Analyticity and symmetry of band extrema in gapped solids: when does the effective mass approximation hold?

TL;DR

This paper proves that band extremum analyticity in gapped solids is guaranteed at non-degenerate points for common ab initio methods, linking non-analyticity to degeneracy, and uses symmetry to constrain effective mass tensors.

Contribution

It establishes conditions for the validity of the effective mass approximation in gapped solids and provides symmetry-based constraints on effective mass tensors from first principles.

Findings

Analyticity holds at non-degenerate band extrema in standard ab initio Hamiltonians.

Band warping is intrinsically linked to degeneracy.

Effective masses at the K point in monolayer MoS2 are strictly isotropic.

Abstract

The effective mass approximation is widely used across models of carrier transport, optical response, and excitons in semiconductors and insulators, but its validity hinges on the assumption that the band dispersion $E_n(\mathbf{k})$ at the relevant extremum is analytic. We prove that analyticity holds at any non-degenerate extremum for the standard ab initio Hamiltonians, including density functional theory with local or hybrid exchange-correlation functionals and for band-edge $G_0W_0$ quasiparticle energies in gapped systems. Band non-analyticity (or warping) in these settings is therefore intrinsically tied to degeneracy. We then use group theory to determine the symmetry-allowed form of the effective mass tensor for each of the 32 crystallographic point groups, providing a stringent consistency check on first-principles calculations. As a representative application, we show that the electron and hole effective masses at the $K$ point of monolayer MoS$_2$ must be strictly isotropic at the DFT and $G_0W_0$ levels.