Le sujet de cette thèse est la théorie à partir des premiers principes de la structure électronique de matériaux présentant de fortes corrélations électroniques. D’importants progrès ont été faits dans ce domaine grâce aux implémentations modernes de Théorie de la Fonctionelle de Densité (DFT). Néanmoins, la méthode DFT a certaines limitations. D’une part, elle est faite pour décrire les propriétés de l’état fondamental mais pas des états excités des matériaux, bien que ces derniers soient également importants. D’autre part, les approximations de la fonctionnelle employées en pratique réduisent la validité de la DFT, conceptuellement exacte : en particulier elles décrivent mal les matériaux aux effets de corrélations les plus importants.Depuis les années 1990, différentes théoriques quantiques à N corps ont été utilisées pour améliorer ou compléter les simulations à base de DFT. Une des plus importantes est la Théorie du Champ Moyen Dynamique (DMFT), dans laquelle un modèle sur réseau est relié de manière auto-cohérente à un modèle plus simple d’impureté, ce qui donne de bons résultats à condition que les corrélations soient principalement locales. Nous présentons brièvement ces théories dans la première partie de cette thèse. Les progrès récents de la DMFT visent, entre autres, à mieux décrire les effets non-locaux, à comprendre les propriétés hors équilibre et à décrire de vrais matériaux plutôt que des modèles.Afin d’utiliser la DMFT pour décrire de vrais matériaux, il faut partir d’un calcul de structure électronique traitant tous les électrons au même niveau, puis appliquer une correction traitant les effets à N corps sur un sous-espace de basse énergie d’orbitales autour niveau de Fermi. La définition cohérente d’un tel sous-espace nécessite de tenir compte de la dynamique des électrons en-dehors de cet espace. Ces derniers, par exemple, réduisent la répulsion de Coulomb entre électrons dans le sous-espace. Néanmoins, combiner la DFT et la DMFT n’est pas aisé car les deux n’agissent pas sur la même observable. Dans la deuxième partie de cette thèse, nous étudions les modèles de basses énergies, comme la technique échange écranté + DMFT récemment proposée. Nous analysons l’importance de l’échange non-local et des interactions de Coulomb retardées, et illustrons cette théorie en l’appliquant aux états semi-cœur dans les métaux d10 Zn et Cd.Dans la dernière partie, nous utilisons ces méthodes pour étudier trois matériaux corrélés importants d’un point de vue technologique. Dans un premier temps, nous nous intéressons à la physique des mono-lacunes dans la phase paramagnétique du fer. De façon surprenante pour un défaut aussi simple, son énergie de formation n’a toujours pas été obtenue de manière cohérente par la théorie et l’expérience. Nous démontrons que cela est dû à de subtils effets de corrélations autour de la lacune dans la phase paramagnétique à haute température : cette phase est plus fortement corrélée que la phase ferromagnétique, où des calculs de DFT ont été faits.Dans un deuxième temps, nous étudions la transition métal-isolant dans la phase métastable VO2 B. Nous montrons que cette transition ressemble à celle entre la phase conventionnelle rutile et la phase M2 de VO2, mettant en jeu à la fois des liaisons covalentes dans les dimères et une transition de Mott sur les atomes V restants. Nous étudions également l’effet de lacunes d’oxygène sur la structure électronique de VO2.Enfin, nous proposons une technique au-delà de la DFT pour calculer le champ cristallin dans les oxydes et alliages de terres rares. Bien que l’amplitude de ce champ soit faible pour les orbitales localisées 4f des lanthanides, il est crucial pour leur caractère d’aimant permanent. En modifiant l’approximation Hubbard I pour résoudre les équations de DMFT, nous évitons une erreur d’auto-interaction faible en valeur absolue mais physiquement importante, démontrant l’importance de modèles de basse énergie correctement définis. / The topic of this thesis is the first-principles theory of the electronic structure of materials with strong electronic correlations. Tremendous progress has been made in this field thanks to modern implementations of Density Functional Theory (DFT). However, the DFT framework has some limits. First, it is designed to predict ground state but not excited state properties of materials, even though the latter may be just as important for many applications. Second, the approximate functionals used in actual calculations have more limited validity than conceptually exact DFT: in particular, they are not able to describe those materials where many-electron effects are most important.Since the 1990's, different many-body theories have been used to improve or complement DFT calculations of materials. One of the most significant non-perturbative methods is Dynamical Mean-Field Theory (DMFT), where a lattice model is self-consistently mapped onto an impurity model, producing good results if correlations are mostly local. We briefly review these methods in the first part of this thesis. Recent developments on DMFT and its extensions were aimed at better describing non-local effects, understanding out-of-equilibrium properties or describing real materials rather than model systems, among others. Here, we focus on the latter aspect.In order to describe real materials with DMFT, one typically needs to start with an electronic structure calculation that treats all the electrons of the system on the same footing, and apply a many-body correction on a well-chosen subspace of orbitals near the Fermi level. Defining such a low-energy subspace consistently requires to integrate out the motion of the electrons outside this subspace. Taking this into account correctly is crucial: it is, for instance, the screening by electrons outside the subspace strongly reduces the Coulomb interaction between electrons within the subspace. Yet it is a complex task, not least because DFT and DMFT are working on different observables. In the second part of this thesis, we discuss low-energy models in the context of the recently proposed Screened Exchange + DMFT scheme. In particular, we study the importance of non-local exchange and dynamically-screened Coulomb interactions. We illustrate this by discussing semi-core states in the d10 metals Zn and Cd.In the third and last part, we use the methods described above to study the electronic structure of three fundamentally and technologically important correlated materials. First, we discuss the physics of point defects in the paramagnetic phase of bcc Fe, more precisely the simplest of them: the monovacancy. Surprisingly for such a simple point defect, its formation energy had not yet been reported consistently from calculations and experiments. We show that this is due to subtle but nevertheless important correlation effects around the vacancy in the high-temperature paramagnetic phase, which is significantly more strongly correlated than the ferromagnetic phase where DFT calculations had been done.Second, we study the metal-insulator phase transition in the metastable VO2 B phase. We show that this transition is similar to that between the conventional rutile and M2 VO2 phases, involving both bonding physics in the dimer and an atom-selective Mott transition on the remaining V atoms. Motivated by recent calculations on SrVO3, we study the possible effect of oxygen vacancies on the electronic structure of VO2.Finally, we propose a scheme beyond DFT for calculating the crystal field splittings in rare earth intermetallics or oxides. While the magnitude of this splitting for the localized 4f shell of lanthanides does not typically exceed a few hundred Kelvin, it is crucial for their hard-magnetic properties. Using a modified Hubbard I approximation as DMFT solver, we avoid a nominally small but important self-interaction error, stressing again the importance of carefully tailored low-energy models.

A flexible nanofibrous PVDF-BaTiO3 composite material is prepared for impact sensing and biomechanical energy harvesting applications. Dielectric polyvinylidene fluoride (PVDF) and barium titanate (BaTiO3)-PVDF nanofibrous composites were made using the electrospinning process based on a design of experiments approach. The ultrasonication process was optimized using a 2k factorial DoE approach to disperse BaTiO3 particles in PVDF solution in DMF. Scanning electron microscopy was used to characterize the microstructure of the fabricated mesh. The FT-IR and Raman analysis were carried out to investigate the crystal structure of the prepared mesh. Surface morphology contribution to the adhesive property of the composite was explained through contact angle measurements. The capacitance of the prepared PVDF- BaTiO3 nanofibrous mesh was a more than 40% increase over the pure PVDF nanofibers. A comparative study of dielectric relaxation, thermodynamics properties and impact analysis of electrospun polyvinylidene fluoride (PVDF) and 3% BaTiO3-PVDF nanofibrous composite are presented. The frequency dependent dielectric properties revealed micro structural features of the composite material. The dielectric relaxation behavior is further supported by complex impedance analysis and Nyquist plots. The temperature dependence of electric modulus shows Arrhenius type behavior. The observed non-Debye dielectric relaxation in electric loss modulus follows a thermally activated process which can be attributed to a small polaron hopping effect. The particle induced crystallization is supported with thermodynamic properties from differential scanning calorimetric (DSC) measurements. The observed increase in piezoelectric response by impact analysis was attributed to the interfacial interaction between PVDF and BaTiO3. The interfacial polarization between PVDF and BaTiO3 was studied using density functional theory calculations and atomic charge density analysis. The results obtained indicates that electrospinning offers a potential way to produce nanofibers with desired crystalline nature which was not observed in molded samples. In addition, BaTiO3 can be used to increase the capacitance, desired surface characteristics of the PVDF nanofibers which can find potential application as flexible piezoelectric sensor mimicking biological skin for use in impact sensing and energy harvesting applications.

This work presents a theoretical study of stability of phases in selected metallic systems. We propose a model of structural transformations in transition metal disilicides MoSi2, CrSi2, VSi2 and TiSi2 and in Pd thin films grown on cubic substrates W(001) and Nb(001). The obtained results yield the total energy proles for the structural transformations studied, the activation energies needed for each individual transformation and an estimate of the temperature at which the structure can transform. The total energies are calculated by full-potential linearized augmented plane waves (FLAPW) method incorporated in the WIEN2k code. Both generalized gradient approximation (GGA) and local density approximation (LDA) are employed for the exchange-correlation term. It turns out that temperatures corresponding to the activation energies of structural transformations in transition metal disilicides exceed their melting temperatures. Comparing the resulting total energy proles to those obtained by the semiempirical Bond Order interatomic potentials (BOP) substantially helps to adjust the fitting parameters of the BOPs. The estimated temperature of 168 K needed to transform the hcp structure of an innite Pd crystal into the dhcp structure explains the behavior of the Pd thin lm on W(001) and Nb(001) substrates. Pd lms deposited on W(001) substrate and thicker than about 100 monolayers undergo this transformation already at room temperature. Thinner lms need to be annealed at 400 K rst, due to their stronger interaction with the substrate. The difference between the computed result and a real temperature at which the hcp Pd lm transforms its structure to the dhcp can be explained by both the interaction between the lm and the substrate and by the inuence of the domain topology of the lm. Analyzing different models of transformation of the initial hcp Pd structure to the ground state fcc structure, we identied the optimum model that respects the domain topology of the Pd lm.

No description available.

<p>    </p> <p>Development on instrumentation and algorithms for chemical structure and chemical kinetics are discussed in this thesis. In Chapter 2 and 3, a consensus equilibrium formalism is introduced for the integration of multiple quantum chemical calculations of molecular and electronic structure. In multi-agent consensus equilibrium (MACE), iterative updates in structure optimization are intertwined with the net output, representing an equilibrium balance between multiple computational agents. MACE structure calculations from the integration of multiple low-level electronic structure calculations were compared favorably for small molecules, with results evaluated through comparison with higher level structure (CCSD). Notably, MACE results differed substantially from the average of the independent computational agent outputs, with MACE yielding improved agreement with higher-level CCSD calculations. The primary focus is on the development of the mathematical framework for implementing MACE for molecular and electronic structure determination, these initial preliminary results suggest potential promise for the use of MACE to improve the accuracy of low-level electronic structure calculations through the integration of multiple parallel methods. In Chapter 4 and 5, Fourier- transform fluorescence recovery after photobleaching (FT-FRAP) coupled with periodically comb pattern was demonstrated to monitor the controlled-release mechanisms of microparticles. By monitoring the time-lapse recovery patterns, spatial mobility were decoded in FT domain. Due to the nature of mobility encoded in FT domain, substantial improvements were demonstrated in terms of enhanced signal-to-noise, simplified mathematics, low requirements of sampling, and multiphoton compatibility to probe inside samples. FT-FRAP was able to discriminate and quantify the internal diffusion and exchange to higher mobility in fitting the recovery kinetics within microparticles. Theoretical modeling of exchange and diffusion- controlled release revealed that both RS and RL microparticles exhibited similar exchange decay, with RL having a much higher diffusion decay. The microscopically higher diffusion of RL microparticles is consistent with the dissolution performance of RL microparticles macroscopically. The distinction of controlled release mechanisms provided by FT-FRAP is important to understand and further optimize the design of controlled release systems for GI tract. </p>

This thesis presents a means of enhancing the iterative calculation techniques used in electronic structure calculations, particularly Kohn-Sham DFT. Based on the subspace iteration method of the FEAST eigenvalue solving algorithm, this nonlinear FEAST algorithm (NLFEAST) improves the convergence rate of traditional iterative methods and dramatically improves their robustness. A description of the algorithm is given, along with the results of numerical experiments that demonstrate its effectiveness and offer insight into the factors that determine how well it performs.

<p>Optimization and Machine learning (ML) have emerged as two positively disruptive methodologies and have thus resulted in unprecedented applications in several domains of technology. In recent years, ML has forayed into physical sciences and provided promising outcomes thanks to its ability in representing and generalizing complex functions to reveal underlying relations among variables describing a system. By casting ML as an optimization task, we first focus on its application in solving quantum many-body problems. Leveraging the power of quantum computation, we develop hybrid quantum machine learning protocols and implement benchmark tests to calculate the band structures of two-dimensional materials. We also show how this method can be used to estimate the critical point for a quantum phase transition. One  hurdle in such techniques is related to parameter optimization, wherein to obtain the desired result, the parameters have to be optimized, which can be computationally intensive. For a particular class of problem and a choice of algorithm, we deduce a simple parameter setting rule. This rule is projected as a heuristic and is validated numerically for several problem instances. Finally, by venturing into thermal photonics, a framework that takes advantage of the spectral and spatial information of hyperspectral thermal images to establish a completely passive machine perception, titled HADAR is presented. A conventional deep neural network is developed that utilizes the governing equation of HADAR and its performance in semantic segmentation is demonstrated. Altogether, this report establishes the need for creative algorithms that exploit modern hardware to solve complex problems that were previously deemed unsolvable.</p>

No description available.

<p>Thermoelectric (TE) materials are able to convert heat directly into electricity and vice versa. This special property makes them valuable for a variety of applications involving power generation and refrigeration. In the search for potential high-performance TE materials, a number of rare-earth (<em>RE</em>) antimonide suboxide phases have been investigated.This presentation will focus on two classes of rare-earth antimonide suboxides: the <em>RE</em>₃Sb₃O₃ and <em>RE</em>₈Sb_3-<em>_d</em>O₈ phases (<em>C</em>2/<em>m</em> space group) based on the <em>RE</em>–O frameworks and the <em>anti</em>-ThCr₂Si₂ type <em>RE</em>₂SbO₂ compounds (<em>I</em>4/<em>mmm</em> space group). The physical property measurements on the high-purity bulk samples revealed unexpected semiconducting properties in the non-charge-balanced systems, i.e.<em> RE</em>₈Sb_3-<em>_d</em>O₈ and <em>RE</em>₂SbO₂. Since the electronic structure calculations suggest that the anionic Sb states dominate the valence band at the vicinity of the Fermi level, the local structure of the Sb atomic site is believed to dictate the observed physical properties. The charge transport properties are explained within the framework of Anderson/Mott-type localizations. Ultimately, systematic investigation of the <em>RE</em>₂SbO₂ and Ho₂Sb_1-<em>x</em>Bi<em>_x</em>O₂ series reveal the large variability of the electrical properties caused by the local structural perturbations.</p> / Doctor of Philosophy (PhD)

First principle electronic structure theory is used to describe the effect of crystal binding on radiation detectors, electron transport properties, and structural damage induced by intense irradiation. A large database containing general electronic structure results to which data mining algorithms can be applied in the search for new functional materials, a case study is presented for scintillator detector materials. Inelastic cross sections for the generation of secondary electron cascades through impact ionization are derived from the dielectric response of an electron gas and evolved in time with Molecular Dynamics (MD). Qualitative and quantitive estimates are presented for the excitation and relaxation of a sample irradiated with Free Electron Laser pulses. A study is presented in where the structural damage on covalent bonded crystals following intense irradiation is derived from a Tight Binding approach and evolved in time with MD in where the evolution of the sample is derived from GW theory for the quasiparticle spectra and a dedicated Boltzmann transport equation for the impact ionization.