Molecular dynamics simulations of forced atomic mixing in crystalline binary alloys during plastic deformation at 100 K are performed. Nearly complete atomic mixing is observed in systems that have a large positive heat mixing and in systems with a large lattice mismatch. Only systems that contained a hard precipitate in a soft matrix do not mix. The amount of mixing is quantified by defining a mean square relative displacement of pairs of atoms, sigma(2)(R,t), that were initially separated by a distance R. Analysis of sigma(2)(R,t) and visual inspection of the displacement fields reveal that forced mixing results from dislocation glide, and that it resembles the forced mixing of a substance advected by a turbulent flow. Consideration of sigma(2)(R,t) also provides a rationalization of compositional self-organization during plastic deformation at higher temperatures.

No abstract

No abstract

We derive expressions for the effective diffusion coefficient of Richardson's pairs in plastically strained solids as a function of the pair separation distance R. We predict that a crossover from superdiffusive to diffusive mixing takes place when R becomes comparable to the coherence length of the shearing events underlying the plastic deformation. Molecular dynamics simulations on nanocrystalline and amorphous systems support this analysis, which thus provides new insight on deformation mechanisms in these systems. Superdiffusive mixing is experimentally observable by monitoring the rate of dissolution of precipitates as a function of their initial size.

The initial stages of vapor condensation of Ge in the presence of a cold Ar atmosphere were studied by molecular-dynamics simulations. The state variables of interest included the densities of condensing vapor and gas, the density of clusters, and the average cluster size, while the temperatures of the vapor and the clusters were separately monitored with time. Three condensation processes were explicitly identified: nucleation, monomeric growth, and cluster aggregation. Our principal finding is that both the average cluster size and the number of clusters scale with the linear dimension of the computation cell, L, and Ln, with the scaling parameter n approximately 4, corresponding to a reaction order of nu approximately 2.33. This small value of n is explained by an unexpected nucleation path involving the formation of Ge dimers via two-body collisions.