Thesis advisor: Ziqiang Wang / One of the most unconventional electronic states in high transition temperature cuprate superconductors is the pseudogap state. In the temperature versus doping phase diagram, the pseudogap state straddles across the antiferromagnetic (AF) state near half filling and the superconducting (SC) dome on the hole doped side above the transition temperature Tc. The relationship between the pseudogap state and these two well known states - the AF state and the SC state is believed to be very important for understanding superconductivity and the emergent quantum electronic matter in doped Mott insulators. The pseudogap is characterized by the emergence of a soft gap in the single-particle excitation spectrum in the normal state in the temperature range between Tc and a characteristic temperature T*, i.e. Tc < T < T*. The most puzzling feature of the pseudogap is the nodal-antinodal dichotomy. Observed by ARPES in momentum space, the Fermi surface is gapped out in the antinodal region leaving a Fermi arc of gapless excitations near the nodes. Whether the pseudogap is an incoherent superconducting gap (onegap scenario) or it is a different gap governed by other mechanisms, other than superconductivity, (two-gap scenario) is still under debate. In this thesis I study the particle-particle channel and the particle-hole channel of the valence bond fluctuations away from half filling. Based on a strong-coupling analysis of the t-J model, I argue that the superexchange interaction J induced incommensurate bond centered density wave order is the driving mechanism for the pseudogap state. Low energy density of states (DOS) are eliminated by multiple incommensurate scatterings in the antinodal region at the Fermi level. I show that the interplay between the incommensurate bond centered d-wave density wave instability and the intrinsic electronic inhomogeneity in real cuprate materials is responsible for the observed pseudogap phenomena. Utilizing the spatially unrestricted Gutzwiller approximation, I show that the off-stoichiometric doping induced electrostatic disorder pins the low-energy d-wave bond density fluctuations, resulting in a VBG phase. The antinodal Fermi surface (FS) sections are gapped out, giving rise to a genuine normal state Fermi arc. The length of the Fermi arc shrinks with underdoping below the temperature T* determined by thermal filling of the antinodal pseudogap. Below Tc, the d-wave superconducting gap due to singlet pairing coexists and competes with the VBG pseudogap. The spatial, momentum, temperature and doping dependence of these two gaps are consistent with recent ARPES and STM observations in underdoped and chemically substituted cuprates. The temperature versus doping phase diagram captures the salient properties of the pseudogap phenomena and provides theoretical support for the two-gap scenario. In addition to resolving the complexities of the quantum electronic states in hole-doped cuprates, my unified theory elucidates the important role of the interplay between the strong electronic correlation and the intrinsic electronic disorder in doped transition metal oxides. / Thesis (PhD) — Boston College, 2011. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Physics.

Thesis advisor: Ziqiang Wang / Gutzwiller wave function is an important theoretical technique for treating local electron-electron correlations nonperturbatively in condensed matter and materials physics. It is concerned with calculating variationally the ground state wave function by projecting out multi-occupation configurations that are energetically costly. The projection can be carried out analytically in the Gutzwiller approximation that offers an approximate way of calculating expectation values in the Gutzwiller projected wave function. This approach has proven to be very successful in strongly correlated systems such as the high temperature cuprate superconductors, the sodium cobaltates, and the heavy fermion compounds. In recent years, it has become increasingly evident that strongly correlated systems have a strong propensity towards forming inhomogeneous electronic states with spatially periodic superstrutural modulations. A good example is the commonly observed stripes and checkerboard states in high-$T_\mathrm c$ superconductors under a variety of conditions where superconductivity is weakened. There exists currently a real challenge and demand for new theoretical ideas and approaches that treats strongly correlated inhomogeneous electronic states, which is the subject matter of this thesis. This thesis contains four parts. In the first part of the thesis, the Gutzwiller approach is formulated in the grand canonical ensemble where, for the first time, a spatially (and spin) unrestricted Gutzwiller approximation (SUGA) is developed for studying inhomogeneous (both ordered and disordered) quantum electronic states in strongly correlated electron systems. The second part of the thesis applies the SUGA to the $t$-$J$ model for doped Mott insulators which led to the discovery of checkerboard-like inhomogeneous electronic states competing with $d$-wave superconductivity, consistent with experimental observations made on several families of high-$T_{\mathrm c}$ superconductors. In the third part of the thesis, new concepts and techniques are developed to study the Mott transition in inhomogeneous electronic superstructures. The latter is termed ``SuperMottness'' which is shown to be a general framework that unifies the two paradigms in the physics of strong electronic correlation: Mott transition and Wigner crystallization. A cluster Gutzwiller approximation (CGA) approach is developed that treats the local ($U$) and extended Coulomb interactions ($V$) on equal footing. It is shown with explicit calculations that the Mott-Wigner metal-insulator transition can take place far away from half-filling. The mechanism by which a superlattice potential enhances the correlation effects and the tendency towards local moment formation is investigated and the results reveal a deeper connection among the strongly correlated inhomogeneous electronic states, the Wigner-Mott physics, and the multiorbital Mott physics that can all be united under the notion of SuperMottness. It is proposed that doping into a superMott insulator can lead to coexistence of local moment and itinerant carriers. The last part of the thesis studies the possible Kondo effect that couples the local moment and the itinerant carriers. In connection to the sodium rich phases of the cobaltates, a new Kondo lattice model is proposed where the itinerant carriers form a Stoner ferromagnet. The competition between the Kondo screening and the Stoner ferromagnetism is investigated when the conduction band is both at and away from half-filling. / Thesis (PhD) — Boston College, 2009. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Physics.

Cette thèse vise à utiliser le t-J Hamiltonian de la corrélation forte pour mieux comprendre la micro-fonctionnalité des scénarios de matériau condensé. Un des problèmes qui existe depuis longtemps est que pour ce type de modèle comme Hubbard Hamiltonian ou t-J Hamiltonian avec une corrélation forte ne peut pas être résolu complètement analytiquement. Par conséquent, quand on aborde ces modèles, il est important de les exploiter de façon numérique. Dans cette thése, nous utiliserons la manière qui s'appelle "Renormalized Mean-Field Theory"(RMFT) pour le t-J Hamiltonian. Grâce à M. Gutzwiller, ce que nous devons faire est simplement de chiffrer des paramètres qui incluent l'influence de la corrélation électronique et de les mettre avant chaque partie du Hamiltonian. Après ce calcul, nous calculerons l'Hamiltonian du champ moyen de manière standard. Ceci sera notre façon principale pour aborder des questions physiques. Ensuite, nous l'appliquerons sur deux systèmes. Le premier est la mystique de supraconducteur à haute température. Après sa découverte il y a 30 ans, on ne peut pas encore définir une théorie pour expliquer sa micromécanique de manière appropriée. Cependant, avec des équipements avancés, on peut faire des expériences correctement et obtenir des résultats exacts. Ces preuves nous facilitent l'élaboration d'une bonne théorie, même s'il est aussi très difficile d'inclure tous les phénomènes ensemble. Nous avons obtenu des résultats et par rapport aux expériences, ils sont similaires qualitativement. Nous montrerons les détails dans le texte. Le deuxième système qui nous intéresse est le mouvement d'électron dans un champ magnétique fort. Le papillon d'Hofstadter et son modèle, l'Hamiltonian de Harper-Hofstadter ont obtenu un grand succès à décrire la mécanique d'électrons libres aux treillis. Donc il est ainsi intéressant de se demander ce qu'il se passera si nous remplaçons des électrons libres avec ceux qui s'interagissent. D'ailleurs, t-J Hamiltonian s'utilise comme bon modèle à le découvrir. Nous allons comparer nos résultats avec ceux de la diagonalisation exacte. Nous proposerons des découvertes intéressantes qui désormais seront réalisées par l'expérience d'atome froide. / This thesis is aiming in utilizing the strongly correlated t-J Hamiltonian for better understanding the microscopic pictures of certain condensed matter scenario. One of the long existing issues in the Hubbard model and its extreme version, t-J model, lies in the fact that there is not an analytical way of solving them. Therefore, when dealing with these models, numerical approaches become very crucial. In this thesis, we will present one of the methods called renormalized mean-field theory (RMFT) and exploit it upon the t-J model. Thanks to the concept proposed by Gutzwiller, all we have to do is to try to include the correlation of electrons, which is mainly the most difficult part, with several renormalization factors. After obtaining the correct form of these factors, we can apply the routine mean-field theory in solving for the Hamiltonian, which is the principle methodology throughout this thesis. Next, the physical systems that we are interested in consist of two parts. The mystery of High-Tc superconductivity comes first. After 30 years of its discovery, people still cannot settle down a complete microscopic theory in describing this exotic phenomenon. However, with more and more experimental equipment with higher accuracy nowadays, lots of behavior of copperoxide superconductor (also known as cuprate) have been revealed. Those discoveries can definitely help us better understand its microscopic mechanism. Therefore, from the theoretical side, to compare the calculated data with experiments leads us to know whether our theory is on the right track or not. We have produced tons of data and made a decent comparison which will be shown in the main text. The second system we are curious about is the mechanism of electrons under magnetic field. The Hofstadter butterfly along with its Hamiltonian, the Harper-Hofstadter model has achieved great success in describing free electrons' movement with lattice present. Thus, it will be also interesting to ask the question: what will happen if the electrons are correlated. Our RMFT for t-J Hamiltonian, by adding an additional phase in the hopping term, happens to serve as a great preliminary model for answering this question. We will compare the results of ours with our collaborators, who solved this model by a different approach, the exact diagonalization(ED). Together with our calculations, we proposed several discoveries which might be realized by the cold atom experiments in the future.

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

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.

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

No description available.

In the last decade, a large number of studies have been devoted to the peculiarities of correlated physics found in the quasi-two-dimensional square lattice iridium oxides. It was shown that this 5d family of transition metal oxides has strong structural and electronic similarities to the famous 3d family of copper oxides. Moreover, a delicate interplay of on-site spin-orbit coupling, Coulomb repulsion and crystalline electric field interactions is expected to drive various exotic quantum states. Many theoretical proposals were made in the last decade including the prediction of possible superconductivity in square-lattice iridates emerging as a sister system to high-Tc cuprates, which however met only limited experimental confirmation. One can, therefore, raise a general question: To what extent is the low-energy physics of the quasi-two-dimensional square-lattice iridium oxides different from other transition metal oxides including cuprates? In this thesis we investigate some of the effects which are usually neglected in studies on iridates, focusing on quasi-two-dimensional square-lattice iridates such as Sr2IrO4 or Ba2IrO4. In particular, we discuss the role of the electron-phonon coupling in the form of Jahn-Teller interaction, electron-hole asymmetry introduced by the strong correlations and some effects of coupling scheme chosen to calculate multiplet structure for materials with strong on-site spin-orbit coupling. Thus, firstly, we study the role of phonons, which is almost always neglected in Sr2IrO4, and discuss the manifestation of Jahn-Teller effect in the recent data obtained on Sr2IrO4 with the help of resonant inelastic x-ray scattering. When strong spin-orbit coupling removes orbital degeneracy, it would at the same time appear to render the Jahn-Teller mechanism ineffective. We show that, while the Jahn-Teller effect does indeed not affect the antiferromagnetically ordered ground state, it leads to distinctive signatures in the spin-orbit exciton. Second, we focus on charge excitations and determine the motion of a charge (hole or electron) added to the Mott insulating, antiferromagnetic ground-state of square-lattice iridates. We show that correlation effects, calculated within the self-consistent Born approximation, render the hole and electron case very different. An added electron forms a spin-polaron, which closely resembles the well-known cuprates, but the situation of a removed electron is far more complex. Many-body configurations form that can be either singlets and triplets, which strongly affects the hole motion. This not only has important ramifications for the interpretation of angle-resolved photoemission spectroscopy and inverse photoemission spectroscopy experiments of square lattice iridates, but also demonstrates that the correlation physics in electron- and hole-doped iridates is fundamentally different. We then discuss the application of this model to the calculation of scanning tunneling spectroscopy data. We show that using scanning tunneling spectroscopy one can directly probe the quasiparticle excitations in Sr2IrO4: ladder spectrum on the positive bias side and multiplet structure of the polaron on the negative bias side. We discuss in detail the ladder spectrum and show its relevance for Sr2IrO4 which is in general described by more complicated extended t-J -like model. Theoretical calculation reveals that on the negative bias side the internal degree of freedom of the charge excitation introduces strong dispersive hopping channels encaving ladder-like features. Finally, we discuss how the choice of the coupling scheme to calculate multiplet structure can affect the theoretical calculation of angle-resolved photoemission spectroscopy and scanning tunnelling spectroscopy spectral functions.

In the last decade, a large number of studies have been devoted to the peculiarities of correlated physics found in the quasi-two-dimensional square lattice iridium oxides. It was shown that this 5d family of transition metal oxides has strong structural and electronic similarities to the famous 3d family of copper oxides. Moreover, a delicate interplay of on-site spin-orbit coupling, Coulomb repulsion and crystalline electric field interactions is expected to drive various exotic quantum states. Many theoretical proposals were made in the last decade including the prediction of possible superconductivity in square-lattice iridates emerging as a sister system to high-Tc cuprates, which however met only limited experimental confirmation. One can, therefore, raise a general question: To what extent is the low-energy physics of the quasi-two-dimensional square-lattice iridium oxides different from other transition metal oxides including cuprates? In this thesis we investigate some of the effects which are usually neglected in studies on iridates, focusing on quasi-two-dimensional square-lattice iridates such as Sr2IrO4 or Ba2IrO4. In particular, we discuss the role of the electron-phonon coupling in the form of Jahn-Teller interaction, electron-hole asymmetry introduced by the strong correlations and some effects of coupling scheme chosen to calculate multiplet structure for materials with strong on-site spin-orbit coupling. Thus, firstly, we study the role of phonons, which is almost always neglected in Sr2IrO4, and discuss the manifestation of Jahn-Teller effect in the recent data obtained on Sr2IrO4 with the help of resonant inelastic x-ray scattering. When strong spin-orbit coupling removes orbital degeneracy, it would at the same time appear to render the Jahn-Teller mechanism ineffective. We show that, while the Jahn-Teller effect does indeed not affect the antiferromagnetically ordered ground state, it leads to distinctive signatures in the spin-orbit exciton. Second, we focus on charge excitations and determine the motion of a charge (hole or electron) added to the Mott insulating, antiferromagnetic ground-state of square-lattice iridates. We show that correlation effects, calculated within the self-consistent Born approximation, render the hole and electron case very different. An added electron forms a spin-polaron, which closely resembles the well-known cuprates, but the situation of a removed electron is far more complex. Many-body configurations form that can be either singlets and triplets, which strongly affects the hole motion. This not only has important ramifications for the interpretation of angle-resolved photoemission spectroscopy and inverse photoemission spectroscopy experiments of square lattice iridates, but also demonstrates that the correlation physics in electron- and hole-doped iridates is fundamentally different. We then discuss the application of this model to the calculation of scanning tunneling spectroscopy data. We show that using scanning tunneling spectroscopy one can directly probe the quasiparticle excitations in Sr2IrO4: ladder spectrum on the positive bias side and multiplet structure of the polaron on the negative bias side. We discuss in detail the ladder spectrum and show its relevance for Sr2IrO4 which is in general described by more complicated extended t-J -like model. Theoretical calculation reveals that on the negative bias side the internal degree of freedom of the charge excitation introduces strong dispersive hopping channels encaving ladder-like features. Finally, we discuss how the choice of the coupling scheme to calculate multiplet structure can affect the theoretical calculation of angle-resolved photoemission spectroscopy and scanning tunnelling spectroscopy spectral functions.

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.