Spin Models and Cluster Multipole Method: Application to Kagome Magnets

We present a multi-scale computational approach that combines atomistic spin models with the cluster multipole (CMP) method. The CMP method enables a systematic and accurate generation of complex non-collinear magnetic structures using symmetry-adapted representations. The parameters of the spin model are derived from density functional theory using the magnetic force theorem, with the paramagnetic state as a reference. The energy landscape of CMP-generated structures is inspected at the model Hamiltonian level, and sets of low-energy magnetic structures are identified for each material candidate. The inclusion of relativistic antisymmetric and anisotropic pair interactions lifts partially the degeneracy among these most stable structures. To demonstrate the applicability and predictive capability of the method, we apply it to the non-collinear Mn3X and collinear Fe3X (X = Ga, Ge, and Sn) kagome compounds. The computational efficiency of the method in identifying low-energy structures among multiple CMP configurations highlights its potential for high-throughput screening of complex magnets with unknown magnetic order.