Spin gaps and spin-flip energies in density-functional theory

TL;DR

This paper develops density-functional theory definitions for spin gaps and spin-flip energies, highlighting differences from charge gaps and illustrating the concepts with calculations on the Lithium atom.

Contribution

It introduces new DFT-based definitions for spin gaps and spin discontinuities, emphasizing their relation to excited states and many-body effects.

Findings

Kohn-Sham and many-body spin gaps differ due to derivative discontinuities.

Single-particle calculations overestimate spin gaps but underestimate charge gaps.

Theoretical definitions are validated through calculations on Lithium atom.

Abstract

Energy gaps are crucial aspects of the electronic structure of finite and extended systems. Whereas much is known about how to define and calculate charge gaps in density-functional theory (DFT), and about the relation between these gaps and derivative discontinuities of the exchange-correlation functional, much less is know about spin gaps. In this paper we give density-functional definitions of spin-conserving gaps, spin-flip gaps and the spin stiffness in terms of many-body energies and in terms of single-particle (Kohn-Sham) energies. Our definitions are as analogous as possible to those commonly made in the charge case, but important differences between spin and charge gaps emerge already on the single-particle level because unlike the fundamental charge gap spin gaps involve excited-state energies. Kohn-Sham and many-body spin gaps are predicted to differ, and the difference is related to derivative discontinuities that are similar to, but distinct from, those usually considered in the case of charge gaps. Both ensemble DFT and time-dependent DFT (TDDFT) can be used to calculate these spin discontinuities from a suitable functional. We illustrate our findings by evaluating our definitions for the Lithium atom, for which we calculate spin gaps and spin discontinuities by making use of near-exact Kohn-Sham eigenvalues and, independently, from the single-pole approximation to TDDFT. The many-body corrections to the Kohn-Sham spin gaps are found to be negative, i.e., single particle calculations tend to overestimate spin gaps while they underestimate charge gaps.