Abstract
We show that the spin density generalization of the AM05 density functional [R. Armiento and A. E. Mattsson, Phys. Rev. B 72, 085108 (2005)] predicts the correct ground spin state for iron, a system known to be heavily dependent on proper spin treatment. Using the fundamental assumptions in the subsystem functional scheme, we resolve an ambiguity in how to treat the separate spin densities in AM05 but also show that the other less preferred treatments give no significantly different numerical outcome of the iron body-centered-cubic and face-centered-cubic test cases. Details and formulas are given to aid in the implementation of functionals in general, and the spin-resolved AM05 exchange-correlation potentials in particular, into different types of computer codes.
- Received 4 December 2008
DOI:https://doi.org/10.1103/PhysRevB.79.155101
©2009 American Physical Society