Microscopic origin of magnetism and magnetic interactions in ferropnictides

One year after their initial discovery, two schools of thought have crystallized regarding the electronic structure and magnetic properties of ferropnictide systems. One postulates that these are itinerant weakly correlated metallic systems that become magnetic by virtue of spin-Peierls-type transition due to near nesting between the hole and the electron Fermi-surface pockets. The other argues that these materials are strongly or at least moderately correlated and the electrons are considerably localized and close to a Mott-Hubbard transition, with the local magnetic moments interacting via short-range superexchange. In this Rapid Communication we argue that neither picture is fully correct. The systems are moderately correlated but with correlations driven by Hund’s rule coupling rather than by the on-site Hubbard repulsion. The iron moments are largely local, driven by Hund’s intra-atomic exchange. Superexchange is not operative, and the interactions between the Fe moments are considerably long range and driven mostly by one-electron energies of all occupied states.

Figure 1

Top-down view of the (a) checkerboard, (b) stripe, and (c) double-stripe magnetic patterns for a single FeAs or FeTe layer. The light colored sites have majority up spin, and the darker sites have majority down spin.

Figure 2

The densities of states for ${\text{BaFe}}_2{\text{As}}_2$ in the nonmagnetic configuration in comparison to a) checkerboard magnetic pattern b) stripe (ground state) magnetic pattern and c) double-stripe magnetic pattern

Figure 3

The densities of states for FeTe in the nonmagnetic configuration in comparison to (a) stripe magnetic pattern and (b) double-stripe magnetic pattern.

Figure 4

(Color online) The Fermi surfaces of stripe-ordered ${\text{BaFe}}_2{\text{As}}_2$. The primitive reciprocal zone is shown with the high-symmetry directions of the conventional Brillouin zone designated. The $Z$ point (along the $c$ axis in real space) is in the very middle of the zone (not marked). Top panel shows the “reverse distortion” in which the Fe-Fe distance is lengthened along the like spin direction and shortened along the unlike spin direction. The bottom panels show the fully relaxed calculation (nearly the same as experiment).