Atomic Diffusion in α-iron across the Curie Point: An Efficient and Transferable Ab Initio-Based Modeling Approach.

An accurate prediction of atomic diffusion in Fe alloys is challenging due to thermal magnetic excitations and magnetic transitions. We propose an efficient approach to address these properties via a Monte Carlo simulation, using ab initio-based effective interaction models. The temperature evolution of self- and Cu diffusion coefficients in α-iron are successfully predicted, particularly the diffusion acceleration around the Curie point, which requires a quantum treatment of spins. We point out a dominance of magnetic disorder over chemical effects on diffusion in the very dilute systems.