Stripe Antiferromagnetic Order Research Articles

Crystal symmetry and magnetic order in iron pnictides: A tight-binding Wannier function analysis

To perform a local orbital analysis of electronic and magnetic interactions, we construct the Wannier functions (WFs) of the Fe $3d$ orbitals in the parent compound of the recently discovered iron pnictide superconductors, LaFeAsO, and a comparison material LaFePO. Comparing the WFs for the stripe antiferromagnetic order with those for no magnetic order, the difference is a significant spreading (``{\it de}localization'') of specifically the $d_{xy}$ and $d_{xz}$ (but not $d_{yz}$) WFs, where parallel Fe spins lie along the $x$ direction. The WF basis gives a tight-binding representation of the first principles, density functional based Fe-derived bands. Comparing hopping parameters, it is found that changes due to stripe antiferromagnetism, even if it is weak, enables more isotropic hopping involving spin-majority electrons in the Fe $3d_{xz}$ (but not the $3d_{yz}$) orbital. This change, counterintuitively, actually reinforces electronic anisotropy. Further insight is gained by comparing the WFs of LaFeAsO and LaFePO, identifying how the difference in WFs is related to the difference in hopping integrals and showing how the pnictide atom is influential in forming the stripe antiferromagnetism. Kinetic energy considerations suggest that orbital fluctuation, in addition to spin fluctuation, may contribute to the decrease in observed ordered moment compared to the calculated values.

Magnetic properties of iron pnictides from spin-spiral calculations

The wave-vector (q) and doping dependences of the magnetic energy, iron moment, and effective exchange interactions in LaFeAsO, BaFe2As2, and SrFe2As2 are studied by self-consistent LSDA calculations for co-planar spin spirals. For the undoped compounds, the calculated total energy, E(q), reaches its minimum at q corresponding to stripe anti-ferromagnetic (AF) order. In LaFeAsO, this minimum becomes flat already at low levels of electron-doping and shifts to an incommensurate q at δ=0.2, where δ is the number of additional electrons (δ > 0) or holes (S < 0) per Fe. In BaFe2As2 and SrFe2As2, stripe order remains stable for hole doping down to δ = −0.3. Under electron doping, on the other hand, the E(q) minimum shifts to incommensurate q already at δ=0.1.

Is Fermi-Surface Nesting the Origin of Superconductivity in Iron Pnictides?: A Fluctuation–Exchange-Approximation Study

We study whether Fermi-surface (FS) nesting can give rise to high-temperature superconductivity in iron pnictides. Starting with ab initio construction of an effective four-orbital model, we employ the fluctuation-exchange approximation to show that FS does not necessarily favor the stripe antiferromagnetic order observed in experiments, especially for realistic electronic correlations. If superconductivity in iron pnictides is magnetically mediated and has fully-gapped sign-reversing s-wave symmetry, our results suggest that the pairing interaction does not arise only from FS nesting and exchange interactions between local moments in the Fe 3d orbitals may play a crucial role.

Interplay between magnetic properties and Fermi surface nesting in iron pnictides

The wave-vector $(\mathbf{q})$ and doping $(x,y)$ dependences of the magnetic energy, iron moment, and effective exchange interactions in ${\text{LaFeAsO}}_{1\ensuremath{-}x}{\text{F}}_{x}$ and ${\text{Ba}}_{1\ensuremath{-}2y}{\text{K}}_{2y}{\text{Fe}}_{2}{\text{As}}_{2}$ are studied by self-consistent LSDA calculations for co-planar spin spirals. For the undoped compounds $(x=0,y=0)$, the minimum of the calculated total energy, $E(\mathbf{q})$, is for $\mathbf{q}$ corresponding to stripe antiferromagnetic order. Already at low levels of electron doping $(x)$, this minimum becomes flat in ${\text{LaFeAsO}}_{1\ensuremath{-}x}{\text{F}}_{x}$ and for $x\ensuremath{\gtrsim}5%$, it shifts to an incommensurate $\mathbf{q}$. In ${\text{Ba}}_{1\ensuremath{-}2y}{\text{K}}_{2y}{\text{Fe}}_{2}{\text{As}}_{2}$, stripe order remains stable for hole doping up to $y=0.3$. These results are explained in terms of the band structure. The magnetic interactions cannot be accurately described by a simple classical Heisenberg model and the effective exchange interactions fitted to $E(\mathbf{q})$ depend strongly on doping. The doping dependence of the $E(\mathbf{q})$ curves is compared with that of the noninteracting magnetic susceptibility for which similar trends are found.