Computational Studies of Model Disordered and Strongly Correlated Electronic Systems

Abstract

The theory of non-interacting electrons in perfect crystals was completed soon after the advent of quantum mechanics. Though capable of describing electron behaviour in most simple solid state physics systems, this approach falls woefully short of describing condensed matter systems of interest today, and designing the quantum devices of the future. The reason is that nature is never free of disorder, and emergent properties arising from interactions can be clearly seen in the pure, low-dimensional materials that can be engineered today. In this thesis, I address some salient problems in disordered and correlated electronic systems using modern numerical techniques like sparse matrix diagonalization, density matrix renormalization group (DMRG), and large disorder renormalization group (LDRG) methods.

The pioneering work of P. W. Anderson, in 1958, led to an understanding of how an electron can stop diffusing and become localized in a region of space when a crystal is sufficiently disordered. Thus disorder can lead to metal-insulator transitions, for instance, in doped semiconductors. Theoretical research on the Anderson disorder model since then has mostly focused on the localization-delocalization phase transition. The localized phase in itself was not thought to exhibit any interesting physics. Our work has uncovered a new singularity in the disorder-averaged inverse participation ratio of wavefunctions within the localized phase, arising from resonant states. The effects of system size, dimension and disorder distribution on the singularity have been studied. A novel wavefunction-based LDRG technique has been designed for the Anderson model which captures the singular behaviour.

While localization is well established for a single electron in a disordered potential, the situation is less clear in the case of many interacting particles. Most studies of a many-body localized phase are restricted to a system which is isolated from its environment. Such a condition cannot be achieved perfectly in experiments. A chapter of this thesis is devoted to studying signatures of incomplete localization in a disordered system with interacting particles which is coupled to a bath.

Strongly interacting particles can also give rise to topological phases of matter that have exotic emergent properties, such as quasiparticles with fractional charges and anyonic, or perhaps even non-Abelian statistics. In addition to their intrinsic novelty, these particles (e.g. Majorana fermions) may be the building blocks of future quantum computers. The third part of my thesis focuses on the best experimentally known realizations of such systems - the fractional quantum Hall effect (FQHE) which occurs in two-dimensional electron gases in a strong perpendicular magnetic field. It has been observed in systems such as semiconductor heterostructures and, more recently, graphene. I have developed software for exact diagonalization of the many-body FQHE problem on the surface of a cylinder, a hitherto unstudied type of geometry. This geometry turns out to be optimal for the DMRG algorithm. Using this new geometry, I have studied properties of various fractionally-filled states, computing the overlap between exact ground states and model wavefunctions, their edge excitations, and entanglement spectra. I have calculated the sizes and tunneling amplitudes of quasiparticles, information which is needed to design the interferometers used to experimentally measure their Aharanov-Bohm phase. I have also designed numerical probes of the recently discovered geometric degree of freedom of FQHE states.