g2 at 25 °C and -14.4x10-4 to -400x10-4 mol.ml/g at 15 °C). In addition, thermodynamic models available in the literature and suitable for the conditions under which experimental measurements were done provided a good fit to the data. Thus, this work showed that there is no universal "crystallization slot" applicable to all systems and that for crystallization agent, volatile salt ammonium carbamate can serve as an alternative to conventional salts. / Doutorado / Desenvolvimento de Processos Biotecnologicos / Doutora em Engenharia Quimica