Theoretical analysis of experimental tracer and interdiffusion data in Cu-Ni-Fe alloys

In this paper, we present strategies to extract fundamental atomistic information from measured diffusion coefficients in a ternary alloy system. The strategies are exemplified with Cu-Ni-Fe alloys at 1271 K where recent extensive interdiffusion coefficients and tracer diffusion coefficients for all three components have become available. We develop new defining phenomenological expressions for the vacancy-wind factors in terms of the diffusion coefficients. We show that the measured tracer diffusion coefficients can be processed using the Manning and Moleko, Allnatt and Allnatt random alloy diffusion kinetics formalisms (with and without the assumption of the Gibbs-Duhem relation between the thermodynamic activities) to give jump frequencies, tracer correlation factors, vacancy-wind factors and phenomenological coefficients. It is shown for example that Cu is generally the most correlated component in its diffusion behavior and that the off-diagonal phenomenological coefficients can be as high as 64% of the smallest of the diagonal phenomenological coefficients. It is also shown that the Darken formalism (which ignores off-diagonal phenomenological coefficients) is in fact a reasonable approximation for expressing the diagonal phenomenological coefficients in terms of the tracer diffusion coefficients. It is then shown how the measured interdiffusivities can be processed with these formalisms to give tracer diffusivities, vacancy-wind factors and phenomenological coefficients. Finally, we show how a straightforward strategy starting with the Darken analysis that is then followed by the Manning or Moleko, Allnatt and Allnatt analysis can be used to gain access to the vacancy-wind factors and the off-diagonal phenomenological coefficients. (c) 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.