Three semi-empirical force field FeCr potentials, two within the formalism of the two-band model and one within the formalism of the concentration dependent model, were benchmarked against a wide variety of density functional theory structures. The benchmarking allows an assessment of how reliable empirical potential results were in different areas relevant to radiation damage modelling. The density functional theory data consist of defect-free structures, structures with single interstitials and structures with small di- and tri-interstitial clusters. All three potentials reproduce the general trend of the heat of formation quite well. The most important shortcomings of the original two-band model potential were the low or even negative heat of formation for Cr-rich structures and the lack of a strong repulsion when moving two solute Cr atoms from being second-nearest neighbours to nearest neighbours. The newer two-band model potential partly solves the first problem. The most important shortcoming in the concentration dependent model potential was the magnitude of the Cr–Cr repulsion, being too strong at short distances and mostly absent at longer distances. Both two-band model potentials do reproduce long-range Cr–Cr repulsion. For interstitials the two-band model potentials reproduce a number of Cr–interstitial binding energies surprisingly well, in contrast to the concentration dependent model potential. For Cr interacting with clusters, the result could sometimes be directly extrapolated from Cr interacting with single interstitials, both according to density functional theory and the three empirical potentials.

Benchmarking FeCr Empirical Potentials Against Density Functional Theory Data. T.P.C.Klaver, G.Bonny, P.Olsson, D.Terentyev: Modelling and Simulation in Materials Science and Engineering, 2010, 18[7], 075004