When does Magnetization Depend on the Boundary?

TL;DR

This paper explores how boundary conditions influence the determination of thermodynamic quantities like magnetization and polarization, showing that bulk properties can be boundary-dependent depending on Berry's phase and surface states.

Contribution

It clarifies the conditions under which bulk calculations of thermodynamic quantities are boundary-independent or boundary-dependent, especially involving Berry's phase and surface states.

Findings

Bulk quantities are unambiguously determined without Berry's phase.

Presence of Berry's phase introduces boundary dependence under certain conditions.

Surface states and symmetry breaking can make boundary details crucial.

Abstract

In this paper, we investigate in general how thermodynamic quantities such as the polarization, magnetization and the magneto-electric tensor are affected by the boundaries. We show that when the calculation with periodic boundary conditions does not involve a Berry's phase, the quantity in question is determined unambiguously by the bulk, even in the presence of gapless surface states. When the calculation involves a Berry's phase, the bulk can only determine the quantity up to some quantized value, given that (i) there are no gapless surface states, (ii) the surfaces do not break the symmetries preserved by the bulk, and (iii) the system is kept at charge neutrality. If any of the above conditions is violated, the quantity is then determined entirely by the details at the boundary. For components of response functions such as the magneto-electric tensor that are entirely determined by the bulk, the value of those components should connect smoothly to finite frequency and momentum.