High-throughput prediction of activation energy for impurity diffusion in fcc metals of Group I and VIII

We develop a high-throughput calculation method that rigorously mines correlations embodied within experimental data and use it to direct quantum mechanical techniques efficiently toward the diffusion activation predictions and assessments. The new diffusion activation energies data of 298 binary systems, covering six faced-centered cubic (fcc) host metals of Group I (Cu, Ag, Au) and Group VIII (Fe, Co, Ni) and 76 elements of periods of 3, 4, 5 and 6, are predicted in this work. Furthermore, a new analytic expression, derived from data-mining, has been developed, which sheds light on the delicate relationship between diffusion activation energy and materials fundamental properties. The present approach can be extended to study the impurity diffusion in metal alloys of other complex structures where diffusion activation energies are difficult to obtain from first-principles calculation due to the complex diffusion mechanism. The present results could be used to provide useful data for other material simulation techniques such as CALPHAD or facilitate the choice of materials for technological applications.