Il est probable qu'à terme les émissions de nanoparticules soient réglementées et ce sont donc les concentrations en nombre qui seront considérées. Il convient donc d'adapter les modèles afin de pouvoir simuler correctement les concentrations en nombre, dans les ambiances confinées comme dans l'atmosphère. Un modèle de dynamique des particules capable de suivre avec autant de précision la concentration en nombre que la concentration en masse, avec un temps de calcul optimal, a été développé. La dynamique des particules dépend de divers processus, les plus importants étant la condensation/évaporation, suivie par la nucléation, la coagulation, et les phénomènes de dépôts. Ces processus sont bien connus pour les particules fines et grossières, mais dans le cas des nanoparticules, certains phénomènes additionnels doivent être pris en compte, notamment l'effet Kelvin pour la condensation/ évaporation et les forces de van der Waals pour la coagulation. Le travail a tout d'abord porté sur le processus de condensation/évaporation, qui s'avère être le plus compliqué numériquement. Les particules sont présumées sphériques. L'effet Kelvin est pris en compte car il devient considérable pour les particules de diamètre inférieur à 50 nm. Les schémas numériques utilisés reposent sur une approche sectionnelle : l'échelle granulométrique des particules est discrétisée en sections, caractérisées par un diamètre représentatif. Un algorithme de répartition des particules est utilisé, après condensation/évaporation, afin de conserver les diamètres représentatifs à l'intérieur de leurs sections respectives. Cette redistribution peut se faire en terme de masse ou de nombre. Un des points clé de l'algorithme est de savoir quelle quantité, de la masse ou du nombre, doit être redistribuée. Une approche hybride consistant à répartir la quantité dominante dans la section de taille considérée (le nombre pour les nanoparticules et la masse pour les particules fines et grossières) a été mise en place et a permis d'obtenir une amélioration de la précision du modèle par rapport aux algorithmes existants, pour un large choix de conditions. Le processus de coagulation pour les nanoparticules a aussi été résolu avec une approche sectionnelle. La coagulation est régie par le mouvement brownien des nanoparticules. Pour cette approche, il a été constaté qu'il est plus efficace de calculer le noyau de coagulation en utilisant le diamètre représentatif de la section plutôt que de l'intégrer sur la section entière. Les simulations ont aussi pu montrer que les interactions de van der Waals amplifient fortement le taux de coagulation pour les nanoparticules. La nucléation a été intégrée au modèle nouvellement développé en incorporant un terme source de nanoparticules dans la première section, commençant à un nanomètre. La formulation de ce taux de nucléation correspond à celle de l'acide sulfurique mais le traitement des interactions numériques entre nucléation, coagulation et condensation/évaporation est générique. Différentes stratégies de couplage visant à résoudre séparément ou en même temps les trois processus sont discutées. Afin de pouvoir proposer des recommandations, différentes méthodes numériques de couplage ont été développées puis évaluées par rapport au temps de calcul et à la précision obtenue en terme de concentration massique et numérique / It is necessary to adapt existing models in order to simulate the number concentration, and correctly account for nanoparticles, in both free and confined atmospheres. A model of particle dynamics capable of following accurately the number as well as the mass concentration of particles, with an optimal calculation time, has been developed. The dynamics of particles depends on various processes, the most important ones being condensation/evaporation, followed by nucleation, coagulation, and deposition phenomena. These processes are well-known for fine and coarse particles, but some additional phenomena must be taken into account when applied to nanoparticles, such as the Kelvin effect for condensation/evaporation and the van der Waals forces for coagulation. This work focused first on condensation/evaporation, which is the most numerically challenging process. Particles were assumed to be of spherical shape. The Kelvin effect has been taken into account as it becomes significant for particles with diameter below 50 nm. The numerical schemes are based on a sectional approach : the particle size range is discretized in sections characterized by a representative diameter. A redistribution algorithm is used, after condensation/ evaporation occurred, in order to keep the representative diameter between the boundaries of the section. The redistribution can be conducted in terms of mass or number. The key point in such algorithms is to choose which quantity has to be redistributed over the fixed sections. We have developed a hybrid algorithm that redistributes the relevant quantity for each section. This new approach has been tested and shows significant improvements with respect to most existing models over a wide range of conditions. The process of coagulation for nanoparticles has also been solved with a sectional approach. Coagulation is monitored by the brownian motion of nanoparticles. This approach is shown to be more efficient if the coagulation rate is evaluated using the representative diameter of the section, rather than being integrated over the whole section. Simulations also reveal that the van derWaals interactions greatly enhance coagulation of nanoparticles. Nucleation has been incorporated into the newly developed model through a direct source of nanoparticles in the first size section, beginning at one nanometer. The formulation of this rate of nucleation corresponds to that of sulfuric acid but the treatment of the numerical interactions between nucleation, coagulation and condensation/evaporation is generic. Various strategies aiming to solve separately or jointly these three processes are discussed. In order to provide recommendations, several numerical splitting methods have been implemented and evaluated regarding their CPU times and their accuracy in terms of number and mass concentrations

La pollution atmosphérique est le produit de fortes émissions de polluants (et de leurs précurseurs) et de conditions météorologiques défavorables. Les particules fines (PM2.5) sont l'un des polluants les plus dangereux pour la santé publique. L'exposition répétée ou prolongée à ces particules entraîne chaque année des maladies respiratoires et cardio-vasculaires chez les personnes exposées ainsi que des morts prématurées. L'évolution du climat dans les années à venir aura un impact sur des variables météorologiques (température, vents, précipitations, ...). Ces variables influencent à leur tour divers facteurs, qui affectent la qualité de l'air (émissions, lessivage par les précipitations, équilibre gaz/particule, ...). Si de nombreuses études ont déjà projeté l'effet du changement climatique sur les concentrations d'ozone, peu se sont intéressées à son effet sur les concentrations de particules fines, en particulier à l'échelle du continent européen. C'est ce que cette thèse se propose d'étudier. La circulation atmosphérique de grande échelle est étroitement liée aux variables météorologiques de surface. Par conséquent, il est attendu qu'elle ait également un impact sur les concentrations de PM2.5. Nous utilisons dans cette thèse une approche statistique pour estimer les concentrations futures de PM2.5 à partir d'observations présentes de PM2.5, de quelques variables météorologiques pertinentes et d'outils permettant de représenter cette circulation atmosphérique (régimes et types de temps). Le faible nombre d'observations journalières de PM2.5 et de ses composants en Europe nous a conduit à créer un jeu de données pseudo-observées à l'aide du modèle de qualité de l'air Polyphemus/Polair3D, puis à l'évaluer de façons opérationnelle et dynamique, afin de s'assurer que l'influence des variables météorologiques sur les concentrations de PM2.5 est reproduite de manière satisfaisante par le modèle. Cette évaluation dynamique d'un modèle de qualité de l'air est, à notre connaissance, la première menée à ce jour.Les projections de PM2.5 sur les périodes futures montrent une augmentation systématique des concentrations de PM2.5 au Royaume-Uni, dans le nord de la France, au Benelux et dans les Balkans, et une diminution dans le nord, l'est et le sud-est de l'Europe, en Italie et en Pologne. L'évolution de la fréquence des types de temps ne suffit pas toujours à expliquer l'évolution de ces concentrations entre les périodes historique et futures, car les relations entre circulation atmosphérique de grande échelle et types de temps, entre types de temps et variables météorologiques, et entre variables météorologiques et concentrations de PM2.5 sont amenées à évoluer dans le futur et contribuent à l'évolution des concentrations de PM2.5. L'approche statistique développée dans cette thèse est nouvelle pour l'estimation de l'impact du climat et du changement climatique sur les concentrations de PM2.5 en Europe. Malgré les incertitudes qui y sont associées, cette approche est facilement adaptable à différents modèles et scénarios, ainsi qu'à d'autres régions du monde et d'autres polluants. En utilisant des observations pour définir la relation polluant-météorologie, cette approche serait d'autant plus robuste / Air pollution is the result of high emissions of pollutants (and pollutant precursors) and unfavorable meteorological conditions. Fine particulate matter (PM2.5) is one of the pollutants of great concern for human health. Every year, a repeated or continuous exposure to such particles is responsible for respiratory and cardiovascular diseases among the concerned populations and leads to premature deaths. Climate change is expected to impact meteorological variables (temperature, wind, precipitation,...). Those variables will influence numerous factors, which will affect air quality (emissions, precipitation scavenging, gas/particle equilibrium,...). A large body of studies have already investigated the effects of climate change on ozone, whereas only a few have addressed its effects on PM2.5 concentrations, especially over Europe. This is the subject we investigate in this thesis. Large-scale circulation is closely linked to surface meteorological variables. Therefore, it is expected that it will impact PM2.5 concentrations too. In this thesis, we develop a statistical algorithm to estimate future PM2.5 concentrations from present PM2.5 observations, selected meteorological variables and tools to represent this circulation (weather regimes and weather types). The lack of daily observations of PM2.5 and its components over Europe prevents us to used observations. Consequently, we have created a pseudo-observed PM2.5 data set, by using the Polyphemus/Polair3D air quality Chemical-Transport Model. Both operational and dynamic evaluations were conducted against EMEP measurements, to ensure that the influence of meteorological variables on PM2.5 concentrations is correctly reproduced by the model. As far as we know, this dynamic evaluation of an air quality model with respect to meteorology is the first conducted to date.Future PM2.5 concentrations display an increase over the U.K., northern France, Benelux, and in the Balkans, and a decrease over northern, eastern, and southeastern Europe, Italy, and Poland compared to the historical period. The evolution of weather type frequencies is not sufficient to explain the PM2.5 changes. The relationships between the large-scale circulation and the weather types, between the weather types and meteorological variables, and between meteorological variables and PM2.5 concentrations evolve with future meteorological conditions and also contribute to PM2.5 changes. The statistical method developed in this thesis is a new approach to estimate the impact of climate and climate change on PM2.5 concentrations over Europe. Despite some uncertainties, this approach is easily applicable to different models and scenarios, as well as other geographical regions and other pollutants. Using observations to establish the pollutant-meteorology relationship would make this approach more robust

[pt] Neste trabalho é realizado o estudo do processo de eletroflotação de partículas finas de hematita em célula modificada de Partridge Smith usando Rhodococcus opacus como biorreagente. Uma amostra mineral com 96 por cento de hematita foi usada neste estudo. Além disso, três frações granulométricas foram escolhidas, sendo estas: menos 53 mais 38; menos 38 mais 20 e menos 20 micrômetros. As variáveis estudadas no processo de eletroflotação foram: o pH, a concentração de reagente, a densidade de corrente e a utilização de bolhas de hidrogênio e oxigênio separadamente. Foram realizadas medições de potencial Zeta e análises por espectroscopia no infravermelho visando avaliar a interação antes e após o contato entre os reagentes e a superfície do mineral. Também foi realizada a medição do diâmetro de bolha de hidrogênio e oxigênio no aparelho Bubble sizer com o intuito de obter um melhor entendimento da influência do diâmetro de bolha na eletroflotação com estes dois gases. O diâmetro médio de bolha (d32) encontrado ficou na faixa de 40 menos 60 micrômetros para as bolhas de hidrogênio e 50 menos 70 micrômetros para as bolhas de oxigênio. Finalmente, foram realizados os ensaios de eletroflotação de hematita com R. opacus e Oleato de sódio. Segundo os resultados obtidos, a redução do tamanho de partícula favoreceu a flotabilidade da hematita, este fato pode ser atribuído ao tamanho das bolhas produzidas no processo e as características do R. opacus de formar flocos mediante interações de Van der Waals. Por outro lado, o aumento da concentração de reagente (R. opacus e oleato de sódio) favoreceu a flotabilidade de hematita. Os valores de pH ótimos encontrados para a flotação de hematita foram entorno de 6 e 7 para o R. opacus e o oleato de sódio, respectivamente. A flotabilidade máxima de hematita obtida com o R. opacus para bolhas de hidrogênio foi em torno de 80 por cento; por outro lado, a flotabilidade obtida com bolhas de oxigênio foi de aproximadamente 65 por cento. O oleato de sódio apresentou um melhor desempenho alcançando uma flotabilidade máxima de aproximadamente 98 por cento para bolhas de hidrogênio e em torno de 90 por cento para bolhas de oxigênio. As características da superfície bacteriana e, principalmente, o diâmetro de bolha menor a 100 micrômetros apresentaram-se como os fatores determinantes no alto valor de flotabilidade das partículas finas de hematita. Desta forma, o processo de eletroflotação com o uso de R. opacus como coletor mostra-se como uma potencial alternativa tecnológica, principalmente, no que diz respeito ao aspecto ambiental devido à biodegradabilidade do microrganismo. / [en] This work aims to evaluate the electroflotation of hematite fine particles in a modified Partridge Smith cell using Rhodococcus opacus. A hematite sample with grade 96 percent was used in this study; additionally, three granulometric fractions were selected: minus 53 plus 38; minus 38 plus 20 and minus 20 micrometers. The variables studied were pH, reagent concentration and current density. Zeta potential measurements and analysis by infrared spectroscopy were carried out to assess the interaction before and after reagents interaction. In order to gain a better understanding of the bubble diameter influence in the process, measurements of the diameter of hydrogen and oxygen bubbles were done. The average bubble diameter (d32) was found in the range of 40 to 60 micrometers for the hydrogen bubbles and of 50 to 70 micrometers for the oxygen bubbles. Finally, electrocoagulation assays the hematite with R. opacus and sodium oleate were carried out. The reduction in particle size favored the floatability of hematite; this fact can be related to the bubble size generated in the process and the ability of R. opacus to form floccules through the interaction of Van der Waals. On the other hand, the increase of reagent concentration (R. opacus and sodium oleate) favored the floatability of hematite. Suitable pH values for hematite flotation were found around 6 and 7 for R. opacus and sodium oleate, respectively. Maximum hematite floatability with R. opacus using hydrogen bubbles was around 80 percent and around 65 percent with oxygen bubbles. Sodium oleate presented a better performance reaching a maximum floatability of approximately 98 percent for hydrogen bubbles and around 90 percent for oxygen bubbles. The characteristics of the bacterial surface and the bubble size were determinant factors in the hematite floatability. Therefore, the electroflotation process using R. opacus as collector is a potential technology regarding environmental issues found with the synthetic reagents.

La poudre de silicium métallurgique (MG-Si, pureté 99 %) ont été étudiées, en se focalisant particulièrement sur les particules fines (taille comprise entre 1 et 10 μm) Ce matériau est utilisé dans l’industrie siliconière pour la synthèse directe du diméthyldichlorosilane et est obtenu par broyage de blocs de silicium. Les propriétés de cette poudre sont cruciales pour le procédé industriel, à la fois en termes de surface spécifique, composition chimique et coulabilité. Comprendre l’influence des particules fines, qui dégradent la coulabilité, et leur origine au cours du broyage est donc d’une importance cruciale. Une nouvelle caractérisation, chimique et cristallographique, des poudres de MG-Si montre que les particules fines sont en moyennes moins chargées en éléments d’alliage que les particules plus grosses. La structure cristalline du silicium est inchangée au cours du broyage, sauf pour les particules superfines (taille inférieure à 1 μm). Celles-ci présentent des zones amorphes : cela montre qu’elles sont soumises à des contraintes plus importantes au cours du broyage, comme cette transformation étant obtenue au-delà d’un seuil de pression. Le comportement du MG-Si en broyage a été étudié pour la première fois. A l’échelle de la particule unique, il est confirmé que les fissures suivent une propagation transgranulaire. De plus, des particules fines peuvent être produites au cours d’un unique événement de broyage, en raison de l’activation simultanée de multiples systèmes de fissures qui peuvent brancher entre elles. La taille critique en-deçà de laquelle la déformation plastique est énergétiquement plus favorable que la propagation de fissure a été estimée à environ 1 μm par une méthode basée sur l’indentation. Ces deux résultats sont cohérents avec la répartition des éléments d’alliages en fonction de la taille de particule. A l’échelle multiparticulaire, une étude pilote en broyeur à tambour tournant a été menée. Les résultats de cette étude ne sont pas disponibles dans cette version publique du manuscrit. Veuillez vous reporter au manuscrit complet. Les conséquences sur la coulabilité de la présence de particules fines dans la poudre de MG-Si produite par broyage ont été caractérisées par mesures d’angle de repos, de dynamique de compaction et en fluidisation. En particulier, un nouveau comportement d’élutriation a été identifié et décrit : l’élutriation séquentielle se produit lorsque des particules fines sont initialement présentes dans le lit fluidisé et se caractérise par l’envolement d’abord des inférieures à environ 30 μm puis seulement des particules de taille supérieure. Ce comportement n’est pas observé en l’absence de fines dans le lit initial. L’explication de ce phénomène pourrait se trouver dans la formation de clusters polydisperses, formés seulement en présence de particules fines. En parallèle de l’élutriation séquentielle, des mesures électrostatiques avec un électromètre externe à la colonne ont montré la présence de potentiels très importants (10 kV), dont le signe correspond à la gamme de taille de particules envolées. Ceci suggère que l’adhésion au sein des clusters pourrait être électrostatique. / Metallurgical-grade silicon (MG-Si, 99 %) powders were extensively investigated, particularly focusing on the fine particles (whose size is between 1 and 10 μm) comprised in these powders. This material is a reactant widely used in the silicones industry for the Direct Synthesis and is obtained by size reduction of millimetric silicon lumps. Powder properties are major stakes of the industrial process. Smaller sizes favor high specific surfaces and high rates of production, but can decrease the lowability, thus inducing poor heat evacuation resulting in hot spots and a decrease in selectivity. Such lowability issues are particularly associated with fine particles, hence understand the generation of these particles during grinding is of critical importance. New chemical and crystallographic characterization of MG-Si is presented, showing that fine particles contain on average less alloying elements than larger particles, yet their crystallographic structure is preserved through grinding. On the contrary, superfine particles (smaller than 1 μm) exhibit amorphous zones: this transformation is pressure induced, showing that these particles experience larger stresses during the grinding step. The behavior of MG-Si in grinding mills has been studied for the first time. At the single particle level, it has been confirmed that transgranular fracture is preferred in MG-Si. Moreover, fine particles can be produced from a single fracture event, due to multiple crack propagation and branching. The critical size under which plastic deformation preferentially occurs over fracture has been evaluated to be approximately 1 μm. These two facts are consistent with a lower level of impurities in fines, yet remaining crystalline, and with superfines exhibiting amorphous areas. At the multiple particle level, pilot scale batch milling experiments have been performed. The results are not included in this public version of the manuscript, please refer to the full manuscript. The consequences of the presence of fine particles in ground MG-Si powder on lowability has been assessed by means of angle of repose, compaction tests and fluidization experiments. A new elutriation behavior has been observed and characterized: for naturally ground MS-Si powders (including fine particles), particles smaller than 30 μm are entrained first, then only larger particles. This was not the case in absence of fine particles. The explanation may probably lie within the presence of polydisperse clusters, formed only in presence of fine particles. Parallel to this elutriation behavior, electrostatic measurements with an external electrometer showed that high potential with sign correlated with the type of particle elutriated are attained during elutriation. This may suggest that electrostatics is responsible for cluster formation.

[pt] Rejeitos da mineração contendo grandes quantidades de finos são constantemente descartados em barragens. A recuperação da água contida nesses rejeitos depende, dentre outros fatores, da remoção de reagentes residuais e material particulado fino, que, devido às suas características de tamanho e cargas superficiais requerem tempos de sedimentação impraticáveis industrialmente. Partículas em suspensão, com diâmetro médio inferior a 2 micrometro, não podem ser removidas por sedimentação simples pois, devido às suas cargas superficiais, permanecem em repulsão mútua, promovendo uma estabilidade eletrostática da polpa. Entretanto, o estado de dispersão dessas partículas está diretamente relacionado ao potencial eletrostático de superfície, sendo necessárias etapas de desestabilização de cargas e agregação, por meio de coagulação e floculação. Na indústria mineral as poliacrilamidas são os polímeros comerciais mais comumente utilizados; porém, com a demanda por floculantes de menor custo e ambientalmente amigáveis, como os polímeros naturais de uso ainda incipiente, a quitosana e o ácido húmico passam a ser atrativos para sua aplicação em polpas minerais. Para este caso de estudo serão avaliados um rejeito de níquel e lama vermelha, ambos provenientes de unidades industriais. O pH é uma variável muito influenciável no processo, porém mediante a análise do potencial zeta foram medidas as cargas superficiais resultantes e se determinou o pH de trabalho para cada uma das polpas. A qualidade dos flocos formados tem forte dependência com as condições operacionais de formação como tipo de reagentes e suas combinações, porém mediante a utilização do planejamento experimental foram estudadas variáveis como concentração de sólidos, dosagens de quitosana e de ácido húmico, velocidade e tempo de agitação. Os parâmetros do processo foram otimizados por meio da obtenção e ajuste de modelos matemáticos. Os ensaios de floculação tiveram como variável de resposta a turbidez do sobrenadante e os resultados foram ajustados atendendo às normas ambientais vigentes para reuso da água. Para o rejeito de Ni, em um pH próximo de 7, obteve-se um modelo quadrático, que conferido experimentalmente com as condições otimizadas deu uma resposta de turbidez de 32 NTU. Para o caso da lama vermelha obtiveramse dois modelos: em pH igual a 12 obteve-se um modelo linear, que gerou uma resposta de turbidez de 33 NTU e em pHigual a 8 obteve-se um modelo quadrático, conferido experimentalmente uma resposta de turbidez de 31 NTU. Dessa forma, nota-se que, para os polímeros estudados, é possível reduzir a turbidez de fluxos de sobrenadantes com eficiência de até 90 por cento. / [en] Mining waste containing large amounts of fines are constantly disposed of in dams. The recovery of water contained in these wastes depends, among other factors, the removal of residuals reagents and fine particulate material, which due to their size and surface charge characteristics require of sedimentation times impractical industrially. Suspended particles with an average diameter of less than 2 micrometre, can not be removed by simple sedimentation by reason of their surface charges, remain in mutual repulsion, promoting an electrostatic stability of the pulp. However, the dispersion state of the particles is directly related to the electrostatic surface potential, and steps necessary to destabilize charges and aggregation, through coagulation and flocculation. Polyacrylamides are the most commonly used commercial polymers at the mineral industry; however, the demand for flocculants of lower cost and environmentally friendly, such as natural polymers have incipient use as chitosan and humic acid, become attractive for application in mineral slurries, for this case study will be evaluated the tailings of Ni and the red mud. pH is a very influenceable variable in the process but by an analysis of the zeta potential were measured of resulting surface charges and determined of pH for each of the slurries. The quality of the flakes have formed strong dependence on operating conditions of formation as a kind of reagents and their combinations, but by using the experimental design were studied variables such as of solids concentration, chitosan, humic acid, stirring speed and stirring time. With the obtained mathematical models, were the optimizations conducted of the process parameters. In tests of flocculation, it had turbidity as response variable of the supernatant and the results were adjusted taking into account the environmental standards for water reuse. To Ni in a pH next to 7, there was obtained a quadratic model, given the conditions optimized experimentally gave a turbidity response of 32 NTU. For the case of red mud was obtained two models at pH 12 was obtained as a linear model given experimentally gave a turbidity response of 33 NTU and pH 8 was obtained a quadratic model, given experimentally has a turbidity response of 31 NTU, which show that it is possible to reduce overflow the turbidity from 500 NTU to less than 50 NTU.

This diploma thesis deals with issues of fine and ultrafine particles formation during thermal decomposition and combustion of biomass. An introduction is devoted to a utilization of biomass as fuel. A following part describes the gas products emerging generally during combustion process. A current emission and air pollution situation in the Czech Republic as well as the health impacts of solid pollutants on human are also mentioned. Within an experimental measurement three basic factors that have impact on fine particles formation during combustion of beech wood (i. e. oxygen content in combustion atmosphere, type of the sample and size of the sample) were thoroughly analyzed. Attention was paid to a progress of thermal decomposition and combustion of sample with respect to the particles formation. Finally, a measurement of fine particles amount was accomplished in a real boiler successively with four different biomass pellet types.

This thesis is focused on nanoparticles produced by internal combustion engines utilized in vehicles. It deals with spatial distribution of nanoparticles within urban areas, impact of alternative fuels usage on particle production and toxicity, and a particle toxicological testing methodology. Monitoring of airborne nanoparticles identified traffic as the main source of airborne nanoparticles in places with heavy traffic load (Prague), as well as in a small city with only local traffic (Čelákovice). Most particles were likely emitted during short episodes of high emissions (e.g. uphill acceleration). During the measurements, high-emission vehicles responsible for a large fraction of the air pollution were also identified. On the other hand, small non-road internal combustion engines, which are not subject to any limit on particle emissions, such as a lawn mower, were operated during the measurement and generated a large number of nanoparticles. The amount and characteristics of the particles produced by combustion depend on the combustion technology and the fuel composition. A large part of the thesis deals with alternative fuels and their effects on the quantity of produced particles and toxicity of organic matter adsorbed on the particles. Hydrotreated vegetable oil (HVO) exhibited the lowest...

Muitos estudos têm sido desenvolvidos com o intuito de descrever a química da fase gasosa na atmosfera da Região Metropolitana de São Paulo (RMSP). Contudo, o tratamento do material particulado (PM) ainda é feito de forma simplificada em modelos de transporte e químicos atmosféricos, apesar do grande conhecimento já adquirido na caracterização da sua composição elementar e da sua estrutura física. Tendo isso em vista, o objetivo do presente estudo é identificar as principais fontes emissoras do material particulado fino, em especial as fontes veiculares que apresentam muitas dificuldades para sua identificação por não haver medidas de traçadores específicos para os combustíveis utilizados. Neste trabalho foram realizadas amostragens, que duravam 24 horas, próximas a uma avenida de intenso tráfego (Avenida Dr. Arnaldo, na Faculdade de Medicina da Universidade de São Paulo) no período de junho de 2007 a agosto de 2008. Com os dados de composição dessas amostras, a identificação das possíveis fontes foi realizada por modelos receptores; mais especificamente foram utilizados: Análise de Fatores (AF) e Positive Matrix Factorization (PMF), uma nova ferramenta estatística, que ainda não havia sido aplicada no estudo do material particulado em São Paulo. O número de fontes identificadas por essas duas ferramentas estatísticas não foi o mesmo: na AF foram extraídos 4 fatores (solo, queima de óleo combustível e dois fatores que se dividiram, identificando a emissão de veículos leves e pesados não diferenciados), enquanto que o PMF identificou 6 (as mesmas fontes identificadas pela AF, com a diferenciação da emissão veicular (leves e pesados) e ainda a queima de biomassa). Houve concordância entre as duas análises que a maior participação para formação de material particulado fino é da emissão por veículos. A comparação entre os modelos mostrou que os resultados obtidos pelo PMF apresentaram uma melhor divisão das fontes, principalmente na identificação das frotas veiculares. Isso se deve ao fato do PMF considerar na análise o erro de cada concentração medida como um peso para cada variável, além de não permitir a ocorrência de fatores negativos, caracterizando melhor as fontes através da presença desses vínculos físicos. / Several studies have been developed in order to describe the gaseous phase of atmospheric constituents in the Metropolitan Region of Sao Paulo (RMSP). However, the aerosol description remains simplified in chemical models, despite the knowledge acquired in its characterization and composition analyses. Facing these limitations, the objective of this work is to identify the main emission sources of fine particulate matter, specially the vehicular ones that present a lot of difficulties due to the fact that the characteristic trace elements are unknown for these sources. It was used in this work 201 samples collected in 24-hour period each at Dr. Arnaldo Avenue, a large and busy avenue in the city of São Paulo, from June 2007 to August 2008. The source identification was accomplished considering the samples composition and using receptor models: Factor Analysis (FA) and Positive Matrix Factorization (PMF) techniques. PMF was a new statistical tool in the study of particulates in the city of São Paulo. The number of sources identified by these two models was different. The FA technique identified 4 factors, (soil, fuel burning, and 2 factors combining in light and heavy-duty vehicles), whereas PMF identified 6, the same as FA (light and heavier vehicles differentiated) and biomass burning. There was concordance between the two techniques, considering that both found that vehicular emission is the major contribution for concentration. The comparison between the models indicated that PMF model present a better source classification, mainly for the vehicular identification. The PMF technique considers the error of each sample in the analysis, weighting the variables and imposing that all the factors must be positive. This mechanism provides a better characterization of sources linking the results with the physics of the process.

Muitos estudos têm sido desenvolvidos com o intuito de descrever a química da fase gasosa na atmosfera da Região Metropolitana de São Paulo (RMSP). Contudo, o tratamento do material particulado (PM) ainda é feito de forma simplificada em modelos de transporte e químicos atmosféricos, apesar do grande conhecimento já adquirido na caracterização da sua composição elementar e da sua estrutura física. Tendo isso em vista, o objetivo do presente estudo é identificar as principais fontes emissoras do material particulado fino, em especial as fontes veiculares que apresentam muitas dificuldades para sua identificação por não haver medidas de traçadores específicos para os combustíveis utilizados. Neste trabalho foram realizadas amostragens, que duravam 24 horas, próximas a uma avenida de intenso tráfego (Avenida Dr. Arnaldo, na Faculdade de Medicina da Universidade de São Paulo) no período de junho de 2007 a agosto de 2008. Com os dados de composição dessas amostras, a identificação das possíveis fontes foi realizada por modelos receptores; mais especificamente foram utilizados: Análise de Fatores (AF) e Positive Matrix Factorization (PMF), uma nova ferramenta estatística, que ainda não havia sido aplicada no estudo do material particulado em São Paulo. O número de fontes identificadas por essas duas ferramentas estatísticas não foi o mesmo: na AF foram extraídos 4 fatores (solo, queima de óleo combustível e dois fatores que se dividiram, identificando a emissão de veículos leves e pesados não diferenciados), enquanto que o PMF identificou 6 (as mesmas fontes identificadas pela AF, com a diferenciação da emissão veicular (leves e pesados) e ainda a queima de biomassa). Houve concordância entre as duas análises que a maior participação para formação de material particulado fino é da emissão por veículos. A comparação entre os modelos mostrou que os resultados obtidos pelo PMF apresentaram uma melhor divisão das fontes, principalmente na identificação das frotas veiculares. Isso se deve ao fato do PMF considerar na análise o erro de cada concentração medida como um peso para cada variável, além de não permitir a ocorrência de fatores negativos, caracterizando melhor as fontes através da presença desses vínculos físicos. / Several studies have been developed in order to describe the gaseous phase of atmospheric constituents in the Metropolitan Region of Sao Paulo (RMSP). However, the aerosol description remains simplified in chemical models, despite the knowledge acquired in its characterization and composition analyses. Facing these limitations, the objective of this work is to identify the main emission sources of fine particulate matter, specially the vehicular ones that present a lot of difficulties due to the fact that the characteristic trace elements are unknown for these sources. It was used in this work 201 samples collected in 24-hour period each at Dr. Arnaldo Avenue, a large and busy avenue in the city of São Paulo, from June 2007 to August 2008. The source identification was accomplished considering the samples composition and using receptor models: Factor Analysis (FA) and Positive Matrix Factorization (PMF) techniques. PMF was a new statistical tool in the study of particulates in the city of São Paulo. The number of sources identified by these two models was different. The FA technique identified 4 factors, (soil, fuel burning, and 2 factors combining in light and heavy-duty vehicles), whereas PMF identified 6, the same as FA (light and heavier vehicles differentiated) and biomass burning. There was concordance between the two techniques, considering that both found that vehicular emission is the major contribution for concentration. The comparison between the models indicated that PMF model present a better source classification, mainly for the vehicular identification. The PMF technique considers the error of each sample in the analysis, weighting the variables and imposing that all the factors must be positive. This mechanism provides a better characterization of sources linking the results with the physics of the process.

La qualité de l'air intérieur est devenue une préoccupation de santé publique du fait notamment de l'augmentation du temps passé dans les environnements intérieurs et les espaces clos. L'objectif de ce travail est de développer des procédés mettant en œuvre la décharge couronne afin de traiter les polluants gazeux chimiques et particulaires dans l'air intérieur. Le premier volet du travail concerne la dégradation des polluants gazeux chimiques présents en très faibles concentrations dans l’air intérieur. Le couplage d’une décharge couronne et d’un catalyseur est mis en œuvre pour dégrader du toluène dans des conditions représentatives de l’air intérieur. Cette étude démontre la très bonne efficacité de la décharge couronne vis-à-vis du toluène avec de faibles densités d’énergie. Elle identifie les conditions opératoires optimales et propose des mécanismes réactionnels suite à l’identification des principaux produits de réaction. Cependant, le plasma génère des espèces indésirables comme l'ozone et les oxydes d'azote qui doivent impérativement être détruits. Nous avons choisi d'adjoindre à la décharge couronne un catalyseur à base d'oxydes de métaux de transition (MnOx/Al2O3). Différents catalyseurs sont synthétisés puis modifiés par greffage d’agents modifiants. Leur efficacité vis-à-vis de l’ozone et des NOx est quantifiée en présence de différentes teneurs en vapeur d’eau. Le deuxième volet concerne la collecte des particules par filtration électrostatique. Trois procédés sont étudiés ; ils associent un étage d’ionisation et un étage de collecte. L’ionisation est assurée soit par un électrofiltre fil-plaque soit par des aiguilles portées à un potentiel de quelques kilovolts. L’efficacité de collecte est mesurée dans la gamme de 10 nm à 20 µm en fonction de différents paramètres opératoires (tension, polarité, vitesse, paramètres géométriques, concentration en particules, humidité, etc.). Les aiguilles ont une efficacité légèrement inférieure à celle de l’électrofiltre mais présentent une très faible production de l'ozone et une consommation énergétique moindre. L’étude paramétrique permet de dimensionner, d'optimiser la géométrie du procédé et de définir les meilleures conditions de fonctionnement / Indoor air quality has become a public health issue because of the increased time spent in indoor environments and confined spaces. The goal of this work is to develop processes using corona discharge to treat chemical gaseous and particulate pollutants in indoor air. The first part of the work concerns the degradation of chemical gaseous pollutants present in very low concentrations in indoor air. The coupling of a corona discharge and a catalyst is implemented to degrade toluene under conditions representative of the indoor air. Tests were carried out under a range of operating and environmental conditions. This study demonstrates the very good efficiency of the corona discharge for toluene with very low specific density. Optimal operating conditions are identified and reaction mechanisms are proposed following the identification of the main reaction products. However, the generations of by-products, ozone and NOx, which can be hazardous compounds, have to be taken into account. The combination of corona discharge with catalysis seems as a promising way to ensure the suitability and the safety of non-thermal plasma as an indoor air cleaner. Different catalysts based on transition metal oxides (MnOx/Al2O3) are synthesized and then modified by grafting modifying agents. Their efficiency for ozone and NOx elimination is quantified in the presence of different water vapor contents. The second part concerns the particle collection by electrostatic precipitation. Three processes are studied; they combine an ionization stage and a collection stage. The ionization is ensured either by a wire-plate electrostatic precipitator (ESP) or by needles brought to a potential of a few kilovolts. The collection efficiency is measured in the range of 10 nm to 20 μm according to different operating parameters (voltage, polarity, velocity, geometrical parameters, particle concentration, humidity, etc.). The needles have a slightly lower efficiency than the ESP but present a very low production of ozone and a lower energy consumption. The parametric study makes it possible to dimension, to optimize the geometry of the process and to define the best operating conditions