High-throughput ab-initio studies of phase stability in multi-component alloy systems have garnered a great deal of research interest in the past decade. However, significantly less attention has been paid to the sometimes startling results of these studies, such as the implied wealth of undiscovered stable phases present in many 3d-5d transition metal binaries. This shortcoming is especially well-demonstrated in the case of Co-Pt, which has a long experimental history demonstrating three stable mixed phases: L10 CoPt and L12 Co3Pt and CoPt3; in contrast, results from density functional theory suggest a set of (yet-unobserved) long-period commensurate phases resembling the β2 superstructure are present at Pt-rich compositions. We have performed an in-depth ab-initio study of the Co-Pt binary, and fit our results to a cluster expansion Hamiltonian appropriate for use in Monte Carlo simulations. Using semi-grand canonical Monte Carlo, we demonstrate that density functional theory implies high-temperature phase behavior wholly incompatible with experimental results[1]. Our case study demonstrates that effective Hamiltonian and statistical mechanic techniques provide a logical â??next stepâ? in ab-initio studies where simulation and experiment disagree.Â
[1] E. Decolvenaere, M.J. Gordon, A. Van der Ven, Phys. Rev. B 92, 085119 (2015)Â