Lithium-ion (Li-ion) batteries are critically important for portable electronics, electric vehicles, and grid-level energy storage. The development of next-generation Li-ion batteries requires high-capacity electrodes with a long cycle life. However, the high capacity of Li storage is usually accompanied by large volume changes, dramatic morphological evolution, and mechanical failures in the electrodes during charge and discharge cycling. To understand the degradation of electrodes and resulting loss of capacity, this thesis aims to develop mechanistic-based models for predicting the chemo-mechanical processes of lithiation and delithiation in high-capacity electrode materials. To this end, we develop both continuum and atomistic models that simulate mass transport, interface reaction, phase and microstructural evolution, stress generation and damage accumulation through crack or void formation in the electrodes. The modeling studies are tightly coupled with in-situ transmission electron microscopy (TEM) experiments to gain unprecedented mechanistic insights into electrochemically-driven structural evolution and damage processes in high-capacity electrodes. Our models are successfully applied to the study of the two-phase lithiation and associated stress generation in both crystalline and amorphous silicon anodes, which have the highest known theoretical charge capacity, as well as the lithiation/sodiation-induced structural changes and mechanical failures in silicon-based multilayer electrodes. The modeling studies have uncovered unexpected electrochemical reaction mechanisms and revealed novel failure modes in silicon-based nanostructured anodes. Our modeling research provides insights into how to mitigate electrode degradation and enhance capacity retention in Li-ion batteries. More broadly, our work has implications for the design of nanostructured electrodes in next-generation energy storage systems.

Philosophiae Doctor - PhD / Fuel cells are long term solution for global energy needs. In current fuel cell technologies, Proton Exchange Membrane (PEM) fuel cells are known for quick start-up and ease of operation compared to other types of fuel cells. Operating PEM fuel cells at high temperature show promising applications for stationary combined heat and power application (CHP). The high operating temperature up to 160°C allows waste heat to be recovered for co-generation or tri-generation purposes. The commercially available PEM fuel cells operating at 160⁰C can tolerate up to 3% CO without significant loss of performance, making HT-PEM fuel cell viable choice when reformate is used. In reality these advantages convert to very little balance-of-plant compared to Nafion® based fuel cells operating at 60°C. However, there are some problems that prevent high temperature fuel cells from large scale commercialization. The cathode is said to have sluggish reaction kinetics and high cell potentials and operating temperature during fuel cell start-up may cause severe degradation. The formation of liquid water during the shut-down can cause the phosphoric acid to leach from the cell during operation. Efforts are being made to reduce the cost and increase the durability of fuel cell components (such as catalyst and membrane) at high temperatures. Apart from degradation issues, the problems are related to cost and performance. The performance of the PEM fuel cells depends on a lot of factors such as fuel cell design and assembly, operating conditions and the flow field design used on the cathode and anode plates. The flow field geometry is one important factor influencing the performance of fuel cells. The flow fields have significant effect on pressure and flow distribution inside the fuel cell. A homogeneous distribution of the reactant gases over the active catalyst surface leads to improved electrochemical reactions and thus enhances the performance of the fuel cell. So, the design of flow fields is one of the important issues for performance improvement of PEM fuel cell in terms of power density and efficiency. There are different types of flow fields available for PEM fuel cells such as serpentine, pin, interdigitated and straight flow fields but the most obvious choice is multiple serpentine. The same can be used for high temperature PEM fuel cell (HT-PEMFC) application with ease because of absence of liquid water during the high temperature operation and no need for complex water management.

Les travaux de these s’inscrivent dans le cadre de deux projets: ReAGaN et ExtremeGaN avec des industriels (UMS, Serma Technologies, Thales TRT) et des laboratoires derecherche (LEPMI, LAAS et l’université de Bristol).Les deux technologies GaN (GH50 et GH25) étudiées dans cette thèse sont fournies parla société United Monolithic Semiconductors (UMS) et elles ont été qualifiées durant cettethèse. Plusieurs composants ont subi des tests de vieillissement accéléré en températureréalisés par UMS, ensuite une campagne de caractérisation électrique approfondie a étéréalisée au laboratoire IMS afin d’étudier les effets parasites et les mécanismes de dégradationqui limitent la fiabilté de cette filière.Le premier chapitre traite les bases du transistor HEMT à base de GaN. Les avantagesdu matériau nitrure de gallium ainsi que les substrats adaptés au HEMT à base de GaN sontprésentés. Une brève description du marché europeen des composants GaN est donnée.Ensuite, la structure ainsi que le fonctionnement du HEMT AlGaN/GaN sont décrit ainsi queles deux technologies d’UMS.Le deuxième chapitre présente les tests de vieillissement utilisés pour l’analyse defiabilité. Ensuite, un état de l’art des effets parasites et des mécanismes de dégradation desHEMTs AlGaN/GaN est donné. Le projet ReAGaN est décrit et les différentes techniques decaractérisation utilisées durant les travaux de cette thèse sont présentées.Le troisième chapitre est divisé en quatre études de cas ; les trois premières sont dans lecadre du projet ReAGaN et la quatrième dans le cadre du projet Extreme GaN. Dans lapremière étude de cas, les mécanismes de conduction qui augmentent les courants de fuitesdes HEMTs AlGaN/GaN issus de la technologie GH50 ont été étudiés. La deuxième étude decas est dédiée à l’étude d’un effet parasite électrique qui apparait après un vieillissementaccéléré en température sur la caractéristique de la diode Schottky en polarisation directe.Dans la troisième étude de cas, l’influence de la variation de la fraction molaire des HEMTsAlGaN/GaN sur les paramètres électriques a été analysée. La dernière étude de cas consiste enla détermination des limites de fonctionnement et l’aire de sécurité de la technologie GH25d’UMS en réalisant les mesures des lieux de claquage en mode diode et en mode transistor. / This thesis is in the framework of two projects: ReAGaN and Extreme GaN withindustrials (UMS, Serma Technologies, Thales TRT) and academics (LEPMI, LAAS andUniversity of Bristol).The studied AlGaN/GaN HEMTs are provided by the society United MonolithicSemiconductors (UMS) from the GH50 and GH25 GaN processes that were qualified duringthis thesis. Many devices were submitted to high temperature accelerated life tests by UMSand characterized at IMS laboratory to study the parasitic effects and degradationsmechanisms that are limiting the electrical reliability of GaN based HEMTs technology.The first chapter gives an overview of the basics of GaN based high electron mobilitytransistors (HEMTs). Gallium Nitride material features are reviewed as well as substratessuited for GaN based devices. GaN market in Europe and the main industrial actors are listed.Furthermore, the structure and operation of GaN based HEMTs are described. In the last part,the two UMS GaN processes are described.The second chapter presents the life tests that are used for reliability studies. State of theart of parasitic effects and degradation mechanisms of AlGaN/GaN HEMTs is given.Furthermore, the ReAGaN project in which the main part of this thesis is involved isdescribed. The electrical characterization techniques used at IMS during this thesis arepresented.The third chapter is divided into four case studies; three case studies are in theframework of ReAGaN project and the fourth one in the Extreme GaN project. In the firstcase study, we investigate the conduction mechanisms inducing the leakage current inAlGaN/GaN HEMTs issued from GH50 process. The second case study is dedicated to thestudy of an electrical parasitic effect that appears on the Schottky diode forward characteristicafter temperature accelerated life tests. In the third case study, we study the influence of Almole fraction on the DC electrical parameters of AlGaN/GaN HEMTs. The last case studyconsists in the determination of the limits and safe operating area (SOA) of UMS GH25 GaNHEMTs by carrying out the two and three terminal breakdown voltages measurements.

Les composés organiques carbonylés et les BTEX (Benzène, Toluène, Éthylbenzène et Xylènes) représentent une classe importante de composés organiques volatils dans l’atmosphère. Ils sont émis par des sources anthropogénique et biogéniques. Leur dégradation atmosphérique conduit à la formation d’ozone, de phooxidants et d’aérosols organiques affectant ainsi la qualité de l’air aux échelles locales et régionales ainsi que la santé humaine. Il est donc important de mesurer leurs concentrations et évaluer leur devenir atmosphérique. Dans la présente thèse, nous avons conduit une étude systématique qui a permis de mesurer les concentrations de ces composés et identifier leurs sources à Pékin (Juillet 2008-Août 2010) et évaluer l’importance des caractéristiques météo. Nous avons aussi mené des études sur la dégradation atmosphérique de trois formates (isoproyle, isobutyle et n-propyle) en utilisant la chambre de simulation atmosphérique d’ICARE (CNRS, Orléans). / Carbonyls and BTEX (Benzene, Toluene, Ethylbenzene, and Xylenes) represent an important class of VOCs (volatile organic compounds) in the atmosphere. They are emitted into the atmosphere through anthropogenic and biogenic sources. Their atmospheric degradation leads to the formation of ozone, photooxidants and organic aerosols affecting the air quality at the local and regional scales and human health. It is, hence, of importance to measure their atmospheric concentrations and investigate their fate. In the present thesis, we have conducted a systematic measurement study of carbonyls and BTEX in Beijing during the period of Jul 2008-Aug 2010 in order to evaluate their ambient levels, possible sources and the influence of characteristic weather conditions. In a separate work, we performed a series of experimental studies on the OH-initiated oxidation of isopropyl formate, isobutyl formate, and n-propyl isobutyrate using the ICARE-CNRS (Orleans) simulation chamber from which we derived the product yields. The data obtained are presented and discussed.

Experimental and theoretical analyses have been conducted upon electrical connector contacts under low frequency fretting conditions. The phenomena of "fretting" -relative micromovements in the components parts of an electrical contact - is known to take a major role in the degeneration of electrical contacts. Low frequency fretting is of particular interest and is typically caused by thermal differential expansion of the component parts due to temperature changes in the environment or the device itself. This thesis begins with a survey of possible failure mechanisms of the contact system. These are analysed and classified into three groups of chemical, physical, and mechanical degradation mechanisms. Fretting has been classified under the mechanical mechanism of degradation but is reported to exacerbate other degradation mechanisms resulting in the phenomena of "fretting corrosion". Developments in contact technology are then surveyed with the emphasis of this study on lubrication of the contact system. A novel fretting simulation apparatus has been developed to study the degradation mechanisms upon the contact system resulting from low frequency micromovements. The study includes investigations on the simulation apparatus of the contact system under different conditions. Particular! y emphasis is given to contacts under the conditions of electrical load and lubrication. Novel trends in the contact performance are reported with respect to electrical contact resistance, corrosion and wear of the contact interface. It is shown that electrically loading or lubricating the contact system has dramatic effects upon the contact performance. Chemical, physical and mechanical mechanisms at the contact interface are presented to explain the contact behaviour under several conditions of low frequency fretting. A "Two Process Model" is proposed which summarises the interaction of these different mechanisms. This model consists of two processes in balance - contact cleaning and contact degradation processes - which either increases or reduces the electrical area of contact. A theoretical computer simulation model is proposed for evaluating contact resistance behaviour of a fretting contact system under several conditions, particularly conditions of electrically loaded and lubricated conditions. Chemical, physical and mechanical mechanisms and their interact are simulated in the model using the Monte Carlo technique.

This work deals with problems of degradation mechanisms of lead-acid batteries. For a better understanding of the diverse effects that cause the degradation are analyzed and the physical explanation of each of the degradation mechanisms. The paper describes the measuring method by which they examined the different manifestations of degradation mechanisms. At the conclusion of works are carried out measurements in which it is possible to trace the manifestations of the various degradation mechanisms in lead battery taking place mainly on the active surface electrodes.

In this thesis we look at one of the aging mechanisms that may have affected current aged Nuclear Power Plants (NPPs). Irradiation as an age-related degradation mechanism is studied for Reinforced Concrete (RC) in NPPs. This problem can be important for aged reactor buildings, radwaste buildings, spent nuclear fuel, research reactors, or accelerators that experience high levels of radiation close to existing thresholds. Mechanical properties of concrete are the most important parameters affected by radiation in NPPs. Compressive strength of concrete is reduced between 80 and 35 \% for radiation fluences between $2\times 10^{19}$ and $2\times 10^{21} n/cm^2$. Tensile strength reduction is more significant than compressive strength. It is reduced between 20 and 80 \% for a radiation fluence equal to $5\times 10^{19}$. We chose three radiation levels $2\times 10^{19}$, $2\times 10^{20}$, $2\times 10^{20}$ based on experimental results as the critical levels of radiation that RC structures in NPPs may be exposed to. Structures susceptible to the problem are mostly RC walls; so the RC panel is chosen as an appropriate representative scale element for the analysis. The effect of radiation on mechanical properties of concrete is considered to analyze degraded scale elements. Material properties, geometry, and loading scenarios of scale elements are selected to be close to actual quantities in existing nuclear power plant. Elements are analyzed under six types of loading combination of shear and axial loading conditions. A nonlinear finite element program, Membrane-2000, based on the Modified Compression Field Theory (MCFT) is used to solve scale elements numerically. Element behaviors are studied considering the factors influence ultimate strength capacity, failure mode, and structural ductility index of members. The results show that ultimate shear capacity of the elements subjected to combinations of shear and tension loading are reduced significantly for highly reinforced elements ($1.35<\rho<1.88$) in $2\times 10^{21} n/cm^2$ radiation. RC panels under shear-biaxial and uniaxial compression also show significant strength capacity reduction in radiation levels $2\times 10^{20} n/cm^2$ and $2\times 10^{21} n/cm^2$, respectively. Failure modes of the elements change from yielding of steel to shear failure by increasing level of degradation for the elements with reinforcement ratio between 0.9 and 1.88. Ductility of the RC panels is reduced significantly in the critical levels of radiation. Ductility of the elements became less than the allowable ductility value by increasing level of radiation.

In this thesis we look at one of the aging mechanisms that may have affected current aged Nuclear Power Plants (NPPs). Irradiation as an age-related degradation mechanism is studied for Reinforced Concrete (RC) in NPPs. This problem can be important for aged reactor buildings, radwaste buildings, spent nuclear fuel, research reactors, or accelerators that experience high levels of radiation close to existing thresholds. Mechanical properties of concrete are the most important parameters affected by radiation in NPPs. Compressive strength of concrete is reduced between 80 and 35 \% for radiation fluences between $2\times 10^{19}$ and $2\times 10^{21} n/cm^2$. Tensile strength reduction is more significant than compressive strength. It is reduced between 20 and 80 \% for a radiation fluence equal to $5\times 10^{19}$. We chose three radiation levels $2\times 10^{19}$, $2\times 10^{20}$, $2\times 10^{20}$ based on experimental results as the critical levels of radiation that RC structures in NPPs may be exposed to. Structures susceptible to the problem are mostly RC walls; so the RC panel is chosen as an appropriate representative scale element for the analysis. The effect of radiation on mechanical properties of concrete is considered to analyze degraded scale elements. Material properties, geometry, and loading scenarios of scale elements are selected to be close to actual quantities in existing nuclear power plant. Elements are analyzed under six types of loading combination of shear and axial loading conditions. A nonlinear finite element program, Membrane-2000, based on the Modified Compression Field Theory (MCFT) is used to solve scale elements numerically. Element behaviors are studied considering the factors influence ultimate strength capacity, failure mode, and structural ductility index of members. The results show that ultimate shear capacity of the elements subjected to combinations of shear and tension loading are reduced significantly for highly reinforced elements ($1.35<\rho<1.88$) in $2\times 10^{21} n/cm^2$ radiation. RC panels under shear-biaxial and uniaxial compression also show significant strength capacity reduction in radiation levels $2\times 10^{20} n/cm^2$ and $2\times 10^{21} n/cm^2$, respectively. Failure modes of the elements change from yielding of steel to shear failure by increasing level of degradation for the elements with reinforcement ratio between 0.9 and 1.88. Ductility of the RC panels is reduced significantly in the critical levels of radiation. Ductility of the elements became less than the allowable ductility value by increasing level of radiation.

Ce travail porte sur l’analyse du comportement en durabilité d’une cellule à oxydes solides en fonctionnement en mode Electrolyse de la vapeur d’eau à Haute Température (EHT). Il s’agit plus particulièrement d’identifier les paramètres influant sur la durabilité de la cellule et de comprendre les mécanismes associés, de manière à établir le meilleur compromis durabilité-performance. Un premier axe de recherche correspond à la caractérisation électrochimique initiale de la mono-cellule choisie pour l’étude, à savoir une cellule à électrode support de type LSFC/YDC/8YSZ/Ni-YSZ dont le comportement électrochimique est étudié sous diverses températures, conditions de gaz et densités de courant. Ensuite, la dégradation des performances est étudiée in situ par spectroscopie d’impédance électrochimique et voltampérométrie lors de diverses études séquentielles par paliers de 200 heures. L’influence du taux de conversion et de la densité de courant sur la dégradation des performances est ainsi analysée. Pour compléter cette étude, des analyses physico-chimiques et microstructurales post tests ont été réalisées sur des échantillons ayant fonctionné 1000 heures afin de faire le lien entre les modifications électrochimiques observées et les changements de structure et de composition des matériaux de cellule. Des hypothèses de mécanismes de dégradation ont ainsi pu être formulées. / This work is dedicated to the study on the long term behavior of a solid oxide electrolysis cell (SOEC). More specifically, the target is to identify and understand the influence of main working parameters on the cell durability in order to find the best compromise between performances and durability. The initial part of this work is to characterize electrochemically the single cell selected for the study, that is to say a cathode supported cell of the type LSFC/YDC/8YSZ/Ni-YSZ at different gas conditions, temperature and current density. Then, the degradation of the cell performances has been studied by Electrochemical Impedance Spectroscopy (EIS) and voltamperometry (i-V) curves during sequential tests of 200h. The influence of the most important working parameters like the current density and the steam conversion has been analyzed. To complete this study, physico-chemical and microstructural analyses have been performed on cells that have been operated over 1000 hours, to make a link between electrochemical degradation observed on the cell and composition and structural changes of cell materials.

De nos jours, les nouvelles technologies ne cessent d'évoluer et de former une partie intégrante dans la vie quotidienne de chacun. Ces dernières profitent du développement de systèmes électroniques complexes qui nécessitent l'utilisation de composants mémoires de plus en plus performants et présentant de grandes capacités de stockage. Ainsi, dans cette course à la miniaturisation, la technologie Flash jusqu'ici prépondérante sur le marché des mémoires non volatiles laisse aujourd'hui entrevoir ses limites. En conséquence, différentes mémoires émergentes résistives sont développées et parmi celles-ci se trouvent les mémoires à changement de phase PCRAM qui présentent un grand intérêt dans le monde des mémoires non volatiles grâce à leur bonne capacité de réduction d'échelle ainsi que leur coût réduit par rapport aux mémoires Flash. Cependant, pour être compétitives face aux autres technologies et pour prétendre à des applications embarquées, elles doivent répondre à plusieurs challenges tels que réduire leur courant de programmation, augmenter leur vitesse de programmation et améliorer leur stabilité thermique. Pour cela, différentes voies sont explorées dans la littérature, notamment l'utilisation d'architectures innovantes ou de matériaux à changement de phase alternatifs. Dans cette thèse, nous nous sommes intéressés à l'investigation des mécanismes de défaillance qui affectent la stabilité thermique et temporelle des mémoires à changement de phase, plus précisément la rétention de l'état RESET et la stabilité des états programmés affectée par le phénomène de « drift ». Le développement de matériaux alternatifs utilisant une stoechiométrie optimisée ou incorporant un dopage nous permet d'obtenir des dispositifs performants d'un point de vue électrique et présentant des propriétés de rétention satisfaisant les spécifications des applications embarquées en particulier l'automobile. De plus, grâce au développement d'une nouvelle procédure de pré-codage, ces dispositifs permettent de conserver les données préprogrammées sur la puce mémoire au cours de l'étape de soudure de cette dernière sur le circuit électronique. Ils constituent une solution prometteuse pour les applications de cartes sécurisées. Enfin, nous avons proposé une procédure de programmation optimisée qui permet de diminuer l'effet du drift de la résistance de l'état SET observé pour les matériaux alternatifs. Ensuite, nous avons montré via des mesures de bruit à basses fréquences que cet effet est dû à la relaxation structurale des zones amorphes présentes dans ces matériaux actifs. De plus, nous avons mis en évidence pour la première fois la diminution du bruit normalisé de l'état SET ainsi que l'influence majeure des défauts d'interfaces sur le bruit à basses fréquences de cet état. / Nowadays, new technologies are rising steadily and forming an integral part in the daily lives of everyone. They take advantage of the development of electronic systems for which the complexity requires the use of memory devices more and more efficient and with large storage capacities. Because of some performance degradation, the scaling of Flash technology who was so far predominant in the non-volatile memories market, is today reaching its limits. As a result, different emerging resistive memories are being developed. Among them, the phase-change memory technology PCRAM is very attractive because of its non-volatility, scalability, as well as reduced cost compared to standard Flash. Nevertheless, to compete with other technologies and to address the embedded applications market, PCRAM still face some challenges, such as decreasing the programming current densities, increasing the programming speed and increasing the thermal stability of the two memory states. For that purpose, different solutions have been tried in the literature, including using new device architectures and optimized phase-change materials. In this work, we are interested in investigating the failure mechanisms that affect thermal and temporal stability of phase change memories, in particular the retention of the RESET state and the stability of the programmed states disturbed by the drift phenomenon. The development of alternative materials using an optimized stoichiometry or incorporating doping allows us to achieve high electrical performance devices and to reach the required retention properties of embedded applications and particularly the automotive one. Moreover, thanks to the development of a new pre-coding procedure, these devices allow to keep stable the preprogrammed data on the memory chip during the soldering step of the latter on the electronic circuit. They represent a promising solution for Smart-Card applications. Finally, we have proposed an optimized programming procedure which enables to reduce the drift effect of the resistance of the SET state observed for optimized materials. This drift phenomenon was investigated by using low frequency noise measurements. Therefore, we have shown that this effect is due to the structural relaxation of amorphous parts in the active material. Besides, we highlighted for the first time the major influence of interface defects on the low-frequency noise of this state.