Efficacy of surface error corrections to density functional theory calculations of vacancy formation energy in transition metals

Abstract

We calculate properties like equilibrium lattice parameter, bulk modulus and monovacancy formation energy for nickel (Ni), iron (Fe) and chromium (Cr) using Kohn-Sham density functional theory (DFT). We describe relative performance of local density approximation (LDA) and generalized gradient approximation (GGA) for predicting such physical properties for these metals. We also make a relative study between the exchange correlation functionals, namely, PW91 and PBE, two different flavors of GGA. Our calculations show that DFT is inherently unable to predict the monovacncy formation energy accurately. We calculate the correction for the surface intrinsic error corresponding to an exchange correlation functional using the scheme implemented by Mattson et al. [Phys. Rev. B 73, 195123 (2006)]. We compare the effectiveness of the correction scheme for the free-electron like Al and 3d-transition metals, namely, Ni, Fe and Cr. The disagreement of the corrected vacancy formation energy with experimental value is found to be less in Al as compared to the transition metals. The reasons for the ineffectiveness of this correction scheme in 3d-transition metals are discussed.