A data-driven approach to study the order-disorder transition in high entropy alloys

We introduce a data-driven approach to construct the effective Hamiltonian from first principles data, and apply it to study the thermodynamics of HEAs through canonical Monte Carlo simulation. This method uses atomic pair interactions as features and systematically improve the representativeness of the dataset using samples from Monte Carlo simulation. This method produces highly accurate effective Hamiltonians that give less than 0.1 mRy test error for all the three refractory HEAs: MoNbTaW, MoNbTaVW, and MoNbTaTiW. From the Monte Carlo results, we identified two order-disorder transition temperatures, each due to different chemical interactions. By comparing with experimental results, we propose that by tuning the chemical composition, the order and disorder phases can be controlled, which further affects the strength and ductility of HEAs.