Ce travail de thèse porte sur l’étude des mécanismes et des bilans de transfert des pesticides vers les cours d’eau sur deux bassins versants agricoles emboités, de tailles différentes en Gascogne (Sud Ouest de la France) : la Save à Larra (1110 km2).et le Montoussé (3,28 km2) à Auradé. Dans cette région, les pratiques d’une agriculture intensive conduisent à des risques importants pour les ressources en eau, notamment pendant les périodes de crue. C’est pour cela que nous avons porté une attention particulière dans cette étude à ces événements au cours desquels une grande quantité de contaminants sont transportés par les cours d’eau. Quatorze molécules de pesticides (herbicides et fongicides) largement utilisées sur ces bassins versants ont été étudiées grâce à un échantillonnage intensif, notamment en périodes de crue, durant deux années hydrologiques (2007/2008 et 2008/2009). La majorité des molécules étudiées présentent des concentrations qui dépassent les limites autorisées par l’Union Européenne pour l’eau potable (0.1 μg.L-1 pour chaque molécule et 0.5 μg.L-1 pour l’ensemble des molécules), notamment en périodes de crue. L’analyse des hystérésis mises en évidence sur les relations entre les concentrations (pesticides, MES, COD, POC) et les débits du cours d’eau permet de mieux comprendre les mécanismes de transfert des pesticides et de leurs paramètres de contrôle. Ces hystérésis montrent pour un même débit des concentrations différentes en montée et en descente de crue ; elles peuvent être dextres (sens des aiguilles d’une montre) ou senestres (sens inverse) suivant l’origine des molécules et l’écoulement responsable de leur transfert des sols vers les eaux. C’est pourquoi nous avons procédé dans cette étude à une séparation des différentes composantes (ruissellement superficiel, écoulement hypodermique, écoulement de nappe) de l’écoulement fluvial pour mieux comprendre la dynamique de transferts des pesticides et de leurs paramètres de contrôle. On peut ainsi mettre en évidence des relations positives entre MES, COD, COP ou certaines molécules de pesticides et les débits des écoulements de surface (ruissellement superficiel ou écoulement hypodermique suivant les caractéristiques physico-chimiques des molécules). Les calculs de flux des différents pesticides exportés par les cours d’eau montrent que 60 à 90% (suivant les molécules) des transports annuels se font durant les épisodes de crue. Les flux spécifiques (masse exportée par unité de surface) calculés sont plus importants sur le bassin versant de la Save que sur celui du Montoussé et les relations flux spécifiques-débits mettent en évidence des concentrations plus élévées sur la Save où l’utilisation des pesticides est globalement plus importante que sur le Montoussé où les pratiques sont raisonnées. L’analyse des pesticides dans l’eau filtrée et non filtrée nous a permis d’estimer la distribution de chaque molécule entre les phases particulaires et dissoutes (Kd). De plus, les valeurs de flux calculés pour chaque crue nous ont permis d’estimer un Kd moyen pour chaque molécule. Ces valeurs de Kd présentent une très bonne relation avec les valeurs de Kow (coefficient de partage octanol-eau) extraites de la littérature. De même, les pourcentages calculés de chaque pesticide exporté sous forme particulaire sont aussi très bien corrélés au Kow de chaque molécule. / The mechanisms of pesticides transport to stream flow were studied in two agricultural nested catchments of different size in Gascogne region (South West of France): the Save river basin at Larra (1110 km2) and the Montoussé experimental watershed at Auradé (3.28 km2). The intensive agricultural practices used in this region lead to an important risk for water resources by pesticides, especially during storm events. This is why we have paid special attention on storm events when a large quantity of contaminant was transported during hydrological periods. Fourteen molecules of pesticides (herbicides and fungicides) were investigated during the study period. Both of these groups are widely used for agricultural purposed in these catchments. The results achieved over the two years monitoring (2007-2009) enable us to emphasize the principal processes, implied in pesticide transfer on these agricultural catchments. The majority of compounds are detected during storm runoff events. And, the average concentrations of some pesticides are exceeded at the authorization limit of the European Union for pesticide concentrations in drinking water (0.1 µg.L-1 for individual pesticides and 0.5 µg.L-1 for total pesticides). To better understand the mechanisms of pesticide transport hysteresis, patterns on the concentration-discharge relationship (result of different concentration of pesticides in rising and falling limb of storm) were studied. However, clockwise or anticlockwise hysteresis patterns could be observed for some molecules of pesticide and their controlling factors such as dissolved organic carbon (DOC), particulate organic carbon (POC) and total suspended matters (TSM) according to their transfer dynamic in the catchment. We proceeded with hydrograph separation of the main stormflow components (surface runoff, subsurface flow and groundwater) so that the main pesticide routing could be traced for its soil-river transfers. We also came to the conclusion that there is a positive relationship between riverine TSM, DOC and pesticide, concentrations and the discharges of surface or subsurface runoffs according to pesticide properties. Pesticide flux calculation shows between 60 to 90% of the molecule transport takes place during storm periods. Specific flux calculation also demonstrated the higher flux value in Save catchment than in Aurade with higher pesticide concentration for a given specific discharge. The latter result may be due to the more consumption of pesticide in Save catchment. The analyses of pesticides both in filtered and unfiltered water enabled us to estimate the distribution of pesticides into particulate and dissolved phases. Moreover, the pesticide flux values allow calculating average partition coefficients kd between dissolved and particulate fractions which present good relationship with Kow values (octanol-water) extracted from literature. The percentage of each pesticide transported as particulate forms is also well correlated to Kow.

The content of analytical part of this thesis is the study of airborne dust samples collected in places with its higher concentration: by the Prague Ring (D0 motorway) and near the Ostrava's iron-mills. In both places, automobile transportation or industrial production are the primary sources of particulate matter emissions in the air. Received samples of PM1, PM2, PM10 a TSP filters were obtained by a standard air pollution monitoring procedure. This study examines the relationship between the magnetic parameters of the samples, their level of concentration of metals frequently present in emissions from these sources, the total mass of the dust fraction in the samples and the meteorological parameters (for Ostrava samples). The magnetic properties of the filters were investigated by magnetometric analyzes using a vibration sampling magnetometer. The concentrations of the metals were monitored by X-ray fluorescence spectrometer, the weight of parts of dust filtres was measured on laboratory scales.

Este trabalho teve como objetivo estudar os transistores de tunelamento por efeito de campo em estruturas de nanofio (NW-TFET), sendo realizado através de analises com base em explicações teóricas, simulações numéricas e medidas experimentais. A fim de avaliar melhorar o desempenho do NW-TFET, este trabalho utilizou dispositivos com diferentes materiais de fonte, sendo eles: Si, liga SiGe e Ge, além da variação da espessura de HfO2 no material do dielétrico de porta. Com o auxílio de simulações numéricas foram obtidos os diagramas de bandas de energia dos dispositivos NW-TFET com fonte de Si0,73Ge0,27 e foi analisada a influência de cada um dos mecanismos de transporte de portadores para diversas condições de polarização, sendo observado a predominância da influência da recombinação e geração Shockley-Read-Hall (SRH) na corrente de desligamento, do tunelamento induzido por armadilhas (TAT) para baixos valores de tensões de porta (0,5V > VGS > 1,5V) e do tunelamento direto de banda para banda (BTBT) para maiores valores tensões de porta (VGS > 1,5V). A predominância de cada um desses mecanismos de transporte foi posteriormente comprovada com a utilização do método de Arrhenius, sendo este método adotado em todas as análises do trabalho. O comportamento relativamente constante da corrente dos NW-TFETs com a temperatura na região de BTBT tem chamado a atenção e por isso foi realizado o estudo dos parâmetros analógicos em função da temperatura. Este estudo foi realizado comparando a influência dos diferentes materiais de fonte. O uso de Ge na fonte, permitiu a melhora na corrente de tunelamento, devido à sua menor banda proibida, aumentando a corrente de funcionamento (ION) e a transcondutância do dispositivo. Porém, devido à forte dependência de BTBT com o campo elétrico, o uso de Ge na fonte resulta em uma maior degradação da condutância de saída. Entretanto, a redução da espessura de HfO2 no dielétrico de porta resultou no melhor acoplamento eletrostático, também aumentando a corrente de tunelamento, fazendo com que o dispositivo com fonte Ge e menor HfO2 apresentasse melhores resultados analógicos quando comparado ao puramente de Si. O uso de diferentes materiais durante o processo de fabricação induz ao aumento de defeitos nas interfaces do dispositivo. Ao longo deste trabalho foi realizado o estudo da influência da densidade de armadilhas de interface na corrente do dispositivo, demonstrando uma relação direta com o TAT e a formação de uma região de platô nas curvas de IDS x VGS, além de uma forte dependência com a temperatura, aumentando a degradação da corrente para temperaturas mais altas. Além disso, o uso de Ge introduziu maior número de impurezas no óxido, e através do estudo de ruído foi observado que o aumento na densidade de armadilhas no óxido resultou no aumento do ruído flicker em baixa frequência, que para o TFET, ocorre devido ao armadilhamento e desarmadilhamento de elétrons na região do óxido. E mais uma vez, o melhor acoplamento eletrostático devido a redução da espessura de HfO2, resultou na redução desse ruído tornando-se melhor quando comparado à um TFET puramente de Si. Neste trabalho foi proposto um modelo de ruído em baixa frequência para o NW-TFET baseado no modelo para MOSFET. Foram realizadas apenas algumas modificações, e assim, obtendo uma boa concordância com os resultados experimentais na região onde o BTBT é o mecanismo de condução predominante. / This work aims to study the nanowire tunneling field effect transistors (NW-TFET). The analyses were performed based on theoretical explanations, numerical simulations and experimental data. In order to improve the NW-TFET performance, it was used devices with different source compositions, such as Si, SiGe alloy and Ge, besides different thicknesses of HfO2 for the gate dielectric. With the aid of numerical simulations it was obtained the NW-TFET energy band diagrams and analyzed the influence of recombination and generation Shockley-Read-Hall (SRH) on the off current, the influence of the trap assisted tunneling (TAT) at low gate voltage bias (0,5V > VGS > 1,5V) and the direct band to band tunneling (BTBT) at higher gate voltage bias(VGS > 1,5V). The predominance of each conduction mechanisms was confirmed by the Arrhenius plot method, being this method adopted in all analysis in this work. The constant current with the temperature in the BTBT region has drawn attention and due to that, this work have studied the NW-TFET analog performance as function of temperature and also the influence of the source composition. The Ge source device shows an improved tunneling current, related to the bandgap narrowing, which leads to higher ION and transconductance. However, due to the strong BTBT dependence with the electric field, the use of Ge as source results in further ION/IOFF degradation. Despite this, the reduced HfO2 thickness in the gate dielectric, results in better electrostatic coupling, which also increases the tunneling current, making this device to present better analog performance when compared to devices with Si source. The use of different materials during the device fabrication leads to an increase of the interface defects. This work presented the influence of the interface trap density on the current, showing a direct relation with TAT and appearance of a plateau region in the IDS x VGS curves. In addition it was shown a strong temperature dependence increasing the current degradation at higher temperatures. Furthermore, the use of Ge has shown an increase of impurities in the oxide, and through the noise study it was observed the flicker noise increase at low frequency, which for TFETs, occurs due to the electrons trapping and detrapping in the oxide region. Once again, the reduced HfO2 thickness leads to better electrostatic coupling, resulting in noise reduction and becoming better when compared to a devices with Si source. In this work was proposed a low frequency noise model for a NW-TFET based on MOSFET models. Minor changes have been done, and thus a good agreement with the experimental results in the region where the BTBT is predominant conduction mechanism was obtained.

Este trabalho teve como objetivo estudar os transistores de tunelamento por efeito de campo em estruturas de nanofio (NW-TFET), sendo realizado através de analises com base em explicações teóricas, simulações numéricas e medidas experimentais. A fim de avaliar melhorar o desempenho do NW-TFET, este trabalho utilizou dispositivos com diferentes materiais de fonte, sendo eles: Si, liga SiGe e Ge, além da variação da espessura de HfO2 no material do dielétrico de porta. Com o auxílio de simulações numéricas foram obtidos os diagramas de bandas de energia dos dispositivos NW-TFET com fonte de Si0,73Ge0,27 e foi analisada a influência de cada um dos mecanismos de transporte de portadores para diversas condições de polarização, sendo observado a predominância da influência da recombinação e geração Shockley-Read-Hall (SRH) na corrente de desligamento, do tunelamento induzido por armadilhas (TAT) para baixos valores de tensões de porta (0,5V > VGS > 1,5V) e do tunelamento direto de banda para banda (BTBT) para maiores valores tensões de porta (VGS > 1,5V). A predominância de cada um desses mecanismos de transporte foi posteriormente comprovada com a utilização do método de Arrhenius, sendo este método adotado em todas as análises do trabalho. O comportamento relativamente constante da corrente dos NW-TFETs com a temperatura na região de BTBT tem chamado a atenção e por isso foi realizado o estudo dos parâmetros analógicos em função da temperatura. Este estudo foi realizado comparando a influência dos diferentes materiais de fonte. O uso de Ge na fonte, permitiu a melhora na corrente de tunelamento, devido à sua menor banda proibida, aumentando a corrente de funcionamento (ION) e a transcondutância do dispositivo. Porém, devido à forte dependência de BTBT com o campo elétrico, o uso de Ge na fonte resulta em uma maior degradação da condutância de saída. Entretanto, a redução da espessura de HfO2 no dielétrico de porta resultou no melhor acoplamento eletrostático, também aumentando a corrente de tunelamento, fazendo com que o dispositivo com fonte Ge e menor HfO2 apresentasse melhores resultados analógicos quando comparado ao puramente de Si. O uso de diferentes materiais durante o processo de fabricação induz ao aumento de defeitos nas interfaces do dispositivo. Ao longo deste trabalho foi realizado o estudo da influência da densidade de armadilhas de interface na corrente do dispositivo, demonstrando uma relação direta com o TAT e a formação de uma região de platô nas curvas de IDS x VGS, além de uma forte dependência com a temperatura, aumentando a degradação da corrente para temperaturas mais altas. Além disso, o uso de Ge introduziu maior número de impurezas no óxido, e através do estudo de ruído foi observado que o aumento na densidade de armadilhas no óxido resultou no aumento do ruído flicker em baixa frequência, que para o TFET, ocorre devido ao armadilhamento e desarmadilhamento de elétrons na região do óxido. E mais uma vez, o melhor acoplamento eletrostático devido a redução da espessura de HfO2, resultou na redução desse ruído tornando-se melhor quando comparado à um TFET puramente de Si. Neste trabalho foi proposto um modelo de ruído em baixa frequência para o NW-TFET baseado no modelo para MOSFET. Foram realizadas apenas algumas modificações, e assim, obtendo uma boa concordância com os resultados experimentais na região onde o BTBT é o mecanismo de condução predominante. / This work aims to study the nanowire tunneling field effect transistors (NW-TFET). The analyses were performed based on theoretical explanations, numerical simulations and experimental data. In order to improve the NW-TFET performance, it was used devices with different source compositions, such as Si, SiGe alloy and Ge, besides different thicknesses of HfO2 for the gate dielectric. With the aid of numerical simulations it was obtained the NW-TFET energy band diagrams and analyzed the influence of recombination and generation Shockley-Read-Hall (SRH) on the off current, the influence of the trap assisted tunneling (TAT) at low gate voltage bias (0,5V > VGS > 1,5V) and the direct band to band tunneling (BTBT) at higher gate voltage bias(VGS > 1,5V). The predominance of each conduction mechanisms was confirmed by the Arrhenius plot method, being this method adopted in all analysis in this work. The constant current with the temperature in the BTBT region has drawn attention and due to that, this work have studied the NW-TFET analog performance as function of temperature and also the influence of the source composition. The Ge source device shows an improved tunneling current, related to the bandgap narrowing, which leads to higher ION and transconductance. However, due to the strong BTBT dependence with the electric field, the use of Ge as source results in further ION/IOFF degradation. Despite this, the reduced HfO2 thickness in the gate dielectric, results in better electrostatic coupling, which also increases the tunneling current, making this device to present better analog performance when compared to devices with Si source. The use of different materials during the device fabrication leads to an increase of the interface defects. This work presented the influence of the interface trap density on the current, showing a direct relation with TAT and appearance of a plateau region in the IDS x VGS curves. In addition it was shown a strong temperature dependence increasing the current degradation at higher temperatures. Furthermore, the use of Ge has shown an increase of impurities in the oxide, and through the noise study it was observed the flicker noise increase at low frequency, which for TFETs, occurs due to the electrons trapping and detrapping in the oxide region. Once again, the reduced HfO2 thickness leads to better electrostatic coupling, resulting in noise reduction and becoming better when compared to a devices with Si source. In this work was proposed a low frequency noise model for a NW-TFET based on MOSFET models. Minor changes have been done, and thus a good agreement with the experimental results in the region where the BTBT is predominant conduction mechanism was obtained.

The thesis investigates methods of examining corneal biomechanics using non-contact tonometry and introduces novel techniques to investigate corneal material properties in vivo. A comprehensive systems analysis of the CorvisST (CST) and Ocular Response Analyser (ORA) was performed. Pressure sensors were used to characterisation the airflow produced by the CST and the ORA. Distinct differences were observed between the central airflow pressures between the two devices: the CST pressure was higher and of shorter duration. Scheimpflug high-speed imaging via the CST allowed components of the corneal deformation to be investigated and the development of a 3D deformation matrix (time, depth and spatial resolution) through tracing of the anterior and posterior corneal surface. Measures of whole eye movement (WEM) with CST were found to be robust. WEM demonstrated an asymmetric profile and a correction method was developed to address the corneal deformation matrix for this asymmetry. Novel methods for characterisation of intrinsic material characteristics of the cornea were developed using numerical and graphical analytical procedures. Application of these parameters was tested on enucleated porcine eyes across a wide range of manometry internal ocular pressure (MIOP). The dynamic E-Modulus was found to be most affected by MIOP change. To investigate the in vivo distribution and heterogeneity of the corneal biomechanics, a novel set-up allowed the mapping of corneal biomechanics across the cornea using the CST (central, paracentral, peripheral) and ORA (central, peripheral). Biometric and demographic grouping of subjects allowed detection of discriminating factors between individuals. The results suggest that the in vivo cornea of healthy human adults can be characterised as a viscoelastic, damped system for longitudinal strain and a highly oscillating system for lateral strain. The cornea is approximately homogenous for measures of rigidity and dynamic E-Modulus but other corneal material characteristics (longitudinal and lateral strain, hysteresis, damping and compressibility) demonstrated regional differences. The experimental design employed allowed for strict control of biometric and biomechanical intersubject variables, based on gold-standard techniques as well as newly-developed methods, thereby creating a normative database for future use.

La corrosion des aciers dans les ouvrages en béton armé due aux chlorures est l’une des principales causes de dégradation des constructions notamment en façade maritime. La prédiction de la durée de vie des constructions vis-à-vis de cette pathologie, nécessite la compréhension des principaux mécanismes liés au transfert de chlorures surtout dans les matériaux insaturés (soumis au marnage ou au brouillard salin). Pour améliorer les travaux qui existent dans la littérature sur le sujet, les principaux objectifs de cette thèse sont : la caractérisation expérimentale du phénomène de double couche électrique (DCE) et l’étude de son influence sur le transfert ionique, l’étude des effets de la température et de l’hystérésis (sorption-désorption) sur le profil de chlorures, l’étude du couplage entre corrosion des aciers et transfert des chlorures. Pour atteindre ces objectifs, le travail est divisé en quatre parties. La première est consacrée à l’approche théorique du problème pour établir les équations permettant de décrire le couplage entre transferts ionique et hygrothermique dans les matériaux insaturés. La prise en compte de la DCE est également explicitée. La deuxième partie est expérimentale, elle est réalisée sur des éco-bétons fabriqués à base de laitier et de cendres volantes. Elle consiste à quantifier les paramètres d’entrée du modèle identifiés lors de la première partie. Il s’agit : de la porosité, de la perméabilité aux gaz, du coefficient de diffusion d’ions chlorures, de la distribution porale, des isothermes d’interactions chimiques et des isothermes de sorption-désorption et du potentiel Zeta caractérisant la DCE. Dans la troisième partie, les équations de transfert traduisant le modèle sont résolues numériquement. L’étude de sensibilité du modèle vis à vis de la DCE, de la température, de la perméabilité à l’eau est réalisée. Une confrontation des résultats du modèle (profil de chlorures) avec des résultats expérimentaux obtenus en situation de marnage a permis de montrer l’intérêt de considérer l’hystérésis des courbes de sorption-désorption. Quant à l’effet de la DCE il se manifeste pour potentiel Zeta dépassant 25mV. La dernière partie concerne l’étude du couplage entre transfert de chlorures et corrosion des aciers. Les estimations de la durée d’incubation, de la perte de section d’acier, les temps requis pour l’apparition des premières fissures et l’éclatement du béton d’enrobage ont été réalisées. Nous avons montré que l’utilisation de la loi de Fick peut sous-estimer ces grandeurs. / The corrosion of steels in the reinforced concrete structures caused by chlorides is one of the main causes of degradation of the constructions in particular in maritime facade. The prediction of the service life of the constructions towards this pathology requires the understanding of the main mechanisms relating to the chlorides transport especially in the unsaturated materials submitted to the tide or to the salt spray. To improve the works of the literature, the main objectives should be accomplished are : to study the experimental characterization of the electrical double layer (EDL) and to study its influence on the ionic transport ; to research on the temperature effect and the hysteresis effect on the chloride profile ; to examine chloride transfer coupled with corrosion of reinforcement. To reach these objectives, the work is divided into four parts. The first one is dedicated to the theoretical approach of the problem to establish the equations allowing describing the coupling between ionic and hydrothermal transfer to the unsaturated materials. The consideration of the EDL is also clarified. The second part is experiment. It is realized on eco-concretes made on base of slag and fly ash. It consists in quantifying the inputs of the model identified during the first part such as: porosity, gas permeability, chloride diffusion coefficients, pore size distribution, chemical interactions, isotherms of sorption-desorption and the Zeta potential characterizing the EDL. In the third part, the transfer equations translating to the model are numerically resolved. The study of model sensibility according to EDL, the temperature, the water permeability is performed. A comparison between the numerical results (chloride profile) obtained with the model with experimental ones obtained in the case of tidal zone allowed shows the interest to consider the hysteresis effects. Also, it is shown that the EDL effect is significant beyond Zeta potential of 25mV. The last part concerns the study of the coupling between chloride transfer and corrosion of steels. The estimations of the duration of incubation, the loss of section of steel, the times required to cover cracking were examined. The result showed that the use of the law of Fick can underestimate these parameters.

Les travaux présentés dans ce manuscrit sont dédiés à l'étude des composants passifs pour l'électronique de puissance à haute température. Des condensateurs et des matériaux magnétiques sont sélectionnés et caractérisés jusqu'à environ 250°C. Les caractéristiques électriques et électromagnétiques montrent, pour certains de ces composants et matériaux, des dépendances significatives en fonction de la température, mais également des non-linéarités et des phénomènes d'hystérésis. Les caractérisations sont ensuite exploitées pour la conception d'un filtre CEM d'entrée d'un onduleur de tension de 2kW. Une démarche et des considérations liées au dimensionnement d'un filtre sont détaillées. Un démonstrateur de filtre CEM est testé en charge et à haute température (200°C). Les résultats montrent une dépendance relativement faible des perturbations conduites entre 150kHz et 30MHz en fonction de la température (environ +6dBµA entre 25°C et 200°C selon la norme DO-160F). Le fonctionnement à haute température de composants passifs au sein d'un filtre CEM pour l'électronique de puissance a été démontré. En complément du filtre à composant discret et pour répondre aux besoins d'atténuation à haute fréquence qui seront accrus pour les convertisseurs à base de semi-conducteurs à grand gap (SiC et GaN) qui commutent plus rapidement que des interrupteurs de type IGBT en Si, nous avons proposé l'intégration de condensateurs de mode commun au sein d'un module de puissance. Les résultats simulés et expérimentaux ont montré une réduction des perturbations conduites grâce à l'intégration de ces condensateurs. Cette solution, compatible avec un fonctionnement à haute température, est positionnée comme une solution alternative à un filtre d'entrée complexe (multi-niveaux) et s'inscrit dans la tendance actuelle des IPEM (Intelligent/Integrated Power Electronics Module) qui recherche l'intégration de fonctions dans le module de puissance. L'ensemble de ces travaux souligne par ailleurs l'importance du packaging pour l'électronique de puissance à haute température. / The study, which is described in this dissertation, is dedicated to passive components in order to be integrated into high temperature power electronic converters. Capacitors and magnetic materials are selected and characterized up to 250°C. Electrical and electromagnetic characteristics are measured. Some components show a significant temperature deviation, but also a non-linear behavior with a hysteresis phenomenon. Based on these characteristics, a high temperature EMI filter for a 2kW voltage inverter is designed. The design procedure and some practical considerations are discussed. Then, the experimental results from the prototype at 200°C under full load conditions are given. The variation of the conducted emissions, from 150kHz and 30MHz, with the temperature is low (about +6dBµA between 25°C and 200°C into a DO-160F setup). The feasibility of a working EMI filter for high temperature power electronics is demonstrated. To meet the high frequency EMI requirements, with wide-band gap semi-conductors devices which are faster than Si IGBT, a solution based on integrated common mode capacitors into the power module is proposed. With this solution, operation at high temperature is also doable. Experimental results show a reduction of the conducted emissions thanks to these integrated capacitors. We consider this solution as an alternative against an increased complexity of the EMI input filter. It follows the present trends toward the integration of functions into a power module, close to the power switches. Moreover, packaging issues are highlighted and remains as a major limitation for high temperature power electronics.

Ce travail de thèse a été consacré à l'étude à la fois théorique et expérimentale de la dynamique de l'aimantation assistée par un champ électrique dans un dispositif à base de (Ga,Mn)As. Une couche de (Ga,Mn)As, dont l’anisotropie magnétique est complexe, est un matériau de choix pour la manipulation de l’aimantation par un champ électrique.J’ai présenté une stratégie de retournement précessionnel de l'aimantation qui tire partie de la réduction transitoire de l'anisotropie cubique provoquée par une courte impulsion de champ électrique. A l’aide d’un modèle macrospin, j’ai démontré notamment qu'une impulsion de champ électrique de quelques ns de durée est suffisante pour basculer l'aimantation entre deux positions d'équilibres.L'aspect expérimental est basé sur l'utilisation d'une jonction p-n tout semi-conducteur dont la région dopée p est formée par une couche mince de (Ga,Mn)As. La déplétion des porteurs de charge dans le canal semi-conducteur provoquée par l'application d'un train d'impulsions de tension de courte durée induit une forte diminution du champ d'anisotropie cubique. L'étude expérimentale du renversement de l'aimantation en champ magnétique a conduit à la mise en évidence d'un retournement de l'aimantation via la nucléation et la propagation de parois et d’une distribution large des champs de piégeage. L'inhomogénéité magnétique au sein de la couche de (Ga,Mn)As a empêché l'observation d'une résonance ferromagnétique induite par un champ électrique. Néanmoins, des mesures de retournement avec une ou plusieurs impulsions de tension de grille ont permis de proposer un processus de renversement de l'aimantation induit par un champ électrique. / This work has been devoted to the study of both theoretical and experimental electric field induced magnetization dynamics in a (Ga,Mn)As based device. A layer of (Ga,Mn)As, whose magnetic anisotropy is complex, is of particular interest for magnetization manipulation by electric fields.I proposed a scheme for the precessional switching of the magnetization using cubic anisotropy field reduction triggered by electric field pulse. Using a model macrospin, I demonstrated that a ns-pulse is sufficient to switch the magnetization between two equilibrium positions.An all-semiconductor epitaxial p-n junction based on low-doped (Ga,Mn)As was fabricated. Gating effects triggered by low voltage pulses induced a strong decrease of the cubic anisotropy field. I demonstrated that magnetization reversal is dominated by the nucleation and the propagation of domain walls. The distribution of pinning fields was found to be broad. Due to a magnetic inhomogeneity of the (Ga,Mn)As layer, it was not possible to observe any ferromagnetic resonance induced by an electric field. However, I proposed a magnetization reversal scenario based on single or multiple gate voltage pulses measurements

Gas, liquid and powder flow in the lower part of a blast furnace is complex phenomena. In order to understand the aerodynamics of the blast furnace properly, these phenomena must be included in their advanced form. Previous studies have shown that the conditions of blast furnace resemble the cold model experiments which have been done in decreasing gas velocities. Also, the recent studies have shown that liquid flow in a blast furnace can be represented more realistically considering it discrete in nature. In the current study, both the phenomena have been considered along with the injection of powder through a nozzle while studying the fluid flow behaviour in a packed bed. The situation resembles the lower part of an ironmaking blast furnace. In this study, gas flow has been modelled using k-ε turbulent model and has been coupled with previously developed stress model to calculate the raceway size. Coal powder is treated as continuum and has been modelled in the similar way as gas flow. After this gas and powder flow model were coupled with previously developed discrete liquid flow model. Liquid flow model has been considered for structured bed only. The governing equations for gas phase were discretized. Finite Volume method was used for the solution. Co-located grid is used for the simulation. Blending of upwind difference scheme and central difference scheme (deferred correction approach) is used to achieve the stability of upwind scheme and accuracy of central difference scheme. Similar treatment was employed for powder phase. For the solution of volume fraction of powder, powder phase continuity equation was used along with pseudo time step scheme. Results obtained from gas and powder models have been validated against published experimental data. Similarly, gas-liquid flow results have been validated against published experimental data on gas-powder flow. Results obtained by gas-powder-liquid model could not be validated against any experimental or theoretical data as they are not available in the literature. The effect of various parameters on the fluid flow (gas/liquid/powder) behaviour have been studied like the effect of increasing and decreasing gas velocities, flow rates of liquid, gas and powder, size of powder and packing etc. It is found that the above mentioned phenomena have significant effect on the fluid flow behaviour in a packed bed.

The parallel active filter (PAF) is the modern solution for harmonic current mitigation and reactive power compensation of nonlinear loads. This thesis is dedicated to detailed analysis, design, control, and implementation of a PAF for a 3- phase 3-wire rectifier load. Specifically, the current regulator and switching ripple filter (SRF) are thoroughly investigated. A novel discrete time hysteresis current regulator with multi-rate current sampling and flexible PWM output, DHCR3, is proposed. DHCR3 exhibits a high bandwidth while limiting the maximum switching frequency for thermal stability and its implementation is simple. In addition to the development of DHCR3, in the thesis state of the art current regulation methods are considered and thoroughly compared with DHCR3. Since the current regulator type determines the SRF topology choice, various SRF topologies are considered and a thorough design study is conducted and SRF topology selection and parameter determination methods are presented via numerical examples. Through a PAF designed for a 10kW diode/thyristor rectifier load, the superior performance of DHCR3 is verified through simulations and experiments and via comparison to other current regulators. The sufficient switching ripple attenuation of the SRF structures for the designed PAF system and the overall performance of the designed and built PAF system are demonstrated via detailed computer simulations and laboratory experiments. This thesis aids the PAF current regulator and SRF selection, design, and implementation.