Regensburg 2019 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MA: Fachverband Magnetismus
MA 55: Electron theory and micromagnetism
MA 55.10: Talk
Friday, April 5, 2019, 12:00–12:15, H38
Extending Liechtenstein's method to strong spin-orbit coupled systems — •Louis Ponet1,2 and Sergey Artyukhin1 — 1Istituto Italiano di Tecnologia, Genova, Italia — 2Scuola Normale Superiore di Pisa, Pisa, Italia
Localized magnetism in transition metal compounds has been very successfully modeled using a variation on the classical Heisenberg model. Calculating the model parameters, so-called magnetic exchange constants, from first-principles has been a notoriously difficult task, usually involving the comparison of total energies from numerous supercell ab-initio calculations. For simple compounds with small magnetic unit cells this approach is feasible and indeed has produced accurate results. A computationally more efficient Green's function-based method developed by Liechtenstein et al., and later adapted for use with Wannier functions by Rudenko et al., remedies this by calculating the exchange coefficients from a single DFT calculation. However this method was only formulated in the situation where there is low-strength atomic SOC present in the material. The crucial assumption is that the spins can rotate without perturbing other degrees of freedom. This picture breaks down when the localized spins are situated on atoms with strong spin-orbit coupling, such as the 5d-transition metal ions. This is because the low-energy degrees of freedom entangle spins and orbitals (describing the charge distribution). This leads to strongly anisotropic magnetic interactions, not contained in the Heisenberg model, leading to new physical phenomena as discussed by Jackeli et al. Here we explore the extension of the Liechtenstein method to this situation.