No description available.

Les propriétés d'électrons en deux dimensions (2D) soulèvent des questions fondamentales qui ont été largement explorées au moyen des techniques théoriques de la matière condensée. L'extension de modèles classiques tel le modèle de Hubbard en 2D, en incluant par exemple plusieurs bandes électroniques, ore la possibilité d'accéder à des phénomèmes plus complexes, comme l'interaction du transport électronique et du magnétisme observé dans les composés de fermions lourds. Ces modèles sont en lien direct avec la question de couches minces couplées, les hétérostructures, qui sont depuis peu l'objet d'intenses recherches et orent la possibilité d'intéressantes applications. Dans ce contexte, nous étudions numériquement diérents syst èmes au moyen de la méthode du Monte Carlo Quantique du Déterminant. Tout d'abord, l'eet de la corrélation électronique dans un isolant de bande est évaluée, montrant en particulier l'absence d'une phase métallique intermédiaire. Un deuxième système est composé de deux bandes électroniques couplées, dans lequel l'eet de la largeur de bande de la partie corrélée est exploré de façon systématique. Finalement, nous étudions une interface métal-isolant, qui présente une phase intermédiaire surprenante lorsque le couplage à l'interface est ajusté. / The properties of electrons in two dimensions (2D) raise fundamental questions that have been extensively explored by condensed matter theory. Extending standard frameworks such as the 2D Hubbard model by accounting for more than one electronic band oers the opportunity to access more complex phenomena, such as the interplay between transport and magnetism found in heavy-fermions materials. Such models are directly connected to the problem of coupled layers in complex materials known as heterostructures, which have been widely studied and synthesized in recent years, and are expected to lead to important applications. In that context, we study numerically several systems, by mean of the Determinant Quantum Monte Carlo Method (DQMC). We rst analyze the eect of electronic correlation in a band insulator, showing in particular the absence of an intermediate metallic phase. A second system consists of two coupled bands that modelize a heavy fermion model, in which the role of the bandwidth of the correlated band is systematically investigated. Finally, we consider the case of a metal-insulator interface, unveiling an intriguing intermediate phase as the interfacial coupling is tuned.

Cette thèse se concentre sur des aspects multiorbitales des systèmes fermioniques fortement corrélés. En particulier, sur l'existence d'une différentiation orbitale dans laquelle la coexistence de caractère itinérant et localisé peut être associée à différentes orbitales. Cette problématique est examinée dans le contexte des atomes froids et des matériaux, offrant un pont entre les deux communautés.Dans la première partie de la thèse, nous donnons un aperçu du problème des corrélations fortes dans les matériaux, et nous introduisons le concept de 'transition de Mott sélective en orbitales'. Nous fournissons également les principaux outils pour comprendre comment les matériaux peuvent être simulés avec des atomes froids, et nous présentons des résultats importants liés à la transition métal-isolant de Mott. Les aspects techniques, basées sur la théorie du champ moyen dynamique sont également discutés, et la solution de deux principaux modèles de systèmes fermioniques fortement corrélés, à savoir le modèle d'Hubbard (HM) et le modèle de Falicov-Kimball (FKM), sont passés en revue.Ensuite, nous étudions en détail la physique de deux espèces fermioniques en interaction forte avec des masses différentes dans un réseau optique. Nous établissons les différentes phases (avec et sans ordre à longue portée) en termes de la force des interactions (U), du rapport des masses et de la température (T), et aussi nous discutons les variables thermodynamiques, qui sont pertinentes pour les expériences d'atomes froids. Nous montrons que dans la phase métallique (U inférieure à une valeur critique) et avec un certain degré de différence de masses, un 'crossover' apparaît entre un état métallique du type de liquide de Fermi à basse T, et un état avec différentiation orbital à haute T, où les fermions lourds se localisent tandis que les fermions légers restent itinérant. Par conséquent, nous proposons ce modèle minimal pour étudier la physique des systèmes qui présentent une différentiation orbitale avec des expériences d'atomes froids.Basé sur les propriétés du modèle étudié, nous proposons la 'chromatographie entropique' comme une nouvelle méthode pour refroidir des atomes fermioniques dans les réseaux optiques. Nous discutons son efficacité et ses limites, et fournissons quelques idées afin de les surmonter.Dans la dernière partie de la thèse, nous généralisons le modèle précédent aux matériaux corrélés à plusieurs bandes qui permet d'afficher la différentiation orbitale. Nous montrons que l'état de Mott sélectif en orbital peut être stable sous les distorsions du réseau, modélisées par une hybridation locale entre les orbitales. Cependant, l'état de Mott est caractérisé par un pseudo-gap, où les fluctuations de charge sont brusquement réduites, mais l'état reste compressible. En relation au modèle précédent, nous discutons le 'crossover' entre l'état métallique et l'état sélectif induit par des effets température, nous comparons nos résultats avec les expériences de photoémission, et nous prédisons ce qui se passerait dans les matériaux qui présentent une hybridation locale entre les bandes. / This thesis focuses on multiorbital aspects of strongly correlated fermionic systems. In particular, it focuses on the existence of orbital differentiation in which coexistence of itinerant and localized character can be associated to different orbitals. This subject is discussed in the context of cold atoms and materials, providing a bridge between both communities.In the first part of the thesis, we give an insight into the problem of strong correlations in materials, and we introduce the concept of 'orbital-selective Mott transition'. We also provide the main tools to understand how materials can be simulated with cold atoms experiments, and we present important related results in the context of the metal-Mott insulator transition. The technical aspects, based on dynamical mean-field theory are also discussed, and the solution of two key models of strongly correlated fermionic systems, i.e., the Hubbard model (HM) and the Falicov-Kimball model (FKM), are reviewed.Then we study in detail the physics of two interacting fermionic species with different masses in an optical lattice. We establish the different phases (with and without long-range order) in terms of the interactions strength (U), mass ratio and temperature (T), and also discuss the thermodynamic variables, which are relevant in cold atoms experiments. We show that in the metallic phase (U below a critical value) and for some degree of mass imbalance, a crossover appears between a Fermi-liquid metallic state at low T, and an 'orbital-selective' state at higher T, where the heavy fermions effectively localize while the light species remain itinerant. Hence, we propose this minimal model for addressing orbital-selective physics with cold atoms experiments.Based on the properties of the studied model, we propose the 'entropic chromatography' as a new method for cooling fermionic atoms in optical lattices. We discuss its efficiency and limitations, and provide some ideas in order to overcome them.In the last part of the thesis we generalize the previous model to a model relevant for multiband correlated materials that can display orbital differentiation. We show that the orbital-selective Mott state can be stable under lattice distortions modeled by local hybridization between the orbitals. However, the Mott state is characterized by a pseudogap, where charge fluctuations abruptly reduce, but the state remains compressible. In connection with the previous model, we discuss the temperature-induced orbital-selective crossover in this problem, we compare our results with photoemission experiments, and predict what would happen in materials that display local hybridization between the bands.

Neste trabalho mostramos que uma grande variedade de modelos exatamente solúveis através do ansatz de Bethe coordenadas podem também ser resolvidos através de um ansatz do produto matricial. Estes modelos são descritos no caso unidimensional por cadeias quânticas, e por matrizes de transferência no caso de sistemas clássicos bi-dimensionais. Diferentemente do ansatz de Bethe, em que as auto-funções do modelo são escritas como uma combinação de ondas planas, no nosso ansatz do produto matricial elas são dadas por produtos de matrizes, onde as matrizes obedecem a uma álgebra associativa apropriada. Estas relações algébricas são obtidas impondo-se que as auto-funções escritas em termos do ansatz satisfaçam à equação de auto-valor do operador Hamiltoniano ou da matriz de transferência. A consistência das relações de comutatividade entre os elementos da álgebra implicam na exata integrabilidade do modelo. Além disso, o ansatz que propomos permite uma formulação simples e unificada para vários Hamiltonianos quânticos exatamente solúveis. Apresentamos nesta tese a formulação do nosso ansatz do produto matricial para uma grande família de redes quânticas, como os modelos anisotrópico de Heisenberg, Fateev-Zamolodchikov, Izergin-Korepin, Sutherland, t-J, Hubbard etc. Mais ainda, formulamos nosso ansatz para processos estocásticos de partículas com tamanhos e classes diferentes difundindo assimetricamente na rede. Por fim, com o objetivo de dar suporte a nossa conjectura de que todos os modelos exatamente solúveis através do ansatz de Bethe coordenadas, associados a Hamiltonianos quânticos unidimensionais ou matrizes de transferência bidimensionais, também podem ser resolvidos através de um ansatz do produto matricial, apresentamos a formulação do nosso ansatz para a matriz de transferência do modelo de seis-vértices com condição de contorno toroidal / In this work we show that a large family of exactly solved models through the coordinate Bethe ansatz can also be solved through a matrix product ansatz. The models are described in the one dimensional case by quantum Hamiltonians, and by transfer matrices in the case of two dimensional classical models. Differently from the Bethe ansatz, where the model\'s eigenfunctions are described by a plane wave combination, in our matrix product ansatz they are given by a matrix product, where the matrices obey a suitable associative algebra. Theses algebraic relations are obtained by imposing that the eigenfunctions described in terms of the ansatz satisfy the eigenvalue equation for the associated Hamiltonian or transfer matrix. The consistency of the commutativity relations among the elements of the algebra implies the exact integrability of the model. Furthermore, the matrix product ansatz we propose allows an unified and simple formulation for the solution of several exact integrable quantum Hamiltonians. We present on this thesis the formulation of our matrix product ansatz for a huge family of quantum chains such as the anisotropic Heisenberg model, Fateev-Zarnolodchikov model, Izergin-Korepin model, Sutherland model, t- J model, Hubbard model, etc. Moreover, we formulated our ansatz for stochastic process of particles with different sizes and classes diffusing asymmetrically on the lattice. Finally, in order to support our conjecture that all exactly solved models through the coordinate Bethe ansatz, associated to unidimensional quantum Hamiltonians or two-dimensional transfer matrices, can also be solved through a matrix product ansatz, we present the formulation of our ansatz, for the transfer matrix of the six-vertex model with toroidal boundary condition

Nesta dissertação estudaremos sistemas de bósons ultrafrios armadilhados em uma rede ótica quadrada bidimensional sem levar em consideração o confinamento harmônico. A dinâmica desses sistemas é bem descrita pelo modelo de Bose-Hubbard, que prevê uma transição de fase quântica de um superfluido para um isolante de Mott a temperaturas baixas, e pode ser induzida variando a profundidade do potencial da rede ótica. Apresentaremos o diagrama de fases dessa transição construído a partir de uma aproximação de campo médio e também com um cálculo numérico usando um algoritmo de Monte Carlo Quântico, denominado algoritmo Worm. Encontramos o ponto crítico para o primeiro lobo de Mott em ambos os casos, concordando com trabalhos anteriores. / This work study the two-dimensional ultracold bosonic atoms loaded in a square optical lattice, without harmonic confinement. The dynamics of this system is described by the Bose-Hubbard model, which predicts a quantum phase transition from a superfluid to a Mott-insulator at low temperatures that can be induced by varying the depth of the optical potential. We present here the phase diagram of this transition built from a mean field approach and from a numerical calculation using a Quantum Monte Carlo algorithm, namely the Worm algorithm. We found the critical transition point for the first Mott lobe in both cases, in agreement with the standard literature.

Die vorliegende Arbeit beschäftigt sich mit der mikroskopischen Modellierung von strominduziertem Schalten der Magnetisierung in magnetischen Tunnelstrukturen. Die Tunnelstruktur besteht aus zwei durch einen nichtmagnetischen Isolator voneinander getrennten Ferromagneten und einem Paramagneten, der als Elektronenreservoir dient. Die Ferromagnete werden beide durch das Hubbard-Modell beschrieben. Durch Anlegen einer Spannung verschieben sich die chemischen Potentiale auf beiden Seiten des Isolators, wodurch ein endlicher Tunnelstrom entsteht. Dieser wird im Rahmen des Modells durch eine Hybridisierung zwischen benachbarten Schichten simuliert. Das Modell muss im Nichtgleichgewicht gelöst werden, da aufgrund der unterschiedlichen chemischen Potentiale thermodynamisches Gleichgewicht nicht angenommen werden darf. Daher wird zur analytischen Auswertung der Keldysh-Formalismus verwendet, der eine Erweiterung der Viel-Teilchen-Theorie ins Nichtgleichgewicht darstellt. Da es sich beim Hubbard-Modell um ein nicht exakt lösbares Viel-Teilchen-Modell handelt, wurde in der Arbeit eine approximative Lösung, der sogenannte Nichtgleichgewichtsspektraldichteansatz, entwickelt. Dieser beruht auf einer Hochenergieentwicklung der retardierten Greenfunktion mit Hilfe der exakt berechenbaren Spektralmomente. Die numerischen Resultate stimmen qualitativ mit dem Experiment überein. Insbesondere gelingt es, das Hystereseverhalten der Magnetisierung des freien Ferromagneten in Abhängigkeit der angelegten Spannung korrekt zu reproduzieren. Es kann somit allein durch Anlegen einer Spannung kontrolliert zwischen paralleler und antiparalleler Ausrichtung der Magnetisierungen geschaltet werden. Dieses Phänomen ist anhand der entsprechenden Quasiteilchenzustandsdichten erklärbar. Weiterhin wird das Verhalten der kritischen Spannung systematisch in Form von Phasendiagrammen dargestellt und diskutiert. / This thesis is concerned with the microscopic modelling of current-induced switching of magnetization in magnetic tunnel junctions. The tunnel junction consists of two ferromagnets which are divided by a nonmagnetic insulator and a paramagnet, which acts as an electron reservoir. The ferromagnets are both described by the Hubbard model. By applying a voltage the chemical potentials on both sides of the insulator are shifted which results in a finite tunneling current. Within the model the current is simulated by a hybridization between neighbouring regions. The model has to be solved in non-equilibrium since thermal equilibrium requires a constant chemical potential for the whole system, which is not the case due to the voltage. Thus the Keldysh formalism will be used for evaluating the model. Since the Hubbard model is not exactly solvable one needs approximations. In this work a non-equilbrium spectral density approach is developed. It is based on a high-energy expansion of the retarded Green''s function and takes interactions beyond the mean field level into account. The numerical results of the theory are in qualitative agreement with experiments. It will be shown that it is possible to correctly get the hysterisis behaviour of the magnetization of the free ferromagnet in dependence on the applied voltage. Thus the relative alignment of the two magnetizations can be switched just by applying an electric field. This can be explained with the corresponding quasiparticle densities of state. Furthermore the behaviour of the critical voltage will be discussed systematically by calculating phase diagrams of the tunnel junction.

No description available.

Für die theoretische Beschreibung verschiedener Substanzen, so z.B. für diverse Kuprate die Anwendungen als Hochtemperatur-Supraleiter finden, spielt das t-J-Modell eine wichtige Rolle. In vielen Fällen kann man Abweichungen der Verbindungen vom idealen translationsinvarianten Festkörper vernachlässigen, für bestimmte Eigenschaften ist jedoch der Einfluß von Störstellen,z.B. Dotieratomen, bedeutsam. Um solche Verunreinigungen einzubeziehen, behandelt die vorliegende Arbeit das t-J-Modell mit einer zusätzlichen on-site-Energie mit über die Gitterpläte zufallsverteilten Werten. Um für dieses Modell die Einteilchen-Greensfunktion zu bestimmen, wird ein Verfahren entwickelt, welches auf der Projektionstechnik basiert und die Einbeziehung des Unordnungsterms ermöglicht. Die notwendige Mittelung über die möglichen Unordnungskonfigurationen erfolgt näherungsweise durch Faktorisierung und ist verwandt mit der sogenannten average T-matrix approximation, wird hier jedoch auf ein stark korreliertes System erweitert. Zur Illustration wird der Grundzustand von La2−xSrxCuO4 und Nd2−xCexCuO4 bei einem zusätzlichen Ladungsträger über Halbfüllung untersucht. Wie Bandstrukturrechnungen zeigen, ruft die Dotierung der elektronendotierten Substanz gerade einen solchen Zufallsterm hervor. Dies wurde in der bisherigen Literatur meist vernachlässigt. Bei der Übertragung der Bandstrukturergebnisse in die Modellparameter des t-t′-t′′-J-Modells zeigt sich, daß der Einfluß der Dotieratome bei La2−xSrxCuO4 um etwa eine Größenordnung geringer ist als in Nd2−xCexCuO4 . Als wichtige Ursache hierfür wird der Einfluß der Apex-Sauerstoffatome angesehen, die im Fall von La2−xSrxCuO4 die Seltenerd- Dotieratome gegenüber der Kupferoxidebene abschirmen. Für das mit diesen Parametern belegte Modell wird anschließend die Einteilchen- Greensfunktion berechnet, die Ausgangspunkt der Berechnung verschiedener Observablen ist. Die für die elektronendotierte Substanz auftretende lokale Mode gibt zu dem Vorschlag Anlaß, daß die unterschiedliche Stabilität der antiferromagnetischen Phase für die beiden betrachteten Substanzen nicht nur auf die Art der Ladungsträger zurückzuführen ist, sondern auch auf die Struktur der Elementarzelle. / The t-J-Modell can be applied to several classes of materials, e.g. high-Tc cuprate superconductors. Often translational invariance can be assumed, but sometimes it is necessary to take into account the effects of the doping atoms at randomly distributed sites. Therefore a t-J-Modell with an additional randomly distributed on-site energy is investigated. To calculate the one-particle Green’s function considering this term of disorder, a method is developed which bases on projection technique. The average over the possible configurations of the dopand atoms is approximated by factorization and is similar to the so-called average T-matrix approximation. Here it is extended to a model with strong correlations. In order to illustrate the methode the single-particle ground state of La2−xSrxCuO4 and Nd2−xCexCuO4 is analyzed. Band-structure calculations exhibit that for the electron-doped case the doping atoms (in first approximation) induce indeed a term of disordered on-site energies. The transformation of the values of this energies at the copper sites into the parameters in the t − t′ − t′′ − J-model shows that the influence of doping in La2−xSrxCuO4 is by about an order of magnitude smaller than in Nd2−xCexCuO4 . The existence of apex oxygen atoms between the rare-earth plane and the copper-oxygen plane in La2−xSrxCuO4 is one important reason for that effect. The single-particle Greens function for the t-t′-t′′-J-model with these parameters is calculated. A local mode appears in the electron-doped case, which suggests that the differences of the stability of the antiferromagnetic phases in both compounds are not only due to the type of charge carriers but also due to the structure of the unit cell.

This thesis analyzes charge properties of (undoped) cuprate compounds from a theoretical point of view. The central question considered here is: How does the dimensionality of the CU-O sub-structure influence its charge degrees of freedom? The model used to describe the Cu-O sub-structure is the three- (or multi-) band Hubbard model. Analytical approaches are employed (ground-state formalism for strongly correlated systems, Mori-Zwanzig projection technique) as well as numerical simulations (Projector Quantum Monte Carlo, exact diagonalization). Several results are compared to experimental data. The following materials have been chosen as candidates to represent different Cu-O sub-structures: Bi2CuO4 (isolated CuO4 plaquettes), Li2CuO2 (chains of edge-sharing plaquettes), Sr2CuO3 (chains of corner-sharing plaquettes), and Sr2CuO2Cl2 (planes of plaquettes). Several results presented in this thesis are valid for other cuprates as well. Two different aspects of charge properties are analyzed: 1) Charge properties of the ground state 2) Charge excitations. (gekürzte Fassung)

The two-dimensional electronic systems (2DESs) have kept surprising physicists for the last few decades. Examples include the integer and fractional quantum Hall effects, cuprate superconductivity, and graphene. This thesis is intended to develop suitable theoretical tools which can be generalized to study new types of 2DESs with strong correlation feature. The first part of this thesis describes the investigation of heterostructures made by Mott insulators. This work is mostly motivated by the significant improvement of techniques for layer-by-layer growth of transition metal oxides in the last few years. We construct a toy model based on generalized Hubbard model complemented with long-ranged Coulomb interaction, and we study it by Hartree-Fock theory, dynamical mean-field theory, and Thomas-Fermi theory. We argue that interesting 2D strongly correlated electronic systems can be created in such heterostructures under several conditions. Since these 2D systems are formed entirely due to the gap generated by electron-electron interaction, they are not addiabatically connected to a noninteracting electron states. This feature makes these 2D systems distinguish from the ones created in semiconductor heterostructures, and they may be potential systems having non-Fermi liquid behaviors. The second part of this thesis is devoted to the study of collective excitations in high-temperature superconductors. One important achievement in this work is to develop a time-dependent mean-field theory for t-U-J-V model, an effective low energy model for cuprates. The time-dependent mean-field theory is proven to be identical to the generalized random-phase approximation (GRPA) which includes both the bubble and ladder diagrams. We propose that the famous 41 meV magnetic resonance mode observed in the inelastic neutron scattering measurements is a collective mode arising from a conjugation relation, which has been overlooked in previous work, between the antiferromagnetic fluctuation and the phase fluctuation of the d-wave superconducting order parameter near momentum ([pi, pi]). Furthermore, we find that this collective mode signals the strength of the antiferromagnetic fluctuations which are responsible for the suppression of the superfluid density in the underdoped cuprates even at zero temperature. Finally, we perform a complete analysis on an effective model with parameters fitted by experimental data of Bi2212 within the GRPA scheme and conclude that the short-range antiferromagnetic interactions which are a remnant of the parent Mott-insulator are more likely the pairing mechanism of the High-T[subscript c] cuprates. / text