Magnetic coupling constants from a hybrid density functional with 35% Hartree-Fock exchange

Abstract

The magnetic coupling constants of ${\mathrm{KCuF}}_{3}$, ${\mathrm{Sr}}_{2}{\mathrm{CuO}}_{2}{\mathrm{Cl}}_{2}$, ${\mathrm{La}}_{2}{\mathrm{CuO}}_{4}$, ${\mathrm{La}}_{2}{\mathrm{NiO}}_{4}$, ${\mathrm{K}}_{2}{\mathrm{NiF}}_{4}$, ${\mathrm{KNiF}}_{3}$, ${\mathrm{NiF}}_{2}$, ${\mathrm{KMnF}}_{3}$, and ${\mathrm{MnF}}_{2}$ are calculated with a hybrid density functional, in which 35% of the nonlocal Hartree-Fock exchange is mixed in the general gradient approximation to the density functional theory. The theoretical magnetic coupling constants for these materials with different structures, spins, and magnetic orderings are in good agreement with experiment. Our results improve significantly over the so-called B3LYP hybrid density functional, which usually overestimates the magnetic coupling constants by about 50%. However, the energy gaps from the B3LYP functional are in better agreement with experiment than the hybrid functional with 35% Hartree-Fock exchange, which means that within the current scheme of hybrid density functionals different functionals are needed to better predict different properties of materials.