ABSTRACT
In order for methods combining ab initio density-functional theory and many-body techniques to become routinely used, a flexible, fast, and easy-to-use implementation is crucial. We present an implementation of a general charge self-consistent scheme based on projected localized orbitals in the projector augmented wave framework in the Vienna Ab Initio Simulation Package. We give a detailed description on how the projectors are optimally chosen and how the total energy is calculated. We benchmark our implementation in combination with dynamical mean-field theory: first we study the charge-transfer insulator NiO using a Hartree-Fock approach to solve the many-body Hamiltonian. We address the advantages of the optimized against non-optimized projectors and furthermore find that charge self-consistency decreases the dependence of the spectral function-especially the gap-on the double counting. Second, using continuous-time quantum Monte Carlo we study a monolayer of SrVO3, where strong orbital polarization occurs due to the reduced dimensionality. Using total-energy calculation for structure determination, we find that electronic correlations have a non-negligible influence on the position of the apical oxygens, and therefore on the thickness of the single SrVO3 layer.
ABSTRACT
The phase stability of group VB (V, Nb, and Ta) transition metals is explored by first-principles electronic-structure calculations. Alloying with a small amount of a neighboring metal can either stabilize or destabilize the body-centered-cubic phase relative to low-symmetry rhombohedral phases. We show that band-structure effects determine phase stability when a particular group VB metal is alloyed with its nearest neighbors within the same d-transition series. In this case, the neighbor with less (to the left) and more (to the right) d electrons destabilize and stabilize bcc, respectively. When alloying with neighbors of higher d-transition series, electrostatic Madelung energy dominates and stabilizes the body-centered-cubic phase. This surprising prediction invalidates current understanding of simple d-electron bonding that dictates high-symmetry cubic and hexagonal phases.