posted on 2015-12-08, 00:00authored byTorben Steenbock, Jos Tasche, Alexander
I. Lichtenstein, Carmen Herrmann
Exchange spin coupling is usually
evaluated in quantum chemistry
from the energy difference between a high-spin determinant and a Broken-Symmetry
(BS) determinant in combination with Kohn–Sham density functional
theory (KS-DFT), based on the work of Noodleman. As an alternative,
an efficient approximate approach relying on Green’s functions
has been developed by one of the authors. This approach stems from
solid-state physics and has never been systematically tested for molecular
systems. We rederive a version of the Green’s-function approach
originally suggested by Han, Ozaki, and Yu. This new derivation employs
local projection operators as common in quantum chemistry for defining
local properties such as partial charges, rather than using a dual
basis as in the Han–Ozaki–Yu approach. The result is
a simple postprocessing procedure for KS-DFT calculations, which in
contrast to the BS energy-difference approach requires the electronic
structure of only one spin state. We show for several representative
small molecules, diradicals, and dinuclear transition metal complexes
that this method gives qualitatively consistent results with the BS
energy-difference approach as long as it is applied to high-spin determinants
and as long as structural relaxation effects in different spin states
do not play an important role.