The single impurity Anderson model (SIAM) is one of the most recurrent model Hamiltonians in condensed matter physics, playing an important role in the description of systems where electron-electron interactions can not be neglected, such as strongly interacting quantum dots and doped metals with magnetic impurities. In order to address the variety of scenarios in which the model can be applied many impurity solvers were developed for the SIAM. It is consensus that there is no universal method that provides reliable results for the model parameter space of experimental interest. The main methods
found in the literature di ers signi cantly on the computational cost involved in the numerical implementation and in the range of the model parameters where the used method is reliable. In this undergraduate dissertation we study the reliability of some impurity solvers for the SIAM, namely, the equations of motion (EOM) method, the numerical renormalization group (NRG), the non-crossing approximation (NCA), and the one crossing approximation (OCA). We study the Green's function of a quantum dot using the aforementioned impurity solvers and then calculate the density of states
as a function of the temperature and the chemical potential of the quantum dot . Our analysis is focused on the Coulomb blockade and the Kondo physics. We report that some standard approximations based on the EOM technique display a rather unexpected poor behaviour in the Coulomb blockade regime even at high temperatures.
In the present work, we calculate the conductance through a single quantum dot weakly coupled to metallic contacts. We use the spin-1/2 Anderson model to describe the quantum dot, while considering a finite Coulomb repulsion. We solve the interacting system using the noncrossing approximation (NCA) and the one-crossing approximation (OCA). We obtain the linear response conductance as a function of temperature and energy position of the localized level. From the comparison of both approximations we extract the role of the vertex corrections, which are introduced in the OCA calculations and neglected in the NCA scheme. As a function of the energy position, we observe that the diagrams omitted within the NCA are really important for appropriately describing transport phenomena in Kondo systems as well as in the mixed valence regime. On the other hand, as a function of temperature, the corrections introduced by the OCA partly recover the universal scaling properties known from numerical approaches such as the numerical renormalization group.
In this thesis we deal with the description of quantum impurity systems. In order to achieve such an aim, we exploit in the first part of this work the Reduced Density Matrix (RDM) formalism, in the second one the Density Matrix-Numerical Renormalization Group (DM-NRG) method. The first half of the thesis deals with the derivation and the application of out-of-equilibrium and non-perturbative approximations for the dynamics of the Single Impurity Anderson Model (SIAM) in the perturbative, intermediate coupling and Kondo cross-over regimes. In this part of the thesis, the main aim is to develop new technical tools useful to extend the RDM formalism from the well established perturbative regime to the intermediate and, in the future, to the strong coupling regimes. In this part, we analyse the advantages and drawbacks of the so-called Dressed Second Order (DSO), Dressed DSO (DDSO) and the Resonant Tunnelling Approximation (RTA). In the DSO we include processes capable to describe the broadening of the Coulomb blockade peaks but the DSO itself, as the RTA, is not capable to describe the Kondo resonance width properly. In the DDSO we propose a prescription to solve this issue and to describe properly the exponential dependence of the Kondo resonance width. However, in the DDSO the vertex corrections are not included and this yields a spurious conductance peak in the empty orbital regime (Coulomb valley with zero electrons). For this reason we turn our attention to the inclusion of the vertex corrections within the RTA framework. In the second part, we describe transport and thermodynamic properties of carbon-nanotubes (CNTs) based quantum dots using the DM-NRG method. We present results on the spectral function and the linear conductance in cooperation with Catalin Pascu Moca and Gergely Zarand from the Universities of Oradea and Budapest, respectively. In order to model the CNTs we consider an extended Anderson Hamiltonian. Exploiting such an Hamiltonian we study the crossover from the SU(4) to the SU(2) Kondo effect. We demonstrate numerically and analytically that the Kondo temperature scales inversely with respect to the SU(4)-symmetry breaking parameter
Δ, tending to the Kondo temperature of the SU(2) Kondo effect in the large
Δ limit. Furthermore we study the CNT quantum dot in the presence of a parallel and perpendicular (with respect to the CNT axis) magnetic field. We find necessary conditions to observe the so-called ``Kondo revivals'' for specific values of the magnetic field. In the final part of this thesis, we model a real world device realised by Jean-Pierre Cleuziou, Ngoc-Viet Nguyen and Wolfgang Wernsdorfer in Grenoble by means of the previously studied Hamiltonian. For this system we search for the best set of parameters to reproduce the experimentally measured linear conductance and we study the underlying Kondo state evaluating the CNT added entropy and specific heat.
We present a detailed comparison of three different methods designed to tackle non-equilibrium quantum transport, namely the functional renormalization group (fRG), the time-dependent density matrix renormalization group (tDMRG) and the iterative summation of real-time path integrals (ISPI). For the non-equilibrium single-impurity Anderson model (including a Zeeman term at the impurity site), we demonstrate that the three methods are in quantitative agreement over a wide range of parameters at the particle–hole symmetric point as well as in the mixed-valence regime. We further compare these techniques with two quantum Monte Carlo approaches and the time-dependent numerical renormalization group method.
Using different techniques, and Fermi-liquid relationships, we calculate the variation with the applied magnetic field (up to second order) of the zero-temperature equilibrium conductance through a quantum dot described by the impurity Anderson model. We focus on the strong-coupling limit $U\ensuremath{\gg}\mathrm{\ensuremath{\Delta}}$, where
U is the Coulomb repulsion and $\mathrm{\ensuremath{\Delta}}$ is half the resonant-level width, and consider several values of the dot level energy
Ed, ranging from the Kondo regime ${\ensuremath{\epsilon}}_{F}\ensuremath{-}{E}_{d}\ensuremath{\gg}\mathrm{\ensuremath{\Delta}}$ to the intermediate-valence regime ${\ensuremath{\epsilon}}_{F}\ensuremath{-}{E}_{d}\ensuremath{\sim}\mathrm{\ensuremath{\Delta}}$, where ${\ensuremath{\epsilon}}_{F}$ is the Fermi energy. We have mainly used the density-matrix renormalization group (DMRG) and the numerical renormalization group (NRG) combined with renormalized perturbation theory (RPT). Results for the dot occupancy and magnetic susceptibility from the DMRG and
NRG+RPT are compared with the corresponding Bethe ansatz results for $U\ensuremath{\rightarrow}\ensuremath{\infty}$, showing an excellent agreement once
Ed is renormalized by a constant Haldane shift. For $U<3\mathrm{\ensuremath{\Delta}}$ a simple perturbative approach in
U agrees very well with the other methods. The conductance decreases with the applied magnetic field for dot occupancies ${n}_{d}\ensuremath{\sim}1$ and increases for ${n}_{d}\ensuremath{\sim}0.5$ or ${n}_{d}\ensuremath{\sim}1.5$ regardless of the value of
U. We also relate the energy scale for the magnetic-field dependence of the conductance with the width of the low-energy peak in the spectral density of the dot.
We generalize the hierarchical equations of motion method to study electron transport through a quantum dot or molecule coupled to one-dimensional interacting leads that can be described as Luttinger liquids. Such leads can be realized, for example, by quantum wires or fractional quantum Hall edge states. In comparison to noninteracting metallic leads, Luttinger liquid leads involve many-body correlations and the single-particle tunneling density of states shows a power-law singularity at the chemical potential. Using the generalized hierarchical equations of motion method, we assess the importance of the singularity and the next-to-leading order many-body correlations. To this end, we compare numerically converged results with second- and first-order results of the hybridization expansion that is inherent to our method. As a test case, we study transport through a single-level quantum dot or molecule that can be described by an Anderson impurity model. Cotunneling effects turn out to be most pronounced for attractive interactions in the leads or repulsive ones if an excitonic coupling between the dot and the leads is realized. We also find that an interaction-induced negative differential conductance near the Coulomb blockade thresholds is slightly suppressed as compared to a first-order and/or rate equation result. Moreover, we find that the two-particle (
n-particle) correlations enter as a second-order (
n-order) effect and are, thus, not very pronounced at the high temperatures and parameters that we consider.