Holistic integration of experimental and computational datasets together with simplified models for rapid establishment of reliable diffusion databases

Discrepancies between experimental and computational datasets are usually hard to reconcile. For the case of diffusion coefficients, large amounts of both kinds of data are available to allow careful assessments of the degree of agreements and disagreements. Such rigorous and systematic assessments allow us to take advantage of the best of both datasets and provide complementary data to be used to establish the most reliable diffusion databases. In addition, simple and yet general mathematical models for binary and multicomponent metallic solid solutions are developed to facilitate future database establishment. This talk will use many examples to illustrate the power of such holistic integration and the new models.