Orbital-resolved beyond-Heisenberg exchange interactions in ferromagnetic bcc Fe: a complete effective spin Hamiltonian and higher-order terms

Ferromagnetic body-centered cubic (bcc) iron is the archetypal itinerant ferromagnet, yet the microscopic structure of its exchange interactions is only partly understood. First-principles work by Kvashnin and coworkers established an orbital-level mechanism separation: the exchange between $T_{2g}$ orbitals is Heisenberg-like, long-ranged, and of RKKY origin tied to Fermi-surface nesting, while the $E_g$ contribution is short-ranged, strongly dependent on the magnetic reference state (non-Heisenberg), and dominated by double exchange. The open problem is to turn this partial picture into a complete description: (i) an orbital-resolved characterization of the exchange interactions of bcc Fe extending beyond the nearest-neighbor couplings and the (111) direction analyzed in the source work, including their non-Heisenberg content; (ii) an effective magnetic Hamiltonian incorporating both Heisenberg and non-Heisenberg terms, with controlled higher-order (beyond bilinear and biquadratic) content, valid from the ground state up to the Curie temperature; and (iii) a quantitative determination of the role of higher-order spin interactions in the finite-temperature magnetism of iron. As of the 2023 review literature none of these components is complete, and no direct experimental test of the mechanism separation exists.

Background

Body-centered cubic (bcc) iron — the room-temperature α phase of Fe — is the prototypical itinerant ferromagnet: its magnetism arises from partially filled, delocalized 3d bands rather than from localized atomic moments. In simulations its magnetic energy is conventionally mapped onto a classical Heisenberg Hamiltonian $H=-\sum_{i\neq j} J_{ij}\,\vec{e}_i\cdot\vec{e}_j$, where $\vec{e}_i$ is a unit vector along the magnetic moment at site $i$ and $J_{ij}$ is the interatomic exchange coupling. These couplings are routinely computed from first principles via the magnetic force theorem (the linear-response approach of Liechtenstein, Katsnelson, Antropov and Gubanov), which evaluates the change of band energies under small rotations of local moments. In the cubic crystal field the five 3d orbitals split into the twofold $E_g$ and threefold $T_{2g}$ manifolds; because the LKAG formula is a trace over orbitals, each $J_{ij}$ can be decomposed into orbital-pair contributions. Several exchange mechanisms compete in such a decomposition: the RKKY (Ruderman–Kittel–Kasuya–Yosida) interaction — an oscillatory, long-ranged indirect exchange mediated by conduction electrons, with periods fixed by Fermi-surface spanning vectors — and double exchange, a mechanism in which electron hopping between sites favors parallel alignment because parallel spins maximize the hopping amplitude and hence lower the kinetic energy. A coupling is called non-Heisenberg when its extracted value depends on the magnetic configuration (reference state) from which it is computed, signaling that the true energy landscape is not a bilinear function of the spin directions; beyond bilinear terms, spin Hamiltonians can contain biquadratic pair terms $\propto(\vec{e}_i\cdot\vec{e}_j)^2$ and more general higher-order or multi-spin (ring) terms. Experimentally, the magnon (spin-wave) dispersion of iron has been measured by inelastic neutron scattering since the late 1960s, and first-principles Heisenberg models reproduce that dispersion and the Curie temperature reasonably well — but such comparisons test only the orbital-summed total coupling. Spin-polarized electron energy loss spectroscopy (SPEELS) measures magnons with surface sensitivity in ultrathin films, and resonant inelastic x-ray scattering (RIXS) can in principle provide element- and orbital-selective access to magnetic excitations at absorption edges. In this context, Kvashnin, Cardias, Szilva, Di Marco, Katsnelson, Lichtenstein, Nordström, Klautau and Eriksson (Phys. Rev. Lett. 116, 217202 (2016)) showed from density-functional theory that in bcc Fe the $T_{2g}$ orbitals dominate the Fermi surface and yield Heisenberg-like, long-ranged exchange with pronounced RKKY oscillations (analyzed in detail along the (111) nearest-neighbor bonding direction), whereas the $E_g$ states form weakly interacting quasi-impurity levels whose exchange contribution is short-ranged, strongly reference-state dependent, and attributable to double exchange (the coupling scaling with the first power of the hopping integral, which also implies it is not primarily biquadratic). Their non-Heisenberg analysis focused on nearest-neighbor couplings, and the calculations describe the zero-temperature ferromagnetic ground state, leaving open how the interactions reorganize as the exchange splitting collapses near the Curie temperature.

Problem Statement

The problem is to replace the partial, nearest-neighbor, zero-temperature picture of exchange in ferromagnetic bcc Fe with a complete beyond-Heisenberg description. The single overarching objective is a quantitative, orbital-resolved account of the exchange interactions of bcc Fe — which terms are Heisenberg-like, which are non-Heisenberg, and which higher-order content is essential — packaged as an effective magnetic Hamiltonian reliable from the ground state up to the Curie temperature. Concretely, the objective comprises the following jointly supporting components: (1) an orbital-resolved characterization ($E_g$–$E_g$, $E_g$–$T_{2g}$, $T_{2g}$–$T_{2g}$, and total) of the exchange interactions of bcc Fe extending beyond the nearest-neighbor and (111)-direction analyses of the source work to further neighbor shells and other crystallographic directions, including the dependence of the non-Heisenberg content on rotation angle, magnetic reference state, and inter-site direction; (2) the effective magnetic Hamiltonian of bcc Fe built from this characterization, incorporating bilinear Heisenberg terms together with the non-Heisenberg terms (for example direction-dependent double-exchange-like contributions), biquadratic terms, and whatever higher-order or multi-spin terms the analysis shows to be required, with a stated and controlled mapping from the underlying electronic structure, and valid from the ground state up to the Curie temperature; and (3) a determination of the role of spin interactions beyond biquadratic order — which higher-order terms materially affect the finite-temperature magnetism of iron (Curie temperature, magnetization curve, magnon renormalization) and which are negligible. These components are interrelated: (1) supplies the microscopic orbital input, (2) is the effective model assembled from it, and (3) establishes the higher-order content that (2) must retain at finite temperature. Each component counts as resolved only against an explicit, checkable standard. Component (1) is resolved when the orbital-resolved couplings ($E_g$–$E_g$, $E_g$–$T_{2g}$, $T_{2g}$–$T_{2g}$, and total) are reported for successive neighbour shells along the principal bcc directions $[100]$, $[110]$, and $[111]$, out to the range at which they have decayed below a stated threshold, and the dependence of the non-Heisenberg content on rotation angle, magnetic reference state, and inter-site direction is resolved on an explicitly declared sampling of those variables. Component (2) is resolved when the Hamiltonian contains the bilinear and non-Heisenberg terms together with exactly those biquadratic, higher-order, or multi-spin terms the analysis shows to be required — the preceding term list delimits the permitted class rather than requiring every listed kind — and when its mapping from the electronic structure is stated with a numerical tolerance over a range of validity spanning the ground state to the Curie temperature. Component (3) is resolved when every higher-order term is shown to shift the Curie temperature, the magnetization curve, or the magnon renormalization by more than that stated tolerance (and is therefore retained) or by less than that tolerance (and is then declared negligible).

References

1. Y. O. Kvashnin, R. Cardias, A. Szilva, I. Di Marco, M. I. Katsnelson, A. I. Lichtenstein, L. Nordström, A. B. Klautau, and O. Eriksson, 'Microscopic Origin of Heisenberg and Non-Heisenberg Exchange Interactions in Ferromagnetic bcc Fe', Physical Review Letters 116, 217202 (2016). DOI: 10.1103/PhysRevLett.116.217202; arXiv:1510.01872, https://arxiv.org/abs/1510.01872
2. A. Szilva, D. Thonig, P. F. Bessarab, Y. O. Kvashnin, D. C. M. Rodrigues, R. Cardias, M. Pereiro, L. Nordström, A. Bergman, A. B. Klautau, and O. Eriksson, 'Theory of noncollinear interactions beyond Heisenberg exchange: Applications to bcc Fe', Physical Review B 96, 144413 (2017). DOI: 10.1103/PhysRevB.96.144413; arXiv:1706.06560, https://arxiv.org/abs/1706.06560
3. A. Szilva, Y. Kvashnin, E. A. Stepanov, L. Nordström, O. Eriksson, A. I. Lichtenstein, and M. I. Katsnelson, 'Quantitative theory of magnetic interactions in solids', Reviews of Modern Physics 95, 035004 (2023). DOI: 10.1103/RevModPhys.95.035004; arXiv:2206.02415, https://arxiv.org/abs/2206.02415
4. R. Cardias, A. Szilva, A. Bergman, I. Di Marco, M. I. Katsnelson, A. I. Lichtenstein, L. Nordström, A. B. Klautau, O. Eriksson, and Y. O. Kvashnin, 'The Bethe-Slater curve revisited; new insights from electronic structure theory', Scientific Reports 7, 4058 (2017). DOI: 10.1038/s41598-017-04427-9, https://doi.org/10.1038/s41598-017-04427-9
5. Jacob B. J. Chapman, Pui-Wai Ma, and Sergei L. Dudarev, 'Effect of non-Heisenberg magnetic interactions on defects in ferromagnetic iron', Physical Review B 102, 224106 (2020). DOI: 10.1103/PhysRevB.102.224106, https://doi.org/10.1103/PhysRevB.102.224106
6. Simon Streib, Attila Szilva, Vladislav Borisov, Manuel Pereiro, Anders Bergman, Erik Sjöqvist, Anna Delin, Mikhail I. Katsnelson, Olle Eriksson, and Danny Thonig, 'Exchange constants for local spin Hamiltonians from tight-binding models', Physical Review B 103, 224413 (2021). DOI: 10.1103/PhysRevB.103.224413, https://doi.org/10.1103/PhysRevB.103.224413
7. Simon Streib, Ramon Cardias, Manuel Pereiro, Anders Bergman, Erik Sjöqvist, Cyrille Barreteau, Anna Delin, Olle Eriksson, and Danny Thonig, 'Adiabatic spin dynamics and effective exchange interactions from constrained tight-binding electronic structure theory: Beyond the Heisenberg regime', Physical Review B 105, 224408 (2022). DOI: 10.1103/PhysRevB.105.224408, https://doi.org/10.1103/PhysRevB.105.224408
8. S. Mankovsky, S. Polesya, and H. Ebert, 'Exchange coupling constants at finite temperature', Physical Review B 102, 134434 (2020). DOI: 10.1103/PhysRevB.102.134434, https://doi.org/10.1103/PhysRevB.102.134434
9. Tomonori Tanaka and Yoshihiro Gohda, 'Impact of electron–spin coupling on exchange coupling parameters: a nonperturbative approach', preprint (2024). arXiv:2410.11256, https://arxiv.org/abs/2410.11256
10. A. A. Katanin, 'Exchange interactions in itinerant magnets: Effects of local particle-hole irreducible vertex corrections and SU(2) symmetry of Hund interaction', Physical Review B 112, 085141 (2025). DOI: 10.1103/179m-rrjc, https://doi.org/10.1103/179m-rrjc; arXiv:2504.08018, https://arxiv.org/abs/2504.08018
11. M. F. Collins, V. J. Minkiewicz, R. Nathans, L. Passell, and G. Shirane, 'Critical and Spin-Wave Scattering of Neutrons from Iron', Physical Review 179, 417–430 (1969). DOI: 10.1103/PhysRev.179.417, https://doi.org/10.1103/PhysRev.179.417
12. J. W. Lynn, 'Temperature dependence of the magnetic excitations in iron', Physical Review B 11, 2624–2637 (1975). DOI: 10.1103/PhysRevB.11.2624, https://doi.org/10.1103/PhysRevB.11.2624
13. R. Vollmer, M. Etzkorn, P. S. Anil Kumar, H. Ibach, and J. Kirschner, 'Spin-Polarized Electron Energy Loss Spectroscopy of High Energy, Large Wave Vector Spin Waves in Ultrathin fcc Co Films on Cu(001)', Physical Review Letters 91, 147201 (2003). DOI: 10.1103/PhysRevLett.91.147201, https://doi.org/10.1103/PhysRevLett.91.147201
14. M. Pajda, J. Kudrnovský, I. Turek, V. Drchal, and P. Bruno, 'Ab initio calculations of exchange interactions, spin-wave stiffness constants, and Curie temperatures of Fe, Co, and Ni', Physical Review B 64, 174402 (2001). DOI: 10.1103/PhysRevB.64.174402, https://doi.org/10.1103/PhysRevB.64.174402