Multiscale Modelling of the Antiferromagnet Mn2Au: From ab-initio to Micromagnetics

Antiferromagnets (AFMs) are strong candidates for the future spintronic and memory applications largely because of their inherently fast dynamics and lack of stray fields, with Mn2Au being one of the most promising. For the numerical modelling of magnetic material properties, it is common to use ab-initio methods, atomistic models and micromagnetics. However, each method alone describes the physics within certain limits. Multiscale methods bridging the gap between these three approaches have been already proposed for ferromagnetic materials. Here, we present a complete multiscale model of the AFM Mn2Au as an exemplar material, starting with results from ab-initio methods going via atomistic spin dynamics (ASD) to an AFM Landau-Lifshitz-Bloch (AFM-LLB) model. Firstly, bulk is modelled using a classical spin Hamiltonian constructed based on earlier first-principles calculations. Secondly, this spin model is used in the stochastic Landau-Lifshitz-Gilbert (LLG) to calculate temperature-dependent equilibrium properties, such as magnetization and magnetic susceptibilities. Thirdly, the temperature dependent micromagnetic parameters are used in the AFM-LLB. We validate our approach by comparing the ASD and AFM-LLB models for three paradigmatic cases; (i) Damped magnetic oscillations, (ii) magnetization dynamics following a heat pulse resembling pump-probe experiments, (iii) magnetic domain wall motion under thermal gradients.