Crystal structure prediction of magnetic materials.

We present a methodology to predict magnetic systems using ab initio methods. By employing crystal structure method and spin-polarized calculations, we explore the relation between crystalline structures and their magnetic properties. In this work, testbed cases of transition metal alloys (FeCr, FeMn, FeCo and FeNi) are study in the ferromagnetic case. We find soft-magnetic properties for FeCr, FeMn while for FeCo and FeNi hard-magnetic are predicted. In particular, for the family of FeNi, a candidate structure with energy lower than the tetrataenite was found. The structure has a saturation magnetization (M s) of 1.2 MA m-1, magnetic anisotropy energy (MAE) above 1200 kJ m-3 and hardness value close to 1. Theoretically, this system made of abundant elements could be the right candidate for permanent magnet applications. Comparing with the state-of-the-art (Nd2Fe14B) hard-magnet, (M s of 1.28 MA m-1 and MAE of 4900 kJ m-3) is appealing to explore this low energy polymorph of FeNi further. Considering the relatively limited number of magnets, predicting a new system may open routes for free rare-earth magnets. Furthermore, the use of the computational algorithm as the one presented in this work, hold promises in this field for which in near future improvements will allow to study numerous complex systems, larger simulations cells and tackled long-range antiferromagnetic cases.