Sluggish diffusion in random equimolar FCC alloys

MS Daw and M Chandross, PHYSICAL REVIEW MATERIALS, 5, 043603 (2021).

DOI: 10.1103/PhysRevMaterials.5.043603

We examine vacancy-assisted diffusion in the 57 random, equimolar alloys that can be formed from Cu, Ag, Au, Ni, Pd, and Pt based on the well- tested embedded atom method functions of Foiles, Baskes, and Daw Phys. Rev. B 33, 7983 (1986). We address the suggestion Yeh et al., Adv. Eng. Mater. 6, 299 (2004) that increasing the number of constituents causes diffusion to be "sluggish" in random, equimolar alloys. Using molecular dynamics (MD) simulations of random alloys with a single vacancy, combined with calculations of vacancy formation, we extract vacancy-assisted diffusivities in each alloy. After developing and applying several possible criteria for evaluating "sluggishness," we find that only a small minority (from 1 to 8, depending on how sluggishness is defined) of the alloys exhibit sluggish diffusion whereas in the large majority of alloys diffusion is faster and in quite a few cases ought to be considered vigorous (that is, faster than in any of the constituents). We correlate diffusivity with a combination of the mean of the constituent diffusivity and a simple function of lattice mismatch. We conclude that simply increasing the number of constituents in such alloys does not systematically alter the diffusion, but that instead lattice mismatch plays a primary factor; sluggish diffusion is more likely to occur in a window of small lattice mismatch (1-3%) even in binary alloys. Quantitatively, our calculated diffusivities correlate with a combination of (1) rule of mixtures of the diffusivities of the constituents, and (2) a simple function of the lattice mismatch; this accounts for the large majority of our calculated diffusivities to within a factor of 2 (over a range of three orders of magnitude). We also find that while lattice mismatch on the order of 1-3% is necessary for sluggish diffusion, it is not sufficient.

Return to Publications page