Segregation to an (A0/2)[11̄0] edge dislocation in Cu0.1Ni0.9

Atomistic simulations of segregation to a dissociated $\text{(a}{}_0\text{/2)[1}\text{1}\text{̄}\text{0]}$ edge dislocation in the solid solution alloy Cu_0.1Ni_0.9 have been performed. Segregation to the stacking fault between the partials is minimal. Results obtained with a general embedded atom method potential and one optimized for the NiCu system differ significantly. Simulations employing the optimized potentials show significantly more Cu segregation to the dislocation cores than do simulations performed with the general potentials. When the general potentials are employed, the Cu concentration around the dislocation is well described using classical segregation isotherms based upon the stress distribution around the dislocation, except in the dislocation core region. Deviations from the theoretically predicted segregation profile around the dislocation core are largest along the slip plane. When the optimized potentials are used, the deviations from the predicted segregation behavior are significantly larger. The large deviations associated with the optimized potentials were traced to the inadequacy of describing the local heat of segregation in terms of the elastic work σ_hΔV. This can be rectified by adding a term to the heat of segregation that explicitly includes the composition dependence. The failure of the classical segregation isotherm to describe the segregation behavior around a dislocation is associated with non-ideal alloy thermodynamics and the inadequacy of linear elasticity to appropriately describe the core region of the dislocation. The failure of the classical segregation isotherm within the core appears to result from the fact that the core atoms have different atomic coordination than those in the bulk material.