Exchange constants for local spin Hamiltonians from tight-binding models
Simon Streib, Attila Szilva, Vladislav Borisov, Manuel Pereiro, Anders, Bergman, Erik Sj\"oqvist, Anna Delin, Mikhail I. Katsnelson, Olle Eriksson,, Danny Thonig

TL;DR
This paper develops a method to derive local spin Hamiltonians from tight-binding models, improving the accuracy of magnetic interactions for small fluctuations and out-of-equilibrium states in itinerant-electron systems.
Contribution
It introduces a linear term in the local spin Hamiltonian to better capture non-collinear magnetic states and provides a practical implementation based on constraining fields.
Findings
Accurate calculation of exchange constants for iron dimers and chains.
Enhanced modeling of magnetic fluctuations near equilibrium states.
Validation of the formalism through numerical examples.
Abstract
We consider the mapping of tight-binding electronic structure theory to a local spin Hamiltonian, based on the adiabatic approximation for spin degrees of freedom in itinerant-electron systems. Local spin Hamiltonians are introduced in order to describe the energy landscape of small magnetic fluctuations, locally around a given spin configuration. They are designed for linear response near a given magnetic state and in general insufficient to capture arbitrarily strong deviations of spin configurations from the equilibrium. In order to achieve this mapping, we include a linear term in the local spin Hamiltonian that, together with the usual bilinear exchange tensor, produces an improved accuracy of effective magnetic Weiss fields for non-collinear states. We also provide examples from tight-binding electronic structure theory, where our implementation of the calculation of exchange…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
