UC Berkeley

Topological insulators (TIs) are a new class of materials which, until recently, have been overlooked despite decades of study in band insulators. Like semiconductors and ordinary insulators, TIs have a bulk gap, but feature robust surfaces excitations which are protected from disorder and interactions which do not close the bulk gap. TIs are distinguished from ordinary insulators not by the symmetries they possess (or break), but by topological invariants characterizing their bulk band structures. These two pictures, the existence of gapless surface modes, and the nontrivial topology of the bulk states, yield two contrasting approaches to the study of TIs. At the heart of the subject, they are connected by the bulk-boundary correspondence, relating bulk and surface degrees of freedom. In this work, we study both aspects of topological insulators, at the same time providing an illumination to their mysterious connection.

First, we present a systematic approach to the classification of bulk states of systems with inversion-like symmetries, deriving a complete set of topological invariants for such ensembles. We find that the topological invariants in all dimensions may be computed algebraically via exact sequences. In particular, systems with spatial inversion symmetries in one-, two-, and three-dimensions can be classified by, respectively, 2, 5, and 11 integer invariants. The values of these integers are related to physical observables such as polarization, Hall conductivity, and magnetoelectric coupling. We also find that, for systems with “antiferromagnetic symmetry,” there is a Z2 classification in three-dimensions, and hence a class of “antiferromagnetic topological insulators” (AFTIs) which are distinguished from ordinary antiferromagnets. From the perspective of the bulk, AFTI exhibits the quantized magnetoelectric effect, whereas on the surface, gapless one-dimensional chiral modes emerge at step-defects.

Next, we study how the surface spectrum can be computed from bulk quantities. Specifically, we present an analytic prescription for computing the edge dispersion E(k) of a tight-binding Dirac Hamiltonian terminated at an abrupt crystalline edge, based on the bulk Hamiltonian. The result is presented as a geometric formula, relating the existence of surface states as well as their energy dispersion to properties of the bulk Hamiltonian. We further prove the bulk-boundary correspondence for this specific class of systems, connecting the Chern number and the chiral edge modes for quantum Hall systems given in terms of Dirac Hamiltonians. In similar spirit, we examine the existence of Majorana zero modes in superconducting doped-TIs. We find that Majorana zero modes indeed appear but only if the doped Fermi energy is below a critical chemical potential. The critical doping is associated with a topological phase transition of vortex lines, which supports gapless excitations spanning their length. For weak pairing, the critical point is dependent on the non-abelian Berry phase of the bulk Fermi surface.

Finally, we investigate the transport properties on the surfaces of TIs. While the surfaces of “strong topological insulators” – TIs with an odd number of Dirac cones in their surface spectrum – have been well studied in literature, studies of their counterpart “weak topological insulators” (WTIs) are meager, with conflicting claims. Because WTIs have an even number of Dirac cones in their surface spectrum, they are thought to be unstable to disorder, which leads to an insulating surface. Here we argue that the presence of disorder alone will not localize the surface states, rather, presence of a time-reversal symmetric mass term is required for localization. Through numerical simulations, we show that in the absence of the mass term the surface always flow to a stable metallic phase and the conductivity obeys a one-parameter scaling relation, just as in the case of a strong topological insulator surface. With the inclusion of the mass, the transport properties of the surface of a weak topological insulator follow a two-parameter scaling form, reminiscent of the quantum Hall phase transition.