Wednesday 12.05.2010, 15:15-16:15 SE3.


Edge states at graphene/graphane interfaces

M. Schmidt, Univ. Basel

We study interfaces between graphene and graphane. These interfaces act as effective edges for the pi-electrons in graphene. If the interface is oriented along a zigzag direction, edge states are found that are very similar to the usual zigzag edge states in graphene nanoribbons. These interface states have some special properties which, in some respects, make them superior to usual edge states in pure graphene: (1) Spin-orbit effects are strongly enhanced at the interface (factors of 100 and more). (2) Effective edges with 'bearded' boundary conditions are stable. (3) The bandwidth of the edge state is tunable by electrostatic gates.
Edge magnetism, well known from graphene edges, is also present at graphene/graphane interfaces. However, the tunability of the edge state bandwidth makes it possible to tune through the quantum phase transition by means of electrostatic gates, i.e. switching edge magnetism on and off. Near the transition, a regime of one-dimensional itinerant ferromagnetism exists. In order to better understand the phase transition, we introduce an effective model for the edge states. Crucially, the electron-electron interaction becomes velocity-dependent in this model. This velocity dependence is responsible for the stabilization of the one-dimensional magnetism.
We incorporate quantum fluctuations around the mean-field theory by means of the bosonization technique. In the non-magnetic regime the edge state behaves as an ordinary spinful Luttinger liquid. In the regime of weak ferromagnetism, the bosonized free Hamiltonian is well behaved only if the proper mean-field theory is used as a basis for the bosonization. The backscattering term, which leads to a sine-Gordon action for the bosonic spin fields, is essentially momentum non-conserving, i.e. not allowed. As a result, the interaction strength in the spin-sector is not renormalized to unity and the spin-rotation invariance is broken. Close to the transition, the usual bosonization technique is not sufficient to describe the system. We discuss how the description may be extended.