We present an approach that combines the local density approximation (LDA) and the dynamical mean-field theory (DMFT) in the framework of the full-potential linear augmented plane waves (FLAPW) method. Wannier-like functions for the correlated shell are constructed by projecting local orbitals onto a set of Bloch eigenstates located within a certain energy window. The screened Coulomb interaction and Hund's coupling are calculated from a first-principle constrained RPA scheme. We apply this LDA+DMFT implementation, in conjunction with a continuous-time quantum Monte-Carlo algorithm, to the study of electronic correlations in LaFeAsO. Our findings support the physical picture of a metal with intermediate correlations. The average value of the mass renormalization of the Fe 3d bands is about 1.6, in reasonable agreement with the picture inferred from photoemission experiments. The discrepancies between different LDA+DMFT calculations (all technically correct) which have been reported in the literature are shown to have two causes: i) the specific value of the interaction parameters used in these calculations and ii) the degree of localization of the Wannier orbitals chosen to represent the Fe 3d states, to which many-body terms are applied. The latter is a fundamental issue in the application of many-body calculations, such as DMFT, in a realistic setting. We provide strong evidence that the DMFT approximation is more accurate and more straightforward to implement when well-localized orbitals are constructed from a large energy window encompassing Fe-3d, As-4p and O-2p, and point out several difficulties associated with the use of extended Wannier functions associated with the low-energy iron bands. Some of these issues have important physical consequences, regarding in particular the sensitivity to the Hund's coupling.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.