The Ni tracer self-diffusion in the B2 phase NiAl over extensive temperature and composition intervals (Frank et al., 2001) was considered. While the Ni tracer diffusivity D*Ni increased notably on the Ni-rich side of the composition in NiAl with increasing Ni content, D*Ni was practically unchanged with increasing Al content on the Al-rich side, in spite of the large number of structural Ni vacancies. When performing atomistic calculations of different diffusion mechanisms in NiAl, these tracer data were well explained by the dominant contribution via the triple defect diffusion mechanism. Different from the tracer diffusion data, the interdiffusion data D̃ in NiAl available in the literature revealed a deep minimum at about Ni50Al50, with a substantial increase of D̃ on both sides of the stoichiometric composition. Recalling the available thermodynamic data, it was shown that the whole set of diffusion data was inconsistent with the Darken-Manning equation applying the generally accepted value of the vacancy wind factor S ≅ 1. To resolve the problem a Monte Carlo simulation of the interdiffusion in the Ni50Al50/Ni58Al42 single-phase couple was performed. The energy barriers were computed by molecular static calculations with empirical embedded atom potentials. The current study was confined to the Ni-rich side of compositions. It was definitively shown that interdiffusion in NiAl could proceed via the triple defect diffusion mechanism. It turned out that the vacancy wind factor S strongly depended on the composition. At the stoichiometric composition S took a very small value, e.g. S ≅ 0.02 at 1273K. Applying the calculated S values the Darken-Manning relation was proved to be valid in Ni-rich NiAl with the anti-structure type of disorder.

Ni Tracer Self-Diffusion, Interdiffusion and Diffusion Mechanisms in NiAl. S.V.Divinski, C.Herzig: Defect and Diffusion Forum, 2002, 203-205, 177-92