A Short Course on Topological Insulators - arXiv

1 J. K. Asb oth, L. Oroszl any, A. P alyiA Short Course onTopological InsulatorsBand-structure topology and edge states inone and two dimensionsSeptember 9, 2015 [ ] 8 Sep 2015 PrefaceThese lecture notes provide an introduction to some of the main concepts of topo-logical Insulators , a branch of solid state physics that is developing at a fast are based on a one-semester Course for MSc and PhD students at the E otv osUniversity, Budapest, which the authors have been giving since aim is to provide an understanding of the core topics of Topological Insulators edge states, bulk Topological invariants, bulk boundary correspondence with assimple mathematical tools as possible. We restricted our attention to one- and two-dimensional band Insulators .

2 We use noninteracting lattice models of topologicalinsulators, and build these up gradually to arrive from the simplest one-dimensionalcase (the Su-Schrieffer-Heeger model for polyacetylene) to two-dimensional time-reversal invariant Topological Insulators (the Bernevig-Hughes-Zhang model forHgTe). In each case we introduce the model first, discuss its properties, and thengeneralize. The prerequisite for the reader is quantum mechanics and not much else:solid state physics background is provided as we go this is an introduction, rather than a broad overview, we try to be self-contained and give citations to the current literature only where it is absolutely nec-essary. For a broad overview, including pointers to the original papers and currenttopics, we refer the reader to review articles and books in the material for these lecture notes in the form of ipython notebooks willbe made available our efforts, the book inevitably contains typos, errors, and less compre-hensible explanations.

3 We would appreciate if you could inform us of any of those;please send your comments We are grateful for enlightening discussions on topoologicalinsulators with Anton Akhmerov, Andrea Alberti, Carlo Beenakker, and AlbertoCortijo. We thank the feedback we got from participants at the courses at E otv osUniversity, especially Vilmos Kocsis. A version of this Course was given by one ofus ( ) in the PhD programme of the University of Geneva, on invitation byMarkus B uttiker, which helped shape the preparation of the first version of these lecture notes were supported bythe grant T We acknowledge financial support from thevviPrefaceHungarian Scientific Research Fund (OTKA), Grant Nos. PD100373, K108676,NN109651, and from the Marie Curie programme of the European Union, GrantNo.

4 CIG-293834. and were supported by the J anos Bolyai Scholarshipof the Hungarian Academy of SciencesBudapest,J anos K. Asb oth, L aszl o Oroszl any, Andr as P alyiSeptember 2015 IntroductionThe band theory of electric conduction was one of the early victories of quantummechanics in the 1920s. It gave a simple explanation of how some crystalline mate-rials are electric Insulators , even though electrons in them can hop from one atom tothe next. In the bulk of a band insulator, the electrons occupy eigenstates that formenergy bands. In a band insulator, there are no partially filled bands: completelyfilled bands are separated by an energy gap from completely empty bands, the gaprepresents the energy cost of mobilizing electrons.

5 In contrast, materials with par-tially filled bands are conductors, where there are plane wave states available totransmit electrons across the bulk at arbitrarily low energy. Although we now knowof situations where band theory is inadequate ( , for Mott Insulators ), it remainsone of the cornerstones of solid states discovery of the Quantum Hall Effect (1980) has shown that the simpledivision into band Insulators and metals is not the end of the story, not even in bandtheory. In the quantum Hall effect, a strong magnetic field confines the motion ofelectrons in the bulk, but the same field forces them into delocalized edge states onthe surface. A two-dimensional metal in strong magnetic field is thus an insulatorin the bulk, but conducts along the surface, via a discrete number of completelyopen edge state channels (in the language of the Landauer B uttiker formalism).

6 The number of edge state channels was linked to the Chern number, a topologicalinvariant of the occupied bands (1982).Over the last twenty years, theoretical progress over artificial systems has shownthat the external magnetic field is not necessary for an insulator to have robust con-ducting edge states: instead, the nontrivial topology of the occupied bands is thecrucial ingredient. The nametopological insulatorwas coined for such systems,and their study became a blossoming branch of solid state physics. Following thetheoretical prediction (Bernevig, Hughes and Zhang, 2006 [5]), electronic transportmeasurements confirmed that a thin layer of HgTe is a Topological insulator (K oniget al, 2007 [21]). Since that time, a host of materials have been shown to be three-dimensional Topological Insulators , and thin films and quantum wires shown to betwo- and one-dimensional Topological Insulators [2].

7 ViiviiiPrefaceThe intense theoretical interest in Topological Insulators has led to signature re-sults, such as the the periodic table of Topological Insulators [29], which showsthat similarly to phase transitions in statistical mechanics, it is the dimensionalityand the basic symmetries of an insulator that decide whether it can be a topologicalinsulator or not. Although it was derived by different ways of connecting topologicalinsulators of various dimensions and symmetries (so-called dimensional reductionschemes), the mathematically rigorous proof of the periodic table is still field of Topological Insulators is very active, with many experimental chal-lenges and open theoretical problems, regarding the effects of electron-electron in-teraction, extra crystalline symmetries, coupling to the environment, get a quick and broad overview of Topological Insulators , with citations for rel-evant research papers, we recommend the review papers [17, 25, 7].

8 For a morein-depth look, there are already a few textbooks on the subject (by Bernevig andHughes [4], by Shen [30], and one edited by Franz and Molenkamp [10]). To seethe link between momentum-space topology and physics in a broader context, wedirect the reader to a book by Volovik[34].There are also introductory courses on Topological Insulators with a broad recommend the lectures by Charles Kane (the video recording of the versiongiven at Veldhoven is freely available online), and the online EdX Course on topol-ogy in condensed matter by a group of lecturers, with the corresponding materialcollected at lecture notesOur aim with this set of lecture notes is to complement the literature cited above:we wish to provide a close look at some of the core concepts of Topological insula-tors with as simple mathematical tools as possible.

9 Using one-and two-dimensionalnoninteracting lattice models, we explain what edge states and what bulk topolog-ical invariants are, how the two are linked (this is known as thebulk boundarycorrespondence), the meaning and impact of some of the fundamental keep things as simple as possible, throughout the Course we use noninteract-ing models for solid state systems. These are described using single-particle latticeHamiltonians, with the zero of the energy corresponding to the Fermi energy. Weuse natural units, with h=1 and length measured by the lattice Su-Schrieffer-Heeger (SSH) model.. The SSH Hamiltonian .. Bulk Hamiltonian .. Edge states .. Chiral symmetry .. Number of edge states as Topological invariant.

10 14 Problems .. 202 Berry phase, Chern number.. Discrete case .. Continuum case .. Berry phase and adiabatic dynamics .. Berry s formulas for the Berry curvature .. Example: the two-level system .. 35 Problems .. 403 Polarization and Berry phase.. Wannier states in the Rice-Mele model .. Inversion symmetry and polarization .. 47 Problems .. 494 Adiabatic charge pumping, Rice-Mele model.. Charge pumping in a control freak way .. Moving away from the control freak limit .. Tracking the charges with Wannier states .. 59 Problems .. 635 Current operator and particle pumping.. Particle current at a cross section of the lattice .. Time evolution governed by a quasi-adiabatic Hamiltonian.