A Bubble is a biphasic object in every sense of the word. First, as a concept. Indeed, its production is within everyone’s reach as long as they have soap, a frame and their breath at their disposal. Yet, the physics behind their very existence and the phenomena associated with them remain puzzling questions that involve complex notions. Second, they are literally made up of two phases: a gaseous and a liquid. The combination, nature and distribution of these phases are responsible for the unique behaviour of bubbles. Among other things, bubbles are easily deformed by external fields, which make them a wonderful object to highlight their effects. In this work, we investigate the different effects that external fields and the body forces related to them have on the bubbles shapes. More precisely, we describe the deformation that bubbles undergo under an electric or a magnetic field. These reshaping are linked both to the applied fields properties and to those of the liquids used to make the bubbles. In addition, the substrate also appears to have an influence on the deformation. We show that these dependencies can be expressed by dimensionless numbers that describe the competition between the forces involved. Moreover, we find that the functions relating the shapes and these dimensionless numbers are linear under weak fields. We obtain these linear functions and, using them, we properly define what weak fields are. With these results, we are able to propose a general guide on how to handle, control and deform bubbles. Furthermore, by com- paring the deformation of bubbles and droplets under similar fields, we demonstrate that bubbles are a wonderful object to illustrate the very nature of the forces acting on them.

Magnetic properties of matter are essential for a wide range of current and future technologies, especially in the domain of microelectronics for which spintronics is largely involved in the next generations of devices. This dissertation, composed of two distinct parts, presents an original research work on microscopic magnetic devices allowing the control and detection of static magnetic fields (part I), and the transmission and sensing of pure spin currents (part II). Firstly, the development of metasurfaces composed of a concentric arrangement of micrometer-wide ferromagnetic petals and allowing the magnetic flux concentration is investigated. Micromagnetic simulations demonstrate the importance of the magnetic domains configuration on the linear response of the device. In the operating regime of the device, a concentrated magnetic field around two times the external field is predicted irrespective of the in-plane applied field direction. The experimental proof-of-concept is demonstrated with 60 nm-thick permalloy structures. The concentration gain is obtained by optically tracking the magnetic vortex at the center of a permalloy disk sensor, using Kerr microscopy. The second study presented in this thesis focuses on the sharp magnetoresistance changes, triggered by out-of-plane magnetic fields, probed in thin permalloy strips grown on monocrystalline lanthanum aluminate substrates. Micromagnetic simulations are used to evaluate the resistance changes of the strips at different applied field values and directions and correlate them with the magnetic domain distribution. The experimentally observed sharp magnetic switching, tailored by the shape anisotropy of the strips, is properly accounted for by the numerical simulations when considering an important substrate-induced uniaxial magnetic anisotropy with a main direction sligthly tilted from the out-of-plane direction. The second part of this thesis is devoted to non-local spin-valves made of ferromagnetic tunnel junctions and implemented for electron spin injection, transport and detection of pure spin currents. We first demonstrate that the non-linear electrical transport occurring in tunnel junctions may lead to a spin-to-charge conversion efficiency larger than 10 times the spin polarization of the tunnel barrier when the latter is under a bias voltage of a few millivolts. The underlying mechanisms are attributed to the tunnel-barrier deformation and the conduction-band shift resulting from a change of the applied voltage. An approximated analytical expression predicting the detector spin sensitivity is suggested. Calculations performed for different barrier shapes show that this enhancement is present in oxide barriers as well as in Schottky-tunnel barriers, and that it depends on the intensity of the spin accumulation generated in the channel. Moreover, although reduced at high temperatures, the spin signal remains superior to the value predicted by the linear model. Finally, we demonstrate that the Hanle precession method as conventionally applied is no longer accurate when the distance between the inner and outer electrodes becomes smaller than 6 times the spin diffusion length, leading to errors as large as 50% for the calculation of the spin figures of merit. We suggest simple but efficient approaches to circumvent this limitation by addressing a revised version of the Hanle fit function and by proposing a refined fabrication process for four-terminal non-local spin valves.

Nowadays, innovative applications such as "lab-on-a-chip", micro-reactors or biological chips are developed for industry, biology or medicine. Capillary issues are encountered which are not yet fully understood. It is therefore essential to open up ways to manipulate tiny amounts of liquid in order to mix them, encapsulate them or to create emulsions. The bouncing droplets on an oscillating liquid interface allow their manipulation. Indeed, under certain conditions, droplets can bounce indefinitely on a bath surface as long as the squeezed air film which separates the drop from the bath is renewed at each bounce. We chose to study deformable droplets on a non-deformable bath. The droplets are deformable because they are large or made of a low viscous oil, the liquid of the bath being highly viscous silicone oil. We investigated how the deforma- tions, the stability and the trajectories of the droplets depends on the forcing parameters such as the frequency and the amplitude of the oscillation. We also studied the possibility of mixing and emulsifying droplets on the bath. Finally, we showed the effect of the deformation of the bath on these bouncing droplets and studied the trajectories of walking droplets [1] which are used as a model for quantum-like particles [2-5]. In this manuscript, we reported an exploration of the droplets behavior as a function of the ability of the bath and/or the droplet to deform. Depending on the forcing frequency, specific deformation modes are excited on the bouncing droplets. We used these modes to rationalize their bouncing stability and to create double emulsions in a compound droplet. Then, we determined numerically the complex bifurcation diagrams of the trajectories of a bouncing droplet thanks to a model based on a spring. On a low viscous bath, we evidenced the importance of its deformation. We showed that a walker exists as soon as the droplet experiences, once every two oscillation periods, a jump high enough to trigger a Faraday wave. As a consequence, we have to take the bouncing droplets into account when looking for an alternative way to manipulate them or as a model of quantum-like particles. The bouncing droplets still exhibit lots of intriguing behaviors which have yet to be explained. They can therefore remain the focus of future works.

La recherche réalisée dans le cadre de cette thèse consiste à développer des électrodes positives pour les accumulateurs (ou "batteries") au lithium sous formes de films minces de quelques centaines de nanomètres d'épaisseur. Afin de stocker un maximum d'énergie dans le volume réduit d'un accumulateur au lithium miniature, une architecture mésoporeuse (dont le diamètre des pores est compris entre 2 et 50 nm), permettant d’augmenter la surface spécifique du matériau, a été réalisée au sein des films. Deux méthodes de soft templating ont été utilisées dans cette thèse. Ces méthodes consistent à ajouter à la solution de précurseurs inorganiques un copolymère amphiphile formateur de micelles, dont l'élimination ultérieure par traitement thermique permet d'obtenir des films contenant des pores dont la dimension varie selon le copolymère utilisé. Dans la perspective du développement d'une électrode pour accumulateur au lithium, deux stratégies sont possibles : (i) la préparation d'un film mésoporeux de l'oxyde non lithié, suivie d'une insertion du lithium par voie électrochimique, ou (ii) la préparation directe d'un film mésoporeux de l'oxyde lithié. Ces deux voies ont été comparées et l’influence du lithium sur la synthèse des films minces mésoporeux par soft templating a donc pu être étudiée. En termes de compositions ciblées, les oxydes de vanadium et de lithium (Li-V-O), tels que LiV3O8 ou LiV2O5, constituent des objectifs évidents en raison de leurs performances électrochimiques reconnues. Etant donné le caractère a priori délicat de la condensation des réseaux à base d'oxyde de vanadium, d'autres systèmes ont également été abordés dans le cadre de la thèse, à savoir Li-V-Nb-O, Li-Nb-O et Li-Ti-O.

L’objectif de cette thèse est d’obtenir un dépôt d’orthoferrite d’yttrium orthorhombique sur substrat conducteur. Ceci, en vue de d’étudier de nouveaux catalyseurs potentiels pour le clivage photoélectrochimique de l’eau. Le clivage photoélectrochomique de l’eau produit dudihydrogène et du dioxygène grâce à l’énergie solaire et un catalyseur. Le dihydrogène et le dioxygène sont les combustibles d’une réaction exothermique. Transformer l’eau en ces deux composés est donc une méthode de stockage de l’énergie. Plusieurs semiconducteurs sont étudiés comme catalyseurs pour ce clivage. Cette thèse se concentre sur l’étude de l’un d’eux : l’orthoferrite d’yttrium orthorhombique (o-YFeO3), et sur deux méthodes de dépôt de ce matériau sur substrat conducteur : la méthode de déposition par plongée (dip-coating) et la pyrolyse d’une vaporisation ultrasoniquement générée (USP). Dans un premier temps, deux méthodes de synthèse del’o-YFeO3, sous forme de poudre, ont été envisagées : l’une basée sur des chlorures de fer et d’yttrium, l’autre basée sur les nitrates de ces mêmes métaux. Les éventuels problèmes de proportion ou de précipitations successives ont été respectivement dissipés par une étude de masse et une analyse de l’évolution du pH. Ces poudres, issues des chlorures ou des nitrates, ont permis de caractériser la proportion et la pureté des phases cristallines en fonction de la température de calcination par l’étude de la diffraction des rayons X (XRD) ; de caractériser l’énergie de la bande interdite (band gap optique) par la spectroscopie de réflectance diffuse (DRS) et d’en observer leur morphologie par les microscopies électroniques à balayage (SEM) ou en transmission (TEM). Toutes ces analyses nous ont poussés à continuer notre étude en nous focalisant sur la synthèse basée sur les nitrates et d’utiliser une température de calcination de 850°C. Ensuite, ces mêmes analyses (pH, XRD, DRS) ont permis d’estimer l’influence de l’ajout de différents composés organiques en cours de synthèsesur la cristallinité, la pureté, la température de calcination et l’énergie de la bande interdite de ces poudres. Cette thèse a montré que l’additif qui améliore au maximum ces propriétés est l’acide citrique. Dans le cas des poudres, dès 450°C – soit 400°C de moins que la même synthèse sans acide citrique – l’o-YFeO3est cristallisé. Les solutions qui seront utilisées pour les méthodes dedéposition seront donc composées à partir, entre autre, de nitrate de fer, de nitrate d’yttrium et d’acide citrique. La méthode dite du « dip-coating » consiste à plonger le substrat qu’on compte recouvrir dans une solution de précurseurs – composée, entre autre, des nitrates métalliques et d’acide citrique – de le ressortir recouvert d’un film mince de cette solution pour ensuite le chauffer et ainsi obtenir un dépôt du matériau désiré. De nombreux problèmes d’adhérence, de stabilité de solution, de quantité de matière, d’homogénéité de dépôt et de cristallisation de composé ont eu raison de nos efforts pour la mise au point de tels dépôts par cette technique. L’autre méthode de déposition utilisée durant cettethèse est l’«USP ». Cette technique vaporise une solution de précurseurs au moyen de vibrations ultrasoniques. Le brouillard ainsi généré est propulsé vers le substrat par du gaz sous pression. Le substrat, pendant ce processus, est placé sur une plaque chauffante. La solution vaporisée chauffera au fur et à mesure qu’elle se rapprochera du substrat. Cette technique implique l’ajustement de plusieurs paramètres qui définissent le type de composés qui atteindront le substrat. Cette thèse a abouti à la mise au point de ces différents paramètres qui ont permis d’obtenir un dépôt d’orthoferrite d’yttrium orthorhombique sur un substrat conducteur. Ce dépôt présente une activité photoélectrochimique en tant qu’anode pour le clivage de l’eau, mais ses performances ne permettent pas, dans son état actuel, de concurrencer les semiconducteurs précédement décrit par d’autres chercheurs. Une poursuite des travaux dans ce domaine devraient se focaliser sur le dopage et/ou la structuration contrôlée de l’orthoferrite d’yttrium.

Today, adhesive bonding is a widespread technology many fields including automotive, aeronautic, surgery, packaging, electronic devices or in the building sector. It enables designing novel (lightweight) materials/products with performances comparable to the ones of systems fixed by mechanical adhesion. Glues, adhesives exist as a large variety of compositions such as cyanoacrylates (superglue ®), (meth)acrylate or epoxy resins, polyesters, polyurethanes… that fit on-demand to the specificity of the glued assembly (nature of the substrate, the thermo- mechanical performances, the resistance against water, acids, bases or solvents…). Due to their easy tunable and versatile properties (soft and flexible to rigid materials, high bonding adhesion, compatibility with numerous substrates…) polyurethanes (PUs) are reference systems. PUs are produced from toxic isocyanates that cause severe health concerns (asthma, skin irritation, DNA mutation). To surpass these issues, the quest for novel isocyanate-free PUs glues and adhesives formulations is essential. This thesis responds to this current trend which aims to develop well-designed innovative sustainable PU adhesives (and coatings) free of isocyanates. It explores the potential of poly(hydroxyurethane)s (PHU) made by step-growth polymerization of CO2-sourced bis- or multi-functional cyclic carbonates with di- or polyamines to construct novel PU glues/adhesives for various substrates (metals, wood, glass, plastics). Three main research axes were investigated and focused i) on the establishment of solvent-free petro-based PHU formulations and their corresponding nanocomposites thermoset PHUs (native or functional silica or ZnO fillers) to tailor high performance adhesives for various substrates; ii) on the increase of the sustainability of the PHU nanocomposites glues by incorporation of renewable monomers (vegetable oils) within the formulations and iii) the development of biomimetic PHU glues inspired from mussels (incorporation of dopamine). All formulations were benchmarked with commercial Terpmix-6700 and Araldite®2000 PU glues and results highlight that these PHU glues represent promising and competitive alternatives to conventional PU glues prepared from the toxic isocyanate chemistry. We believe that this work opens a realistic route to the next generation of PU adhesives.

During his lecture entitled “ There is plenty of room at the bottom ”, Richard Feyn- man foresaw the possibility of manipulating material at the scale of individual atoms and molecules. Although Feynman’s conceptual idea of a nanoworld was evoked in 1959, the nanoscience and nanotechnology revolution began 30 years later with the ability to see at the atomic scale with the invention of electronic microscopy and related tools. The size range that has attracted so much attention over these last 30 years is typically from 100 nanometers down to the atomic level. Subsequently, nanomaterials were defined as materials, which have structured components with at least one dimension in this size range. For instance, material with three nanometric dimensions defines a nanoparticle. Among the variety of core materials available to synthesize nanoparticles, noble metal nanoparticles, i.e. gold and silver, have fascinated people for centuries owing to their bright and intense colors, used in particular as decorative pigments in cathedral stained glasses and artworks. The red and yellow colors displayed by gold and silver nanoparti- cles arise from their interaction with light, which one induces collective oscillations of free electrons at the nanoparticle surface in resonance with the light field. This phenomenon is commonly known as the localized surface plasmon resonance. Their remarkable optical properties and the intense electric field generated by plasmonics nanoparticles have brought these nanomaterials in the forefront of nanotechnology research, ranging from photonics to medicine. Shining light on plasmonic nanoparticles to push back limitations of light-activated therapy and so taking part to the societal challenge of cancer treatment improvement defines the global framework of the thesis. Falling within the nanomedicine topic, this one more precisely deals with the development of efficient plasmonic nano-drugs using light to cure diseases. Clearly, nanoplasmonics, which explores how electromagnetic field can be confined over dimension on the order or smaller than the wavelength of light, has come a long way since the stained glass of Roman times.

In this thesis, we explore the electronic, dynamic and thermoelectric properties of different tellurium-based compounds. We perform ab-initio calculations within the Vienna Ab-initio Simulation Package (VASP) that works in the framework of Density Functional Theory (DFT). For the thermoelectric properties, we use the Boltztrap code that solves the Boltzmann Transport Equations (BTE) for electrons within the Constant Relaxation Time Approximation (CRTA). This computational package allows us to obtain accurate values of the Seebeck coefficient as a function of temperature and carrier concentration (this last with the help of the rigid band approximation). While for the calculation of the lattice contribution to the thermal conductivity, we use the ShengBTE code that solves the BTE for phonons iteratively. The first tellurium-based compound that we study is the best room temperature thermoelectric material, Bi2Te3. We obtain results comparable with experimental data for the Seebeck coefficient at room temperature and pressure. Afterwards, we proceed to explore the evolution of the electronic properties and the thermoelectric performance under pressures up to 5 GPa. We reproduce the overall trend of the Seebeck coefficient as a function of pressure for two different values of doping, however, our results do not reproduce the small improvement found in experiments close to 1 GPa. Nevertheless, we support the experimental evidence of an Electronic Topological Transition (ETT) around 2 GPa and we explain this particular behavior. We also perform calculations on the tellurium-based phase-change materials (GeTe)x(Sb2Te3)1 (with x = 1, 2, 3). We show results for different stacking configurations since for some compositions, the stacking arrangement of the atoms in the primitive cell is still unsettled. We find that the change of the atomic arrangement leads to the systems to go from semiconductors to metals. We find that the semiconductor arrangements systematically overestimate the experimental values for the Seebeck coefficient, whereas the metallic stacking sequences are in very good agreement with the experimental data for the Seebeck coefficient and for the lattice contribution to the thermal conductivity. We show that (GeTe)x(Sb2Te3)1 materials could reach values of ZT=0.5 around 600 K with a proper optimization of S with respect to the carrier concentration. We also report that in the case of x=3, the most accepted stacking configuration is dynamically unstable, therefore we proposed another sequence. Finally, we discuss the discrepancies between our work and recent theoretical reports that claim the existence of a Dirac-cone like band structure for (GeTe)2(Sb2Te3)1. We explain the conditions necessary to obtain such electronic topology.

De nombreuses questions subsistent quant à la nature d’une interface entre une mousse et une solution. Ce travail concerne l’influence des conditions aux limites au niveau d’une telle interface sur le volume de mousse lorsqu’elle est perturbée par une contrainte extérieure. Le lien entre mousse et interface a été étudié dans deux situations différentes : une contrainte normale et une contrainte tangentielle. L’instabilité de Faraday permet de soumettre l’interface à une contrainte normale périodique. L’influence de la géométrie du système a été investiguée pour une surface libre. La modification de la longueur d’onde a été expliquée en terme d’augmentation de l’énergie interfaciale. La perte d’énergie a également été modélisée à l’aide de trois sources : la viscosité de la solution, la présence de molécules de surfactants à la surface et la condition de non-glissement aux parois. L’interaction entre une mousse et l’instabilité de Faraday est ensuite étudiée. La dissipation visqueuse est augmentée par la présence de bulles et a pu être modélisée à l’aide de considérations énergétiques. Il a également été montré qu’un faible nombre de couches de bulles est suffisant pour amortir efficacement toute perturbation de l’interface. La contrainte tangentielle est appliquée à l’interface grâce à un dispositif inspiré des milieux granulaires permettant la rotation d’une cellule de Hele-Shaw autour de son centre. Dans un tel dispositif, les caractéristiques de la mousse et de l’interface varient. Deux modèles prédictifs permettent d’expliquer l’évolution temporelle de la fraction de liquide moyenne. Des outils statistiques ont permis de définir une relation entre les déformations des bulles et les caractéristiques macroscopiques de la mousse. Grâce à la modélisation de l’écoulement, un lien a été établi entre le gradient de pression interne de la mousse et la déformation de l’interface. Finalement, nos résultats sont comparés à ceux obtenus pour des ensembles granulaires dans un dispositif expérimental similaire.

Among network forming glasses, chalcogenide glasses are of great importance not only for their optoelectronic applications, but also for the network structure that displays enhanced structural variability due to the covalent bonding network. In this project, we study the Ge-Se binary alloy as the target system in order to investigate the structure and dynamics of the liquid and glassy phases under pressure by using a combination of ab initio molecular dynamics (AIMD) simulations and X-ray scattering experiments. The wide glass forming range of the GexSe100􀀀x system allows one to tune the stiffness of the network structure by increasing the mean coordination number with the Ge content which affects the macroscopic material properties such as resistance to aging, hardness, conductivity, and fragility. In this respect, we study 10 different AIMD generated and 5 experimentally produced (i.e melt quenching) compositions spanning the flexible-to-rigid elastic phase transitions according to Maxwell’s isostatic stability criterion. As for the liquid state, after having validated the structural models by comparing the experimental findings available, we examine the dynamics of Ge-Se melts at ambient pressure. The investigations on the diffusion coefficients and viscosity at 1050 K showed clear anomalies, departing from the expectation that atomic mobility should decrease as the system becomes more and more rigid. Furthermore, the relaxation behavior at 1050 K also shows similar anomalies when intermediate scattering factors are examined at q vector corresponding to the principal peak position of the total structure factor. To elucidate this anomaly, we discuss the effect of the topological constraints on the dynamics in liquid state. The results show that the isostatic systems have slower dynamics as compared to flexible and stressed rigid phases. Moreover, we speculate that the reason of this anomaly may originate from the distribution of the topological bond bending constraints of the higher coordinated species (i.e. Ge) results. In particular, the flexible, and stressed rigid compositions showed a high variance in the Ge bond bending constraints whereas the isostatic composition forms a network in which the bond bending constraints are homogeneously distributed. We link this behavior with a global fragility concept for network forming liquids in such a way that fragility minima are obtained both by experimental findings and the calculated fragility values of the AIMD generated compositions (i.e. VFT or MYEGA fits) when scaled to the isostatic composition for a number of different systems. As for the glassy phase, the AIMD generated structural models show good agreement in both real and reciprocal space. The equation of state and normalized stress-strain curves are compared to the available experiments in order to verify the pressure behavior of the simulations. The results show a good agreement. In addition to the simulations under pressure, we show results of X-ray absorption spectroscopy (XAS) and X-ray diffraction (XRD) experiments under pressure. Both simulations and the experiments show that there are no sign of crystallization during compression up to 42 GPa. One of the main important finding is the evolution of the Ge-Se bond length for the compositions studied. We observe a bond compression in the early stages of densification in the phase identified as low density amorphous (LDA), which is followed by an abrupt jump starting in pressures around 10-15 GPa, both in simulations and experiments. Furthermore, a semiconductor to metal transition is identified with the red shift in Ge K edge energy. The features of polyamorphism was also detected from the pressure evolution of the principal peak position of the structure factor which show two distinct slopes indicating different structural response to the applied pressure. In order to have a deeper understanding of the densification mechanisms, we apply neighbor analyses to our atomic trajectories and show that the tetrahedral to octahedral transformation (i.e LDA to HDA) starts to take place when the fifth and sixth neighbors effectively become the part of first shell neighbors, where the bond angles adapt themselves to 90 degrees. We furthermore speculate the effect of network rigidity of the glasses at ambient conditions onto the kinetics of the amorphous-amorphous transitions. It appears that the polyamorphic transitions are more sluggish as the network rigidity (i.e Ge content) increases. Finally we show that there is a universal threshold value in coordination change from LDA to HDA phases GexSe100􀀀x (where x 25) when scaled to reduced densities.

Acute kidney injury (AKI) is a frequent complication of sepsis that can increase mortality as high as 70%. The pathophysiology of this kidney failure was previously believed to be secondary to decreased global renal perfusion causing hypoxia-induced injury. However, new research suggests this paradigm is overly simplistic, and injury is now considered multifactorial in origin. Mechanisms that contribute to kidney injury mainly include inflammation, alterations in microvascular renal blood flow and changes in bioenergetics. To study the mechanism of oxygen regulation in acute kidney injury during sepsis, we developed a sepsis-induced in vitro model using proximal tubular epithelial cells (HK-2) exposed to a bacterial endotoxin (lipopolysaccharide, LPS). Our first investigation, by using both high-resolution respirometry and electron spin resonance spectroscopy, showed that HK-2 cells exhibit a decreased oxygen consumption rate when treated with LPS. Surprisingly, this cellular respiration alteration persists even after the stress factor is removed. We suggested that this irreversible decrease in renal oxygen consumption after LPS challenge is related to a pathologic metabolic down-regulation such as a lack of oxygen utilization by cells for ATP production. In the long term, this metabolic disturbance leads cells to a predominantly apoptotic death. To confirm this hypothesis of cytopathic hypoxia, we demonstrated that this alteration in the renal respiratory function is mainly due to an impairment in the metabolic activity of HK-2 cell mitochondria. Following LPS treatment, the oxidative phosphorylation is interrupted because of the inhibition of cytochrome c oxidase activity. As a consequence, disruptions in the electron transport and the proton pumping across the system occur, leading to a decrease of the mitochondrial membrane potential, the release of apoptotic-inducing factors and a decrease in ATP production. To clarify the mechanism by which the LPS induces mitochondrial alterations, we studied the oxidative stress generation in HK-2 cells. Interestingly, we revealed that the induction of a cytosolic oxidative stress is an event that appears before mitochondrial dysfunction in the LPS-treated HK-2 cells. This primary redox state is notably due to the activation of the two enzymes NADPH oxidase 4 and inducible NO synthase. The simultaneous production of anion superoxide and nitric oxide strongly suggests the formation of peroxynitrite, a relative stable powerful oxidant that can diffuse through mitochondrial compartments and undergo cytotoxic reactions. To our knowledge, our model reveals for the first time the role of NADPH oxidase-derived cytosolic ROS in triggering tubular cell damage. Moreover, after being first target of the oxidative stress, mitochondria become in turn producer of reactive oxygen species that carry on mitochondrial dysfunction. It seems thus that a mechanism of oxidative stress-induced redox cycling is a main cause of the mitochondrial dysfunction of LPS-treated HK-2 cells. The role of oxidants in mitochondrial dysfunction was further confirmed by the use of iNOS inhibitors or antioxidants that preserve cytochrome c oxidase activity and block mitochondrial membrane potential dissipation. Overall, these results suggest that sepsis-induced AKI should not only be regarded as failure of energy status but also as an integrated response, including transcriptional events, ROS signaling, mitochondrial activity and metabolic orientation such as apoptosis.

The Frank-Starling mechanism (FS) is an essential feature of the heart, which allows for a beat-to-beat adaptation of cardiac output to hemodynamic conditions. This intrinsic adaptative mechanism to preload variations lies in the cellular components that make up the cardiac muscle. It is believed that lengthdependent activation (LDA), a cardiac cellular property, is responsible for the FS mechanism observed at the heart scale. However, LDA is essentially highlighted in cellular experiments that do not reproduce the in vivo conditions of a beating heart. The connexion between LDA and the FS mechanism is actually difficult to unravel experimentally, as two very different scales (cellular and ventricular) are involved. This thesis is devoted to the analysis of this connexion between a cellular mechanism (LDA) and its manifestation at the cardiovascular scale (FS mechanism). This analysis is performed in silico with a multiscale model of the cardiovascular system (CVS), where ventricular contraction is described at the cellular scale. Such models help overcome the experimental difficulties of linking two different scales, while providing a formal framework to integrate the experimental observations coming from both scales. Our multiscale model is first used to study the relevance of some cardiac contractility indices. Then, an analysis of the FS mechanism is proposed. Attention is paid to providing rigorous definitions and numerical protocols so that the correlation between LDA and the FS can be established without any ambiguity. LDA is shown to underlie the macroscopic (ventricular) response to preload variations, but in a highly dynamical way, in contrast with what is generally presented in the literature. In addition to these physiological considerations, the relationship between the FS mechanism and clinical therapies is also addressed. The FS mechanism is commonly presented as the founding principle for vascular filling, but we challenge this theory and introduce the concept of lengthdependent fluid response (LDFR). We show that LDA underlies LDFR, but it is not the only factor that drives the macroscopic (ventricular) response to fluid infusions. The afterload also comes into play and the global CVS response results from a balance between a cellular LDA-driven mechanism and a hemodynamic resistance to blood ejection. Finally, the role of the FS mechanism regarding stroke volumes equilibrium is also investigated. We conclude that LDA indeed underlies the FS mechanism in vivo, but in a way that implies a complex dynamical interaction of cellular and hemodynamical variables. The FS mechanism is thus really a multiscale phenomenon, where the cellular variables and the hemodynamic variables influence each other during the whole heartbeat. It is hoped that our multiscale CVS model could be developed and used for further studies that aim at linking cellular properties and organ behaviors, either in healthy or in pathological conditions.

Nous présentons les résultats expérimentaux obtenus avec un système modèle expérimental dédié à l’étude des structures et des transitions de phases à 2D. Le système est composé d’une monocouche de billes soft-ferromagnétiques de taille millimétrique confinées dans une cellule 2D horizontale. Les billes sont plongées dans un champ magnétique vertical et homogène qui induit une interaction dipôle-dipôle entre elles. L’effet combiné du confinement et des interactions répulsives ordonne les billes. Celles-ci sont athermales, une agitation mécanique est donc utilisée afin de leur conférer un mouvement Brownien qui crée du désordre dans le système. Ajuster l’importance relative entre les effets de l’interaction et de l’agitation permet de contrôler l’ordre du système. Grâce à ce dispositif expérimental, nous pouvons donc étudier la transition d’un état figé et très ordonné, dit cristallin, vers un état désordonné et dynamique semblable à un liquide. Nous montrons que cette transition se déroule en deux étapes, avec une phase intermédiaire appelée hexatique, comme cela est prédit par la théorie KTHNY de la fusion 2D. De plus, les structures observées sont identiques à celles obtenues à partir de systèmes colloïdaux et de simulations numériques. Notre expérience semble donc tout à fait adaptée à l’étude des systèmes 2D thermiques. Dans la suite de ce travail, nous avons décidé de nous concentrer sur l’étude des défauts topologiques car ils sont d’une importance cruciale en physique à basse dimensionnalité. Nous avons d’abord forcé les défauts topologiques dans le système en induisant une frustration venant du confinement. Nous montrons que la taille et la géométrie du confinement affectent significativement l’ordre global, ainsi que le type et le nombre de défauts observés dans le système. Ensuite, nous forçons des défauts ponctuels dans la structure en y introduisant des billes de tailles différentes, appelées impuretés. A l’échelle globale, l’augmentation du taux d’impureté permet d’empêcher la cristallisation du système. A l’échelle locale, nous observons que les impuretés induisent des frustrations géométriques très localisées, ce qui mène à l’apparition de défauts topologiques dont la nature est directement liée à la taille de l’impureté.

The evaporation of colloidal droplets is an area of intensive research. From paint coating to blood analysis on crime scene, applications of patterning from evaporation of colloids are numerous and various. In our work, we aimed to bring highlight on how interactions between colloidal particles influence the eventual deposit's pattern. To do this we used superparamagnetic colloids as a way to have a tunable interaction with these particles. We first studied the influence of dipolar interactions on the suspension of particles. We performed systematic experiments to characterize the thermodynamic equilibrium reached by the suspensions. We showed that tuning the viscosity parameter could be used to speed-up numerical simulations. We used this process in sped-up simulations in order to study a new range of volume fraction. We showed that high volume fraction conditions led to higher chains' length than expected. We proposed a modification of current models. We then focused on evaporating droplets. We began by reviewing the liquid flows in our suspensions. We showed there was a competition between coffee-ring flow and Marangoni instability. We then characterized the deposits left after evaporating suspensions under magnetic fields, and showed the influence of both this parameter and the Marangoni instability on the eventual deposits. We evidenced a transition in the deposit's behaviour when the DLVO interaction between the particles and the substrate becomes attractive.

Rare-earth orthoferrite perovskites (RFeO3, where R is a rare-earth element, i.e., La, Gd, Dy, …) are a family of materials that have attracted a lot of attention due to their original magnetic properties and large nonlinear magnetoelectric responses (ME). A large magnetoelectric response will allow controlling magnetic properties using an electric field which can bring in a plethora of applications and improvements of the current technologies. In my thesis, I have studied magnetic properties of rare-earth orthiferrites and I have made a Heisenberg model to explain the origin of different unique magnetic behaviors (spin reorientation and magnetization reversal) present in these rare-earth orthoferrites. We have then used the model to explain large ME responses observed in these materials. We have shown that the nonlinear response present in these materials arises from the fact that the antiferromagnetic ordering changes nonlinearly with an applied magnetic field. Through a collaboration with experimentalists, we have also studied ultrafast manipulation of the magnetic phase of DyFeO3 using laser pulses in an ultrafast time scale which allows the use of these materials in memories with ultrafast time responses.

In the last fifteen years, multifunctional materials, and more specifically, multi- functional oxides have been widely studied due to its wide range of properties. Properties that go from superconductivity to ferroelectricity passing through mag- netism and multiferroism have been reported. Nonetheless, the fluoride family was left aside and little information is known about its possible ferroelectricity or mul- tiferroism. In this Ph.D thesis, we explored the electronic, vibrational, structural and magnetic properties of fluoride perovskite-based compounds. To such pur- poses, We performed ab-initio calculations based in the density-functional theory (DFT) as implemented in VASP and CRYSTAL codes. Our first step was to perform vibrational analyses in a large set of fluoroper- ovskites ABF3. Based on the results, we proposed a model that establishes an A-site geometrically driven ferroelectric vibrational instability in fluorides. Our studies reveal a different behavior as a function of isotropic pressure for NaBF3 with respect to oxides (e.g. BaTiO3) with B = Ca, V, Mn, and Zn. For these compounds we found an increase of the ferroelectric instability as a function of hydrostatic pressure. This probably due to the “transformation” of eigendisplace- ments responsible for the mode that creates the corresponding instability. In particular, an increase of ionic A-site radii present a strong influence in FE-polar instability. We also have shown, based on our first-principles calculations and symmetry theory analysis that all post-perovskites ABX3 with an active magnetic B-site cation can exhibit a noncollinear magnetic configuration, which happens to be allowed by symmetry. With these findings, the magnetic properties found exper- imentally were clarified for this particular high-pressure phase perovskite found at the Earth’s mantle. Additionally, We have predicted that NaMnF3 suffers a structural phase transition under pressure to a post-perovskite phase, where non- collinear ferromagnetism and large magnetic moment components are obtained within this high-pressure phase.. Going beyond, We have shown that it is possible to achieve multiferroic-induced state in NaMnF3 under epitaxial strain at compressive or tensile strain. We found a nonlinear behavior of the ferroelectric instability as well as a non-linear piezo- electric response as a function of epitaxial strain. The later completely different as the one found in oxide perovskites. Similarly, an out-of-plane polarization was observed, a property that has not been observed in oxides. We observed a Na + Mn sites cooperative ferroelectric ordering for compressive strain against a pure A-site geometrically driven ferroelectricity at tensile values of the ac-strain. Magnetic ordering reveals a non-collinear ground state with the GzAxFy repre- sentation. Even more interesting, and non-linear magnetoelectric coupling was found under the strained Pna1 ground state becoming the first known multifer- roic/magnetoelectric perovskite fluoride. Later, in order to go further, We studied the electronic and structural proper- ties of novel heterostructures based on oxyfluorides (KTaO3)n/(KBF3)l B = Zn and Ni interfaces. We found that the orbital levels splitting at the interfaces is strongly modified by the O–B–F coordination. The polar catastrophe phenomena also takes place in the oxyfluoride interfaces similarly to oxide heterostructure, however, we found that less number of layers are needed in order to achieve the insulator-to-metal transition when comparing to SrTiO3/LaAlO3 superlat- tices. We observed that the magnetism in the KTaO3/KNiF3 exhibits a moment magnitude modulations. Nevertheless, the magnetic structure keeps the G-type antiferromagnetism such as in the bulk former compound. Surprisingly, we ob- served a large k3-Rashba type splitting in at the oxyfluoride interfaces, at least four times larger than the one reported in SrTiO3/LaAlO3 interface and twice of the KTaO3-based transistor. In conclusion, we observed that fluorides-perovskites are good prototypes for multifunctional properties as oxides. Therefore, based on the results reported in this thesis, we expect that experimentalist and theoreticians can be motivated in characterization of fluorides, which can lead to a new set of unexplored materials with potential novel applications in electronics.

This thesis presents investigations on the interplay of coherence and many-body effects in the quasi one-dimensional transport of Bose-Einstein condensates (BEC) through scattering potentials. Such configurations can be realized with guided atom lasers that provide a coherent atomic beam. An exact theoretical description of the dynamics is out of reach due to the presence atom-atom interactions. Different levels of approximations are nevertheless possible with their strengths and weaknesses. The mean-field approximation, where the dynamics of the BEC is governed by the Gross-Pitaevskii equation, is most commonly used in the field of ultracold atoms. In this thesis the truncated Wigner method is used to go beyond the standard Gross-Pitaevskii description. This method is adapted in order to study the scattering of Bose-Einstein condensates in one-dimensional waveguides where atom-atom interactions and external potentials are nonvanishing only in a finite region of space. In this case, the truncated Wigner method is combined with the smooth exterior complex scaling method and incorporates quantum noise that originate from the vacuum fluctuations in the waveguide. Inelastic scattering is shown to play a major role in the resonant transport of BEC through a symmetric double potential barrier effectively forming an atomic quantum dot. Indeed, fully resonant transmission is prohibited and incoherent atoms as well as collective oscillations are detected in the transmitted beam. It is also shown that inelastic scattering destroys Anderson localization in the case of transport through disordered potentials. The classical (incoherent) ohmic transmission is recovered for finite atom-atom interactions. The validity of the truncated Wigner method is then assessed using the semiclassical van Vleck-Gutzwiller propagator in the Fock space of the many-body system. It is shown that the truncated Wigner method corresponds to the so-called diagonal approximation, and it is possible to identify the leading correction to the truncated Wigner results, which is provided by the so-called coherent backscattering (CBS) contribution. Coherent basckattering in Fock space is a genuine quantum many-body effect that lies beyond the reach of any mean-field approach. For the case of closed Bose-Hubbard models, the relevance of CBS is confirmed by numerically comparing the (classical) truncated Wigner evolution probabilities to the exact quantum probabilities in Bose-Hubbard models: While a CBS-induced enhancement of the return probability to the initial state is clearly seen in the exact quantum simulations of the bosonic many-body system, this enhancement is absent in the classical calculations. The magnitude and dependence of the CBS contribution on gauge fields, which break time-reversal invariance, is numerically confirmed. For the case of disordered open systems, it can be shown that this contribution as well as next-to leading order contributions vanish thereby confirming the validity of the truncated Wigner method.

Quantum simulation with ultracold atoms gained a lot of traction recently by proposing a framework with a lot of flexibility, versatility and tunability to emulate diverse quantum ef- fects. It indeed provides the ideal playground to study many–body effects in a well–controlled environment and is particularly useful in the domain of quantum coherent transport of waves in random media. The purpose of this thesis is to study several configurations of coher- ent transport within random media with Bose–Einstein condensates and to investigate the interplay between coherence and interaction effects. In particular, we start by numerically studying Aharonov–Bohm oscillations in the transmission of particles across the eponymous rings in a 1D configuration. When exposed to a suitably chosen disorder potential, those rings yield oscillations with double frequency, which are routinely encountered in solid–state physics where they are referred to as Al’tshuler–Aronov–Spivak oscillations, similar in essence to coherent backscattering and weak localisation. We then study the behaviour of those os- cillations in the presence of interaction within Aharonov–Bohm rings and find that in the mean–field regime, they are inverted for finite interaction. Truncated Wigner simulations are then carried out in the same scenario and indicate that the inversion should be observ- able for realistic atomic and experimental parameters with 39 K atoms, although dephasing of the oscillations is observed at strong interaction owing to interaction–induced inelastic scattering. A first–order nonlinear diagrammatic theory is then presented and benchmarks our numerical findings. The question of the inversion prevalence is then investigated in a 2D scenario, following state–of–the–art observations in the literature. It has indeed been nu- merically observed that coherent backscattering is inverted in the mean–field approximation for finite interaction strength. We numerically confirm this observation with our study and extend it beyond the mean–field approximation by applying the truncated Wigner method. These simulations show that the inversion prevails beyond the mean–field regime and should moreover be observable experimentally with 87 Rb atoms for realistic parameters, despite a partial dephasing. This dephasing however completely eclipses interference effects and washes out this signature of antilocalisation for stronger interaction.

The present dissertation theoretically investigates the generation of entangled states with ultracold bosonic atoms. Specifically, that focuses on the NOON states in a double-well potential, which are the coherent and equivalent superposition of |N,0> and |0,N> with $N$ atoms, and on the triple-NOON states in a three-site optical trap, which are the coherent and equivalent superposition of |N,0,0>, |0,N,0> and |0,0,N>. These states can be seen as large manifestations of entanglement. The collective tunneling in the self-trapping regime is made possible by the atom-atom interactions. For example, the NOON state is formed after half the tunneling time, i.e. the time needed to obtain a total transfer of population to the other site. The main message of this dissertation is that the timescale required to generate this transition can be considerably reduced by means of an external periodic driving without qualitatively altering the quantum dynamics. Moreover, indications of this speedup are available in the corresponding classical phase space. The presence of nonlinear resonances at the classical level induces perturbative couplings at the quantum level. The subsequent reorganization of the eigenspectrum enables one to explain the modifications of the tunneling time. These modifications can also be produced by prominent chaotic layer known to welcome strongly connected states. Built upon the phase space features, resonance- and chaos-assisted tunneling is a semiclassical theory which can be used as a guideline in the quest of suitable parameters.

The use of transparent conductive materials (TCMs) has rapidly increased in the last two decades as a result of the increasing demand for personal electronic devices and the development of thin film based solar cells. To date, the most commonly used TCM is indium tin oxide (ITO). However indium is a rare earth metal with a complex geopolitical environment surrounding its supply and production. Furthermore, the oxide family suffers from poor mechanical properties such as brittleness and generally requires either high temperature synthesis (>400°C) or vacuum processes for their deposition. For these reasons, research in recent years has focused on the discovery or the design of a TCM to replace ITO. This thesis applies a dual approach combining simulations and experiments to explore the fabrication and optimisation of silver nanowire networks for use as a TCM and to improve the understanding of their physical properties. The simulation contribution focuses on the application of percolation modelling to 2D nanowire networks while the experimental part explores the electrical and optical properties of silver nanowire networks and their electrical behaviour under thermal annealing. We present in this work the modelling of 2D stick percolation systems initially composed of perfect idealised sticks, and then, investigate the influence of parameters such as length distributions, angular distributions or shape (curved nanowires). We address the divergence of the critical density for the onset of percolation observed for small system sizes and introduce some preliminary work on simulating the collection (or injection) efficiency of charges by a nanowire network. The experimental component provides a discussion of the impact of wire length, wire diameter, network density and fabrication technique on the optical and electrical properties of silver nanowire networks. An in-depth study of the effect of thermal annealing on the networks properties was undertaken, which revealed several mechanisms responsible for the initial reduction of resistance and the observed final loss of conductivity. An original observation enables the revelation of geometrical quantized percolation for rather sparse networks. Finally we conclude that silver nanowire networks are an excellent prospect as a TCM to replace ITO: these materials have superior mechanical properties and enable comparable and even superior electro-optical properties.

The present thesis aims to contribute to the quest for increasingly complex functional polymers while respecting a simpler and more efficient chemistry. For this purpose, we took advantage of multicomponent reactions which involve at least three compounds and yield complex structures containing almost all atoms of the reactants. First, stimuli-responsive and biocompatible materials were synthesized using the Ugi four-component polymerization of amino acids. A variant of this reaction was then adapted for step-growth polymerization leading to a series of unprecedented poly(α-amino amide)s. In a second part, the Radziszewski reaction and the emulsion-templating polymerization method were combined giving access to macroporous poly(imidazolium)s networks of interest in catalysis and CO2 capture.

In this thesis, we present investigations on the cooperative dynamics of (ultra)cold atoms coupled to the electromagnetic field in vacuum. The main objective is to identify the consequences of the quantization of the atomic center-of-mass motion on the atoms-field dynamics, more particularly the atomic \emph{internal} dynamics including dipole-dipole interactions and cooperative spontaneous emission processes. For this purpose, we derive a Markovian master equation for the internal state of a collection of two-level atoms coupled to vacuum accounting for all effects related to the quantization of their motion. These effects depend on three characteristic lengths that can be tuned experimentally: the interatomic distance $r$, the wavelength $\lambda$ of the emitted radiation, and the typical size $\ell$ of the atomic wave packets. This leads to a rich panel of phenomena depending on the hierarchy between these characteristic lengths. Our master equation provides a unifying picture of the consequences of recoil and quantum statistics (bosonic or fermionic) on both the internal dissipative and conservative dynamics, and applies equally well to distinguishable and indistinguishable atoms. It is valid even beyond the Lamb-Dicke regime, i.e.\ for large recoil effects. We give general expressions for the decay rates and the dipole-dipole shifts entering the master equation for arbitrary motional states, and we find closed-form formulas for a number of relevant states (Gaussian states, harmonic oscillator eigenstates, Bose-Einstein condensates motional states). In particular, we show they can be strongly influenced by the motional state of the atoms, which suggests the possibility to quantum program their internal dynamics through motional state engineering. Based on the derived master equation, we investigate in full detail the super- and subradiance phenomena arising from a collection of indistinguishable atoms. Due to the symmetry (for bosons) or antisymmetry (for fermions) of the global state under exchange of atoms, the internal dynamics is restricted to the permutation invariant subspace whose dimension grows as $N^2$ with $N$ the number of atoms. In this particular case, the master equation involves only three rates: a single-atom decay rate $\gamma_0$, a cooperative decay rate $\gamma$ and a collective dipole-dipole shift $\Delta_\mathrm{dd}$. We solve the dynamics exactly for $2$ atoms, numerically for up to $30$ atoms, and obtain the large-$N$ limit by a mean-field approach. We show that a superradiant enhancement of the radiated intensity is always observed for a sufficiently large number of atoms. As regards subradiance, we show that exact decoherence free subspaces that protect against spontaneous emission through destructive interference of individual spontaneous emission amplitudes exist only in the limit of classically localized atoms, i.e.\ for atoms in infinitely steep traps. We assess the validity of our master equation in the optical domain through the study of the exact atoms-field dynamics. Among others, this complementary approach gives access to the vibrational state populations of the atoms after a photon emission and to the angular dependence of the emission spectrum. It also highlights the complex interplay between the atomic internal and motional states, such as the entanglement transfer that can occur during the collective decay of the atoms. Moreover, it opens the door for the study of the non-Markovianity of the internal dynamics. In conclusion, our thesis provides a comprehensive picture of the consequences of the quantization of the atomic motion on cooperative spontaneous emission processes and contributes to the development of theoretical tools for the study of the internal dynamics of (ultra)cold atoms.

In the agricultural field, conventional farmers use Plant Protection Products (PPP) to control crop enemies as for instance weeds, diseases and pests. In practice, PPP application techniques are based on droplet clouds to carry the spray mixture containing the active ingredient and the adjuvant to the surface target. The PPP application efficiency consists in maximizing the deposition of the mixture on the target while reducing the environmental losses. However, the droplet characteristics within the spray affect drastically the treatment efficiency. For instance, small droplets (< 200 µm) are prone to drift while big droplets (> 300 µm) have the tendency to splash on a specific target surface. The widely used agricultural nozzles produce a liquid sheet that disintegrates into ligaments leading to droplets with various diameters and velocities. Hence, the generated spray is characterized by a wide droplet size distribution (RSF = 1) resulting in potential drifting or efficiency losses due to splashing phenomena. The spray must deliver an optimal droplet distribution in term of diameters and velocities by reducing the extent of the droplet size distribution. The design of new agricultural nozzles is a challenge for the practitioners in the field of agricultural nozzles. As the simplicity, the robustness, the low cost and the high flow rate ranges are required for the agronomic application, the Savart configuration namely a round jet impacting vertically a motionless disk is the ideal candidate for the massive production of droplets. The Savart sheet develops in the air (outside the disk) and it results in random breakup leading to wide droplet size distribution as in the case in hydraulic nozzles. As the used flow ranges are high in the PPP application, the obtained sheet on the disk is turbulent. The challenge is to tame the turbulent sheet. I propose to split the sheet into individual jets using textured disks by acting on the semi-free film or by inserting the right structures directly in the free sheet. Then, the jets break up according the Plateau-Rayleigh mechanism and lead to a narrower droplet size distribution. Therefore, this thesis aims to study experimentally the effect of the disk geometry on the sheet break up. This study is seen as a practical guide for specialists in fluid mechanics who desire developing the generation of drops with controlled sizes. One firstly detailed the experimental setup based on the impact of a turbulent jet on a non-textured disk. The disk configuration constitutes our reference case to which results on textured disks are to be confronted. The sheet was characterized through several parameters: The mean velocity U, the mean thickness h and the ejection angle φ that depends on the disk geometry and on the flow rate Q. As the jet flow is turbulent at the impact, local disturbances in the film triggered downstream the disk edge the appearance of random holes at a distance R′ (from the jet axis) in the liquid sheet. These holes lead to the disintegration of the sheet into droplets. Furthermore, the produced droplets are characterized in term of diameters. The droplet size distribution is clearly wide (RSF values are close to 1) that is similar to the case of hydraulic nozzles. Moreover, one perturbed the semi-free film evolving on the disk surface. The case of a turbulent round water jet impacting a disk engraved along its circumference by a number N of radial grooves is addressed. By the insertion of grooves, one controls the turbulent flow and the film splits into a number n of liquid jets before reaching the disk edge. The phase diagrams presented as a function of the inside gap between grooves d1 and the flow rate Q illustrate the transition between jet regimes. Droplets were characterized in term of their diameters and velocities. For all configurations including an engraved disk, the obtained droplet size distribution is narrower compared to the ungraved disk and to the standard flat fan nozzle Teejet TP 65 15. The V50 is reduced towards smaller droplets in the case of the engraved disk configurations but it is still coarse compared to that of the Teejet nozzle. Furthermore, one perturbed directly the Savart sheet, i.e. the free sheet evolves in the air. A number N of triangular prisms are set in the sheet at a radial distance r from the jet axis. This radial distance is strictly greater than D/2 (D = disk diameter) and less than or equal to R’ (for which the holes appear in the sheet). Once the number of prisms and the distance from the injector is fixed, the geometry is determined by the size of the prism and the distance b between two successive prisms. When increasing the flow rate Q, the jet numbers are n = 0, n = N and n = N* (elastic coalescence of jets) for large Q. One used a geometrical model that explains the generation of individual jets through these structures. Then, the emitted droplets are characterized in term of diameters and velocities. The droplet size distribution is narrower compared to the non-textured disk and the hydraulic nozzle TP 65 15 with the same spray class. The V50 decreased by decreasing the external gap b between two neighboring textures at the same radial distance r from the jet axis. Also, the V50 decreased by increasing the radial distance from the jet axis as the sheet thickness decreases. However, we are limited by R’ due to holes appearances. Finally, one concludes on the fundamental findings and on the role that the developed nozzle could play in applications. Also, one proposed some original perspectives to the present thesis such as testing the triangular prism textures in the case of the liquid sheet produced by a standard hydraulic nozzle.

Methylene heterocyclic compounds (MHCs) are an attractive monomer class for radical polymerization allowing the introduction of heteroatoms and functionalities in the side groups or in the main chain of polymers. The vast majority of the MHCs considered in polymer synthesis, however, are conjugated derivatives, so well adapted to the modification of macromolecules derived from conjugated monomers but not suited to copolymerize with some industrially important non-conjugated less activated monomers (LAMs) such as vinyl acetate (VAc), vinyl amides, ethylene, etc. Moreover, most of them are petroleum-based compounds and the development of biomass-derived MHCs still need to be addressed. The present thesis, carried out in the frame of the BioFact-EOS program dedicated to the development of bio-based chemicals and polymers, addresses the above mentioned limitations by developing non- conjugated MHCs, including CO2-derived compounds, able to copolymerize with LAMs and to produce a new range of functional copolymers. In particular, we developed the synthesis of ?-methylene-?-butyrolactone (?M?BL) and of the CO2-sourced 4,4-dimethyl-5- methyleneoxazolidin-2-one (DMOx), and explored their potential as building blocks for the preparation of functional copolymers. Their reversible deactivation radical copolymerization with VAc, N-vinyl caprolactam (NVCL) and ethylene, was notably studied leading to a series of ?M?BL or DMOx containing copolymers with controlled molar mass and compositions. Various post-polymerization modifications of these MHC-based copolymers gave access to unique stimuli responsive materials including pH-sensitive acid-functional poly(vinyl alcohol), dual pH/thermo-responsive acid-bearing PNVCL, pH/thermo/metal ion-responsive amino alcohol-functional poly(vinyl alcohol) as well as unprecedented pH-responsive amino- functional ethylene/vinyl alcohol copolymers. Overall, this work broadens the scope of the radical polymerization of MHCs which will further develop and become a powerful tool in the hands of polymer chemists.

Research such as the famous Millikan experiment or the studies concerning thunderclouds have shown that droplets can be considerably influenced by an excess of electric charges. Indeed, an excess of charges can affect the intrinsic properties of a droplet, such as its natural oscillation frequency or its internal pressure. Moreover, the electric charges in excess in droplets can also interact with external electric fields. In this thesis, we investigated the influence of electric charges on millimetric droplets that are electrically isolated. In literature, research on isolated charged droplets are mainly focused on droplets with a micrometric size. The lack of studies concerning millimetric charged droplets is explained by the difficulty storing them while avoiding charge leakage. In order to answer to this issue, we examined three storage systems limiting the charge leakage: the microgravity, the vibrating bath method and the Leidenfrost effect. Through these systems, we studied the influence of electric charges on the droplet physical properties, but also the interaction between the charged droplet and its storage system. Furthermore, we investigated the interaction of charged droplets with external electric fields. More precisely, we studied the interaction between two electrically charged droplets and the interaction between one charged droplet and an external homogeneous electric field. A first set of experiments on electrically charged droplets allowed us modeling the charge migration process in liquids and the charge leakage from a millimetric droplet. In particular, we identified and modeled a new mechanism of charges leakage occurring at a time scale of several minutes. Moreover, we confirmed the influence of the electric charges on the droplet surface energy previously deduced from experiments on micrometric droplets. Concerning the three storage systems, the experiments performed in microgravity allowed us describing the influence of the electric interaction on the impact between two charged droplets. The diverse behaviors observed were compared to the cases of impacts between two neutral drops. On a different note, the study of a charged droplet moving on the surface of a vibrating bath because of the influence of an external electric field gave new insights on the interaction between a bouncing droplet and a viscous liquid bath. For example, we observed a ``go-stop" motion during which the droplet horizontally moves when it bounces away and is stopped during its interaction with the liquid bath. We showed that this motion occurs when large droplets are influenced by a weak electric field. Droplets with this kind of motion move with a constant average speed, which makes them easily manoeuvrable. Therefore, the control of the droplet motion led to the development of a new microfluidic prototype. Via this new setup, basic microfluidic tasks can be performed without polluting droplet via contacts with solids or liquids. With these results in mind, we also examined the interaction between two charged droplets bouncing on the vibrating bath. This study brought new insights on the interaction between two charged droplets. Indeed, we observed that two drops with the same charges tend to remain at an equilibrium distance. Our study showed that this equilibrium distance is due to the compensation of the electric repulsion by capillary attraction at the surface of the vibrating bath. Finally, our study of charged droplet in Leidenfrost state on a liquid bath led to a better understanding of the interaction between charged liquid interfaces. Indeed, we showed that electric charges cause the early coalescence of charged droplet because of an increase in the vapor layer drainage. We conclude from our results that an excess of electric charges influences ostensibly the intrinsic behavior of a droplet and its interaction with the environment. Furthermore, each storage system studied brought answers to specific issues. The study of the impact between charged droplets in microgravity outlines new explanations on the behavior of thunderclouds. The results accumulated on the micromanipulation of charged droplet bouncing on a vibrating bath opens the way to a new kind of microfluidic system. Finally, the study on the charged Leidenfrost droplets describes new ways to investigate the influence of electric charges on liquid interfaces.

In the first part of this thesis, we investigate the dynamics of open quantum systems coupled in a purely dephasing way to their environment, which is a paradigmatic example of decoherent dynamics. We focus more specifically on the emergence of non-Markovian effects in the dynamics of such open systems. After reviewing different measures of non-Markovianity, we obtain an analytical expression for the reduced density operator of a finite-dimensional open system coupled with its environment through time-dependent purely dephasing interactions. When the system is initially uncorrelated with its environment, we obtain the time-local master equation that describes its dynamics. In this case, we also derive an exact expression for the witness of non-Markovianity based on the volume of accessible states of the open system. These results are then applied to spin systems coupled to either a spin or a bosonic environment. For spin environments, we study non-Markovianity for different ranges of interaction and in different thermodynamical limits. Moreover, the generation of entanglement between the open system and its environment as well as between spins within the open system is investigated. For bosonic environments, we study the effect of the initial state of the environment on non-Markovianity. We also show that time-dependent modulations of the system-environment coupling can dramatically alter the non-Markovian behavior of the open system. Finally, we consider the case of an open quantum system simultaneously coupled to spin and bosonic baths. In that case, we study how the coupling between the different baths and of the presence of initial correlations in the environment affect the non-Markovianity displayed by the open quantum system. The second part of this thesis is a contribution to the investigation of how quantum systems can be used to process information. More precisely, we focus on the implementation of gate-based quantum computing using neutral atoms trapped in optical or magnetic lattices and interacting through dipole-dipole interactions when excited to Rydberg states. In this context, we present protocols implementing two-qubit quantum gates between atoms far apart in the lattice. To this end, we propose to use a chain of non-coding ancilla atoms to transfer the state of the control qubit to an atom near the target qubit. Before discussing entangling gates between distant qubits, we present two protocols that implement state transfer along a chain of two-level atoms relying respectively on single- and two-atom laser pulses combined with dipole blockade. After reviewing protocols implementing CZ and CNOT gates between nearby qubits using dipole-dipole interactions, we show how the state-transfer protocols previously presented can be used to perform entangling gates between distant qubits. The performance of the state-transfer and gate protocols are assessed using the process fidelity. Moreover, we provide a theoretical explanation for the behavior of the process fidelity with respect to dissipation and imperfect blockade.

DSSC is a 3rd generation photovoltaic technology with potential to economically harvest and efficiently convert photons to electricity. Full solid state-DSSC based on solid polymer electrolyte prevents the solvent leaking and evaporation during cell fabrication and operation, which will effectively prolong the cell life time. However, it suffers from low ionic conductivity and poor pore infiltration. The present thesis is dedicated to the concomitant development of polysiloxane-based polymer electrolytes on one side, and TiO2 photoanodes with tuned porosity on the other side, and their incorporation in solid-state dye-sensitised solar cell (ss-DSSCs), with the aim to improve their photovoltaic efficiency and the long term stability. To best of our knowledge, DSSCs comprising bimodal TiO2 layers and polysiloxane electrolytes have never been reported. The ionic conductivity and tri-iodide diffusion coefficient of the polysiloxane-based poly(ionic) liquids (PILs) were largely improved by adding of ionic liquids (ILs) or et hylene carbonate (EC), achieving ionic conductivities of 10−4 -10−3 S cm−1. The DSSCs fabricated with the optimized electrolytes showed efficiencies up to 6%, with long term stability for 250 days. Bimodal TiO2 films with dual porosity (meso- and macro-porosity) were fabricated by spin-coating, by using soft and hard templating. The dual templated films benefit from increased pore size while maintaining high surface area for dye adsorption. Bimodal films were shown to be more efficient when tested with polymer electrolytes, having comparable efficiencies with liquid electrolyte when in DSSCs, despite lower dye uptake. This thesis brings a significant contribution to the field of DSSCs as efficient and stable solar cells were prepared from newly synthesized polymer electrolytes and bimodal films.

Compounds with the general formula A2BO6, where A is a transition metal cation and B is a cation with a high nominal charge, have attracted interest due to the variety of their structures and the complexity of their physical properties. Within this family, Fe2WO6 is an interesting compound because of its three polymorphic structures (α, γ, and more recently β) and its magnetic and photoelectrochemical properties. However, the studies reported so far presented only a partial view of Fe2WO6, since they mostly focused on one polymorph (γ) and did not investigate the influence of the structures on the functional properties. In this manuscript, we present the study of the crystal chemistry of the three polymorphs of Fe2WO6 and report about their optical and magnetic properties. The developments of repeatable spray-drying synthesis procedures from aqueous Fe/W solutions and quantification by ICP-OES allowed us to prepare homogeneous powders with well-controlled Fe/W ratios. A solid solution domain was identified for Fe/W ratios in the [1.91-2.00] range and the general formula Fe2-2xW1+xO6 was proposed to describe the off-stoichiometry. The unknown nuclear and magnetic structures of polymorph β were solved by combining synchrotron X–ray diffraction, neutron diffraction and physical properties measurements. The structure of polymorph α was revised and the impact of the synthesis-dependent cationic ordering in polymorph γ is discussed. The polymorphic transitions and the structural and magnetic relationships with other related compounds (FeWO4 and Fe2TeO6) are also presented.

Superconducting Bi2Sr2CaCu2O8+δ tubes and rods manufactured at Nexans SuperConductors gmbh are known to achieve excellent performances despite apparently weak texture. A full texture characterization of the bulks has been performed at different scales, using X-ray diffraction and neutron diffraction. The texture experiments confirm that samples exhibit weak textures and reveal some differences between local and macroscopic texture. When investigating how the physical properties might be affected by texture, it was observed that the critical current density Jc is not much influenced by the direction of measurement, while the normal state electrical resistivity displays a ratio of 2.3 between the resistivity along the sample axis and the resistivity along the radial direction. These observations are qualitatively consistent with the texture results. A quantitative study based on the geometric mean of the resistivity tensor compares the experimental data with values calculated from the measured orientation distribution and literature single crystal data. It turns out that, despite the weak texture, a large fraction of the observed resistivity anisotropy is explained by texture effects, because of the very strong anisotropy between the resistivity along the c-axis and the resistivity in the ab-planes. Lastly, the influence of sample oxygenation on Tc and Jc has been studied. In these bulks, Tc as a function of oxygen content follows a bell-shaped curve, confirming literature results for other superconducting samples. Jc is also influenced by the oxygen content; interestingly, the maximum Jc is not obtained for the same oxygen content as the maximum Tc.

The shaping of functional materials and the control of their texture at all length scales are sine qua non conditions for the improvement of current systems. This PhD project consists in creating complex solid architectures using interdisciplinary methods such as sol-gel chemistry or complex fluids physics. Therefore, it is possible to synthesize Titanium Dioxide macroscopic fibers or films which possess a hierarchical porosity. This organization allows the optimization of the matter transport (liquid/gaz) for air depollution application (photocatalysis) or dye-sensitized solar cells. In another project, we were able to control the alignment of zinc oxide nanorods within a macroscopic fiber. This alignment provides to the fiber an anisotropic photoluminescence behavior which can be useful for switching devices application. Finally, we synthesized anisotropic particles and nano-sheets of polypyrrole (conducting polymer) in order to obtain smooth thin films presenting interesting electrical properties. The objective was to use them as electrolyte and/or electrode in dye-sensitized solar cells.

Worldwide metallic aluminium production involves the Hall-Héroult process where the metal is electro-deposited from aluminium oxide solubilised in a molten NaF-AlF3-CaF2 mixture at around 950°C. The cryolitic melt is conveniently characterised by both the molar NaF/AlF3 ratio and the Al2O3 content. Nowadays the Hall-Héroult process remains the more economically efficient process even if it still suffers from a high consumption of energy. In particular the overvoltage required by the electrolysis is strongly dependent on the melt composition, especially regarding the Al2O3 content. Controlling the industrial baths composition during the process is therefore critical to reduce the energy loss. Unfortunately there is, up to now, no in situ direct analytical method to do so. Considering our experience in the study of such highly corrosive media by Raman spectroscopy, that technique has been applied to directly determine the melt composition. Three sets of reference spectra are considered in this study, each of them recorded with a different setup. The employed setups were developed to reach progressively, at the laboratory scale, a design that is suitable for a plant application. Eventually, a high quality spectrum can be recorded by the top of the melt, in less than 20 seconds. The employed apparatus is found to influence significantly the shape and quality of the spectra, and consequently their involvement in the quantification. A complex digital treatment of the spectral data acquired is necessary because all Raman bands of interest strongly overlap and some are situated close to the Rayleigh decay. Two main quantitative procedures for the melt composition determination are studied. The first one, the AutoAnalysis procedure, developed in the past and adapted here to the new data, gives reliable predictive results for both the NaF/AlF3 molar ratio and the alumina content. They can be determined with an absolute deviation of 0.06 molar ratio unit and 0.5 wt% respectively. However, the intensity normalisation, required for comparing the intensities of different spectra, relies on the Rayleigh decay that is likely to change with the experimental conditions in the plants. In our second quantitative procedure, the NormaAnalysis procedure, the intensity normalisation is based on the equilibria taking place in the melt. Since those equilibria do no differ with the experimental setup, the NormaAnalysis procedure can be imported to the industrial field. The predicted composition is also evaluated with a good precision: the NaF/AlF3 molar ratio and the alumina content can be determined with an absolute deviation of 0.08 cryolitic ratio unit and 0.3 wt% respectively. It is concluded that the composition of the melt can now be determined with our NormaAnalysis procedure, from a single Raman spectrum, recorded with a Raman apparatus exportable for an in situ measurement on the industrial cells.

The ionic liquids are salts having the particularity of being molten at temperatures lower than 100 °C. Consequently, these new solvents mainly made up of ions have original physicochemical properties. Within a few years, the ionic liquids passed from a laboratory curiosity to a true field of research impossible to circumvent, currently in full rise. Indeed, the replacement of usual organic solvents in catalytic and/or separation processes by these neoteric solvents offers many advantages but also new opportunities for the “Green Chemistry”. However, the systematic exploitation of the ionic liquids as reactional media rests in particular on the understanding of their chemical properties, which for some of them are still scarcely known, such as the acidity for example. This thesis thus aims to undertake a study of the acido-basic properties of (and in) these solvents, and more particularly to determine the accessible levels of acidity for acid solutions (with added HOTf or HNTf2) in second generation ionic liquids such as [HNEt3][NTf2], [BMIm][NTf2], [BHIm][NTf2], [BMIm][BF4], [BMIm][OTf], [BMIm][PF6] and [BMIm][SbF6]. In order to evaluate these acidity levels, we propose two different methods, each one resting on an extra-thermodynamic assumption. The first, the Hammett acidity function H0, is based on the protonation equilibrium of indicators whose pKa's are proposed as solvent independent. The second, the Strehlow potentiometric function R0(H+), consists in measuring, in a given solvent, the electrochemical potential of the proton compared to the ferricinium/ferrocene redox couple whose potential is supposed to be independent of the solvent, and then to refer it versus the Normal Hydrogen Electrode (NHE) in water. The two methods lead to the same conclusions. Firstly, the ionic liquids are generally contaminated by residual basic impurities (from solvents needed for the synthesis…) which need to be neutralized before reaching the acidity characteristic of the medium. The levels of acidity then obtained are very high and can reach values as high as R00(H+) = -10 in the case of [BMIm][BF4]. Then, the accessible level of acidity in an ionic liquid depends mainly on the nature of its anion, and not of that of its cation. We thus obtain the following classification, by decreasing acidity: [PF6-] > [BF4-] > [NTf2-] > [OTf-], indicating that the triflate is the more solvating anion It was found however that the Hammett acidity function led, for the same concentration in acid, to different levels of acidity, depending on the indicator used. The ionic liquids would consequently be media less dissociating than estimated in the literature and the Hammett function would then be related to an apparent acidity (H0)app, underestimating the real acidity. Finally, a difference in acidity between HNTf2 and HOTf is observed in [BMIm][NTf2] and [BMIm][OTf], HNTf2 showing an acidic character stronger than HOTf. On the other hand, in [BMIm][OTf] these two acids show the same acidity since that of HNTf2 has been leveled by solvent.

Superconductivity is an electronic state of matter arising from the existence of a common wave function with a coherent phase extending on a truly macroscopic scale. One major manifestation of this striking quantum phenomenon is the dissipationless transport of electrical current, an asset deserving particular attention in the present times where the efficient energy distribution has become of utmost importance. Unfortunately, the motion of quantum units of magnetic flux (so-called vortices or fluxons), which is an unavoidable side-effect found in superconductors in the presence of transport currents and magnetic fields, severely limits the conditions to preserve dissipationless transport. This poses a challenge for achieving the functionalization of superconducting materials and threatens their spectrum of applications. It is widely known that any inhomogeneities (either material imperfections, or ones made artificially), which locally suppress superconductivity on the scale comparable to the core of the vortex, can pin the vortex and delay the onset of the vortex motion to higher applied currents. In recent years a substantial effort has been made to minimize the effects of current-induced vortex motion by tailoring arrays of artificial pinning centers. Besides improving the critical parameters of the superconducting state, a pinning matrix can be used for the manipulation of vortex matter, thus directly affecting the vortex dynamics, such as rectification of vortex motion under an ac drive (vortex diode) by introducing asymmetric pinning landscapes. In the literature one can find that the realization of the anchoring of the vortices can be based on nanostructured arrays of perforations, chemically grown defects, permanent nanomagnets, or even pinning sites produced by heavy ion bombardment. All of those realizations are based on a permanent imprint on the superconductor, without any possibility for subsequent modifications in the distribution and strength of the pinning. The principal objective of this thesis is to investigate the dynamical behavior of vortex matter under an entirely new kind of pinning landscape consisting of spatial and temporal modulation of the superconducting condensate. A particular case of spatial modulation is considered in a constricted structure where current lensing can cause extremely high vortex velocities. Subsequently, a time-dependent thermal potential introduced to the superconducting condensate will cause stroboscopic resonances during the vortex motion - a phenomenon that cannot be observed in the systems with static pinning imprints. Finally, a study of electronic gating is presented, where the local properties of superconductor, such as mean free path, or electronic band structure in general, can be influenced electronically. This is a completely unexplored interdisciplinary research topic, which will eventually allow one to manipulate individual vortices in superconducting materials by means of spatially confined and temporally controlled thermal and electromagnetic excitations. Furthermore, such techniques can provide one fundamental insight in different states of the vortex matter with respect to variation of the transport current, highly relevant for understanding the resistive state of superconducting materials and their applications.

Quantum entanglement is a key property of quantum information theory, that is at the heart of numerous promising applications in fields such as quantum cryptography, quantum computing or quantum sensing. In the past decades, the advent of such innovative technologies has reinforced the need for a better understanding of entanglement. The aim of this thesis is to contribute to this effort through the development of new tools targeting the characterization of several features of entanglement. Concerning the issue of entanglement detection, we present an optimization of the approach that exploits the concept of generalized concurrences to solve the separability problem for pure states. We then reformulate the separability question of mixed states into a matrix analysis problem, from which we obtain general separability criteria for multipartite states of ranks two and three. We also briefly discuss some properties of separable states. In particular, we characterize optimal separable decompositions of symmetric (i.e. permutation invariant) states of two and three qubits with maximal rank properties. Regarding the quantification of entanglement, we propose a function to quantify the entanglement of symmetric multiqubit states within classes of entangled states gathering states that are stochastically equivalent through local operations assisted with classical communication. This function establishes a link between the amount of entanglement of a symmetric state and the distribution of its Majorana points on the Bloch sphere. We finally investigate the robustness of entanglement with respect to particle loss and provide a full description of all multiqubit states that are fragile for the loss of one of their qubits. For symmetric states, the fragility for the loss of one qubit is shown to be related to a particular symmetry of the Majorana points.

Among current frequency standards, hydrogen masers still exhibit the best stability perfor- mances and are therefore very interesting in the context of the global positioning system. The IPNAS institute (University of Liège) is currently part of a consortium whose purpose is to implement and characterise a compact passive hydrogen maser for space environment. The aim of this thesis is to perform original and innovative studies on the physical package of a hydrogen maser in order to define the guidelines of the design and to propose a concrete implementation scheme that takes into account the desired performances, as well as constraints imposed by the space environment. We first give a state-of-the-art review of hydrogen masers. Then, we re- port the modelling of the main components of a passive hydrogen maser for space environment. The resonant cavity is the key component of hydrogen masers since it determines their overall dimensions. We develop finite element method (FEM) models of a magnetron cavity and we benchmark them against an existing cavity. These models allow us to compute the filling factor, the quality factor, as well as the frequency temperature coefficient (FTC) of the cavity and bulb assembly. We show that, for a proper set of cavity and bulb parameters, it is possible to reduce the FTC easily within mechanical tolerances. We use these FEM models in order to suggest a design of magnetron cavity and bulb assembly that takes into account constraints imposed by the storage volume, the filling factor, the quality factor and the FTC. We then carry out a study of the residual and static magnetic fields within the bulb of a hydrogen maser. The related models include a 6-pole state selector, a solenoid and a set of magnetic shields. For all these components, we suggest a design that is compliant with maser operation at a mag- netic field intensity of 1 mG, which yields good stability performances. Afterwards, we perform Monte-Carlo simulations of the hydrogen beam optics system used to implement the state se- lection that generates a population inversion within the bulb. Combining state selection and magnetic field studies, we determine an acceptable range of distances between the storage bulb entrance and the state selector output. Then, we report the theoretical study of the molecular hydrogen source and radio-frequency dissociator for a spaceborne hydrogen maser. We develop an analytical model for a plasma of pure hydrogen gas in order to design the discharge cell used to dissociate the molecular hydrogen gas. Molecular hydrogen is supplied thanks to a metal hydride storage container. We estimate the required hydrogen flux for maser operation, we se- lect a commercial metal hydride container and then evaluate the requirements on the container output pressure in order to guarantee a 15 years operating lifetime. At last, we sum up the main parameters of the proposed maser design and we report the expected values of the maser operational parameters as well as the predicted Allan deviation.

Many applications of modern electronic devices are based on thin film geometries including sharp turns, holes and exhibiting inhomogeneities. The inevitable detour of current streamlines around such obstacles cause an inhomogeneous current density profile giving rise to current crowding. The goal of my research is to highlight the current crowding effects in micro and nanopatterned superconducting films.

Le présent travail est dédié à la conception, la modélisation et l'implémentation d'un dispositif expérimental visant, pour la toute première fois, à piéger et à refroidir par laser des atomes de fer. Dans ce cadre, une contribution majeure au développement de ce dispositif est rapportée ici. Cette contribution comprend notamment la mise en place de dispositifs de stabilisation laser sub-MHz adaptés aux transitions spécifiques de l'atome de fer à 372 et 358 nm, ainsi que l'implémentation d'un ralentisseur Zeeman sur des faisceaux de ces atomes issus d'un four à haute température. Par ailleurs, des déterminations à ultra-haute résolution de grandeurs spectroscopiques fondamentales liées aux effets isotopiques et de structure hyperfine des transitions à 372 nm et 373.7 nm du fer sont rapportées pour la toute première fois. Une comparaison extrêmement précise de la position de la raie du fer à 358 nm par rapport à la raie de l'iode moléculaire R(90)3-10 à 13957.8542 cm-1 est également présentée. Cette comparaison met à disposition avec la précision requise la dernière donnée spectroscopique qui était inconnue dans la littérature scientifique pour une mise au point optimale du refroidissement laser d'atomes de fer. Elle ouvre la voie à l'implémentation finale du dispositif présenté dans ce travail.

Complex oxides exhibit a wide range of physical properties making them very attractive for future electronic and device applications. Although more and more studied, additional scientific investigations are required, especially in oxide interfaces, where new and amazing phenomena can arise. A prototypical example is the LaAlO3-SrTiO3 interface that appear to be conducting, magnetic or even supra-conducting while these properties are not present in the LaAlO3 and SrTiO3 bulk insulator compounds. The conductivity arises from the formation of a highly localized electron gas at the interface which exhibits a different behavior than the one at semiconductor interfaces. Even nowadays, its exact origin, intrinsic versus extrinsic, is still intensively debated. The existence of an electric field in LaAlO3 used as a key feature of models based on an intrinsic origin is highly controversial. In these models, the closing of the band gap with increasing LaAlO3 film thicknesses finally results to a Zener breakdown and to the metal/insulator transition. In this Ph.D. thesis we aim to investigate the various consequences of the presence of an electric field in LaAlO3 through first-principles calculations in pristine LaAlO3-SrTiO3 interfaces. First, using both experimental and theoretical structural distortions in LaAlO3, we predicted a lattice expansion via an electrostrictive effect, supporting the existence of an electric field in LaAlO3. Second, the metal/insulator transition was tuned with regards to the intensity of the electric field in the film, which was controlled by the composition of a solid solution between LaAlO3 and SrTiO3. The theoretical results match the experimental one where, nevertheless, extrinsic origin mechanisms are allowed and defects are present. These two works are in favors of an intrinsic origin of the electronic gas observed in LaAlO3-SrTiO3 heterostructures. In addition, a relationship between the sheet carrier density and spatial extension of the gas was established and thus setting an intrinsic threshold to the sheet carrier concentration. At lower density the electrons are strictly localized close to the interface while above this value the carriers start to spill into the SrTiO3 substrate.

Multiferroic materials have attracted much interest during the recent years. Our study is devoted to a prototypic system: yttrium manganite. In particular, we focus on the ferroelectric properties in bulk and in thin film forms. Yttrium manganite belongs to the class of ABO3 compounds. Most theoretical studies of ferroelectricity to date were concentrated on cubic perovskite ABO3. Yttrium manganite is hexagonal and is an improper ferroelectric. We were interested to study theoretically and experimentally how these two features behave in thin film form. Our study is organized as follows.

The present thesis focuses on the physics of various solid-state systems sharing the common feature of involving 3d electrons with a low-dimensional aspect for transport, and studied using Density Functional Theory. Exploiting an original hybrid functional approach for the exchange-correlation energy, with improved accuracy compared to local/semi-local functionals, we present the seminal two-dimensional electron system (2DES) at the (001) interface of band insulators SrTiO3 and LaAlO3 , and review two of the most popular hypotheses about its origin, namely the electric-field driven Zener breakdown model and polarity-induced surface oxygen vacancies model. This analysis is extended to the interface between SrTiO3 and the (Sr1−xLax)(Ti1−xAlx)O3 alloy. We also study, based on experiments and theoretical modelling, how the composition of the alloy overlayer affects the charge density of the 2DES. We then address the effect of structural confinement on the 2DES when the host layer thickness is reduced toward the very-thin limit, and how such effects are witnessed in angle-resolved photo-emission spectroscopy experiments. We study the effects of capping the SrTiO 3 /LaAlO 3 het- erostructures with SrTiO 3 , highlighting how experiments may be interpreted from the aforementioned electric-field driven models. This work also focuses on the thermoelec- tric properties of layered oxides, specifically Ca3Co4O9 and SrTiO3 -based superlattices, discussing the relevance of their layered structure for improving the thermoelectric prop- erties. Importing the concepts of low-dimensional transport found in SrTiO3 -based sys- tems to the iron-based Heusler Fe2YZ family, we explore the effect of electron doping, highlighting magnetic instabilities related to their Fe 3d orbitals, which impact signif- icantly the thermoelectric properties. Finally, shifting our attention on Fe2TiSn, we rationalize experimental results provided by collaborators from first-principles, address- ing the role of native defects and their relevance for tailoring transport.

Thermoelectricity has been regarded as one of the most promising strategies for clean, low-cost and environmental friendly sustainable energy for several decades. Perovskite oxides, like SrRuO3, are considered as a potential thermoelectric material for low-cost and large-scale thermoelectric applications due to their good thermal and chemical stability in a wide temperature range, great flexibility for structural and compositional manipulating, and environmental friendliness. This thesis is devoted to a theoretical study of the lattice dynamics and thermoelectric properties of materials, like SrRuO3 perovskite and MgAgSb-based materials. Firstly, to obtain insight into the lattice dynamics of the SrRuO3, the phonon-related properties are presented and contributes to rationalize better why many ABO3 perovskites, including metallic compounds, exhibit an orthorhombic ground state. Then the thermoelectric properties of SrRuO3 are investigated by combining first-principles calculations and Boltzmann transport theory, revealing the relationship between the exchange-correlation functionals and the thermoelectric quantities. Furthermore, based on the first-principles calculations, effective model potentials for SrRuO3 are constructed providing access to the finite-temperature properties and phase-transitions. Additionally, the electronic structure and thermoelectric properties of a class of new emerging MgAgSb-based materials, which are promising for room-temperature thermoelectric applications, are also studied and the optimization strategies are proposed for the improvement of thermoelectric performance.

In the present thesis we present theoretical studies of perovskite compounds under uniaxial mechanical constraints combining first-principles DFT calculations approach and phenomenological Landau theory. ABO$_3$ perovskites form a very important class of functional materials that can exhibit a broad range of properties (e.g., superconductivity, magnetism, ferroelectricity, multiferroism, metal-insulator transitions\ldots) within small distortions of the same simple prototype cubic structure. Though these compounds have been extensively studied both experimentally and computationally, there are still unresolved issues regarding the effect of pressure. In recent years, strain engineering has reported to be an original approach to tune the ferroelectric properties of perovskite ABO$_3$ compounds. While the effect of epitaxial biaxial strain and hydrostatic strain is rather well understood in this class of materials, very little is yet known regarding the effect of uniaxial mechanical constraints. Our study is motivated by the little existing understanding of the effect of uniaxial strain and stress, that has been up to now almost totally neglected. Two prototype compounds are studied in detail: PbTiO$_3$ and SrTiO$_3$. After a general introduction on ABO$_3$ compounds and calculations techniques (ab initio and phenomenological Landau model), we studied the effect of mechanical constraints in these compounds in our thesis. PbTiO$_3$ is a prototypical ferroelectric compound and also one of the parent components of the Pb(Zr,Ti)O$_3$ solid solution (PZT), which is the most widely used piezoelectrics. For PbTiO$_3$, we have shown that irrespectively of the uniaxial mechanical constraint applied, the system keeps a purely ferroelectric ground-state, with the polarization aligned either along the constraint direction ($FE_z$ phase) or along one of the pseudo-cubic axis perpendicular to it ($FE_x$ phase). This contrasts with the case of isotropic or biaxial mechanical constraints for which novel phases combining ferroelectric and antiferrodistortive motions have been previously reported. Under uniaxial strain, PbTiO$_3$ switches from a $FE_x$ ground state under compressive strain to $FE_z$ ground-state under tensile strain, beyond a critical strain $\eta_{zz}^c \approx +1$\%. Under uniaxial stress, PbTiO$_3$ exhibits either a $FE_x$ ground state under compression ($\sigma_{zz} < 0$) or a $FE_z$ ground state under tension ($\sigma_{zz} > 0$). Here, however, an abrupt jump of the structural parameters is also predicted under both compressive and tensile stresses at critical values $\sigma_{zz} \approx$ $+2$ GPa and $- 8$ GPa. This behavior appears similar to that predicted under negative isotropic pressure and might reveal practically useful to enhance the piezoelectric response in nanodevices. The second compound of interest is SrTiO$_3$. It has been widely studied in the past decades due to its unusual properties at low temperature. In this work, we have extended our previous investigations on PbTiO$_3$ by exploring theoretically the pressure effects on perovskite SrTiO$_3$ combining the first-principles calculations and a phenomenological Landau model. We have discussed the evolution of phonon frequencies of SrTiO$_3$ with the three isotropic, uniaxial and biaxial strains using first-principles calculations. We also reproduce the previous work done in SrTiO$_3$ with epitaxial strain and hydrostatic strain. Finally, we have calculated the phase diagram of SrTiO$_3$ under uniaxial strain, as obtained from Landau theory and discussed how it compares with the first-principles calculations.

High-performance piezoelectrics are key components of various smart devices and, recently, it has been discovered that (Ba,Ca)(Ti,Zr)O 3 (BCTZ) solid solutions show appealing electromechanical properties. Nevertheless, the microscopic mechanisms leading to such features are still unclear and theoretical investigations of BCTZ remain very limited. Accordingly, this thesis analyzes the properties of various compositions of (Ba,Ca)TiO3-Ba(Ti,Zr)O3 solid solutions by means of first-principles calculations, with a focus on the lattice dynamics and the competition between different ferroelectric phases. We first analyze the four parent compounds BaTiO3, CaTiO3, BaZrO3 and CaZrO3 in order to compare their properties and their different tendency towards ferroelectricity. Then, the core of our study is a systematic characterization of the binary systems (Ba,Ca)TiO3 and Ba(Ti,Zr)O3 within both the virtual crystal approximation (VCA) and direct supercell calculations. When going from BaTiO3 to CaTiO3 in (Ba,Ca)TiO3, the main feature is a gradual transformation from B-type to A-type ferroelectricity due to steric effects that largely determine the behavior of the system. In particular, for low Ca-concentration we found out an overall weakened B-driven ferroelectricity that produces the vanishing of the energy barrier between different polar states and results in a quasi-isotropic polarization. A sizable enhancement of the piezoelectric response results from these features. When going from BaTiO3 to BaZrO3 in Ba(Ti,Zr)O3, in contrast, the behavior is dominated by cooperative Zr-Ti motions and the local electrostatics. In particular, low Zr-concentration produces the further stabilization of the R3m-phase. Then, the system shows the tendency to globally reduce the polar distortion with increasing Zr-concentration. Nevertheless, ferroelectricity can be locally preserved in Ti-rich regions. We also found out an unexpected polar activation of Zr as a function of specific atomic ordering explained via a basic electrostatic model based on BaZrO3/mBaTiO3 superlattice. A microscopic factor behind the enhanced piezoelectric response in BCTZ, at low concentration of Ca and Zr, can thus be the interplay between weakened Ti-driven and emerging Ca-driven ferroelectricity, which produces minimal anisotropy for the polarization. In addition, our comparative study reveals that the specific microscopic physics of these solid solutions sets severe limits to the applicability of the virtual crystal approximation (VCA) for these systems.

In this thesis we present an original ab-initio study of the evolution of antiferrodistortive (AFD), anti-polar electric (APE), and ferroelectric (FE) instabilities in various ABO3 oxides of perovskite structure, as well as their structural and dynamic properties. The main goal is to understand better the microscopic origin of the antiferroelectricity exhibited in these compounds. Three prototypical compounds are studied in detail : PbZrO3 , NaNbO3 , and SrZrO3. After a general introduction on ABO3 compounds, and the ab-initio techniques, we review the concept of antiferroelectricity in perovskites, highlighting some ambiguities in the usual definition and the necessity of turning to what we call a modern definition of antiferroelectricity. First, we highlight that it is the rigidity of the oxygen cage that tends to favor the FE distortion compared to the APE instability. Although illustrated on BaTiO3 , this argument is general, and confirmed by the inspection of the phonons dispersion curves of the ABO3 compounds in whom the strongest instability of the FE/APE branch is systematically at Γ. We show that the emergence of a stable or meta-stable APE distortion appear naturally through a coupling with other instabilities. The presence of AFD modes turns out to be a concrete way to create mixed FE/AFD and APE/AFD phases, crucial for the emergence of antiferroelectricity (AFE). This clarifies why the known AFE compounds systematically include AFD distortions. In this context, since the FE, APE and AFD instabilities are usually in competition, the coexistence of FE, APE and AFD instabilities of strong amplitudes seems required to create mixed phases combining them. This establishes the context convenient to the development of FE and AFE metastable phases close in energy. Another important element concerns the need of a first order AFE-FE transition under electric field producing a double hysteresis loop, typical of AFE compounds. Here also the AFD modes could play a key-role by allowing the emergence of FE/AFD and APE/AFD phases close in energy and developing distinct tilt patterns. These various elements give a new perspective on AFE and allow us to have a more precise idea of the origin of the AFE behavior in perovskites. We identify some key intrinsic characteristics allowing the prediction of materials with the propensity of developing an AFE behavior.

The Metal-Insulator transition (MIT) upon cooling at elevated temperatures is a fascinating effect that is observed in some ABX 3 perovskite compounds with specific electronic configurations on the transition metal B cation d states. These compounds behave completely oppositional to ordinary metals who become better conductors upon cooling. At the MIT temperature these specific perovskite compounds show electron localization accompanied by cooperative lattice distortions that deform the BX 6 corner shared octahedral network. Since the discovery of these MITs, theoretical concepts about its origin have been proposed and debated. In this work we study from ab initio density functional theory (DFT) calculations LaMnO 3 and the alkaline earth ferrite series AFeO 3 who all have a formal d 4 occupation. As a basis to our discussion we introduce canonical notations, that were missing until now, for lattice distortions in perovskites that distort the X anion octahedra and are connected to MITs. While LaMnO 3 shows electron localization through orbital ordering - the appearance of a static order of Mn d orbital occupations-, CaFeO 3 shows electron localization through charge ordering - the appearance of a static order of formal Fe charges-. From DFT calculations we can show that both mechanisms are compatible with the concept of a Peierls transition and are enabled by octahedral rotations. From Monte-Carlo simulations we show furthermore that spin disorder in the paramagnetic phase is a key ingredient to explain the high MIT temperatures. Last we study the influence of external epitaxial strain on these compounds. Here, our canonical notations help to discriminate internal relaxations and external constraints. Our calculations show that in LaMnO 3 epitaxial strain alone can provoke a change from an antiferromagnetic to an ferromagnetic order without the necessity of Oxygen vacancies as has been often speculated. In CaFeO 3 epitaxial strain can provoke a change from the bulk charge ordering MIT to an orbital ordering MIT, which explains the experimental finding of a strongly elevated MIT in CaFeO 3 thin films. In the second part of this work, we present methodological developments for generating effective lattice potentials by a polynomial expansion that describe electronic potential energy surfaces in which atomic nuclei move. This is the so-called second-principles approach. The aim thereby is to replace the computational intensive self consistent DFT procedure by an lightweight evaluation of a polynomial depending on nuclear positions. If successful, this approach provides a mean to achieve a scale-transfer retaining the accuracy of ab initio calculations applicable to large scale systems with many thousand atoms and statistical simulations. In a proof of concept study we apply this approach to the perovskite CaTiO 3 . The retained effective potential reproduces with good accuracy a number of ab initio quantities. Moreover, in the low to average temperature range (T<=1000K) the lattice dilatation of CaTiO 3 is well described. In the highest temperature range the effective potential deviates from the experimentally measured lattice dilatation and proposed phase transition sequence that is itself, however, debated. We conclude that the lattice dilatation properties can be refined by extending the lattice expansion and that a reexamination of the high temperature phase sequence of CaTiO 3 due the the information of our effective potential might be meaningful. Finally, we also highlight that there exists a strongly ferroelectric low energetic phase of CaTiO 3 whose stabilization through external constraints is part of ongoing research.

ABO3 perovskites, with a B cation in a eg1 electronic configuration, have long been the focus of extensive research due to their promising metal-insulator transition (MIT) and related technological applications. However, because of the complex interplay between charge, orbital, spin and lattice degrees of freedom, distinct families of compounds like RMnO3, RNiO3, and AFeO3 perovskites, with R = rare-earth and A = alkaline-earth element, can exhibit quite different behaviors showing either orbital ordering and Jahn-Teller distortions or charge ordering and breathing distortions. What makes eg1 perovskites even more interesting is that the MIT property is extremely sensitive to external stimuli (e.g., strain, pressure, dimensionality, and stoichiometry). This provides an ideal platform to purposely tune related properties for versatile applications and has recently fueled significant research interest. In particular, epitaxial thin films and superlattices are two effective and promising approaches to tune MIT temperature (TMI ) in device applications. The great and recent advances in the synthesis of high-quality epitaxial thin films and superlattices spark rapidly increasing interest in the control of TMI. However, despite the effectiveness of tailoring TMI, their actual influence is quite complex, and different studies may give inconsistent conclusions. Over the past years, continuous efforts have been devoted to understanding the microscopic mechanism behind the evolution of TMI in eg1 perovskites thin films and superlattices. Nevertheless, for the thin films, there is still a lack of a reliable and general explanation to simultaneously describe the evolution of TMI under different strain conditions. In terms of superlattices, the roles of interface effect including the structural and electronic coupling, dimensionality, and confinement effect remain to be understood. In this thesis, combining first-principles calculations and Landau theory analysis, we explore comprehensively the interplay between lattice, strain, spin, and charge degrees of freedom in eg1 perovskites thin films and superlattices. We firstly investigate the origin of breathing distortions and charge ordering induced MIT in CaFeO3 from lattice coupling and electron-phonon coupling. Another aim is to rationalize the origin of a huge increase of TMI recently observed in CaFeO3 thin film under large tensile strain. We reveal that the breathing distortions, which are the origin of MIT, are triggered by the mode coupling with oxygen octahedral rotations. We further highlight that epitaxial strain can tune the balance between charge ordering and orbital ordering phases. More especially, epitaxial tensile strain favors the Jahn-Teller distortions and drives the phase transition from charge ordering to orbital ordering. Since the orbital ordering system like LaMnO3 tends to have a high value of TMI, we infer that the new orbital ordering phase results in the unusual increase of TMI observed in thin film under tensile strain. Then, we revisit the fundamental role of strain and pressure in several charge ordering perovskites. We propose that the effect of strain and pressure can be unified by the activation or tuning of specific strain modes, which in turn affect material properties through the direct and indirect strain-phonon-phonon coupling. Taking widely investigated NdNiO3 thin film as an example, we develop a new perspective on the mechanical control of TMI and formulate a reliable and general theoretical guidance for the rational control of TMI in experiments. Finally, we explore the superlattices composed of eg1 perovskites. The first purpose is to induce MIT in metallic eg1 perovskites like SrFeO3 through the superlattices strategy. The roles of cation ordering induced confinement effect, lattice distortions and strain have been carefully investigated, allowing us to reveal that both charge ordering and orbital ordering phases are unstable and capable to induce MIT by collaborating with cation ordering. Furthermore, we point out that the stability of the orbital ordering phase with Jahn-Teller distortions is very sensitive to epitaxial strain, while the charge ordering phase is insensitive to strain. Basing on this knowledge, we highlight that strain engineering can be utilized to switch between orbital ordering and charge ordering phases. More interestingly, we can purposely design charge ordering and orbital ordering ferromagnetic and ferroelectric multiferroics by selecting specific superlattices and strain conditions. Another aim is to rationalize the experimental findings of a collaborated work on (NdNiO3)m/(SmNiO3)m superlattices, which show layer thickness dependent MIT behaviors. From the view of first-principles calculations, we firstly helped to rule out the structural coupling at the interface as the key factor that controls the TMI evolution in the superlattices. Then we helped to verify the phase boundary energy between metallic and insulating phases as the dominant factor affecting the TMI proposed by our collaborators. Combining first-principles calculations, Landau model, and the experimental results, we convincingly propose a new paradigm to tune the TMI through controlling the layer thickness in the superlattices. Our findings not only extend the knowledge of tuning MIT in eg1 perovskites but also are important for practical applications.

Magnetocapillary self-assemblies form when soft-ferromagnetic particles are deposited on a liquid surface under an induction field. They emerge from the competition of a magnetic dipolar interaction and the Cheerios effect, an attractive force due to the deformation of the meniscus. Under dynamic fields, these self-assembled structures swim along the interface. Due to their conceptual simplicity, they allow to verify and expand our general knowledge of microswimmers. Different arrangements of particles can be built with different goals in mind: from simple one-dimensional swimmers to complex bio-inspired structures; from fast, controllable swimmers to cargo transporters and fluid mixers.

Granular materials can produce spectacular phenomena due to the dissipation that occurs when the grains collide. In microgravity environments, a granular material can adopt different behaviors, mainly depending on the packing fraction of the system and the intensity of the energy that is injected into the system. By regulating the packing fraction, a granular gas or a dense and slow aggregate can be observed. These aggregates are called "clusters" and their emergence has been studied both theoretically and numerically in the last decades. In addition, the parameters triggering the gas-cluster transition were highlighted during these studies. Nevertheless, these theoretical and numerical considerations have never been verified experimentally and the cluster, which always coexist with granular gases in the case of excited systems, requires still a lot of attention. That's why the SpaceGrains project has been launched by the European Space Agency (ESA). This project contributes to the study of granular phenomena in microgravity thanks to the development of the VIP-Gran-PF instrument, containing of a cell in which bronze beads can be excited by two oscillating pistons. The cell can be quasi-two or three dimensional. The instrument was developed for parabolic flights and the first part of this work has consisted in verifying experimentally the gas-cluster transition and to observe some dynamics which have been predicted by numerical simulations. These simulations constitute the second part of our work. Perfect knowledge of the capabilities and performance of the VIP-Gran-PF instrument as well as advanced training were required before starting to use it in parabolic flights. Our experimental work began with the study of the instrument itself. Once the handling of the VIP-Gran-PF device acquired, we focused on several topics. The first concerned the experimental verification of the gas-cluster transition. This study required a large number of parabolic flight campaigns (5 to be exact) since it was necessary to explore an entire phase diagram, that was derived from numerical simulations. Granular gases and clusters, as well as an unexpected "bouncing aggregate" regime, have been observed during these campaigns. By developing an original image processing, we have been able to reconstruct the distributions of the three dimensional positions of the particles in the VIP-Gran cell on the basis of two dimensional pictures. Adapting an existing model, we also predicted the possibility of emergence of the "bouncing aggregate" regime. In parallel with this study, we performed numerical simulations in order to determine the mechanism of birth and growth of clusters within the instrument. By sorting the grains as gaseous or clustered with a local criterion, we have discovered that the cluster was born in the corners near the lateral walls of the cell, as far as possible from the oscillating pistons. We also show that the cluster grows because the gas has to keep a critical packing fraction. Other numerical simulations allowed us to identify a specific dynamic of the cluster. By asymmetrically exciting the particles, it is possible to displace the position of the center of mass of the system. Moreover, a natural oscillation of the cluster, linked to its size, the asymmetry and the driving frequency, has been highlighted. Finally, we were interested in the structure of clusters in the case of a mixture of two different types of particles. As a function of their difference in mass and/or in volume, we showed that different structures are observed and that a phase segregation could take place in the system. Finally, we realized experiments in order to validate all of our numerical simulations, except for the growth of the clusters, which is impossible to observe in the current configuration of the VIP-Gran-PF instrument. We also investigated the clustering of elongated particles, the behavior of intruders in a granular gas, such granular osmosis by using a semi permeable wall and the behavior of a mobile wall placed in between both pistons. At this point, the data of these studies still has to be analyzed.

When a granular material is driven in microgravity environment, one can assist to the formation of dense and slow regions in the system. Indeed, given the dissipative character of the collisions in the media, energy is lost at each particle interaction and the grains begin to clump locally. The phenomenon has been observed for the first time in the late nineties during sounding rocket experimentation by Falcon and his coworkers and has attracted the interest of many scientists since then. However, precise laws describing the formation and the dynamics of such clusters are still lacking. In order to allow an intensive study of the phenomenon, the European Space Agency set up the SpaceGrains project. Small bronze spheres are enclosed in a rectangular cell and vertically driven by to pistons oscillating in phase opposition. Our work consists in the preparation of the SpaceGrains experiment via molecular dynamics simulations and the elaboration of models predicting the behaviour of the system. Before we started our study concerning SpaceGrains, we reproduced and extended Falcon’s sounding rocket experiments. We showed that, in addition to the granular gas and the cluster, another dynamical regime can be observed in the system. Indeed, for higher filling fractions, the entire granular media behaves like one single completely dissipative particle called the bouncing aggregate. Bouncing modes are observed and can be explained considering the bouncing ball paradigm. Moreover, we highlighted the role of the packing fraction φ as well as the size of the particles R on the different observed dynamics. Within the frame of the SpaceGrains device, we studied the impact of all tunable parameters of the experiment on the dynamics of the system. Thanks to an appropriate scaling all transition points that we obtained by varying the driving amplitude A, the packing fraction and the dimen- sions of the cell L fall along a same theoretical curve. The latter is explained regarding the energy transfer from the piston towards the center of the cell. Once the clustering was controlled, we investigated the handling of the agglomerate. By compartmentalizing the container, local trapping can be achieved and a granular pendant of Maxwell’s demon can be observed in microgravity. Based on the measured particle flux between the compartments, we realized a theoretical model predicting the asymptotic steady state of the system depending on the total number N of particles. In a clustered system, we investigated the impact of asymmetrical driving on the system’s dy- namics. We showed that the mean position of the cluster can be fully controlled via the amplitude ratio a. Moreover, the natural fluctuations of the agglomerate around its equilibrium position are dictated by the driving frequency f and the mass of the cluster. Finally, we realized simulations of driven bi-disperse gases and investigated the segregation phenomena in the system. We showed that clustering and segregation are strongly linked and that the size and the mass of the particles impact the segregation dynamics in different ways.

Dans le présent manuscrit, notre travail est focalisé dans un premier temps, sur l’étude de la dynamique de propagation des épidémies sur un modèle de réseaux complexes qui est une variante du modèle de Barabási-Albert (BA). Les simulations numériques révèlent une transition d’une loi exponentielle à une loi de puissance de la distribution du nombre de liens par nœud en fonction du paramètre de précision ω. Nous avons prouvé aussi que la synchronisation collective σ induite par le processus épidémiologique Susceptible-Infected-Recovered-Susceptible (SIRS ) dépend de la structure topologique du réseau. Dans un second temps, nous nous intéressons à l’étude numérique de ce modèle épidémique en introduisant une source permanente d’infection I_0, afin d’éviter la disparition de la maladie, sur un réseau exponentiel généré par le modèle (BA) d’attachement préférentiel à précision finie. Les simulations montrent que la valeur moyenne de la fraction des personnes infectées dépend de la période naturelle du cycle d’infection τ_0 et du temps d’infection τ_I . Un maximum de synchronisation des personnes infectées est atteint lorsque le rapport τ_I /τ_0 est légèrement plus petit que 1/2. Un autre résultat important est la dépendance linéaire de la valeur moyenne de la fraction d’éléments infectés ⟨n_inf ⟩ en fonction de τ_I /τ_0 bien que, la synchronisation est une fonction non triviale de τ_I /τ_0.

Active particles, and more generally active matter, are known for their ability to move in a given medium by harnessing energy in their surrounding or by carrying their own energy reservoir. A large field of interest regarding active particles is their collective behaviour due to the interactions between individual components of the active system. Examples can be found in biology, medicine, microfluidic or chemistry. In this thesis, the role of individuals in active matter is investigated in two peculiar systems: walking droplets and magnetocapillary microswimmers. Each system lies at an liquid-air interface and relies on the deformation of the liquid surface in their dynamics. Walking droplets are known to propel themselves thanks to the standing capillary waves they generate at each impact on the liquid interface. The persistence time of those waves can be controlled which allows to keep images of the droplet on the interface and to alter the particle motion. This is the memory of the walking droplets. Changing this persistence time allows to change the number of images of the droplet and to explore different dynamics. The limit of extremely large persistence time is considered in this manuscript. In free space, this unique wave memory dynamics allows to generate the first example of deterministic run and tumble dynamics widely encountered in biology. This behaviour finds its origin in the wavefield which traps temporarily the walking droplets. The properties of this run and tumble dynamics are shown to by directly related to the memory stored in the wavefield. If placed in an harmonic potential, the walking droplet is forced to continuously interact with this own wavefield. It is shown that the waves self-organise. In this case, the energy stored in the wavefield mimics an equipartition of energy as well as a minimisation principle. Magnetocapillary microswimmers use the liquid interface in order to self-assemble and the liquid underneath in order to move thanks to hydrodynamic interactions and non-reciprocal deformation. This thesis models two different experimental microswimmers: the linear microswimmer better known as the NajafiGolestanian microswimmer and the triangular magnetocapillary microswimmer. In each case, the non-reciprocal deformation required for the swimming dynamics is at the centre of the discussion. For the linear structure, non-reciprocity is produced by breaking the spatial symmetry of the swimmer. We also discuss the importance of the particles inertia in this low Reynolds dynamics. For the triangular structure, a new swimming mechanism is highlighted where the particles rotation and the structure deformation act cooperatively to generate the translation of the swimmer along the interface. This findings constitute the first step towards the modelisation of larger structures and more efficient swimmers for application in microfluidic.

Electric fields have been used to modify properties of dilute and dense granular materials. For the dilute case, we showed that electric fields can be used to inject kinetic energy on the material. We studied the dynamics of a single bead, two beads and three beads systems. The dynamics is described by a mathematical model which is in quantitative agreement with experiments. A synchronization effect is observed and it is described by a phase coupled oscillators model. We also show that electric fields can be used to tune the cohesion of dense granular media. A transition of the geometry of the flow is observed. Intermittences are induced by the electric field.

Un empilement de grains soumis à une vibration ou à une série de secousses voit sa compacité augmenter. Ce phénomène, appelé compaction, peut être utilisé pour densifier les matériaux granulaires ou comme une méthode de mesure pour caractériser les propriétés d'écoulement d'une poudre. Cette méthode est d'ailleurs populaire dans l'industrie pharmaceutique en raison de sa facilité d'utilisation. Bien que le phénomène de compaction soit étudié depuis un certain temps, il reste actuellement mal compris et les modèles disponibles pour décrire sa dynamique sont, pour la plupart, empiriques. Ce travail se concentre sur le phénomène de compaction et la manière dont certains paramètres, comme les conditions de vibration et la granulométrie du milieu granulaire, peuvent impacter la dynamique de ce processus. Cette thèse se présente sous la forme d'une collection d'articles, liés entre eux par la recherche de paramètres de contrôle et de l'optimisation du processus.

The functionalization of biomaterial surfaces for the design and exploration of novel properties has attracted increasing attention in research. In the field of dentistry, fighting bacterial colonisation is largely associated to the use of antibiotics. More recently, the deposition of coatings on dental implants by wet chemical processes has been studied and it involves complicated protocols and leads to a high amount of produced wastes. Surface modification and coating of the implants through a plasma polymerization can offer a new, greener perspective for surface treatments with good control over surface morphology and chemistry. Atmospheric pressure plasma technique would undoubtedly be a good alternative for the deposition of fully organic thin films designed for the immobilization of biomolecules. Operating at low temperature and atmospheric pressure, such processes are easily up-scalable. This thesis aims at the development of a novel surface treatment to confer antibacterial and antibiofouling properties while maintaining osseointegration on Ti dental implants. To achieve this, different mild treatments in dry and aqueous media were combined (Fig.1.). The dry technique of Atmospheric Pressure Dielectric Barrier Discharge (AP-DBD) plasma deposition was used to deposit a coating containing highly chemical-reactive groups (catechol/quinone moieties), which can further serve for bioconjugate reactions. The AP-DBD was used to perform a liquid assisted, plasma-initiated polymerization setup including monomer spraying zone and subsequent plasma zone generated between two electrodes. Based on the good stability, chemical passiveness, mechanical resistance and reports of bio and cytocompatibility for polymethacrylates, the PhD research focused on the preparation of fully organic methacrylic-based coatings. The challenge was then to develop a coating that presents a sufficient number of reactive chemical groups for bioconjugation and, at the same time, a resistance to aqueous media at different pH. During the project, the liquid-assisted AP-DBD approach allowed for the design and successful production of three different methacrylate-based coatings, stable in liquid media and bioactive after bioconjugation. As proof of concept, plasma copolymerization of ethylene glycol dimethacrylate (EGDMA) polymer layers with methyl 3-(3,4-dihydroxyphenyl)-2-(2-methylprop-2-enamido) propanoate (DOMAm) was then performed with the possibility to tune the number of functional catechol and quinone moieties. All the plasma-deposited films have later served as versatile platforms for efficient immobilization of virtually any kind of biomolecules containing in its structure: thiol, amine or imidazole groups. Indeed, these nucleophilic groups promote the quinone-nucleophilic reactions. Here, Dispersin B1, Ranaspumin-2 derivative and Lysozyme were tested in order to create multifunctional biomaterial surfaces. Primarily, to prove the concept of functional biomaterial in the field of dental implants, an antibiofilm activity was obtained by immobilization of Dispersin B1 derivative on EGDMA-DOMAm coating. After, the bioconjugation of bioengineered of Ranaspumin-2, another antifouling protein, was tested onto the robust EGDMA-MMA-DOMAm plasma coating. Lastly, to obtain multifunctional biomaterial surface, a controlled co-immobilization of two different biomolecules onto quinone-enriched EGDMAMMA-DOMAm surface was performed. Wherein Lysozyme, a natural bactericidal enzyme, and Ranaspumin-2 were both successfully immobilised on the polymer surface. Depending on the used biomolecule, biomaterials were tested against microorganisms’ adhesion, mortality as well as cytocompatibility in in vitro studies.

Polycarbonates (PCs) belong to some of the world-leading polymers that are widely used in aircraft or automotive applications (windows, etc.), in safety equipment (helmets, bullet proof glass, etc.),… However, their industrial production requires the use of toxic compounds such as phosgene, high temperatures, and is not compatible with the introduction of some functional groups along the polymer backbone. Making plastics more sustainable by valorising CO2 as a cheap, inexhaustible and renewable feedstock imposes itself as a strategic driver for developing low carbon footprint materials. In 2017, our research group reported a new process for the preparation of a novel family of regioregular PCs (i.e. poly(oxo-carbonate)s) by the facile organocatalysed polyaddition of CO2-sourced bis(α-alkylidene cyclic carbonate)s (bisαCCs) with diols under ambient conditions. These bisαCCs were prepared by the organocatalysed carboxylative coupling of CO2 to bis(propargylic alcohol)s. The goal of my PhD thesis work was to investigate the fundamentals to permit the synthesis of these new PCs in a one-pot process, wherein bisαCCs are produced in-situ and directly involved in polymerisation. Firstly, we designed various organic salts that were tested as catalysts for the carboxylative coupling of CO2 with propargylic alcohols. We investigated the influence of the structure of the organocatalyst (mainly the type and structure of the cation and anion) on the catalytic performances. Optimum activity resulted from the best compromise between ion-pair separation controlled by steric effects and the basicity of the anion. Organic salts with too basic anions increased the rate of the reaction but at the expense of the selectivity. Then we optimised the activity of the organic catalyst by incorporating a metal cocatalyst (CuI or AgI) that allowed to synthesise the bisαCCs with a high selectivity (> 95%) under mild conditions (25-40 °C, 15 bar). Eventually, we implemented the one-pot cascade approach to prepare oxo-alkylcarbonate scaffolds and poly(oxo-carbonate)s from CO2, propargylic alcohols, and mono-alcohols under moderate operating conditions (PCO2 = 1-15 bar, T = 40-80 °C). By varying the nature of the bis(propargylic alcohol) and the diol, various poly(oxo-carbonates) were successfully prepared. Lastly, we prepared new alkyne-1-n-diols and studied their reactivity for the carboxylative coupling with CO2. By carefully choosing the reaction conditions and the diols, we prepared diverse keto-cyclic carbonates, elusive tetrasubstituted carbonate scaffolds and poly(oxo-carbonate) oligomers. This thesis includes operando FT-IR/ATR and DFT computational studies that allowed to understand and explain the mechanism of the reactions investigated. In conclusion, this work contributes to the quest to develop simple catalytic systems for the transformation of CO2 into useful compounds (cyclic carbonates and poly(oxo-carbonate)s).

Cobalt-mediated radical polymerization (CMRP) is an emerging class of controlled radical polymerization technique based on the reversible deactivation of the propagating radicals by a cobalt complex. This thesis aims at broadening the nature, the structure, the properties and the end-use of (co)polymers accessible by CMRP. The first part of this work is devoted to the synthesis of novel statistical, graft and block copolymers by CMRP. First, the statistical copolymerization of conjugated (n-butyl acrylate) with a non-conjugated one (vinyl acetate) has been successfully controlled by bis(acetylacetonato)cobalt(II). The optimal experimental conditions were then implemented to the copolymerization of vinyl acetate with poly(ethylene glycol) acrylate, which yielded (quasi-)diblock amphiphilic graft copolymers for potential biomedical applications in one step. CMRP also revealed successful for the synthesis of well-defined thermo-responsive copolymers based on poly(N-vinylcaprolactam), either amphiphilic or double-hydrophilic at room temperature. In the second part of this thesis, we investigated the potential of CMRP for the synthesis of new agents for cancer photodynamic therapy. The synthesis of such photosensitizers relied on the chain-end functionalization by [60]fullerene of water-soluble (co)polymers (or precursors) prepared by CMRP. The photodynamic activity of poly(vinyl alcohol)/C60, poly(N-vinylpyrrolidone-co-vinyl acetate)/C60 and poly(PEG acrylate-co-vinyl acetate)/C60 nanohybrids were assessed through photophysical and in vitro experiments. Also, the ability of the nanohybrids to avoid recognition by the immune system was evaluated by a protein adsorption test. Finally, some hints on how to improve the synthesis of the nanohybrids - and therefore, their photodynamic activity - were provided by a mechanistic study of the grafting onto C60 of macroradicals released by cobalt end-capped polymers

The lightweight of porous nanocomposites makes them attractive materials for various applications such as thermal and sound barriers, shock absorbers, insulation, packaging, and their porous structure is very interesting in bone tissue engineering. Moreover, the incorporation of appropriate carbonaceous nanoparticles into polymeric foams contributes to the reinforcement of their mechanical performances but also renders them electrically conductive, consequently extending their potential interest in electromagnetic shielding (EMI) and electrostatic discharge (ESD) applications for instance. In this PhD thesis, we aim at designing various polymeric foams containing a conductive nanofiller (carbon nanotubes) and to identify the main morphological parameters (pore size, cell density, cell wall thickness,…) that affect and govern the final properties of the foams. In this work, the electrical conductivity of the foams is the main property investigated because it is governing their performances as materials for EMI absorbers, the main application targeted in this work. These important morphology/electrical conductivity relationships would indeed be very useful to guide the foam development towards the material with the best performances for the targeted applications. Two different foaming methods are used in this work: (i) the supercritical CO2 (scCO2) foaming technology and (ii) the freeze-drying process. The first technique enables to produce isotropic foams with spherical closed cells structures and the second one, oriented anisotropic foams with cylindrical open cells. The variation of the foaming parameters allows preparing foams with a large panel of morphologies required for the establishment of the structure/properties relationships. In parallel to this main objective, an improvement of the overall conductive performances of the nanocomposites foams is also investigated through the optimization of the foam morphology and the content in conductive nanofillers.

Among all materials, plastics have become one of the most important products of our daily life. Thanks to their ease of production and processing, low cost and customizable properties, plastics are indeed found in pretty much all sectors with applications, for example, in packaging, automotive, pharmaceutical industries or for electrical and electronical devices. However, they are mainly produced from petroleum and contaminate the environment when not properly recycled, which leads to the environmental disaster that we all know. The utilization of carbon dioxide (CO2) as a substitute to petroleum is expected to limit our dependency to this depleting fossil resource and to avoid to be dramatically affected by the associated geopolitical issues that we are currently facing. CO2 is indeed available at large quantities and low cost, and it might contribute to a more sustainable plastic sector by enabling the on-site production of plastics. To reach this ambitious goal, efficient transformations operating under mild conditions and producing limited waste have to be developed, and end- of-life scenarios of the new polymers have to be considered. In this perspective, the aim of this PhD thesis is to merge innovative CO2 transformations with bio-based molecules to produce functional polymers under mild conditions. This work was carried out in the framework of the Excellence of Science (EoS) “BIOFACT” project that deals with the transformation of lignin into high-value chemicals and polymers, starting with the very challenging selective depolymerization of lignin. My contribution to this project was to develop innovative routes to copolymerize a variety of lignin- (or sugar)-based diols or diamines with novel highly reactive CO2-based cyclic carbonates with the objective to produce non- phosgene-based polycarbonates (with diols) and non-isocyanate-based polyurethanes (with diamines) at room temperature. The influence of the structure of the amine/alcohol and reaction conditions on the polymer structure and its thermal properties was investigated. With the aim of circular economy in mind, special attention was also devoted to the exploration of end-of-life options for the CO2-based polymers by investigating chemical degradations via solvolysis or via novel approaches of polymers skeletal-editing.

Renewable monomers have the potential to replace petroleum-derived monomers for reversible deactivation radical polymerisations (RDRP) for a variety of applications, such as adhesives and coatings. Yet, challenges in the polymerisation of non-activated and often internal double bonds found in natural molecules still remain. Moreover, functionalisation pathways attaching renewably-sourced double bonds to natural molecules are rare and sustainable strategies using catalytic or enzymatic reactions are sought after. This thesis aims to introduce a set of renewable monomers for reversible deactivation radical polymerisation (RDRP), namely organometallic-mediated radical polymerisation (OMRP) using a cobalt complex, in the quest for renewable and functional (co)polymers. This particular type of OMRP was chosen because it controls the polymerisation of a large range of non-activated monomers with excellent control over the chain growth process under mild experimental conditions. The monomers prepared in this thesis were obtained from plant oils and/or CO2 via catalytic reactions and contain ester and carbonate functionalities of interest for post-polymerisation modifications. Successful copolymerisations with monomers bearing non-activated double bonds, namely vinyl acetate and ethylene, were performed using OMRP under mild conditions and the comonomer content was tuned via the initial polymerisation feed or the ethylene working pressure. The introduction of carbonate moieties into vinyl acetate copolymers allowed for the synthesis of three discrete functional poly(vinyl alcohol) copolymers. Moreover, the ability to incorporate such carbonate functionalities into polyethylene copolymers by OMRP was shown for the first time. Highly linear ethylene copolymers over a broad range of carbonate content were obtained with significantly altered properties compared to homo-polyethylene. Particularly the ability to finely tune the molecular copolymer parameters, such as molecular weight and copolymer architecture, allows a systematic study of their influence on the compatibilisation capability of such copolymers. The potential of polyethylene copolymers bearing functional groups as compatibilisers was highlighted for poly(ethylene-co-vinyl acetate) copolymers. Finally, a fully renewable compatibiliser, based on starch and high oleic sunflower oil, obtained by non-radical means was applied to cellulose/LDPE composites. An improved cellulose dispersion within the matrix was observed by rheology, while the mechanical properties, notably Young’s modulus, was increased. This work aims to highlight the unexplored potential of renewable resources for the synthesis of functional polymers for their application in polyolefin composites.

Mesoporous thin films are promising architectures for positive electrodes in Li-ion battery applications. A particular challenge in this field has been successful templating of vanadium-based oxides, materials known for their ability to host lithium, since their thermal instability and complex vanadium chemistry hinder templating through traditional soft-chemistry approaches. To address these technical problems we develop the soft-templating of vanadium and niobium mixed oxides based on Evaporation Induced Micelles Packing using thermally stable polystyrene-b-polyethyleneoxide structuring agents. In-situ thermal monitoring via ellipsometry allows successful navigation of the thermal stability landscape. TEM and AFM analyses reveal homogeneous wormlike mesoporous structures whose pore and inorganic wall sizes can be tuned from 15 to 100 nm by changing the hydrophobic/hydrophilic surfactant chain lengths. Ellipsometric porosimetry shows that 100 nm thick films with a 15 nm pore size displays 30% electrolyte accessible porosity. The interconnected tridimensional mesoporous network has been highlighted by electronic tomography. Thicker films up to 1.3 µm are prepared by a multidipping process. The superiority of such nanoarchitectures compared to non porous materials in terms of electrochemical properties such as capacity are revealed using cyclic voltammetry.

Electrical energy consumption has increased dramatically and its mode of consumption has diversified over the years, with the development of mobile energy storage devices such as lithium batteries. The present thesis is dedicated to the synthesis and study of the compound Li4Ti5O12 as negative electrode in Li-ion batteries. The main objective is the development of innovative synthesis techniques based on soft chemistry in order to control and study the effect of structural and microstructural parameters on the electrochemical properties of the compound. Because of its ability to intercalate/deintercalate Li+ ions in its structure without stress-generating volume changes of the structure, Li4Ti5O12 is considered as a so-called "zero-strain" electrode, favouring excellent cycling performance. Therefore, this compound appeared to us a most interesting candidate in this study. The first part of this work investigates a method involving the freeze-drying of a gel precursor followed by self-ignition. Hydroxypropylmethylcellulose is used as an additive and the synthesis of Li4Ti5O12 takes place in four steps : Gelling – Freeze-drying – self-ignition – Calcination. Ammonium nitrate was also added to the reactants to determine the impact on the self-ignition process and the consequences on the structure, microstructure and electrochemical properties of the compound. Indeed, a major issue usually encountered in self-ignition is controlling the amount of heat released, and therefore the temperature during self-ignition. These factors greatly influence the structural and microstructural properties of the compound. The results show that addition of ammonium nitrate leads to an excessive crystal growth which is unfavorable to electrochemical properties of the compound. In the second part of the thesis, the spray-drying method has been investigated using both a lab-scale and a pilot-scale equipment, to take into account possible developments on an industrial scale. For both devices, the influence of concentration on particle morphology was observed. The lab spray-dryer allowed us to obtain a convex to mushroom morphology. For the pilot spray-dryer, the morphologies vary from cauliflower to smooth hollow spheres, and can be further modified by addition of a carbonate source which decomposes during calcination. Depending on the particle morphology and its surface for intercalation of Li+ ions, the electrochemical performances have been improved, sometimes considerably. The drying temperature also affects the performance of the compound through the agglomeration rate of the particles. In conclusion, a wide panel of microstructures has been obtained by investigating two synthesis methods with various sets of experimental conditions. Although all these powders are Li4Ti5O12, there is a significant variation of their electrochemical performance and we have attempted to correlate these with particle morphology. Some prominsing results have been obtained and suggest that performance can be further improved through future work, on both fundamental and applied aspects.

Dans le contexte général du développement, au Laboratoire de Génie Chimique, de nouveaux photocatalyseurs à base de TiO2 présentant une activité sous lumière UV-Vis améliorée, l’objectif de ce travail consiste à étudier les mécanismes de formation mis en jeu lors de leur synthèse et de leur dopage par la méthode sol-gel de cogélification. Plusieurs dopants, contenant du silicium, du phosphore ou de l’argent, ont été utilisés. L’influence des variables opératoires de synthèse sur les propriétés physicochimiques, et de là sur les propriétés photocatalytiques des xérogels obtenus, a été examinée en détail. La compréhension de cette relation a permis d’ouvrir la voie à la synthèse de photocatalyseurs à base de TiO2 combinant différents dopages et présentant des activités élevées. Le développement de nouveaux dopants à base de porphyrines fonctionnalisées par des groupements phosphonatés a également été réalisé en vue de leur incorporation ultérieure dans les gels de TiO2. L’obtention d’un gel ou d’un précipité dépend des cinétiques des réactions d’hydrolyse et de condensation du tétraisopropoxyde de titane (TIPT). Le recours à un solvant chélatant a permis de ralentir la cinétique des ces réactions et donc d’obtenir un gel. Ces xérogels sont cependant non poreux après séchage. L’insertion, lors d’une synthèse par cogélification, d’un additif silylé ou phosphonaté contenant une éthylènediamine ralenti également la cinétique des réactions via un mécanisme de complexation du TIPT. Dans ce cas, il est possible d’obtenir un gel poreux dans un solvant non chélatant, et non un précipité. Ce résultat est très intéressant car il a été prouvé que lors du dopage du TiO2 avec du phosphore par cogélification, la formation du gel est importante car elle s’accompagne d’un degré élevé d’hétérocondensation (liaisons P-O-Ti), contrairement à la formation des précipités. Afin d’éliminer les résidus organiques encore présents à l’issue du séchage et de cristalliser le TiO2 en anatase, les échantillons doivent être calcinés. L’influence de la température de calcination sur les propriétés des photocatalyseurs a été examinée en détail. Dans les échantillons calcinés, des propriétés telles que la taille des cristallites d’anatase, la surface spécifique et le spectre d’absorption UV-Vis sont principalement influencés par la teneur en phosphore. Pour maîtriser cette dernière, il est important de maximiser le degré d’hétérocondensation entre le TIPT et l’additif phosphonaté (EDAP). En effet, lors de la calcination, l’EDAP non intégré au réseau de TiO2 via des liaisons Ti-O-P, se décompose en dessous de 350 °C et du phosphore est perdu par sublimation de P2O5. Parmi les dopages étudiés, seul celui au P permet de déplacer le spectre d’absorption du TiO2 vers les plus grandes longueurs d’onde. Lors du test de dégradation du p-nitrophénol sous lampe halogène, les xérogels de TiO2 dopé au P calcinés donnent de meilleurs résultats que le xérogel de TiO2 pur ou encore que ceux dopés à l’Ag uniquement. L’influence intrinsèque du phosphore sur l’activité photocatalytique a été confirmée par une étude statistique. Cette étude montre également que l’activité dépend de la teneur en phosphore via son influence sur la taille des cristallites d’anatase et sur la surface spécifique des xérogels. Cette activité est d’autant plus élevée que la taille des cristallites est petite et que la surface spécifique est élevée. Afin de maximiser les performances photocatalytiques du TiO2 sous lampe halogène, un co-dopage à l’Ag et au P a été réalisé pour essayer de combiner l’effet bénéfique du P sur l’activité sous lumière visible et celui de l’Ag sur la vitesse de recombinaison des paires électron-trou. L’utilisation d’une molécule comme l’EDAP, capable de complexer l’ion Ag+, permet d’obtenir dans le matériau final des nanoparticules d’Ag finement dispersées dans les particules de TiO2. Le co-dopage à l’Ag et au P permet d’augmenter de manière très significative l’activité des xérogels de TiO2 pour la dégradation du p-nitrophénol sous lampe halogène. Cette augmentation résulte d’un effet synergique entre les deux additifs. Le dernier chapitre de ce travail est une étude préliminaire pour ouvrir la voie au dopage du TiO2 avec des porphyrines fonctionnalisées par des groupements phosphonatés ou silylés via la méthode de cogélification. Cette étude a permis de mettre au point des protocoles de synthèse de porphyrines phosphonatées et de porphyrines aminées. Ces porphyrines aminées serviront de précurseurs pour l’obtention de porphyrines silylées.

In this work the roughness and topography evolution of optical materials sputtered with low energy ion beam (≤1 keV) has been investigated. These materials (bulk or thin layer) are used in the manufacturing of mirrors for scientific (ground or space) instruments or for other optical applications. In the first part of the work, the roughness evolution of optical surfaces under sputtering has been investigated in the frame of the industrial process known as Ion Beam Figuring. This technique consists in removing shape errors on optical surfaces with a low energy ion beam (≤1 keV). One disadvantage of this process is a potential increase of roughness for surfaces under treatment. The roughness evolution of some materials relevant to the optical industry has been accurately characterized as function of etching depth down to 5 µm. These sputtering experiments have been carried out at normal incidence, mainly with argon ions (but also in a lesser extent with krypton and xenon ions), ion current density of ~1 mA/cm2 and ion beam energy ranging from 200 eV to 1000 eV. The roughness evolution under sputtering is low for materials with amorphous (glass, electroless nickel), monocrystalline (silicon) or even polycrystalline structure (CVD silicon carbide, PVD gold or nickel film), whereas it is considerably more important for some other metallic materials such as electroplated nickel and aluminium. This work has shown small differences in the roughness evolution of CVD silicon carbide as function of the ion beam energy. The roughness increase is faster at low ion energy (<500 eV) than at higher ion energy (650-1000 eV). The grain structure of this material is less revealed at higher energy, which is supposed to be due to a larger amorphization of the sputtered layer in this case. The influence of the ion mass on CVD silicon carbide and gold films on nickel substrates has been also illustrated. Our measurements have been also compared to scaling laws. Various growth and roughness exponents have been found, sometimes rather different from those foreseen by the KPZ equation. In the second part, we focus on periodically modulated structures (ripples) which developed on many solids when sputtered by an off-normal ion bombardment. In this work, we first observed these ripples on gold films deposited on electroplated nickel (materials used as reflective surfaces for X-ray space telescope) sputtered at grazing incidence. We studied the influence of sputtering parameters (ion beam incidence angle, energy and flux) on the characteristics of ripples induced on gold and silver thin film (~0.2 µm). Ion-induced ripples have also been observed on CdS, an interesting semiconductor crystal for optical applications. The ripples orientation and dimensions (spatial wavelengths from 0.13 µm to 0.29 µm) have been confronted to the Bradley-Harper (B-H) linear model. We used the SRIM software to evaluate the deposited energy and the surface tension coefficient distributions. Our results can be in great part explained by the current theories (Bradley-Harper, Makeev) on morphology of ion-sputtered surfaces. These results can be summarized hereunder: • Clear development of ripples for angle of incidence equal or higher than 60° on gold film and 70° on silver film. • In this work the ripples wave vector is always perpendicular to the ion beam direction for all angles, whereas the change in ripple orientation beyond a critical angle is usually reported in literature. This is a due to the different shape of the energy distribution function for our sputtering conditions. • Different regimes for roughness and topography evolution (grains, ripples) have been observed in function of the angle of incidence. 3 different areas can be distinguished, as predicted by Makeev non-linear model. • The diminution of ripple wavelength with ion energy shows that thermal diffusion is the main relaxation mechanism.

The steel industry is constantly looking for innovations and solutions to improve production processes as well as product properties. However, current technologies result from decades of development and thus have already reached maturity. Therefore, bright innovations have to arise from technological breakthroughs. The objective of those novelties is to induce drastic changes in terms of process or product. Magnetic heat treatment might be a solution to reach both objectives. Indeed, recent researches have shown that magnetic fields can significantly modify the transformation kinetics in steels. This can turn into very positive impacts on the metallurgical processes. However, researches on this topic are in the early stage and a lot of fields have not been studied yet. This work has been carried out in this challenging context. The main objective is to study the effect of magnetic fields on cementite spheroidization and ferrite recrystallization. So far, these two transformations involve long thermal treatments at high temperatures. Thus, a reduction of the processing time or an improvement of the mechanical properties of the steels by using magnetic field processing would be a significant improvement for this kind of thermal treatment. The transformations that are studied imply large microstructure modifications: lamellae breaks into spheroids while ferrite recrystallizes. As a consequence, the developments of dedicated microstructure characterization techniques are concomitant objectives. As it will be shown latter, we decided to develop and optimize image analysis tools. Practically, the completion of this work has required the pursuit of four objectives: Objective 1: develop an image analysis tool dedicated to pearlite spheroidization study, Objective 2: study the effect of magnetic field on cementite spheroidization, Objective 3: develop an image analysis tool dedicated to ferrite recrystallization study, Objective 4: study the effect of magnetic field on ferrite recrystallization. This work is divided in four parts described below. A general introduction constitutes the first part. It contains two chapters. Chapter I focuses on a literature review of the effect of magnetic fields on steels. The first section of this chapter deals with the basics of metallurgy and magnetism required to understand this work. Then, the reader will find a literature review about the effects of homogeneous magnetic fields on steel transformations. As it has been said before, the specificity of this work lies in the fact that the transformations studied were analysed and characterized using image analysis. The Chapter II is dedicated to this technique. The basics of image analysis are summarized in the first section of this chapter. Then, specific sections are dedicated to each step of image analysis: pre-treatment, image segmentation and characterizations. The limits of this technique, as well as its applications are described in the two last sections of this chapter. The second part of this work is divided in three chapters and deals with cementite spheroidization under magnetic field. Chapter III provides a detailed introduction to cementite spheroidization. The first section of this chapter introduces the pearlite as well as its microstructure. Using these concepts, the mechanisms and the kinetic of cementite spheroidization and cementite ripening will be introduced. Finally, the last section of this chapter will focus literature results which indicate that a potential effect of magnetic field on cementite spheroidization might be expected. After this detailed introduction, the tools used to study the cementite spheroidization under magnetic field will be described in Chapter IV. The first section deals with the characterization techniques used to study pearlite spheroidization. Then, the furnace and the heat treatments that have been performed will be described. With these tools, we will describe, in Chapter V, the results that have been achieved about cementite spheroidization under magnetic field. First, we will deeply analyse the effect of temperature and heat treatment duration on cementite spheroidization. This will be the opportunity to study in details the mechanism of cementite spheroidization and spheroids ripening. Microstructure evolutions induced by these two transformations will also be analysed. The effects of magnetic field on cementite spheroidization are described in the last section of this chapter. The analysis of the effect of magnetic field on ferrite recrystallization constitutes the third part of this work. It is divided in three chapters. Chapter VI provides a detailed introduction to ferrite recrystallization. The first section of this chapter deals with the crystalline structure defects induced by the steel forming. The detailed description of the mechanisms and the kinetics of defects elimination by the recovery, primary recrystallization and secondary recrystallization constitute the three next sections of this chapter. The last section of this chapter summarizes the results of different relevant studies on the effects of magnetic fields on these three processes. The chapter VI is followed by a detailed description of the characterization techniques as well as the heat treatment performed to study ferrite recrystallization (Chapter VII). Chapter VIII describes the results that have been obtained about ferrite recrystallization under magnetic field. We will study the effect of temperature and heat treatment duration on ferrite recrystallization. The involved transformations will be studied in detail. Finally, the effect of magnetic field on ferrite recrystallization will be discussed in the last section of this chapter. The general conclusions as well as the prospects of this work will be addressed in the fourth part of this work, respectively in Chapters IX and X.

A brief introduction to the thermoelectric effects, the studied materials, and the experimen- tal methods is given in the first chapter. The introduction is not intended to be exhaustive, but only to summarize important basic information for the reader. The introduction is followed by four chapters dedicated to detailed experimental studies of the lattice dynamics in selected thermoelectric Zintl phases. First, the lattice dynamics in the unfilled and filled skutterudites FeSb3 , CoSb3 , and YbFe4 Sb12 were studied by nuclear inelastic scattering, inelastic neutron scattering, and several macroscopic methods. These studies reveal that FeSb3 exhibits softer Sb bonds than CoSb3 , that the density of phonon states is shifted towards lower energies and the velocity of sound is lower in FeSb3 as compared to CoSb3 . It appears thus that the soft [Fe4 Sb12 ] framework dynamics might play an important role in the thermoelectric properties of filled skutterudites. The observed anomalous temperature dependence of the elastic constants and the rearrangement of the spectral weight of the Yb phonon states in YbFe4 Sb12 can be explained by a change of the Yb valence state with the temperature. Second, the lattice dynamics in the Sr8 Ga16 Ge30 clathrate was investigated by inelastic neutron scattering measurements on a single crystal. We found that several mechanisms contribute to the low thermal conductivity in this system and that the reduction of the heat capacity contribution to the thermal conductivity plays a significant role to the low thermal conductivity, besides the reduction in the phonon lifetime and the phonon group velocity that is related to the guest atom. Third, the lattice dynamics of the Zintl phase Yb14 MnSb11 was studied by inelastic neu- tron scattering and nuclear inelastic scattering measurements. All phonon modes of these complex crystal systems are in a narrow energy range below ∼25 meV and the Debye tem- perature, the velocity of sound and the mean force constants are small compared to those of other thermoelectric materials such as Zn4 Sb3 . By comparing the lattice dynamics in Yb14 MnSb11 and Zn4 Sb3 different mechanisms which lead to the low thermal conductivity in Zintl phases have been identified. Between 300 and 1200 K no softening of the velocity of sound in Yb14 MnSb11 was observed by temperature dependent inelastic neutron scattering measurements, which is in line with its large melting temperature.

Dans le cadre de la réutilisation des eaux usées et de l’accès limité à l’eau potable dans certaines régions du globe, de nouvelles techniques d’épuration nécessitant de faibles consommations d’énergie sont étudiées. Parmi ces nouvelles techniques, la photocatalyse hétérogène utilise des réactions d’oxydation et de réduction pour la dégradation de polluants organiques et de micro-organismes. Le photocatalyseur le plus employé est le TiO2. Lorsqu’il est illuminé par un rayonnement UV, des paires électron/trou sont générées et réagissent avec l’eau et l’oxygène pour former des radicaux libres qui vont dégrader les polluants. Le procédé sol-gel permet de contrôler facilement la texture, la cristallinité, l’introduction de dopants et la mise en forme de matériaux. Il a été utilisé dans ce travail pour la synthèse de photocatalyseurs à base de TiO2, (xérogels). À partir de sols en cours de gélification, des films de TiO2 ont été déposés par spin-coating. Après calcination, ces films sont actifs pour la dégradation du bleu de méthylène sous UV-A. Afin d’améliorer l’activité photocatalytique de xérogels de TiO2, des particules d’argent ont été introduites dans la matrice par cogélification. Ces particules d’argent améliorent l’activité du TiO2 pour la dégradation du bleu de méthylène sous UV-A. En plus de l’amélioration de l’activité photocatalytique, l’extension de l’activation du TiO2 au domaine visible du spectre solaire a été étudiée par sensibilisation avec des porphyrines. Ces porphyrines sont introduites d’une part, par greffage sur le catalyseur Degussa P25 et d’autre part, in situ dans la matrice de xérogels de TiO2. Le greffage des porphyrines ne permet pas l’activation du catalyseur Degussa P25 sous lumière halogène. Par contre, l’introduction in situ des porphyrines permet l’activation du TiO2 avec des longueurs d’onde du domaine visible et améliore ainsi l’activité des xérogels sous lumière halogène pour la dégradation du p-nitrophénol. L’activité antibactérienne de plusieurs catalyseurs préparés au cours de la thèse a également été évaluée pour la dégradation de deux espèces de bactéries : Escherichia coli et Lactobacillus rhamnosus. Cette activité antibactérienne est corrélée à l’activité dépolluante des photocatalyseurs. L’étude cinétique de la dégradation du p-nitrophénol sur un xérogel de TiO2 dopé in situ avec une porphyrine a été réalisée. Cette étude montre que la dégradation implique un type de site actif correspondant aux trous h+ créés lorsque le TiO2 est exposé à la lumière. L’étape limitante correspond à la réaction de surface entre le p-nitrophénol adsorbé et les radicaux OH• adsorbés. Cette étude cinétique utilise une loi sigmoïdale plutôt que linéaire pour modéliser les courbes de dégradation du p-nitrophénol.

Depuis quelques années, la recherche portant sur l’administration ciblée de nouvelles formes galéniques solides a trouvé un nouvel essor et ne cesse de progresser. Dans ces formulations, très souvent, le principe actif est associé à un agent de transport : l’excipient. Cette approche est notamment employée dans le cas de l’insuline, de bon nombre d’antibiotiques, de facteurs de croissance développés sur des ciments utilisés en ingénierie tissulaire osseuse, d’acide nucléique ou de médicaments encapsulés dans des nanoparticules utilisés pour le traitement des cancers. Les avantages de tels systèmes sont une meilleure biocompatibilité, une plus grande biodisponibilité et une résorption plus importante. La recherche d’un agent de délivrance optimal a ainsi conduit à développer de nouvelles voies de synthèse reposant essentiellement sur l’optimisation des propriétés architecturales et de surface du matériau. La première partie de ce travail consiste à développer un nouvel agent de transport de protéines destiné à la vaccination orale. Cet agent est construit à partir de phosphate de calcium (hydroxyapatite). L’albumine bovine est employée comme protéine de référence. Elle permet de mesurer la capacité de charge des particules de phosphate de calcium. Le chapitre 2 donne un aperçu général des principes de vaccination orale et du type de matériaux transporteurs utilisés. Les chapitres 3 à 6 présentent les résultats des quatre stratégies utilisées dans cette étude pour préparer le phosphate de calcium. Une caractérisation complète de la surface des poudres sera réalisée de manière à mieux comprendre le mécanisme d’adsorption qui prend naissance entre la protéine et le phosphate de calcium. Le chapitre 3 montre le potentiel d’une méthode de préparation classique de poudres (traitement ultrasonique des suspensions et/ou broyage mécanique des poudres). Elle a pour objet de modifier les caractéristiques de surface des poudres commerciales. Une augmentation de l’intensité du broyage permet de diminuer la taille des particules en dessous du micron. L’augmentation de la surface spécifique des poudres qui en résulte conduit à une augmentation de la quantité de protéine adsorbée par unité de surface. Des tests préliminaires in vitro ont démontré la possibilité d’intégrer des microparticules de phosphate de calcium chargées en protéine dans des cellules dendritiques sans observer d’effet toxique notoire. L’importance que revêt la chimie de surface de l’agent de transport dans le mécanisme d’adsorption des protéines est prouvée au travers de la fonctionnalisation de la surface des particules. Le chapitre 4 présente l’utilisation d’acides aminés (ou de dérivés) comme molécules attractives dans l’optique d’améliorer l’adsorption des protéines à la surface des particules d'hydroxyapatite. Intensifier les interactions électrostatiques entre la protéine et la surface des poudres en les fonctionnalisant avec de la lysine ou de l’arginine conduit à une augmentation de la quantité de protéines adsorbées (jusqu'à 66% en plus par rapport à la poudre d’hydroxyapatite non fonctionnalisée). Le type d’interaction observé entre la molécule attractive et la surface des poudres a été déterminé par RMN. Un processus d’échange rapide se produit à la surface des poudres d’hydroxyapatite. Les mêmes expériences ont été réalisées avec de la dihydroxyphényl alanine et de la dopamine. Cette dernière conduit à une augmentation significative de la quantité de protéine adsorbée en raison de la formation d’un lien covalent.Les poudres ont également été traitées par Plasma Atmosphérique. Les résultats obtenus sont présentés au chapitre 5. L’influence du design du réacteur et de la nature de l’atmosphère gazeuse sur la chimie de surface des poudres d’hydroxyapatite a été étudiée. Bien que l’identification des groupes fonctionnels de surface suite au traitement par Plasma Atmosphérique n'a pas été possible, on constate néanmoins une augmentation jusqu’à 39% de la quantité de protéines adsorbées en surface des particules traitées. Une optimisation du procédé est néanmoins à envisager et des analyses plus poussées complémentaires devront être réalisées afin d'identifier la nature de la surface des poudres traitées. Au chapitre 6, la stratégie envisagée consiste à synthétiser les poudres d’hydroxyapatite par co-précipitation. Les poudres obtenues sont de dimensions nanométriques et présentent les quantités de protéines adsorbées les plus élevées reportées jusqu’à présent dans ce travail. La co-précipitation permet de contrôler la taille des particules, leur degré de pureté et le taux de cristallinité. La deuxième partie de la thèse envisage la formation d'un dépôt électrophorétique de particules d’hydroxyapatite sur des implants métalliques utilisés en réparation osseuse. La présence d’hydroxyapatite en surface de l’implant permet une meilleure et plus rapide ostéointégration et offre également la possibilité de délivrer des substances biologiquement actives telles que des antibiotiques ou des facteurs de croissance. Le chapitre 7 décrit l’influence du solvant, de la durée du dépôt électrophorétique ainsi que la nature et l’intensité du champ électrique appliqué sur la microstructure du dépôt. L’éthanol et le butanol ont été utilisés comme solvant de manière à accroître, d’une part, la stabilité de la suspension et, d’autre part, afin d’éviter la formation de pores dans le film déposé suite à l’éventuelle électrolyse de l’eau. Les films obtenus présentent, pour la plupart, des fractures. Le film le moins "craquelé" a été obtenu par déposition électrophorétique dans du butanol en courant alternatif. Le chapitre 8 évalue enfin le taux de pénétration et l’épaisseur du revêtement d’hydroxyapatite à l’intérieur d’un implant tridimensionnel d’architecture poreuse. On remarque une diminution de l’épaisseur du revêtement d’hydroxyapatite au fur et à mesure que l’on pénètre dans la structure 3D de l’implant. En fonction de la taille des pores de l’implant, le dépôt peut atteindre 400 microns (pour des pores de l’ordre de 0,4mm) jusqu’à 1400 microns (pour des pores atteignant 0,9mm). Des études comparatives ont été réalisées avec de l’hydroxyapatite recouverte d’alginine.

Photocatalytic processes possess favourable features that could address the various issues concerning environmental pollution. Among these issues, treatment of polluted wa- ter and water splitting for renewable hydrogen production are extensively studied but are still confronted to limitations for achieving high photocatalytic efficiencies that could be suc- cessfully commercialized. Investigations on powder materials have been widely reported for pollutant degradation/water treatment, but difficulties are prevailing in the re-usability of the material. Moreover, there is the need for finding a suitable heterostructured photo- catalyst that could provide better charge kinetics, an ultimate goal in photocatalyst design. Therefore, in this work, we have investigated thin-film based heterostructure photocatalysts, for improving the photocatalytic activity, especially towards pollutant degradation. For this purpose, we have investigated the surface and interfacial properties of semiconductor/semiconductor (p-n type, NiO/ZnO) and metal/semiconductor (metal/n- type, RuO2/ZnO) heterostructures using systematic (step-by-step) interface studies, in order to gain knowledge regarding the influence of ZnO surface cleaning in the interfacial band bending, thereby analyzing the possibilities of their use as photocatalysts. Furthermore, we have explored the electrical, optical and interfacial properties of ZnO nanorods (n-type) with NiO coating (p-type) by varying the NiO deposition parame- ters, to identify an optimized heterostructure. We examined the photocatalytic performance of these films for pollutant (Rhodamine B) degradation. In parallel, we explored the inter- action of water with heterostructured (NiO/ZnO) photocatalysts, to interpret the surface reactions and their influence on interfacial band bending, a strategy for understanding the heterostructured photocatalysts, which was not explored before. Finally, we tested the ZnO nanorod film in an industrial research context for Rhodamine B degradation, to investigate the upscaling perspectives of the materials developed in this project.

Epitaxy of metal oxides is of great interest since it provides a way to obtain desired novel properties for the applications such as electronics and energy. However, earlier epitaxy research's have been restricted because of the limited range of compositions and low-index of commercially available single crystal substrates. Consequently, novel epitaxy synthesis methods need to be developed in order to go beyond the present demands of of single crystal substrates in terms of phase, composition, size, orientation and symmetry. In this research work, we have developed a high-throughput synthesis process, called combinatorial substrate epitaxy (CSE), where an oxide film is grown epitaxially on a polycrystalline substrate. As a proof-of-concept, we firstly fabricated Ca3Co4O9 films on Al2O3 ceramics. Films have a good local epitaxial registry, and the Seebeck coefficient is about 170 µV/K at 300 K. High quality BiFeO3/La0.7Sr0.3MnO3 thin film heterostructures were secondly deposited on dense LaAlO3 ceramics prepared by spark plasma sintering. Piezoforce microscopy was used to confirm the local ferroelectric properties. Thirdly, we investigate the growth of of metastable monoclinic Dy2Ti2O7 epitaxial films on polycrystalline La2Ti2O7 substrates. We conclude that CSE approach opens the way towards unexpected electronic properties in oxide films.

The goal of this work is to characterize the structure and lattice dynamics of complex chalcogenide alloys. Particular interest is paid to the system AgPbmSbTem+2 (LAST-m), AgSbTe2 and the binaries PbTe, SnTe and GeTe. Synchrotron radiation studies including high energy X-ray di raction and nuclear inelastic scattering, and macroscopic measurements of heat capacity and elastic constants were performed. A new resonant ultrasound spectroscopy setup with capable of performing measurements from room temperature to 1073K was built for mechanical characterization of the thermoelectric alloys at their working temperatures. The rst chapter presents a brief review of relevant information on thermoelectricity and on the materials under study. The characterization methods including heat capacity, resonant ultrasound spectroscopy, X-ray di raction and nuclear inelastic scattering are introduced. Not as an exhaustive review, but rather in order to give the reader a basic level of understanding and a sense of the acessible information. The introduction is followed by three chapters which address the experimental studies of lattice dynamics in chalcogenide alloys. Chapter 2 describes the lattice dynamics in the compounds GeTe, SnTe and PbTe studied by 119Sn and 125Te nuclear inelastic scattering. The obtained partial density of phonon states were compared with published theoretical calculations, and the resulting vibrational properties were found to be in good agreement with these reports. Additionally, the phase purity and structure were characterized by high energy X-ray di raction. The atomic arrangement, rhombohedral for GeTe and cubic for SnTe and PbTe, is seem to a ect the density of phonon states, with the NaCl-type structure having a softer character in comparison with the rhombohedral structure. In Chapter 3, the lattice dynamics of a polycrystalline AgSbTe2 sample was investigated by 121Sb and 125Te nuclear inelastic scattering, at low temperatures. For this compound, the phonon modes have energies below 25meV and a sound velocity of vs =1490(30) m/s was determined. A simple temperature independent estimation of the lattice thermal conductivity of AgSbTe2 yielded L =0.50 0.05Wm􀀀1K􀀀1. The low Debye temperature, D =150(15)K combined with the short phonon lifetime and the low sound velocity are found to be key factors for the low thermal conductivity in AgSbTe2 and are related to the good thermoelectric performance in AgSbTe2 and AgSbTe2containing systems. Chapter 4 is dedicated to the study of the average and local structure in bulk AgPb18SbTe20 alloy, by a combined Rietveld and Pair Distribution Function analysis. The strong in uence of the synthesis conditions on the lattice parameters and on the composition and the concentration of nanoclusters in LAST-18 is con- rmed. Moreover, the 121Sb and 125Te partial density of phonons states were obtained by nuclear inelastic scattering in order to separately characterize the lattice dynamics from the matrix and the nanoinclusions. Additional characterization of the elastic properties and lattice governed properties were performed by resonance ultrasound spectroscopy, heat capacity and thermal expansion measurements. The nal chapter is dedicated to the resonant ultrasound spectroscopy technique, and the process of building up this bu er-rods high-temperature measurement system are presented. Advantages and disadvantages, as well as limitations and di culties are discussed. Using the \mode-tracking" method, the mechanical behavior of a PbTe and a Niobium sample, from room temperature to 523K and from room temperature to 973 K, respectively, were investigated.

In the last few years, nanosphere lithography emerged as an inexpensive, material specific and high-output nanostructure fabrication process to manufacture arrays of periodic structures. The goal of this thesis was centered on both parts of the nanosphere lithography process, namely first the optimization of monolayer colloidal masks prepared by spin coating of monodisperse polystyrene (PS) nanospheres and secondly the use of these masks to develop new attractive applications in various fields. In order to assess the quality of the manufactured colloidal crystal masks, we developed a computerized image analysis procedure (Matlab) based on SEM micrographs. We successfully performed the different stages of the image analysis in such a way to discriminate and identify each nanosphere. As a quantification of order in the self-organized nanospheres, we chose to determine the percentage of hexacoordinated nanospheres by computing the distances between each of them. We applied experimental design to spin coating to evaluate the efficiency of this method to extract and model the relationships between the experimental parameters and the degree of ordering in the particles monolayers. We identified adequate spin coating parameters to synthesize large defect-free domains, reaching up to 200 μm2, which is the highest value ever reported for samples prepared by spin coating. Statistical analysis highlighted that rapid substrate acceleration and high rotation rates are necessary to get large, well-ordered areas. We also studied the surfactant concentration usually added to the beads suspension or the use of reactive ion etching (RIE) process to modify the masks. By using PS nanosphere templates (490 nm or 250 nm diameter), we successfully manufactured large arrays of L10-Fe50Pt50 and Co nanotriangles with uniform sizes. In addition to crystallographic and microstructural characterizations, we evaluated the magnetic properties of the nanostructures both from a qualitative (MFM) and quantitative (SQUID) point of view. The magnetic stability of the single-domain FePt nanodots was evidenced by focused MOKE analysis. This is of major importance for further use in magnetic storage applications and has never been reported yet. The soft magnetic Co nanodots displayed either single domain or vortex domains states, depending on the magnetization direction. The MOKE hysteresis loops revealed an increased coercive field compared with thin films. This is probably due to a specific magnetization reversal process caused by the shape of the nanodots. Oxide nanostructures were then manufactured. The polystyrene templates (490 nm diameter) were used for the guided hydrothermal growth of well-aligned ZnO nanowires. The control of the spacing between the nanowires combined with high c-axis preferred orientation led to higher dye loading values compared with continuous unpatterned films. This was undoubtedly attributed to an increased accessible surface area due to the patterning. Moreover, the increased roughness due to the patterning induced a higher water contact angle compared with an unpatterned ZnO nanowire array. Reversible superhydrophylicity to hydrophobicity was observed and controlled by successive UV illumination and O2 annealing. The achievements attained in this work have brought a significant contribution to the field of nano- and microfabrication. New pathways were opened for interesting future work with respect to continued fundamental and applied research.

Ce travail de thèse porte sur l’étude des relations procédé, micro/nanostructures et propriétés thermoélectriques de composés oxydes de formulation In2-xGexO3. Dans ce contexte, des techniques de synthèse des poudres par chimie douce (procédé citrate), de mise en forme en milieu liquide (coulage en moule poreux) et de frittage non conventionnel des céramiques (frittage micro-ondes) ont été développées. Par comparaison à des procédés d’élaboration classiques, le développement du procédé citrate a permis la préparation de microstructures homogènes permettant d’optimiser les propriétés de transport. La figure de mérite ZT des composés In2-xGexO3 atteint ainsi des valeurs supérieures à 0,3 à 1000 K. Afin de mieux contrôler la densification pendant le frittage, l’utilisation d’une technique de mise en forme par coulage en moule poreux a été expérimentée et des densités après frittage proches de la densité théorique ont pu être obtenues grâce à une mise en ordre optimisée des particules dans les composés crus. Le développement du frittage micro-ondes à ce type de matériau a par ailleurs permis d’obtenir des microstructures très fines (taille de grains inferieure à 500 nm). Un dispositif thermoélectrique (surface de 35 × 40 mm², 56 jambes) de formulations Ca3Co4O9 et In1.994Sn0.006O3 a été également réalisé. Celui-ci délivre une puissance de 480 mW pour un gradient de température de 550 K. Ces différentes études ouvrent des perspectives intéressantes dans l’élaboration de composés et de dispositifs thermoélectriques à base d’éléments oxydes présentant des performances accrues.

Nuclear resonance scattering of synchrotron radiation enables probing hyperfine interactions and element specific vibrational modes of nuclei that exhibit a Mössbauer transition. A prerequisite for this method is a monochromator with narrow bandwidth. Silicon, as the most commonly used crystal in monochromators, is not suitable for experiments above 30 keV photon energy. Using a sapphire single crystal in backscattering geometry is an alternative. An x-ray beam with narrow bandwidth can be obtained from a back-reflection with a Bragg angle of a few arcsec to a few arcmin smaller than =2. The development of a sapphire backscattering monochromator with high energy resolution, better than 1 meV, would permit detailed lattice dynamics characterization of novel functional materials. Sapphire is a very rigid material with desirable optical properties, high chemical resistance, and high heat conduction. Crystals of large size with high quality are of interest for many industrial applications, and suited for optics and optoelectronics operating under ambient condition or extreme conditions. The purpose of the work in this dissertation is twofold. First, study the quality of sapphire single crystals by modern high-resolution characterization techniques in order to acquire microstructural information and thereby a better understanding of the origin of lattice defects. The second goal is the study of the lattice dynamics in materials based on tellurium and antimony with Mössbauer energies of 35.49 keV and 37.13 keV using nuclear resonance scattering with energy resolution given by one of the highest quality sapphires. White beam topography of more than thirty crystals, grown at the Shubnikov Institute of Crystallography in Moscow, revealed qualitative information about the distribution of lattice defects of which linear defects, i.e. dislocations, are the majority type. The lowest dislocation density of 10^2 -10^5 cm^-2 was found in C-plane crystals grown by the Kyropoulos and Bridgman techniques. We carried out rocking curve imaging in backscattering geometry to estimate the lattice parameter variation and energy resolution from back-reflections. Minimum variations of the lattice parameters on the order of 10^-8 were observed from spots with an edge length of 0.2-0.5 mm. There are a few spots with such a quality in one crystal which makes it suitable as backscattering monochromator in nuclear resonance scattering or as analyzer in resonant inelastic x-ray scattering. The use of very high energy resolution nuclear inelastic scattering with 0.7 meV at the energy of the nuclear transition in 121Sb and 125Te enables valuable insight into the phonon scattering of thermoelectric materials, that convert heat to prvide electricity and vice versa, composed of Sb and Te. A careful study is done on heat carrying acoustic modes in the partial density of states of (PbTe)mAgSbTe2, so called LAST-m alloys. An impressive mismatch in the phonon group velocities in the 2-5 meV range and di erence in the force constants for the Sb and Te density of phonon states is observed, a phonon mismatch predicted to be responsible for low lattice thermal conductivity in LAST-m. An in-depth understanding of the element-specific dynamic properties of cubic and orthorhombic antimony trioxides was achieved using nuclear resonance scattering with an energy resolution of 1 meV at the nuclear transition energy of 121Sb. A softening of the Sb bonds upon transformation from cubic molecular structured in alpha-Sb2O3 to chain structured orthorhombic in beta-Sb2O3 is observed. Furthermore, results on the lattice dynamics on alpha-TeO2, with quasi molecular structure, demonstrate strong interatomic Te bonds, comparable with the strong bonds in molecular structured alpha-Sb2O3. The nuclear resonance data is complemented with inelastic neutron scattering data that reveals the oxygen vibrational modes. In addition, the experimental results validate the calculations of the vibrational modes in these types of materials and serve as benchmark for the calculation.

Drug delivery systems in the size range of ~ 10-250 nm are enabling tools for the site-specific targeting and controlled release applications. To take advantage of these capabilities, various nanocarriers e.g., micelles, dendrimers, liposomes, nanoparticles, nanocapsules, nanotubes, and nanogels have been designed for drug delivery. Specifically, micelle-based drug carrier systems have emerged as promising tools for site-specific delivery and controlled release applications. Despite several advantages over conventional drugs, some limitations of micelle-based drug delivery have also been reported. These drawbacks include low stability in vivo, poor penetration, modest accumulation in tumor tissues, and inadequate control over drug release. To overcome these limitations, stimuli-responsive or smart polymeric nanocarriers have been developed for drug delivery and tumour therapy, previously. The most well-known internal stimuli in cancerous regions include higher acidity associated with dysregulated metabolism in tumour tissues, elevated levels of glutathione in the cytosol and nucleus of cancer cells, and altered degradative enzymes in the lysosomes, and reactive oxygen species in the mitochondria. These intrinsic microenvironments can be exploited as internal stimuli to attain active drug release in the tumor tissues or cancer cells. Particularly, the reducing potential inside the cancer cells is considerably higher than found in the extracellular environment and bloodstream. Therefore, such varying redox potential can be exploited for tumor specific drug delivery and controlled release applications. Various types of redox-responsive micelles have been developed previously. Generally, redox responsive micelles have disulfide linkages that undergo rapid cleavage in the presence of reducing agents in the intracellular components, however, are stable at oxidizing extracellular environment. The redox-responsive disulfide bridges can be incorporated into nanocarriers by placing multiple disulfide bonds in the hydrophobic backbone or by conjugating therapeutic agents to the side chain of the polymer via a disulfide linker. Another strategy to construct redox-responsive linkages is to crosslink the polymeric nanocarriers with a disulfide crosslinker. It has been studied that polymeric micelles can dissociate, especially upon administration when they are diluted below their critical micelle concentration. The stability of polymeric micelles can be enhanced by chemical crosslinking. Various types of crosslinked micelles can be prepared subjected to the localisation of the crosslinking, e.g. shell crosslinked micelles, and core crosslinked micelles. Introducing redox-responsive bridges by disulfide crosslinker may not only provide stability to nanocarriers against dilutions during circulation, but also render them responsive to reduction conditions. Specifically, redox-responsive core-crosslinked micelles have demonstrated good stability and better ‘stealth’ properties, however, the hydrophobic cores of most of the existing core-crosslinked micelles have been based on non-degradable polymers such as polyacrylamide or polyacrylate. The non-degradable constituent of the block copolymer may cause complications in clinical applications. Therefore, reduction-responsive core-crosslinked micelles comprising entirely of biologically inert or biocompatible and biodegradable polymers would be better candidates for drug delivery and controlled release application. To overcome these limitations, micelles based on polyesters (a class of aliphatic biodegradable polymers) can used for drug delivery application. In the last few decades, various FDA approved aliphatic polyesters e.g. poly(lactic-co-glycolic acid) (PLGA), poly(ε-caprolactone), and poly(lactic acid), have been intensively studied to exploit their potential in drug, gene and protein delivery and controlled release applications. However, most of these polyesters lack functional groups which make it difficult to incorporate redox-responsive linkages to core-crosslink their micelles. To address these issues, we have synthesized functional biodegradable and biocompatible block copolymers based on mPEG-b-poly(εCL-co-αClεCL). The pendent chloro groups of the block copolymer were converted into azides using nucleophilic substitution reaction to produce mPEG-b-poly(εCL-co-αN3CL) block copolymer as a precursor of reactive polymeric micelles. The synthesized polymers were characterized by NMR, FT-IR and size exclusion chromatography (SEC). Micelles were prepared using dialysis method and methotrexate (an anticancer drug) was loaded into the hydrophobic core of the reactive micelles. Micelles were subsequently crosslinked by a redox-responsive bis-alkyne ethyl disulfide crosslinker. The size distributions and morphology of core-crosslinked micelles were assessed using dynamic light scattering (DLS) and transmission electron microscopy. The drug release studies were performed under simulated non-reducing and reducing conditions. Cellular uptake studies in human breast cancer cells (MCF7 cells) were performed using Oregon-green loaded core-crosslinked micelles. The MTX-loaded core-crosslinked micelles were assessed for their cytotoxicity in human breast cancer cells by MTT assays. The apoptosis inducing potential of MTX-loaded core-crosslinked micelles was analysed using Hoechst/PI assays and was further probed by annexin-V/PI assays. The data from these studies indicate that drug release from these cross-linked micelles can be controlled and that the redox-responsive micelles are more effective carriers for MTX than non-cross-linked analogues in the cell-lines tested. In another strategy, a multifunctional amphiphilic block copolymer based on α-amine-PEG-b-poly(εCL-co-αN3εCL) was synthesized and subsequently was used to conjugate methotrexate on the hydrophilic block for receptor mediated targeting of breast cancer cells. Cellular uptake studies revealed 2.3-fold higher uptake of MTX-conjugated micelles as compared with un-conjugated micelles. The blank micelles showed low cytotoxicities in breast cancer cells, however, MTX-conjugated micelles exhibited greater antitumor activities in contrast to free-MTX. We hypothesize that these functional micelles could be potentially powerful nanocarriers for stimuli-responsive controlled release, active tumour targeting and therapy.

The development of nano-sized particles is motivated by their optical, electronic and magnetic behavior related to quantum confinement resulting from the nanometric size. To prevent aggregation in solution, the nanoparticles are covered with stabilizing molecules. The aim of this thesis is to develop a new generation of functional copolymers with different architectures to improve the stability of various synthesized NPs. Two types of nanoparticles are considered, gold NPs for the optical properties and iron oxide NPs for the magnetic properties. The copolymers considered in this study are synthesized following a controlled radical polymerization process, i.e. Reversible Addition - Fragmentation Chain Transfer (RAFT) and confer novel properties to the coated nanoparticles. Stealth NPs are obtained when they are covered by the poly(ethylene oxide), and thermosensitive NPs when they are stabilized by the poly (N-isopropyl acrylamide). These properties have been exploited in applications in the biomedical field. Another challenge in this work is the synthesis and functionalization of the surface of carbon NPs, and thus carbon nano-capsules were synthesized by graphitization of poly (acrylic acid)-poly (acrylonitrile) micelles and carbon nanotubes have been decorated by magnetite NPs allowing their orientation in a magnetic field.

Glioblastoma (GBM) is the most common and lethal form of brain cancer. The diffusive nature of GBM means the neoplastic tissue cannot be removed completely by surgery. Often, residual GBM cells can be found close to the border of the resection cavity and these cells can multiply to cause tumor recurrence in ≥90% of GBM patients. An implant that can sustainably release chemoattractant molecules called stromal cell-derived factor-1α (SDF- 1α), which bind selectively to CXCR4 receptors on the surface of GBM cells, may be useful for inducing chemotaxis and recruitment of the residual GBM cells. This may then give access to selective killing of the cells and ultimately reduce tumor recurrence. In this work, SDF-1α was initially encapsulated into poly-lactic-co- glycolic acid (PLGA)-based nanoparticles. A high encapsulation efficiency (76%) could be achieved using a simple phase separation process. The SDF-1α-loaded nanoparticles were then incorporated into a chitosan-based scaffold by electrospinning to obtain nanofibrous implants that mimic the brain extracellular matrix structure. In vitro release study revealed that the implant could provide sustained SDF-1α release for 5 weeks. The gradual SDF-1α release will be useful for establishing SDF-1α concentration gradients in the brain, which is critical for the chemotaxis of GBM cells. A 7-day in vivo biocompatibility study revealed evidence of inflammation at the implantation site without any visible signs of clinical deterioration in the animal subjects. A long-term study (100 days) aiming to confirm the in vivo safety of the implants before proceeding to efficacy studies in a suitable GBM resection model is currently underway.

Over the last decades, polymer micelles have attracted an increasing interest in drug pharmaceutical research because they could be used as efficient drug delivery systems. The goal of this thesis was centered on the design of new smart nanocarriers and more particularly on the basis of reversibly redox-cross-linked polymer micelles. The first part of that work was dedicated to the synthesis of new macromolecular architectures associating biodegradable hydrophobic polymers such as polyester (e.g. PCL), polycarbonate (e.g. PTMC) or also polyphosphate (e.g. PBODOP) and the water soluble poly(ethylene oxide) (PEO) frequently used due to its biocompatible properties. Well-defined block copolymers have been synthesized by ring-opening polymerization. The second part of that work focused on the cross-linking of the hydrophobic block in order to obtain well stabilized micelles. The copolymerization of α-chloro-ε-caprolactone (αClεCL) allows to easy functionalize the hydrophobic block in order to reversibly cross-link the future micelle core by the addition of a disulfide bearing cross-linker. The self assembly of theses copolymers and redox-dependent micellization behaviours have been studied by diffusion light scattering and transmission electronic microscopy. Finally, the potential of these redox-sensitive micelles as active drug delivery system has been analysed by investigating their stealthy behaviours using the complement activation (CH50) test, their cytotoxicity, their cellular internalization and also the redox-sensitive profile of a hydrophobic dye.

This research project falls within the framework of CO2 valorization for the development of new monomers and CO2-based polymers (chemical fixation of CO2). Specifically, this project targets the study of the coupling of alcohols and CO2 for the fabrication of low carbon footprint carbonates and polycarbonates. This thesis work revolves around two main stages. The first step is to propose a mechanistic study on the synthesis of (a)cyclic carbonates from alcohols and CO2 in mild conditions (low temperature and pressure). Using kinetic monitoring by in-situ ATR-IR spectroscopy supported by molecular modelling (DFT calculations), a careful comparison of dual organic systems, used as activators of these thermodynamically unfavorable reactions, and their performances on the synthesis of (a)cyclic carbonates is described in details . In particular, reaction mechanisms are identified to better understand the differences observed in reactivity and selectivity. In the second step, the development of CO2- and bio-sourced polycarbonates is studied, especially for the synthesis of isosorbide polycarbonates as promising materials with properties similar to conventional petroleum-based polycarbonates. The direct coupling between isosorbide and CO2 is first studied by varying the experimental conditions. Since this direct copolymerization is very difficult due to the thermodynamic limitations encountered during the alcohol / CO2 coupling, a strategy for synthesizing new monomers from bio-based alcohols and CO2 is then proposed. This new platform of molecules then allows for the synthesis of polycarbonates with tunable chemical structures by melt polycondensation. In this context, an in-depth study of the reactivity of these monomers for obtaining (co)polycarbonates is carried out as well as the physico-chemical characterisations of the synthesized materials.

This study focused on the design and development of new structuring agents of silica constituted of induced and reversible assemblies of original copolymers, the double hydrophilic block copolymers (DHBC). The first system studied consists of a neutral-anionic DHBC PAPEO-b-PAA ou poly(acrylate methoxy poly (ethylene oxide))-b-poly (acrylic acid). The PAA block is a weak polyacid with a degree of ionization depending on the pH. In aqueous solution and in a right pH range, the association of this copolymer with a weak polybase, an oppositely charged polyamine, such as an oligochitosan, leads to the formation of polyion complex micelles (PIC) with a core/corona structure. These micelles can direct the structure of highly organized inorganic materials with different types of mesostructures. In a second step, by adjusting the conditions of pH, ionic strength, it is possible to "control" the extraction of organic species to get functional porous materials able to trap species of charge opposite to the functionality. Organized materials are obtained because of a favourable balance of the interactions between organic and inorganic species. If a polyamine/silica interaction occurs at the expense of the interaction polyamine/DHBC, the mesostructuring process by the micelles is limited. A neutral-cationic DHBC PEO-b-PDMAEMA poly(ethylene oxide)-b-poly(2-(dimethylamine)ethyl) associated with an anionic PVS poly(vinyl sulfonic acid) polymer can play a dual role in the synthesis of silica materials: firstly managing the growth of silica particles by interacting with the silicates and secondly acting as a structuring agent in association with PVS, confering a mesostructuration to the material. Finally, a very promising approach allowed to encapsulate water-soluble and charged drugs in a material by using as silica complexing agent a complex between the drug and a DHBC.

Tumor acidity has been shown to play a major role in resistance to chemotherapy. The use of nanomedicines, as lipid nanocapsules (LNC), allows to protect drugs from this acidic environment. They can also improve the biodistribution of therapeutics and to target the tumor environment. The aim of this thesis concerns the evaluation and characterization of stealth and pH-sensitive LNC in the context of melanoma. Firstly, these works consisted in characterizing the vascularization of human and mice melanoma. These studies allowed to compare different tumors (density, size and structure), and determine if the used of nanocarrier is suitable in the context of melanoma. The second part of this thesis described the development and the characterization of new copolymers, combining stealth and pH-sensitive properties. These copolymers, composed of N-vinylpyrrolidone (NVP) and vinylimidazole, were synthesized by RAFT polymerization and were post-inserted onto LNC surface. These modifications allowed to obtain charge reversal nanocarriers, leading to increase their melanoma cell uptake under acid pH. Finally, biodistribution of these modified nanoparticles was studied in vivo and highlighted the interest of NVP in the development of stealth nanocarriers. To conclude, the developed copolymers able to extend nanocarrier circulation time and to provide pH-responsive properties which should increase the tumor internalization of LNC in vivo and potentiate the effect of anticancer drugs.

Colloids are known since the early 19th century but they have mainly sparked interest since the last few decades thanks to their use in biomedical applications or in the design of new materials. The aim of this thesis is the study of stimuli-responsive colloids by Small-Angle Neutron Scattering (SANS). The stimuli may be the pH, the temperature, the addition of a ligand or of metallic ions or a combination of them. SANS is a useful technique which provides information about the internal structure of nano-objects but also, if appropriate conditions are met, about the organization of the objects in solution. Several samples built from the auto-assembly of block copolymers and liposomes have been prepared. The macroscopic cross sections have been modeled with the aim to infer the main structural parameters of the samples: the global size, the polydispersity, the structure and volume occupied by the hydrophilic and hydrophobic components inside the nano-objects. The SANS measurements have been performed as a function of the intensity of the stimulus, in order to quantify the evolution of the structural parameters. The first experimental part focuses on micellar samples built from sequenced block copolymers. The blocks may be biodegradable (e.g., poly(ε-caprolactone)), biocompatible (e.g., poly(ethylene oxide)), pH-sensitive (e.g., poly(2 vinylpyridine), poly(acrylic acid)) or temperature-sensitive (e.g., poly(N-isopropylacrylamide)). The influence of the concentration and the formation of bridges between micelles have been investigated through the analysis of metallo-supramolecular micellar gels resulting from the self-assembling of polystyrene-block-poly(tert-butylacrylate) PS-b-PtBA-tpy (tpy stands for terpyridine) block copolymers in the presence of transition metal ions. The second part focuses on the analysis of liposomes interacting with an increasing concentration of Randomly Methylated β-cyclodextrins (RAMEB). This part is divided into two chapters. The first one deals with liposomes mainly composed of dimyristoylphosphatidylcholine (DMPC) and the second one investigates the effects of cholesterol doping on the same DMPC liposomes interacting with RAMEB.