<p>In the first part of this work, the influence of hydrogen on the structural and magnetic properties of Fe/V(001) superlattices was studied. The local structure of the vanadium-hydride layers was determined by extended x-ray absorption fine structure (EXAFS) measurements. The magnetic ordering in a weakly coupled Fe/V(001) superlattice was investigated using the magneto-optical Kerr effect (MOKE). The interlayer exchange coupling is weakened upon alloying with hydrogen and a phase with short-range magnetic order was observed.</p><p>The second part is concerned with first-principles calculations of magnetic materials, with a focus on magnetic recording applications. The uniaxial magnetic anisotropy energy (MAE) of Fe, Co, and Ni was calculated for tetragonal and trigonal structures. Based on an analysis of the electronic states of tetragonal Fe and Co at the center of the Brillouin zone, tetragonal Fe-Co alloys were proposed as a material that combines a large uniaxial MAE with a large saturation magnetization. This was confirmed by experimental studies on (Fe,Co)/Pt superlattices. The large uniaxial MAE of L1₀ FePt is caused by the large spin-orbit interaction on the Pt sites in connection with a strong hybridization between Fe and Pt. Furthermore, it was shown that the uniaxial MAE can be increased by alloying the Fe sublattice with Mn. The combination of the high-moment rare-earth (RE) metals with the high-<i>T</i>_C 3<i>d</i> transition metals in RE/Cr/Fe multilayers (RE = Gd, Tb, Dy) gives rise to a strong ferromagnetic effective exchange interaction between the Fe layers and the RE layer. The MAE of hcp Gd was found to have two principal contributions, namely the dipole interaction of the large localized 4<i>f</i> spins and the band electron magnetic anisotropy due to the spin-orbit interaction. The peculiar temperature dependence of the easy axis of magnetization was reproduced on a qualitative level.</p>

<p>The subjects of the thesis are theoretical first principles calculations of X-ray absorption (XA) spectra and current induced spin transfer torque. XA spectra calculated from atomic multiplet theory and from band structure calculations, based on density functional theory for La_0.7Sr_0.3MnO₃ have been compared to experiment. The comparison shows that the effect of the core hole created in the XA process must be considered in the calculation. The theory by Mahan, Nozières and De Dominicis (MND) of dynamical core hole screening is generalised to multiband systems and implemented in first principle calculations. Calculations of the XA spectrum of graphite, including dynamical core hole screening, are shown to better reproduce the relative intensity of the peaks in the experimental spectrum compared to static calculations based on the local density of state of a core excited atom. In combination with experiments the developed method to calculate XA spectra is used to investigate the electronic structure of mixed valent Yb, hydrogen storage in carbon nanotubes and the structure of liquid water. Moreover, a method to calculate the current induced spin transfer torque in materials with a helical spin density wave from first principles has been developed. The method is applied to rare earth metals and it is shown that a current along the axis of spin rotation induces a torque which gives rise to a rotation of the magnetisation direction.</p>

The subjects of the thesis are theoretical first principles calculations of X-ray absorption (XA) spectra and current induced spin transfer torque. XA spectra calculated from atomic multiplet theory and from band structure calculations, based on density functional theory for La0.7Sr0.3MnO3 have been compared to experiment. The comparison shows that the effect of the core hole created in the XA process must be considered in the calculation. The theory by Mahan, Nozières and De Dominicis (MND) of dynamical core hole screening is generalised to multiband systems and implemented in first principle calculations. Calculations of the XA spectrum of graphite, including dynamical core hole screening, are shown to better reproduce the relative intensity of the peaks in the experimental spectrum compared to static calculations based on the local density of state of a core excited atom. In combination with experiments the developed method to calculate XA spectra is used to investigate the electronic structure of mixed valent Yb, hydrogen storage in carbon nanotubes and the structure of liquid water. Moreover, a method to calculate the current induced spin transfer torque in materials with a helical spin density wave from first principles has been developed. The method is applied to rare earth metals and it is shown that a current along the axis of spin rotation induces a torque which gives rise to a rotation of the magnetisation direction.

The solid oxide fuel cell (SOFC) technology has been under research since thelate 1950s, and most of the research has been on designs utilizing yttria stabilized zirconia (YSZ) as the electrolyte of choice. However, the SOFC technology has the major drawback of requiring high operation temperatures (up to 1000 degrees Celcius), so research of alternative materials have come into interest that would possibly require a lower working temperature without any significant loss of conductivity.One such material of interest for the electrolyte is compounds of ceriumdioxide (ceria). Ceria is well known for its ability to release oxygen by formingoxygen vacancies under oxygen-poor conditions, which increases its oxygen ionconductivity, and works at a lower temperature than the YSZ compounds whenproperly doped. Conversely, ceria is also able to absorb oxygen under oxygen-rich conditions, and those two abilities make it a very good material to use in catalytic converters for reduction of carbon monoxide and nitrogen oxide emission. The ability for the oxygen ions to easily relocate inbetween the different lattice sites is likely the key property of oxygen ion transportation in ceria. Also, in oxygen-rich conditions, the absorbed oxygen atom is assumed to join the structure at either the roomy octrahedral sites, or the vacant tetrahedral sites. Following that, the oxygen atom may relocate to other vacant locations, given it can overcome a possible potential barrier. This thesis studies how those interstitial oxygen vacancies (defects) affect theenergy profile of ceria-based supercells by first principles calculations. The system is modeled within the density functional theory (DFT) with aid of (extended) local density approximation (LDA+U) using the software VASP. Furthermore, it is studied how those vacancies affect neighbouring oxygen atoms, and wether or not it is energetically benificial for the neighbouring atoms to readjust their positions closer or further away from the vacancy. The purpose of this thesis is to analyze wether or not it is theoretically possible that interstitial oxygen vacancies may cause neighbouring oxygen atoms to naturally relocate to the octahedral site in ceria, and how this affects the overall energy profile of the material.

In the first part of this work, the influence of hydrogen on the structural and magnetic properties of Fe/V(001) superlattices was studied. The local structure of the vanadium-hydride layers was determined by extended x-ray absorption fine structure (EXAFS) measurements. The magnetic ordering in a weakly coupled Fe/V(001) superlattice was investigated using the magneto-optical Kerr effect (MOKE). The interlayer exchange coupling is weakened upon alloying with hydrogen and a phase with short-range magnetic order was observed. The second part is concerned with first-principles calculations of magnetic materials, with a focus on magnetic recording applications. The uniaxial magnetic anisotropy energy (MAE) of Fe, Co, and Ni was calculated for tetragonal and trigonal structures. Based on an analysis of the electronic states of tetragonal Fe and Co at the center of the Brillouin zone, tetragonal Fe-Co alloys were proposed as a material that combines a large uniaxial MAE with a large saturation magnetization. This was confirmed by experimental studies on (Fe,Co)/Pt superlattices. The large uniaxial MAE of L10 FePt is caused by the large spin-orbit interaction on the Pt sites in connection with a strong hybridization between Fe and Pt. Furthermore, it was shown that the uniaxial MAE can be increased by alloying the Fe sublattice with Mn. The combination of the high-moment rare-earth (RE) metals with the high-TC 3d transition metals in RE/Cr/Fe multilayers (RE = Gd, Tb, Dy) gives rise to a strong ferromagnetic effective exchange interaction between the Fe layers and the RE layer. The MAE of hcp Gd was found to have two principal contributions, namely the dipole interaction of the large localized 4f spins and the band electron magnetic anisotropy due to the spin-orbit interaction. The peculiar temperature dependence of the easy axis of magnetization was reproduced on a qualitative level.

Afin de pouvoir utiliser les nouveaux matériaux semi-conducteurs dans les domaines de l’électronique et de l’optique, il faut parvenir à comprendre leur «structure électronique», ou plus précisément le positionnement des niveaux d’énergie des électrons impliqués dans l’absorption / émission d’un photon. Les propriétés électroniques, sensibles à la composition chimique et à la structure du matériau, sont théoriquement accessibles en résolvant les équations de la mécanique quantique sur ordinateur. Ce travail porte sur des simulations théoriques de la structure électronique de semi-conducteurs binaires constitués d'indium (ou du gallium) et de sélénium, ainsi que de leurs "dérivés" à base de cuivre. La stabilité relative des phases cristallographiques de certains composés In-Se et Ga-Se a été évaluée, ce qui a permis d’expliquer certaines tendances connues et de formuler des prédictions. Les résultats obtenus seront particulièrement utiles dans le domaine du photovoltaïque. Les simulations numériques ont été réalisées dans le cadre de la théorie de la fonctionnelle de la densité (DFT), visant les structures cristallines d'équilibre et les propriétés électroniques de quelques semi-conducteurs binaires ou (pseudo)ternaires à base de Cu, In, Ga et Se. Les systèmes étudiés possèdent la même structure à courte portée (environnement tétraédrique des cations et anions) mais diffèrent à longue portée. Les composés binaires (Ga/In)Se, (Ga/In)2Se3 constituent des références importantes dans les diagrammes de phases des systèmes à base de (Cu, In, Se) et (Cu, Ga, Se), au sein desquels figurent les phases potentiellement utiles dans le domaine du photovoltaïque. Le travail comprend deux chapitres d'introduction et trois chapitres exposant des résultats nouveaux / In order to optimally use new semiconductor materials in electronics or optics, one needs to understand their “electronic structure”, that is, the mutual placement of the electron energy levels concerned by the processes of absorption / emission of a photon. The electronic properties, which depend on the material’s chemical composition and crystal structure, may be assessed by theory via solving quantum-mechanical equations on a computer. The present work deals with theory simulations of electronic structure done for several binary semiconductors consisting of indium (or gallium) and selenium, moreover for their “derivatives” containing copper. As a result, the relative stability of crystallographic phases of some Ga-Se and In-Se compounds has been assessed, explaining the known trends and making predictions. The results are expected to be useful for current works in photovoltaics. The numerical simulations have been performed within the density functional theory (DFT), aimed at the equilibrium crystal structures and electronic characteristics of several binary or (pseudo)ternary semiconductors based on Cu, In, Ga and Se. The compounds under study share similar short-range order features (tetrahedral environment of both cations and anions), differently assembled on a long-range scale. The binary compounds (Ga/In)Se, (Ga/In)2Se3 mark important end points at the phase diagrams of the (Cu,In,Se) and (Cu,Ga,Se) systems that cover a number of phases relevant, e.g., for applications in photovoltaics. The work comprises two chapters of introduction and three outlining novel results

<p>Combination of the Calphad method and theoretical calculations provides new possibilities for the study of materials science. This work is a part of the efforts within the CCT project (Centre of Computational Thermodynamics) to combine these methods to facilitate modeling and to extend the thermodynamic databases with critically assessed volume data. In this work, the theoretical calculations refer to first-principles and Debye-Grüneisen calculations. The first-principles (i.e. ab initio) electronic structure calculations, based on the Density- Functional Theory (DFT), are capable of predicting various physical properties at 0 K, such as formation energy, volume and bulk modulus. The ab initio simulation software, VASP, was used to calculate the binding curves (i.e. equation of state at 0 K) of metallic elements, cubic carbides and nitrides. From the binding curves, the equilibrium volumes at 0 K were calculated for several metastable structures as well as stable structures. The vibrational contribution to the free energy was calculated using the Debye-Grüneisen model combined with first-principles calculations. Two different approximations for the Grüneisen parameter, γ, were used in the Debye-Grüneisen model, i.e. Slater’s and Dugdale-MacDonald’s expressions. The thermal electronic contribution was evaluated from the calculated electronic density of states. The calculated thermal expansivities for metallic elements, cubic carbides and nitrides were compared with Calphad assessments. It was found that the experimental data are within the limits of the calculations using the two approximations for γ. By fitting experimental heat capacity and thermal expansivity around Debye temperatures, we obtained optimal Poisson’s ratio values and used them to evaluate Young’s and Shear moduli. In order to reach a reasonable agreement with the experiments, it is necessary to use the logarithmic averaged mass of the constitutional atoms. The agreements between the calculations and experiments are generally better for bulk modulus and Young’s modulus than that for shear modulus. A new model describing thermodynamic properties at high pressures was implemented in Thermo-Calc. The model is based on an empirical relation between volume and isothermal bulk modulus. Pure Fe and solid MgO were assessed using this model. Solution phases will be considered in a future work to check the model for compositional dependence.</p>

Materials intended for disposal of spent nuclear fuel require a particular combination of physical and chemical properties. The driving forces and mechanisms underlying the material’s behavior must be scientifically understood in order to enable modeling at the relevant time- and length-scales. The processes that determine the mechanical behavior of copper canisters and iron inserts, as well as the evolution of their mechanical properties, are strongly dependent on the properties of various defects in the bulk copper and iron alloys. The first part of the present thesis deals with precipitation in the cast iron insert. A nodular cast iron insert will be used as the inner container of the spent nuclear fuel. Precipitation is investigated by computing effective interaction energies for point defect pairs (solute–solute and vacancy–solute) in bcc iron using first-principles calculations. The main considered impurities in the iron matrix include 3sp (Si, P, S) and 3d (Cr, Mn, Ni, Cu) solute elements. By computing interaction energies possibility of formation of different second phase particles such as late blooming phases (LBPs) in the cast iron insert is evaluated. The second part is devoted to the fundamentals of dislocations and their role in plastic deformation of metals. Deformation of single-crystal copper under high strain rates is simulated by employing dislocation dynamics (DD) method to examine the effect of strain rate on mechanical properties as well as dislocation microstructure development. Creep deformation of copper canister at low temperatures is studied. The copper canister will be used in the long-term storage of spent nuclear fuel as the outer shell of the waste package to provide corrosion protection. A glide rate is derived based on the assumption that at low temperatures it is controlled by the climb rate of jogs on the dislocations. Using DD simulation creep deformation of copper at low temperatures is modeled by taking glide but not climb into account. Moreover, effective stresses acting on dislocations are computed using the data extracted from DD simulations. / <p>QC 20170428</p>

Les nanostructures, et en particulier les nanofils semi-conducteurs, ont suscité ces dernières années un très grand intérêt pour de nombreuses applications comme les systèmes de récupération d'énergie ou les capteurs de très haute précision. Dans de telles structures des expérimentations et des calculs théoriques ab-initio ont mis en évidence des effets de taille, pouvant modifier significativement les propriétés électromécaniques pour des diamètres de fils en dessous de 10 nm. L'objectif de ce travail de thèse est de proposer des modélisations multi échelle des nanostructures électromécaniques, telles que les nanofils ioniques et des nanocomposites stratifiés, permettant de reproduire les effets de taille associés à l'échelle nanométrique dans un cadre continu, en se basant sur des calculs ab-initio pour identifier et valider les modèles. Dans une première partie, les effets de surface dans des nanofils piézoélectriques isolés homogènes sont modélisés. Une approche multi échelle est développée, incluant une modélisation continue des nanofils en prenant en compte une énergie de surface supplémentaire dans un cadre piézoélectrique, dont les paramètres associés sont identifiés par calculs ab-initio. Pour cela, une procédure basée sur un modèle de films minces est développée, permettant au travers de calculs ab-initio sur des films d'épaisseurs successives d'isoler l'énergie volumique et de surface, et d'en déduire les coefficients élastiques et piézoélectriques de surface. Les équations du modèle continu sont ensuite résolues par une méthode d'éléments finis incluant des éléments de surface adaptés. Le modèle multi échelle continu est comparé à des calculs ab-initio impliquant des modèles atomistiques complets de nanofils de différents diamètres (de 0,6 à 3,9 nm) pour valider les effets de taille des propriétés électromécaniques. Dans une deuxième partie, des modèles multi échelles sont construits en vue de modéliser les effets de taille pour des nanostructures hétérogènes. Ces structures incluent des nanofils revêtus, ou des nanocomposites stratifiés. Pour les nanofils avec hétérogénéités radiales, l'approche précédemment développée est étendue au cas des surfaces revêtues, et le modèle continu fait intervenir une énergie de surface incluant les effets du revêtement. Pour les nanocomposites stratifiés AlN/GaN, les effets de taille observés par calculs ab-initio sont dus à des effets d'interface et induisent des propriétés élastiques dépendantes des épaisseurs des couches. Un modèle de matériau homogénéisé continu est proposé, incluant un modèle d'interface imparfaite, permettant d'inclure les effets de taille, identifié par calculs ab-initio. Dans une dernière partie, des applications à des systèmes de nanogénérateurs à base de nanofils sont proposées, faisant intervenir des ensembles de nanofils alignés dans une matrice polymère et surmontés par une feuille de graphène. Les approches précédemment développées sont utilisées pour modéliser ces structures par éléments finis / Nanostructures, and more specifically semiconductor nanowires, have drawn special attention in recent years for many applications such as energy harvesting systems or sensors of very high precision. Many recent experiments and theoretical ab-initio calculations have evidenced size effects, which can significantly modify the electromechanical properties of nanowires for diameters below 10 nm. The objective of this thesis is to provide multi-scale modeling of electromechanical properties of nanostructures, such as ionic nanowires and laminated nanocomposites, to reproduce the size effects associated with nanoscale in a continuum model, based on ab-initio calculations to identify and validate the models. In a first part, the surface effects in isolated homogeneous piezoelectric nanowires are modeled. A multi-scale approach is developed, including continuous nanowires modeling taking into account an additional surface energy in the piezoelectric laminates where the associated parameters are identified by ab-initio calculations. For this, a procedure based on slabs is developed, allowing through first-principles calculations on successive slabs thicknesses to isolate the surface energy and to deduce the surface elastic and piezoelectric coefficients. The equations of the continuous model are then solved by a finite element method including appropriate surface elements. The continuous multi-scale model is compared with ab-initio calculations involving full atomistic models of nanowires with different diameters (from 0.6 to 3.9 nm) to validate model regarding size effects of electromechanical properties. In the second part, multi-scale models are constructed to describe the size effects for heterogeneous nanostructures. These structures include coated nanowires or laminated nanocomposites. For nanowires with radial heterogeneity, the previously developed approach is extended to the case of coated surfaces, and involves a continuous surface energy incorporating the effects of the coating. For laminated AlN/GaN nanocomposites, size effects observed by ab-initio calculations are caused by the presence of the interfaces and induce size-dependent elastic properties with respect to the layer thickness. A continuum model based on an imperfect interface is proposed to describe the size dependent effective elastic properties of the overall composite, which are identified by ab-initio calculations. In the last part, nanogenerators system based on nanowires are modeled, involving nanowires arrays aligned in polymer substrates with graphene electrode. The previously developed finite element models are used to simulate the electromechanical properties of such systems

Les monofeuillets de dichalcogénures de métaux de transition (TMDC) tels que MoS2 ou WSe2 sont des semiconducteurs bidimensionnels à gap direct, dont les allées K et K' sont inéquivalentes dans la première zone de Brillouin : la levée de dégénérescence induite par le couplage spin-orbite entre les bandes de spin up et dawn est inversée entre les vallées K et K'. Des contacts métalliques magnétiques devraient permettre une injection de spin efficace depuis une électrode magnétique vers un TMDC. Les indices de vallée (Kou K') et de spin (up ou dawn) étant fortement couplés, cela permettrait de sélectionner électriquement l'une ou l'autre des vallées et de réaliser des dispositifs à base de TMDC pour la spintronique (exploitant le spin des électrons) ou pour la valléetronique (exploitant l'indice de vallée des électrons). Dans cette thèse, nous explorons les propriétés physiques des interfaces Co(OOOl)/MoS2 et Ni(lll)/WSe2 par des méthodes de calcul ab-initia basées sur la théorie de la fonctionnelle de la densité. Nous démontrons la nature covalente des liaisons à l'interface entre les monofeuillets de TMDC et les surfaces magnétiques Co(OOOl) et Ni(lll). Nous décrivons la structure atomique de ces interfaces, ainsi que la modification des moments magnétiques induite par des transferts de charge électrique entre atomes. Les liaisons covalentes aux interfaces confèrent aux monofeuillets de MoS2 et de WSe2 un caractère métallique. Nos calculs donnent finalement accès à la polarisation en spin au niveau de Fermi du TMDC connecté à ces électrodes magnétiques, ainsi qu'à la hauteur de la barrière Schottky (différence entre le niveau de Fermi dans la phase métallique du TMDC situé sous le contact magnétique et le bas de la bande de conduction du TMDC pur dans le canal). / Transition metal dichalcogenide (TMDC) single layers like MoS2 or WSe2 are direct band gap two-dimensional semiconductors, with non-equivalent K and K' valleys in the first Brillouin zone. The degeneracy liftingbetween spin-up and spin-down energy bands induced by spin-orbit coupling is inverted between the K and K' valleys . Magnetic metallic contacts should allow spin-injection from a magnetic electrode to a TMDC single layer. The valley (K or K') and spin (up or down) indexes being strongly coupled, this should also allow to electrically select one of the valleys in TMDC-based spintronic or valleytronic deviees. In this Thesis, we have studied the physical properties of the Co(OOOl)/MoS2 and Ni(lll)/WSe2 interfaces with first-principles methods based on the density functional theory. We demonstrated that the TMDC single layers are covalently bound to the Co(OOOl) and Ni(lll) surfaces. We describe the atomic structure of these interfaces and the modification of the magnetic moments induced by charge transfer between interface atomes. The MoS2 and WSe2 single layers become metallic when they are covalently bound to the magnetic metals. We also calculated the spin-polarization at the Fermi level of the TMDC layers connected to th Co and Ni electrodes and the Schottky barrier height (difference between the Fermi level in the metallic phase of the TMDC below the magnetic contact and the bottom of the conduction band in a pure TMDC channel).