L'industrie du béton est une source importante d'émissions de gaz CO2 lors de la production de ciment. Une des solutions proposées à l'industrie de la construction est de remplacer le ciment par des matières premières secondaires moins polluantes (ex : filler calcaire, fumée de silice), qui peuvent également améliorer les caractéristiques rhéologiques du liant et accroître la durabilité du béton.Le but principal de cette thèse est donc de : (1) standardiser la formulation de béton issue des différents travaux de recherche publiés tout en assurant une bonne compréhension de ses propriétés rhéologiques; (2) adapter ces principes aux conditions de réalisation (prise en compte des caractéristiques des matériaux locaux utilisés); et (3) vérifier dans le laboratoire ou sur le terrain la viscosité plastique et le seuil d’écoulement des bétons.Outre les moyens expérimentaux classiques nous avons utilisé, dans notre étude, un autre moyen de caractérisation rhéologique : c’est l’écoulement de béton dans un V-Funnel couplé à un canal horizontal en plexiglas. Dans un premier temps nous cherchons à comprendre la corrélation entre la viscosité et les paramètres rhéologiques de béton en se basant sur des modèles proposés dans la littérature. À partir de là, nous cherchons à présenter le modèle qui décrit le comportement de béton. Pour ce faire, nous proposons une corrélation théorique entre la viscosité plastique du béton frais et le temps d’écoulement dans le V-Funnel, puis nous comparons cette solution théorique avec des mesures expérimentales rapportées dans la littérature. / The concrete industry is an important source of carbon dioxide gas emissions during cement production. One of the proposed solutions to the construction industry is to replace the cement by less polluting secondary raw materials (e.g. limestone filler, silica fume), which can also improve the rheological properties of binder and increase the concrete durability.Thus, the development of self-compacting concrete can be seen as an attempt to reduce the environmental impact of the construction. Similarly, the structure optimization also aims at reducing the quantity (in volume) of concrete and is reflected in the use of high performance concrete and ultra performance fiber-reinforced concrete (HPC and UPFRC).The most widely adopted approach to quantify these rheological properties is to experimentally measured the shear stress versus strain rate using a concrete rheometer.In addition to traditional experimental tests such as the V-funnel, spread, sieve stability and L box, in our study we used another test of rheological characterization: it is the flow of concrete in a V-Funnel and then in a horizontal channel Plexiglas.The main purpose of this thesis is to: (1) standardize the concrete formulation outcome of various research works published while ensuring a good understanding of its rheological properties, (2) adapt these principles to the realization conditions (taking into account characteristics of local materials used), and (3) check, in the laboratory or in the field, the plastic viscosity and the flow threshold of concrete.

Freeze-drying of protein pharmaceuticals is a procedure frequently used to obtain stability of the active pharmaceutical ingredientduring distribution and storage. It can be performed in pre-filled syringes, with a lubricous coating of silicone on the inside, to enable the piston moving. The coating changes the environment potentially affecting the features of the freeze-dried cake since the wetting behavior and adhesion to the inner wall is affected.This project aimed to investigate the effect of the siliconization of the cakes. Three different formulations were freeze-dried in nonsiliconized (NS) and siliconized vials using different siliconization protocols. Analysis was done using differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA),scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS) and an embedding method, intended to give information about the cake’s shrinkage, cracking and pore-structure. The water content in the bottom of the cakes was consistently higher than in the top, a difference decreasing over time. Increased surface hydrophobicity lead to increased shrinkage of the cake’s volume and a decrease in fogging. The bottom of the protein cake in the vial siliconized with a commercial silicone emulsion consisted of pores with regularly equal pore size and thick pore walls, a structure not seen in any other cake. All cakes in the silicone emulsion siliconized vials had lower water content than the cakes in the vials using the other siliconization method. The XPS-analysis showed that the cakes in the emulsion siliconized vials contained silicon, indicating an excess of silicone when siliconizing and/or an unstable coating. / NextBioForm

<p>Dry Powder inhalers (DPIs) have been one of the most promising developments in pulmonary drug delivery systems. In general, DPIs are more effective than systemic administrations and convenient to use. However, delivering high-dose antibiotics through a DPI is still a challenge because high powder load may need a very large inhaler or increase the incidence of local adverse effects. Spray drying has been increasingly applied to produce DPI formulations for high-dose antibiotics; nevertheless, many spray-dried particles are amorphous and physically unstable during storage, particularly under the humid environment. </p> <p> </p> <p>My research focuses on addressing critical challenges in physical stability of DPIs for spray-dried high-dose antibiotics. The effects of moisture-induced crystallization on physical stability and aerosol performance of spray-dried amorphous Ciprofloxacin DPI formulations stored at different humidity conditions were studied. Our study not only provided a mechanistic understanding in the impact of crystallization on aerosol performance but also developed novel approaches for improving stability of spray-dried formulations used in DPI.</p> <p> </p> <p>Our work has shown that recrystallization of amorphous spray-dried Ciprofloxacin led to significant changes in aerosol performance of DPIs upon storage, which cause critical quality and safety concerns. These challenges have been solved through co-spray-drying Ciprofloxacin with either excipient such as leucine or synergistic antibiotic like Colistin. Co-spray-drying Ciprofloxacin with Colistin not only improved physical and aerosol stability but also enhanced antibacterial activity which is a great advantage for treating ‘difficult to cure’ respiratory infections caused by multidrug resistant bacteria.</p> <p> </p> <p>My research work is a sincere effort to maximize the utility and efficacy of high-dose DPI, an effective delivery tool for treating severe resistant bacterial respiratory infections.</p>

<p>FDA has long served as the front safeguard to the U.S. citizen public health, is also perceived as one of the world-leading drug regulators. Despite the tremendous efforts and progress have been made to promote the public health, FDA was criticized for putting the agency’s trust icon at stake and was questioned of its ability to serve the agency’s ultimate mission to protect the public. In the wake of the arousing concerns, FDA sought the transformation the oversight model of the medicinal products. One of the actions is to launch quality metrics program. However, this program has been unanimously opposed by the industry. Instead of the current conventional approach, which is constrained by the high dependence on industry cooperation, we try to explore</p><p>the measurement of company and product quality risk with public domain data, try to help in visualizing quality and risk. To that end, we develop conceptual frameworks for both company and product quality, examine some of the factors (education, local authority intensity, historical inspection results, physiochemical, physiological, formulation factors, etc.), further developed a warning letter and product recall prediction model with machine learning method referenced to the data analysis outcome.</p>

In the last decade Process Synthesis has been an emerging active area of research. Since 1999 at COMPS, we have continuously applied this technique to many studies and recently have carefully started to consider the integration of this technique with others to simplify multilevel process design. In this regard the need for new tools that do not require very accurate data or experimental technique has become an imperative for the initial analysis of systems. This thesis presents some of these results.Process synthesis requires new rapid tools for the assessment and comparison of new technologies (possibilities) in the process development stage.This thesis examines and formulates process synthesis problems in the case of separation processes (separation with mixing , separation with reaction).Using simple generic vectorial tools, the thesis addresses the folowing issue: How the topology and mapping of the process vector field defines patterns that may be used for alternative process synthesis solution and design (structure selection ,stability analysis or controllability etc...) . The tools presented are intended for preliminary process design and feasibility studies and will allow for simple comparisons of various options.The key principle used in the study is the exploitation of the generic vectorial representation of fundamental physical phenomena (separation, mixing , reaction) in the process models as used for process synthesis and analysis.The demonstrated power of this approach is that it allows one to exploit the formal mathematical similarities between the different processing modes. Extensive topological Reactive Column Profile has been generated and analysed, offering new insights for manipulation of process behaviour in process synthesis and design. The significance of the contribution of this research is in offering extended fundamental insigths in creating process synthesis options for reactive-separations systems, creating the essential structural basis for subsequent mathematical optimisation of the performanceof process designs.

O presente trabalho trata do desenvolvimento de uma ferramenta computacional para a análise de sólidos bidimensionais contendo múltiplas fissuras utilizando uma nova estratégia mediante a combinação do método da partição (spliting method) e formulação híbrido-Trefftz de tensão. A primeira consiste em um método de decomposição para a análise de sólidos contendo múltiplas fissuras pela partição do problema original (PG) em três subproblemas: subproblema global (PG(0)), onde o sólido é analisado sem a presença de fissuras, subproblemas locais (PL(k)), onde cada fissura é analisada individualmente e inserida em domínio arbitrário, e subproblemas globais (PG(k)), onde são computados os efeitos de interação entre fissuras. A solução de PG é obtida impondo a condição de nulidade do somatório das tensões nas faces das fissuras de todos os subproblemas. Por simplicidade, os subproblemas PG(0) e PG(k) são analisados via método dos elementos finitos clássico (MEF). Em contrapartida, os PL(k) são analisados via formulação híbrido-Trefftz visando uma alta eficiência na avaliação desses subproblemas contendo fissura. Essa formulação promove a aproximação dos campos de tensão e deslocamento no domínio e contorno do elemento, respectivamente, de maneira independente. As bases aproximativas do campo de tensão são formadas por funções solução da equação de Navier adicionadas às funções analíticas da mecânica da fratura. Assim, essa formulação proporciona uma boa solução do problema utilizando malha grosseira. Além disso, é possível obter os fatores de intensidade de tensão (FIT) da fissura diretamente da solução do sistema linear do problema. Resultados numéricos são apresentados a fim de ilustrar a aplicação da estratégia e sua eficiência ao alcançar soluções precisas aliado a um baixo custo computacional na análise. / This paper presents the development of a computational framework to the analysis of two dimensional solids containing multiple cracks using a new strategy with the combination of splitting method and hybrid-Trefftz stress formulation. The first consists of a decomposition method to the analysis of solids containing multiple cracks through the split of original problem (PG) into three subproblems: global subproblem (PG(0)), where the solid is analyzed without the cracks, local subproblems (PL(k)), where each crack is analyzed individually and insert in an arbitrary domain, and global subproblems (PG(k)), where the effects of the interaction between cracks are evaluated. Solution of PG is obtained using the condition that the sum of tractions on crack faces of all subproblems are null. For simplicity, subproblems PG(0) and PG(k) are analyzed by classical finite element method (FEM). On the other hand, PL(k) are analyzed by hibrid-Trefftz formulation aiming a high efficiency on the evaluation of cracked subproblems. This formulation promote the approximation of stresses and displacements fields on domain and boundary of elements, respectively, on an independent manner. Bases of appoximation of stress field are constructed by functions that solves the Navier equation with the addition of analitical functions of mechanic of fracture. Hence, this formulaition provides good solution of the problem with coarse mesh. Moreover, the stress intensity factors (SIF) of cracks may be obtained directly from the solution of linear system of the problem. Numerical results are presented in order to ilustrate the use of the strategy and its efficiency on the evaluating of accurate solutions with a low computational cost of the analysis.

Este trabalho propõe a aplicação de dois diferentes conceitos para a resolução do Problema de Estivagem de Unidades de Celulose - PEUC, que de acordo com Ribeiro e Lorena (2008) pode ser definido como um problema que busca alocar a máxima quantidade de unidades de celulose ao porão de cargas de um dado navio, respeitando as restrições físicas de dimensões, de posicionamento, de não-sobreposição das unidades e de capacidade máxima do porão do navio. Esse tipo de problema se encaixa, no contexto da Pesquisa Operacional, na classe de Corte e Empacotamento (Cutting and Packing - C&P) e pode ser classificado, de acordo com a tipologia de Wäscher, Haußner e Schumann (2007), como sendo um Single Large Object Placement Problem (SLOPP). Em última instância, o objetivo do PEUC é definir o melhor plano de estivagem para o carregamento de unidades de celulose em um dado porão de um navio, maximizando a área ocupada pelas unidades de celulose. Trata-se de um problema NP-Completo (DOWSLAND; DOWSLAND, 1992; BISCHOFF; WÄSCHER, 1995; MALAGUTI; DURáN; TOTH, 2013) e por isso foram propostas duas abordagens para buscar a melhoria das soluções encontradas e/ou redução do tempo computacional necessário. As abordagens propostas, o Modelo Matemático Modificado e o Método Iterativo de Solução, apresentaram bons resultados para instâncias experimentais, confirmando a efetividade de suas aplicações. Os resultados foram melhores tanto na qualidade das soluções (ocupação total do objeto), como no tempo computacional necessário. Também foram avaliadas quatro instâncias reais, com a comparação dos planos de estivagem resultantes da aplicação dos modelos matemáticos com os planos reais, elaborados manualmente por especialistas. Em três dos quatro casos os resultados das abordagens aqui propostas se mostraram melhores que os planos reais. / This work proposes the application of two different concepts to tackle the Woodpulp Stowage Problem - WSP, that according to Ribeiro e Lorena (2008) can be defined as a problem that seeks the allocation of the maximum quantity of woodpulp units inside the hold of a cargo vessel, always respecting the physical constraints, positioning constraints, non-overlapping of units and also the hold capacity. This kind of problem fits, in the context of Operational Research, into the class of Cutting & Packing and can be classified, according to Wäscher, Haußner e Schumann (2007) typology, as a Single Larga Object Placement Problem (SLOPP). Ultimately the objective of the WSP is to define the best stowage plan for the loading of woodpulp units inside a given hold of a given cargo vessel, maximizing the total area occupied by the woodpulp units. As it\'s a NP-Complete problem (DOWSLAND; DOWSLAND, 1992; BISCHOFF; WÄSCHER, 1995; MALAGUTI; DURáN; TOTH, 2013) two approaches were proposed to improve the quality of the resulting solutions and/or the reduction of the computational time needed. The proposed approaches, the Modified Mathematical Model and the Iterative Solution Method, showed good results for experimental instances, confirming the effectiveness of these approaches. The results were better regarding the quality of the solutions (total occupied area of the object) and also regarding the computational time needed. Also, four real instances were evaluated, comparing the results of the mathematical models with the real stowage plans, manually created by specialists. In three of the four instances, the proposed approaches showed better results than the real stowage plans.

Este trabalho aborda a formulação geométrica das teorias clássicas de calibre, ou Yang-Mills, considerando-as como uma importante classe de modelos que deve ser incluída em qualquer tentativa de estabelecer um formalismo matemático geral para a teoria clássica dos campos. Tal formulação deve vir em (pelo menos) duas variantes: a versão hamiltoniana, que passou por uma fase de desenvolvimento rápido durante os últimos 10-15 anos, levando ao que hoje é conhecido como o ``formalismo multissimplético\'\', e a mais tradicional versão lagrangiana utilizada nesta tese. O motivo principal justificando tal investigação é que teorias de calibre constituem os mais importantes exemplos de sistemas dinâmicos que são altamente relevantes na Física e onde a equivalência entre a versão lagrangiana e a versão hamiltoniana, que no caso de sistemas não-singulares é estabelecida pela transformação de Legendre, deixa de ser óbvia, pois teorias de calibre são sistemas degenerados do ponto de vista lagrangiano e são sistemas vinculados do ponto de vista hamiltoniano. Esta propriedade característica das teorias de calibre é uma consequência direta do seu alto grau de simetria, isto é, da sua invariância de calibre. No entanto, numa formulação plenamente geométrica da teoria clássica dos campos, capaz de incorporar situações topologicamente não-triviais, a invariância sob transformações de calibre locais (transformações de calibre de segunda espécie) e, surpreendentemente, até mesmo a invariância sob as transformações de simetria globais correspondentes (transformações de calibre de primeira espécie) não podem ser adequadamente descritas em termos de grupos de Lie e suas ações em variedades, mas requerem a introdução e o uso sistemático de um novo conceito, a saber, fibrados de grupos de Lie e suas ações em fibrados (sobre a mesma variedade base). A meta principal da presente tese é tomar os primeiros passos no desenvolvimento de ferramentas matemáticas adequadas para lidar com este novo conceito de simetria e, como uma primeira aplicação, dar uma definição clara e simples do procedimento de ``acoplamento mínimo\'\' e uma demonstração simples do teorema de Utiyama, segundo o qual lagrangianas para potenciais de calibre (conexões) de primeira ordem (i.e., que dependem apenas dos próprios potenciais de calibre e de suas derivadas parciais até primeira ordem) que são invariantes sob transformações de calibre são necessariamente funções dos campos de calibre (i.e., do tensor de curvatura) invariantes sob as transformações de simetria globais correspondentes. / This thesis deals with the geometric formulation of classical gauge theories, or Yang-Mills theories, regarded as an important class of models that must be included in any attempt to establish a general mathematical framework for classical field theory. Such a formulation must come in (at least) two variants: the hamiltonian version which has gone through a phase of rapid development during the last 10-15 years, leading to what is now known as the ``multisymplectic formalism\'\', and the more traditional lagrangian version studied in this thesis. The main motivation justifying this kind of investigation is that gauge theories constitute the most important examples of dynamical systems that are highly relevant in physics and where the equivalence between the lagrangian and the hamiltonian version, which for non-singular systems is established through the Legendre trans% formation, is far from obvious, since gauge theories are degenerate systems from the lagrangian point of view and are constrained systems from the hamiltonian point of view. This characteristic property of gauge theories is a direct consequence of their high degree of symmetry, that is, of gauge invariance. However, in a fully geometric formulation of classical field theory, capable of incorporating topologically non-trivial situations, invariance under local gauge transformations (gauge transformations of the second kind) and, surprisingly, even invariance under the corresponding global symmetry transformations (gauge transformations of the first kind) cannot be described adequately in terms of Lie groups and their actions on manifolds but requires the introduction and systematic use of a new concept, namely Lie group bundles and their actions on fiber bundles (over the same base manifold). The main goal of the present thesis is to take the first steps in developing adequate mathematical tools for handling this new concept of symmetry and, as a first application, give a simple clear-cut definition for the prescription of ``minimal coupling\'\' and a simple proof of Utiyama´s theorem, according to which lagrangians for gauge potentials (connections) that are gauge invariant and of first order, i.e., dependent only on the gauge potentials themselves and on their partial derivatives up to first order, are necessarily functions of the gauge field strengths (i.e., the curvature tensor) invariant under the corresponding global symmetry transformations.

Uma forma de reduzir a difusão de agentes agressivos para o interior do concreto é através da aplicação de sistemas de pintura, que funcionam como uma barreira ao ingresso destes agentes. A difusão dos agentes através de películas de pintura está relacionada com a sua porosidade, que, por sua vez, depende principalmente do teor e estrutura química da resina e do teor e morfologia dos pigmentos. Esta tese teve por objetivo principal demonstrar que, com uma formulação adequada, as tintas à base de resina acrílica podem ser utilizadas como barreira de proteção eficiente contra a penetração de água e vapor de água, anidrido carbônico (C\'O IND.2\') e íons cloreto, em estruturas de concreto ou em outros tipos de substratos. As formulações das tintas foram definidas em função do tipo de acabamento final e obtidas pela variação da relação pigmento/resina. Foram realizados ensaios de caracterização nas tintas líquidas e nas resinas utilizadas na preparação das tintas estudadas, bem como ensaios de desempenho nas películas livres e em concretos pintados, que apresentavam resistência à compressão de 20 e 40 Mpa. Para a medida da resistência à penetração dos agentes agressivos no concreto, foram desenvolvidos ensaios acelerados, com base em metodologias de avaliação de argamassas e concretos, adaptados a substratos pintados. A microestrutura das películas de tinta foi correlacionada com as propriedades físicas das películas e dos concretos pintados. Os resultados obtidos mostram a importância da formulação das tintas nos fenômenos de transporte e, além disso, mostram que tintas acrílicas e as estirenoacrílicas com diferentes temperaturas vítreas (Tg) apresentam comportamento físico distinto. Tintas de acabamento semi-brilho conferem boa proteção contra a penetração de agentes agressivos no concreto, enquanto que as de acabamento fosco apresentam função mais decorativa do que de proteção. / The diffusion of aggressive agents into concrete can be reduced by the application of painting systems which work as barriers to these agents. The capacity of these agents to pass through the paint is related to its porosity, which in turn is mainly dependent on the structure and content of the resin and on the morphology and content of the pigments. The main objective of this thesis is to demonstrate that with adequate formulation, paints based on acrylic resin polymers can be used as efficient barriers against the penetration of water and water vapour, carbon dioxide and chloride in concrete structures or any other porous substrate. Paint formulations were defined based on the final appearance (gloss) and obtained by varying the pigmentbinder ratio. Characterization tests on the liquid paints and polymers prepared for this study were done, as were performance tests on free dry films and painted concrete substrates, with compressive strength of 20 MPa and 40 MPa. In order to determine the effectiveness of paint in inhibiting the diffusion of aggressive agents into concrete, accelerated tests based on methodologies for the evaluation of mortars and concrete were developed. The films microstructures were correlated with their physical properties and with painted concrete. The results showed that the formulation of the paints has a important effect on the transport mechanism and also that acrylic and styrene-acrylic paints with different transition temperatures (Tg) do not present the same physical behavior. Glossy paints presented a high level of protection against the penetration of aggressive agents whereas matte finishes offered a more aesthetic function rather than protection.

A utilização de bioinseticida à base de Bacillus thuringiensis tem se mostrado eficaz no combate às larvas de insetos veiculadores de doenças como a dengue, elefantíase e malária. O bioinseticida apresenta alta especificidade, ausência de resistência nos insetos alvos e baixo impacto ambiental comparado aos inseticidas químicos. Porém, a difusão do uso de bioinseticida apresenta alguns desafios como o custo da produção e da formulação. O levedo excedente da indústria cervejeira pode tornar a produção do bioinseticida economicamente viável devido ao seu baixo custo, assim como o glicerol, resíduo da produção de biodiesel. A escolha adequada dos componentes do meio de cultivo é essencial para o sucesso de um produto e deve ser realizada buscando-se obter uma alta produtividade com o menor custo. Além disso, no caso do bioinseticida, os componentes do meio exercem uma grande influência no crescimento e na síntese de toxinas pela bactéria. No presente trabalho, propôs-se formular um meio de cultivo para Bacillus thuringiensis var. israelensis utilizando levedo de cerveja e glicerol proveniente da fabricação de biodiesel, visando obter um meio fermentado com elevada toxicidade e de menor custo, comparado aos obtidos atualmente. Inicialmente foi definido um procedimento de preparo do extrato de levedo de cerveja. Constatou-se que o melhor método consiste da esterilização da suspensão de levedo de cerveja (121 ºC por 20 minutos) seguida de filtração qualitativa. Foi definida também a faixa de concentração de extrato de levedo de cerveja que seria utilizada no estudo de composição do meio (10 g/L a 30 g/L). Os ensaios foram realizados em frascos Erlenmeyer de 1000 mL em incubadora de movimento recíproco. Os componentes do meio de cultivo (glicerol residual, extrato de levedo de cerveja, KNO3, CaCl2.2H2O, MgSO4.7H2O, MnSO4, K2HPO4 e KH2PO4) foram avaliados segundo um planejamento experimental e, nas condições estudadas, não exerceram influência significativa sobre a produção de toxinas. Definiu-se então a composição do meio de fermentação em termos de fonte de carbono e nitrogênio. O glicerol e o extrato de levedo de cerveja não exerceram influência significativa sobre o consumo de substrato pela bactéria. Quanto à concentração celular máxima, o aumento da concentração de glicerol apresentou efeito negativo, enquanto o aumento da concentração de extrato de levedo de cerveja apresentou efeito positivo. Embora as concentrações das fontes de carbono e nitrogênio não tenham sido significativas para a atividade larvicida, a interação entre os fatores foi significativa em seu nível negativo. O melhor resultado quanto à atividade larvicida do meio foi apresentado pelo meio composto por 10 g/L de glicerol residual, 30 g/L de extrato de levedo de cerveja, 5 g/L de KNO3, 0,12 g/L de CaCl2.2H2O, 1,5 g/L de MgSO4.7H2O, 0,09 g/L de MnSO4, 1,5 g/L de K2HPO4 e 1,5 g/L de KH2PO4. Essa mesma composição de meio de cultivo foi testada em um ensaio em biorreator, obtendo-se um meio fermentado mais eficiente que o ensaio realizado em incubadora, e com CL50 de 2,43 &#181;L. / The use of bioinsecticide based on Bacillus thuringiensis has been shown to be effective against larvae of insects able to spread diseases such as dengue, malaria and elephantiasis. The bioinsecticide has high specificity, absence of resistance in target insects and low environmental impact compared to chemical insecticides. However, there are some challenges in the way of widespread use of bioinsecticide, such as the costs of production and formulation. The surplus yeast from the brewing industry can make the production of bioinsecticide economically feasible due to its low cost, as well as the residual glycerol from the biodiesel production. The proper choice of the components of the culture medium is essential for the success of a product and must be carried out in order to obtain a high productivity at the lowest cost. Furthermore, in the case of bioinsecticide, medium components have a major influence on the growth and synthesis of toxins by the bacteria. In the present study it was proposed the formulation of a fermentation medium for Bacillus thuringiensis var. israelensis using brewer\'s yeast and glycerol from the biodiesel manufacturing, seeking to obtain a fermented broth with high toxicity and lower cost compared to those currently obtained. Initially it was defined a procedure to prepare the extract of brewer\'s yeast. It was found that the best method is the sterilization of the brewer\'s yeast suspension (121 °C for 20 minutes) followed by qualitative filtration. The range of brewer\'s yeast extract concentration to be used in the study of the medium composition was also defined (10 g/L to 30 g/L). Assays were performed in 1000 mL Erlenmeyer flasks in a reciprocal motion shaker. The components of the medium (glycerol waste, brewer\'s yeast extract, KNO3, CaCl2.2H2O, MgSO4.7H2O, MnSO4, K2HPO4 e KH2PO4) were evaluated according to an experimental design and, considering the studied conditions, they did not have significant influence on the toxin production. Thus, the composition of the fermentation medium was defined in terms of nitrogen and carbon source. Glycerol and brewer\'s yeast extract did not have any significant influence on the substrate consumption by the bacteria. For the maximum cell concentration, the increase of glycerol concentration showed a negative effect, while increasing the concentration of brewer\'s yeast extract showed a positive effect. Although the concentrations of carbon and nitrogen sources were not significant for the larvicidal activity, the interaction between such factors was significant in its negative level. The best result of larvicidal activity was shown by the following medium: 10 g/L of residual glycerol, 30 g/L of brewer\'s yeast extract, 5 g/L of KNO3, 0.12 g/L of CaCl2.2H2O, 1.5 g/L of MgSO4.7H2O, 0.09 g/L of MnSO4, 1.5 g/L of K2HPO4 and 1.5 g/L of KH2PO4. This same composition of the culture medium was tested in a bench fermenter, resulting in a more efficient fermented broth than the one obtained in shaker flask, and with LC50 of 2.43 &#181;L.