Le procédé de refusion sous laitier électroconducteur (ESR =Electro Slag Remelting) permet de produire des alliages à haute valeur ajoutée utilisés pour des applications critiques. Les mesures in-situ sur les fours industriels étant coûteuses, la simulation numérique est essentielle à la maitrise et à la compréhension de ce procédé complexe. L’Institut Jean Lamour a développé depuis plusieurs années un modèle 2D axisymétrique qui permet de simuler la refusion d’une électrode consommable au sein d’une lingotière considérée comme totalement isolée électriquement du reste du système. Sont alors décrit en régime transitoire les transferts couplés de chaleur et de quantité de mouvement, ainsi que le passage du courant électrique dans le laitier et le lingot lors de la croissance et de la solidification de ce dernier. Les connaissances acquises au cours des dernières années sur le procédé ESR remettent en cause l’hypothèse d’isolation électrique du moule au cours de la refusion. L’objet de cette thèse est d’intégrer et d’étudier la possibilité de passage de courant entre le laitier et la lingotière au cours de la refusion. Un premier modèle a été développé. Il consiste en un calcul électromagnétique complet dans l'ensemble du système pour une géométrie simplifiée. Il a permis de simplifier la mise au point du modèle global, tout en fournissant de premiers résultats. Par la suite, un modèle totalement couplé a été finalisé puis les modifications du code ont fait l’objet de validations avec des mesures expérimentales. Des études de sensibilité ont été menées pour tester l’influence des propriétés du laitier et des paramètres opératoires sur la qualité du lingot final. / The ElectroSlag Remelting process (ESR) is widely used to produce high added value alloys for critical applications (aerospace industry, nuclear plants, etc.). Trial-and-error based approaches being expensive, numerical simulation is fundamental to improve the knowledge and the understanding of this complex process. The Institut Jean Lamour has been developing for several years a numerical code to simulate the melting of a consumable electrode, supposedly perfectly cylindrical, within a mold assumed to be perfectly electrically insulated from the electrode-slag-ingot system. Based on these assumptions, the 2-D axisymmetrical transient-state numerical model accounts for electromagnetic phenomena and coupled heat and momentum transfers, to simulate the continuous growth of the electroslag remelted ingot and the solidification of the metal and slag. Recent studies on the ESR process are challenging the insulated mold hypothesis. Therefore, the main objective of the thesis is to acknowledge and study the existence of a mold current during an ESR remelting. A first model has been set-up, aimed to simulate the electromagnetic phenomena in the whole system for a simplified geometry. The possibility of the existence of such a mold current was confirmed. Based on this work, a fully-coupled model has then been developed and the results have been compared with experimental data to check the validity of the modifications. The influence of slag properties and operating parameters on the final quality of the ingot has been tested.

Un moteur Ericsson est un moteur alternatif à apport de chaleur externe fonctionnant selon un cyclethermodynamique de Joule. Ce type de moteur est particulièrement intéressant pour la valorisationde certaines "sources chaudes" (l'énergie solaire, la biomasse, les effluents gazeux chauds...).Dans cette thèse, une configuration innovante de moteur est proposée. Celle-ci permet d'une part,de s’affranchir du problème d’étanchéité autour des pistons des moteurs Ericsson, par l'usage despistons liquides en lieu et place des pistons mécaniques et d'autre part, de simplifier le système dedistribution mécanique des moteurs conventionnels par l'usage de soupapes commandéesparticulières, dont l'ouverture est déclenchée par contact avec le piston. Ce type de moteur estadapté pour la production d'électricité de petite puissance (jusqu'à... 10 kW).Trois lois différentes de commande des soupapes du cylindre de détente du moteur à air chaudproposé sont étudiées et leurs influences sur la conception et les performances énergétiques dusystème sont présentées. En raison de la masse importante d'eau dans le système, un modèle quiprend en compte la dynamique des colonnes de liquide est développé. Ce modèle permet de prédireles performances d'un premier prototype expérimental.Un prototype de machine de détente a ensuite été conçu et réalisé dans notre laboratoire. Uneprésentation du prototype et du banc d'essai est faite, et les résultats d'un essai préliminaire sontprésentés. Malgré le caractère préliminaire de ces résultats, ils sont très encourageants car d'unepart, ils n'ont pas révélés de problèmes techniquement délicats à résoudre, et d'autre part ils ontpermis de tirer de nombreux enseignements pour la suite des travaux à mener sur le prototype. / An Ericsson engine is a reciprocating external heat input engine operating on a Joule thermodynamiccycle. This type of engine is particularly interesting for the valorization of certain "hot sources" (solarenergy, biomass, hot gaseous effluents ...).In this thesis, an innovative engine configuration is proposed. This allows, on the one hand, toovercome the problem of tightness around the pistons of Ericsson engines, by the use of liquidpistons instead of mechanical pistons and on the other hand, to simplify the mechanical distributionsystem of conventional engines by the use of particular actuated valves, whose opening is triggered by contact with the piston. This type of motor is suitable for the production of electricity of smallpower (up to... 10 kW).Three different valve command laws for the proposed hot air motor expansion cylinder are studiedand their influences on the design and energy performance of the system are presented. Due to thelarge amount of water in the system, a model that takes into account the dynamics of the liquidcolumns is developed. This model makes it possible to predict the performances of a firstexperimental prototype.A prototype of the expansion machine was then designed and built in our laboratory. The prototypeand the test bench are presented together with the preliminary test results. Despite the preliminarynature of these results, they are very encouraging because, on the one hand, they have not revealedany technically difficult problems to be solved, and on the other hand they have made it possible todraw many lessons for the further work to be done lead on the prototype.

Les inductances et interconnexions des composants passifs intégrés pour la téléphonie mobile, sont sujettes à des défaillances dues à l’électromigration et l’auto-échauffement. L’électromigration n’est pas un risque majeur, au regard des grandes dimensions de cuivre de la technologie étudiée et de l’application. L’auto-échauffement est, en revanche, le principal phénomène qui limite le choix des dimensions des inductances et interconnexions lors de leurs conceptions.L’effet Joule pour les interconnexions et les inductances, a été étudié par le biais de caractérisations et de simulations électrothermiques. La méthodologie des plans d’expériences a été utilisée afin de modéliser le comportement électrothermique des inductances et des interconnexions. Un modèle prédictif de l’auto-échauffement en fonction des dimensions et de l’intensité, a été déterminé permettant d’étudier et de déterminer l’influence de chaque facteur dimensionnel, en régime continu. En radiofréquence, une méthodologie de mesure de l’auto-échauffement a été déterminée permettant sa caractérisation sur des composants sur plaquette. Une corrélation entre les régimes continus et alternatifs ne donnant pas de résultat concluant, une méthodologie de couplage faible, entre un simulateur électromagnétique et électrothermique a été effectuée, permettant la simulation du phénomène d’auto-échauffement sous contrainte radiofréquentielle. / Electrothermal and electromigration failure are likely to occur on copper inductor and interconnection in integrated passive devices for wireless telephony application. Electromigration is not a concern considering the high thickness of the copper and the application but the Joule heating is the main restriction on the dimensions during the component design. Joule heating on interconnections and inductors has been studied by electrothermal characterization and simulation. We have shown that Joule heating depends of several parameters such as material layers parameters and component dimensions. Design of experiments methodology has been used in order to model the inductor and interconnection electrothermal behavior. A self-heating predictive model has been determined allowing the study and the determination of dimensions impact in direct current.In radiofrequency, a self-heating measurement methodology has been determined allowing its characterization at wafer level. A correlation between direct current and radiofrequency is not satisfactory and a weak coupling between an electromagnetic and electrothermal simulator has been performed, providing the self-heating simulation under radiofrequency stress.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The possibility of exploring the magnetic properties of amorphous microwires in technological applications has attracted the attention of the scientific community in the last years. Beyond, the magnetoimpedance effect has been established as a powerful tool to study these properties. In this work we show a study of the magnetic properties of Joule heated glasscovered amorphous microwires, CoFeSiB, under applied stress. The stress modifies the anisotropy and the domain structure of the sample being studied by the magnetoimpedance measurements and the FMR dispersion relations. It was established a method to determine the permeability from the magnetoimpedance measurements, valid for a wide frequency range. From the study of the permeability curves parameters associated to the domain walls dynamic can be determined. From the FMR dispersion relations, the value and the direction of the transverse anisotropy field, as well as an estimative of the magnetic domain structure in these microwires were obtained. It can be concluded that the longitudinal anisotropy has the main role in the magnetic behavior of the microwire without stress. Also, it was verified the presence of an inner core with longitudinal anisotropy surrounded by an outer shell with circumferential anisotropy under applied stress. These magnetic configurations can be explained in terms of the frozen stress in the production process of the wire and the additional applied stress. / A possibilidade de explorar as propriedades magnéticas de microfios amorfos em aplicações tecnológicas tem atraído a atenção da comunidade científica nos últimos anos. Além disso, a magnetoimpedância foi estabelecida como uma ferramenta poderosa para estudar estas propriedades. Neste trabalho apresentamos um estudo das propriedades magnéticas de microfios amorfos, CoFeSiB, recobertos por vidro, tratados termicamente e sob aplicação de tensão. A tensão modifica a anisotropia e a estrutura de domínios presente na amostra de tal maneira que podem ser estudadas a partir das medidas de magnetoimpedância e da relação de dispersão de FMR (Ressonância Ferromagnética) extraída delas. Foi estabelecido um método para a determinação da permeabilidade a partir de medidas de magnetoimpedância, válido para uma ampla faixa de freqüências. A partir do estudo dessas curvas de permeabilidade pode-se determinar, por exemplo, parâmetros relativos à dinâmica de paredes de domínios. Do ajuste das relações de dispersão de FMR foram obtidos o valor e a orientação do campo de anisotropia transversal, bem como uma estimativa da estrutura de domínios dos microfios. Destes estudos, concluiu-se que a anisotropia longitudinal domina o comportamento magnético do microfio sem aplicação de tensão. Aplicando-se tensão aos microfios, verificou-se a presença de um núcleo interno com anisotropia longitudinal envolto por uma casca externa com anisotropia circunferencial. Estas configurações magnéticas podem ser explicadas em termos da tensão residual que surge no processo de produção do fio e da tensão adicional aplicada.

Metallic materials play important roles in providing electrical, thermal, and mechanical functions in electronic devices and systems. The understanding of the electrical-thermal-mechanical interaction caused by the passage of electric current with high density is important to improve the performance and reliability of electronic assembly and packaging. The electromechanical interaction on the deformation behavior of copper and tin is studied in this work. The electromechanical response of Cu strips was studied by passing a DC electric current. The electric resistance linearly increased with time before the occurrence of electric fusing. The electrothermal interaction led to the buckling of the Cu strips with the maximum deflection increasing with the increase of the electric current density. The total strain was found to be proportional to the square of the electric current density. A power law relation was used to describe the dependence of the time-to-fusing on the electric current density. Using the nanoindentation technique, the effect of electric current on the indentation deformation of copper and tin was studied. The reduced contact modulus of copper and tin decreased with increasing the electric current density. With the passage of a DC electric current, the indentation hardness of copper increased slightly with increasing electric current density. With the passage of an AC electric current, the indentation hardness of copper decreased with increasing the indentation deformation. With the passage of a DC electric current, the indentation hardness of tin decreased with increasing the indentation load, showing the normal indentation size effect. Both the limit of infinite depth and the characteristic length were dependent on the electric current density. Using the tensile creep technique, the creep deformation of pure tin was studied with the passage of a DC electric current. The steady state creep rate increased with the increase in temperature, tensile stress and electrical current density. For the same tensile stress and the same chamber temperature, the steady state creep rate increased linearly with the square of the electric current density. The electric current density has no significant effect on the stress exponent and activation energy of the tensile creep of tin for the experimental conditions.

La mise en œuvre des réfrigérateurs à dilution implique la consommation d'4He liquide. Ces réfrigérateurs ne sont pas autonomes et leur utilisation devient ainsi contraignante et onéreuse lorsque les laboratoires ne disposent pas de système de reliquéfaction.<br />A la fin des années 1990, les tubes à gaz pulsé bi-étagés commerciaux ont atteint des performances telles, que l'idée de les utiliser à la place des bains d'4He classiques commençé à émerger. Quelques prototypes de laboratoire furent ainsi développés, dont un exemplaire commercialisé par l'Air Liquide en collaboration avec le CNRS-CRTBT.<br />Cette thèse, réalisée au CRTBT et cofinancée par l'Air Liquide, a permis d'étudier le couplage d'un réfrigérateur à dilution avec un tube à gaz pulsé afin de réaliser un appareil commercial offrant des performances comparables, en volume expérimental et puissance froide, aux réfrigérateurs classiques de moyenne puissance.<br />Dans un premier temps, nous avons porté notre attention sur la réalisation du cryostat. L'absence de fluides cryogéniques permettant la thermalisation homogène des diverses entités, ainsi que la puissance frigorifique limitée par les performances du tube à gaz pulsé et les vibrations générées par ce dernier constituaient les principales caractéristiques de notre étude.<br />Nous avons réalisé un cryostat offrant un volume expérimental de 750 mm sur un diamètre de 200 mm. Le temps de prérefroidissement est de 12 heures entre 300 K et 4,2 K.<br />Dans un second temps, nous avons étudié l'intégration du réfrigérateur à dilution.<br />Le prérefroidissement de l'injection d'3He entre 300 K et 4,2 K nous a amenés à développer un système de thermalisation spécifique permettant d'optimiser l'utilisation des ressources du tube à gaz pulsé. Ce développement a abouti à un brevet CNRS-Air Liquide. <br />Le prérefroidissement entre 4,2 K et la température de l'évaporateur a constitué un travail important concernant l'étude du principe de refroidissement par échangeurs Joule-Thomson.<br />Nous avons ainsi atteint des températures inférieures à 10 mK et obtenu des débits de circulation allant de 90 à 200 micromoles par seconde avec des temps de condensation de l'ordre de 2 heures pour un réfrigérateur à deux échangeurs discrets.

Étude du traitement des siloxanes par adsorption sur matériaux poreux : application au traitement des biogaz Les biogaz sont des mélanges à forte teneur en méthane utilisés dans la production de chaleur ou d'électricité. Ils contiennent des quantités plus ou moins importantes de siloxanes qui sont interdits pour de nombreuses utilisations du biogaz. La possibilité d'élimination des siloxanes par adsorption est étudiée. Une étude en réacteurs statiques permet d'évaluer les capacités d'adsorption des charbons actifs, de zéolithes et du gel de silice. L'influence sur les capacités de traitement des siloxanes, de la présence de CH4, de CO2, de l'humidité, et d'autres composés organiques volatils dans la matrice gazeuse, est étudiée. Des bonnes capacités d'adsorption de certains adsorbants sont retrouvées. Dans les conditions d'adsorption les plus défavorables on obtient une diminution de 20 % sur les capacités d'adsorption. Les capacités de traitement restent modestes. L'influence des conditions opératoires du procédé est étudiée dans le but d'augmenter les capacités de traitement et réduire la quantité d'adsorbant. Les résultats nous ont permis de définir, valider et dimensionner un système de traitement à échelle réduite. Le procédé envisagé consiste en une adsorption sur tissu de carbone activé, alternant avec des phases de régénération thermique. Une unité pilote a permis de réaliser et d'évaluer le vieillissement du procédé. L'ensemble de ce travail laisse entrevoir des possibilités d'applications industrielles du procédé, même si une phase d'essai sur site en conditions réelles est encore nécessaire.

Infrared (IR) imaging is a key technology in a variety of military and civilian applications, especially for night vision and remote sensing. Compared with cryogenically cooled IR sensors, uncooled infrared imaging devices have the advantages of being low cost, light weight, and superior reliability. The electro-thermal analysis of a microbolometer pixel is critical to determine both device performance and reliability. To date, most microbolometer analysis research has focused on performance optimization and computation of thermal conductance directly from the geometry. However, modeling of the thermal distribution across the microbolometer pixel is critical for the comprehensive analysis of system performance and reliability. Therefore, this thesis investigates the electro-thermo-mechanical characteristics of a microbolometer pixel considering the effects of joule heating and incoming IR energy. The contributions of the present research include the electro-thermal models for microbolometer and methods of validating thermal distribution using experimental results. The electro-thermal models explain the effect of microbolometer material properties and geometry on device performance and reliability. The research also contributes methods of estimating the thermal conductivity of microbolometer, which take into account different heat transfer mechanisms, including radiation and convection. Previous approaches for estimating the thermal conductance of uncooled microbolometer consider heat conduction via legs from the geometry of the pixel structure and material properties [2]. This approach assumes linear temperature distribution in the pixel legs structure. It also leaves out the various electro-thermal effects existing for multilayer structures. In the present research, a different approach is used to develop the thermal conductance of microbolometer pixel structure. The temperature distribution in the pixel is computed from an electro-thermal model. Then, the average temperature in the pixel microplate and the total heat energy generated by joule heating is utilized to compute the thermal conductance of the structure. The thesis discusses electro-thermal and thermo-mechanical modeling, simulation and testing of Polysilicon Multi-User MEMS Process (PolyMUMPs®) test devices as the groundwork for the investigation of microbolometer performance and reliability in space applications. An electro-thermal analytical and numerical model was developed to predict the temperature distribution across the microbolometer pixel by solving the second order differential heat equation. To provide a qualitative insight of the effect of different parameters in the thermal distribution, including material properties and device geometry, first an explicit formulation for the solution of the electro-thermal coupling is obtained using the analytical method. In addition, the electro-thermal model, which accounts for the effect of IR energy and radiation heat transfer, spreading resistance and transient conditions, was studied using numerical methods. In addition, an analytical model has been developed to compute the IR absorption coefficient of a Thin Single Stage (TSS) microbolometer pixel. The simulation result of this model was used to compute absorbed IR energy for the numerical model. Subsequently, the temperature distribution calculated from the analytical model is used to obtain the deflections that the structure undergoes, which will be fundamental for the reliability analysis of the device. Finite element analysis (FEA) has been simulated for the selected device using commercial software, ANSYS® multiphysics. Finite element simulation shows that the electro-thermal models predict the temperature distribution across a microbolometer pixel at steady-state conditions within 2.3% difference from the analytical model. The analytical and numerical models are also simulated and results for a temperature distribution within 1.6% difference. In addition, to validate the analytical and numerical electro-thermal and thermo-mechanical models, a PolyMUMPs® test device has been used. The test results showed a close agreement with the FEM simulation deflection of the test device.

Abstract Dynamic electrical resistance during resistance spot welding has been quantitatively modeled and analyzed in this work. A determination of dynamic resistance is necessary for predicting the transport processes and monitoring the weld quality during resistance spot welding. In this study, dynamic resistance is obtained by taking the sum of temperature dependent bulk resistance of the workpieces and contact resistances at the faying surface and electrode-workpiece interface within an effective area corresponding to the electrode tip where welding current primarily flows. A contact resistance is composed of constriction and film resistances, which are functions of hardness, temperature, electrode force, electrical resistivity and surface condition. Unsteady, axisymmetric transport of mass, momentum, energy, species, and magnetic field intensity with a mushy-zone phase change in workpieces and temperature, and magnetic fields in electrodes during resistance spot welding, are systematically investigated. Electromagnetic force, joule heat, heat generation at the electrode-workpiece interface and faying surface between workpieces, different properties between phase, and geometries of electrodes are taken into account. The predicted nugget thickness and dynamic resistance versus time show quite good agreement with available experimental data. Excluding expulsion, the dynamic resistance curve can be divided into four stages. A rapid decrease of dynamic resistance in stage 1 is attributed to decreases in film resistances at the faying surface and electrode-workpiece interface. In stage 2, the increase in dynamic resistance results from the primary increase of bulk resistance in the workpieces and an increase of the sum of contact resistances at the faying surface and electrode-workpiece interface. Dynamic resistance in stage 3 decreases, because increasing rate of bulk resistance in the workpieces and contact resistances decrease. In stage 4 decrease of dynamic resistance is mainly due to the formation of the molten nugget at the faying surface. The molten nugget is found to occur in stage 4 rather than stage 2 or 3 as qualitatively proposed in the literature. The effects of different parameters on the dynamic resistence curve are also presented. Besides, electromagnetic force effect on velocity field of molten nugget was proven to be crucial. Higher current, smaller magnetic diffusivity and decreasing the radius of electrode tip will lead to high current density around the corner between electrode and workpiece. Sometimes the corner of electrode and surface of workpieces will be melted due to local high current density.

This research presents a detection scheme for analyzing the temperature distribution produced by the Joule heating effect nearby the channel wall in a microfluidic chip utilizing a temperature-dependent fluorescence dye. An advanced optical microscope system¡Xtotal internal reflection fluorescence microscope (TIRFM) is used for measuring the temperature distribution on the inner channel wall at the point of electroosmotic flow in an electrokinetically driven microfluidic chip. In order to meet the short working distance of the objective-type TIRFM, microscope cover glass are used to fabricate the microfluidic chips. The short fluorescence excitation depth from a TIRFM makes the intensity information obtained is not sensitive to the channel depth variation which ususally biases the measured results while using conventional epi-fluorescence microscope (Epi-FM). Therefore, a TIRFM can precisely describe the temperature profile of the distance within hundreds of nanometer of the channel wall where consists of the Stern layer and the diffusion layer for an electrokinetic microfluidic system. In order to investigate the temperature distribution produced by the Joule heating effect for electrokinetically driven microchips, this study not only measures the temperature on the microchannel wall by the proposed TIRFM but also measures the temperature inside the microchannel by an Epi-FM. In addition, this research presents a method to reduce the Joule heating effect and enhance the separation efficiency of DNA biosamples in a chip-based capillary electrophoresis (CE) system utilizing pulse DC electric fields. Since the average power consumption is reduced by the pulse electric fields, the Joule heating effect can be significantly reduced. Results indicate the proposed TIRFM method provides higher measurement sensitivity over the Epi-FM method. Significant temperature difference along the channel depth measured by TIRFM and Epi-FM is experimentally observed. In addition, the measured wall temperature distributions can be the boundary conditions for numerical investigation into the Joule heating effect. The proposed method gives a precise temperature profile of microfluidic channels and shows the substantial impact on developing a simulation model for precisely predicting the Joule heating effect in microfluidic chips. Moreover, in the research of reducing the Joule heating effect and enhancing the separation efficiency in a chip-based CE system utilizing pulse electric fields, the experimental and numerical investigations commence by separating a mixed sample comprising two fluoresceins with virtually identical physical properties. The separation level is approximately 2.1 times higher than that achieved using a conventional DC electric field. The performance of the proposed method is further evaluated by separating a DNA sample of Hae III digested £XX¡V174 ladder. Results indicate the separation level of the two neighboring peaks of 5a (271 bp) and 5b (281 bp) in the DNA ladder is as high as 120% which is difficult to be achieved using a conventional CE scheme. The improved separation performance is attributed to a lower Joule heating effect as a result of a lower average power input and the opportunity for heat dissipation during the zero-voltage stage of the pulse cycle. Overall, the results demonstrate a simple and low-cost technique for achieving a high separation performance in CE microchips.