Microscopic mechanism of the ferromagnetic exchange coupling in insulating EuO

EuO (europium monoxide) is a rock-salt ferromagnetic semiconductor in which localized Eu$^{2+}$ $4f^7$ moments order below a Curie temperature of about 69 K while the material remains insulating. The 1961 discovery paper by Matthias, Bozorth, and Van Vleck reported the ferromagnetism and explicitly declined to settle the precise mechanism of the exchange coupling: the Yosida mechanism of conduction-electron polarization is excluded because EuO is an insulator, direct $4f$ overlap appeared conceivable, and the authors considered indirect exchange via excited states more likely. Later literature has converged on the qualitative answer - the dominant coupling is an indirect exchange driven by virtual excitation of $4f$ electrons into the empty Eu $5d$ and $6s$ conduction states, with direct $4f$-$4f$ overlap negligible and pure $4f$-O-$2p$-$4f$ superexchange only a second-order effect - but the quantitative, orbital-resolved decomposition of the exchange channels in EuO remains unsettled: first-principles studies disagree on the role of the anion $p$ states and on the sign of the next-nearest-neighbor coupling $J_2$, and the combined spectroscopic-plus-first-principles decomposition achieved for the sibling compound EuS in 2021 has not been demonstrated for EuO. The open problem is to establish, with experimentally validated precision, which microscopic channels set the sign and magnitude of the ferromagnetic exchange in EuO, thereby enabling predictive control of its Curie temperature under pressure, strain, and doping.

Background

EuO is the prototypical localized-moment ferromagnetic semiconductor. It crystallizes in the rock-salt (NaCl) structure, and each Eu ion is divalent with a half-filled $4f$ shell ($4f^7$, ground term $^8S_{7/2}$, spin $S=7/2$): the $4f$ orbitals are deeply localized and shielded by the filled $5s$ and $5p$ shells, so the Eu$^{2+}$ moments behave as nearly free-ion spins of $7\,\mu_B$ each ($\mu_B$ is the Bohr magneton, the unit of atomic magnetic moment). EuO orders ferromagnetically below a Curie temperature $T_c \approx 69$ K (77 K in the slightly off-stoichiometric samples first reported) while remaining an insulator: the occupied, nearly flat Eu $4f$ band lies roughly 1.1 eV below an empty, dispersive Eu $5d$/$6s$-derived conduction band, with the O $2p$ valence band several eV deeper. Its siblings EuS, EuSe, and EuTe share the same lattice and moments but order at progressively lower temperatures, ending in antiferromagnetism in EuTe. All four are conventionally described by a Heisenberg Hamiltonian $H=-\sum_{ij} J_{ij}\,\mathbf{S}_i\cdot\mathbf{S}_j$, dominated in the rock-salt lattice by the nearest-neighbor coupling $J_1$ (12 Eu neighbors along the $\langle 110\rangle$ family of crystallographic directions) and the next-nearest-neighbor coupling $J_2$ (6 neighbors along the $\langle 100\rangle$ axis directions, connected by a linear Eu-O-Eu chain - the classic 180-degree superexchange geometry). Across the series the ferromagnetic $J_1$ weakens and $J_2$ becomes antiferromagnetic, producing the trend from ferro- to antiferromagnetic order.

The microscopic origin of $J_1$ is the puzzle at issue. In a metal, localized moments can be coupled indirectly through the spin polarization of conduction electrons - the exchange between itinerant $s$-like conduction electrons and localized $4f$ moments, integrated out to the long-ranged oscillating Ruderman-Kittel-Kasuya-Yosida (RKKY) interaction - the mechanism analyzed by Yosida and explicitly excluded by the 1961 discoverers because EuO is an insulator. In an ionic compound the textbook alternative is superexchange (the Kramers-Anderson mechanism): virtual hopping of electrons through the intervening anion's $p$ orbitals generates an effective coupling whose sign follows the Goodenough-Kanamori rules (the standard semi-empirical rules relating the sign of the coupling to the orbital geometry of the virtual hopping path). For EuO both pictures strain: the $4f$ wave functions are too contracted for appreciable direct Eu-Eu overlap or strong $4f$-$2p$ hybridization, yet $J_1$ is ferromagnetic and sizeable. The remaining candidate, anticipated already in 1961, is indirect exchange via excited states: a $4f$ electron is virtually excited across the gap into an empty Eu $5d$ (or $6s$) orbital, aligned with the $4f^6$ residue by intra-atomic Hund exchange, and the itinerant $5d$/$6s$ electron then couples to the neighboring $4f$ shell - a virtual-excitation channel usually called $f$-$d$ exchange (Kasuya's proposal for the europium chalcogenides).

The question acquired quantitative content through decades of competing assignments. Kasuya's model analysis attributed $J_1$ to indirect $4f$-$5d$ exchange and dismissed superexchange as negligible; band-structure work in the 1980s and LDA+U-based Wannier analyses in 2005 emphasized instead the anion valence band and $4f$-$2p$ hybridization (LDA+U is density-functional theory in the local-density approximation supplemented by an on-site Coulomb parameter $U$ for correlated states; Wannier functions are localized orbitals constructed to reproduce the ab initio bands). Spectroscopy pulled in several directions at once: X-ray absorption spectroscopy (XAS, probing unoccupied states via core-level absorption) saw a sizeable $4f$-$5d$ exchange splitting; neutron spin-wave analysis stressed the anion $p$ shells; Moessbauer spectroscopy (nuclear gamma-ray resonance probing the local electronic environment) supported $6s$-band $s$-$f$ exchange. Controlled channel-by-channel tests became possible with modern first-principles methodology: exchange constants are computed via the magnetic force theorem (evaluating the total-energy change of a magnetic rearrangement from single-particle band-energy changes) with the correlated $4f$ shell treated by an atomic-like self-energy (the Hubbard-1 approximation) or by hybrid functionals (functionals mixing in a fraction of exact exchange), and individual orbital levels can be artificially shifted to switch specific exchange channels on and off. This culminated in Wan, Dong, and Savrasov's 2011 conclusion that $5d$-$4f$ and $6s$-$4f$ indirect exchange dominates in EuO while anion-$p$ superexchange is numerically negligible, and in a 2021 combined resonant-photoemission-plus-first-principles decomposition for EuS - angle-resolved photoemission spectroscopy (ARPES, measuring band dispersion versus momentum) at a photon energy resonant with the Eu $4f$ levels - in which the $f$-$d$ channel, itself mediated by Eu $4f$-anion-$p$ hybridization, is the primary driver of ferromagnetism. For EuO itself, however, the quantitative decomposition is not settled: studies disagree on the anion-$p$ contribution and on the sign of $J_2$, and the orbital-resolved, spectroscopically validated decomposition of the EuO exchange - the precise mechanism the 1961 discovery paper left open - remains an active research question.

Problem Statement

Identify the precise microscopic mechanism of the ferromagnetic exchange coupling in insulating EuO, as posed at the close of the 1961 discovery paper: which interaction couples the localized Eu$^{2+}$ $4f^7$ moments to produce a Curie temperature near 69 K? That qualitative mechanism is established by the literature cited below (see Background); the open problem addressed here is therefore not the identification of that qualitative mechanism but its quantitative, channel-resolved form. The Yosida mechanism (conduction-electron polarization) is excluded because stoichiometric EuO is an insulator. A complete answer must report, for the nearest-neighbor coupling $J_1$ and the next-nearest-neighbor coupling $J_2$, the separate contribution of each of the following channels - direct $4f$-$4f$ overlap; $4f$-O-$2p$-$4f$ superexchange; and indirect exchange via virtual excitation of $4f$ electrons into the empty Eu $5d$ and $6s$ conduction states, including the mediating role of $4f$-$2p$ hybridization - rather than a qualitative restatement, establishing which channel(s) determine the sign and magnitude of each coupling and being consistent with the measured Curie temperature $T_c \approx 69$ K, the neutron spin-wave dispersions of Dietrich et al., and the measured Eu $5d$ (Steeneken et al.) and O $2p$ (Heider et al.) exchange splittings.

References

1. B. T. Matthias, R. M. Bozorth, and J. H. Van Vleck, "Ferromagnetic Interaction in EuO", Physical Review Letters 7, 160 (1961). DOI: 10.1103/PhysRevLett.7.160, https://doi.org/10.1103/PhysRevLett.7.160
2. J. Callaway, "Indirect exchange ferromagnetism in insulators", Il Nuovo Cimento 26, 625 (1962). DOI: 10.1007/BF02771834, https://doi.org/10.1007/BF02771834
3. T. R. McGuire, B. E. Argyle, M. W. Shafer, and J. S. Smart, "Ferromagnetism in divalent europium salts", Applied Physics Letters 1, 17 (1962). DOI: 10.1063/1.1777353, https://doi.org/10.1063/1.1777353
4. T. Kasuya and A. Yanase, "Anomalous Transport Phenomena in Eu-Chalcogenide Alloys", Reviews of Modern Physics 40, 684 (1968). DOI: 10.1103/RevModPhys.40.684, https://doi.org/10.1103/RevModPhys.40.684
5. T. Kasuya, "Exchange Mechanisms in Europium Chalcogenides", IBM Journal of Research and Development 14, 214 (1970). DOI: 10.1147/rd.143.0214, https://doi.org/10.1147/rd.143.0214
6. O. W. Dietrich, J. Als-Nielsen, and L. Passell, "Neutron scattering from the Heisenberg ferromagnets EuO and EuS. III. Spin dynamics of EuO", Physical Review B 14, 4923 (1976). DOI: 10.1103/PhysRevB.14.4923, https://doi.org/10.1103/PhysRevB.14.4923
7. V.-C. Lee and L. Liu, "Exchange mechanism in europium compounds", Physical Review B 30, 2026 (1984). DOI: 10.1103/PhysRevB.30.2026, https://doi.org/10.1103/PhysRevB.30.2026
8. A. Mauger and C. Godart, "The magnetic, optical, and transport properties of representatives of a class of magnetic semiconductors: The europium chalcogenides", Physics Reports 141, 51 (1986). DOI: 10.1016/0370-1573(86)90139-0, https://doi.org/10.1016/0370-1573(86)90139-0
9. P. G. Steeneken, L. H. Tjeng, I. Elfimov, G. A. Sawatzky, G. Ghiringhelli, N. B. Brookes, and D.-J. Huang, "Exchange Splitting and Charge Carrier Spin Polarization in EuO", Physical Review Letters 88, 047201 (2002). DOI: 10.1103/PhysRevLett.88.047201, https://doi.org/10.1103/PhysRevLett.88.047201
10. J. Kunes, W. Ku, and W. E. Pickett, "Exchange Coupling in Eu Monochalcogenides from First Principles", Journal of the Physical Society of Japan 74, 1408 (2005). DOI: 10.1143/JPSJ.74.1408, https://doi.org/10.1143/JPSJ.74.1408
11. N. J. C. Ingle and I. S. Elfimov, "Influence of epitaxial strain on the ferromagnetic semiconductor EuO: First-principles calculations", Physical Review B 77, 121202(R) (2008). DOI: 10.1103/PhysRevB.77.121202, https://doi.org/10.1103/PhysRevB.77.121202 (arXiv:0711.4586)
12. N. M. Souza-Neto, D. Haskel, Y.-C. Tseng, and G. Lapertot, "Pressure-Induced Electronic Mixing and Enhancement of Ferromagnetic Ordering in EuX (X = Te, Se, S, O) Magnetic Semiconductors", Physical Review Letters 102, 057206 (2009). DOI: 10.1103/PhysRevLett.102.057206, https://doi.org/10.1103/PhysRevLett.102.057206
13. H. Miyazaki, T. Ito, H. J. Im, S. Yagi, M. Kato, K. Soda, and S. Kimura, "Direct Observation of Momentum-Dependent Exchange Interaction in a Heisenberg Ferromagnet", Physical Review Letters 102, 227203 (2009). DOI: 10.1103/PhysRevLett.102.227203, https://doi.org/10.1103/PhysRevLett.102.227203
14. X. Wan, J. Dong, and S. Y. Savrasov, "Mechanism of magnetic exchange interactions in europium monochalcogenides", Physical Review B 83, 205201 (2011). DOI: 10.1103/PhysRevB.83.205201, https://doi.org/10.1103/PhysRevB.83.205201 (arXiv:1003.2039)
15. P. Liu, J. A. Colon Santana, Q. Dai, X. Wang, P. A. Dowben, and J. Tang, "Sign of the superexchange coupling between next-nearest neighbors in EuO", Physical Review B 86, 224408 (2012). DOI: 10.1103/PhysRevB.86.224408, https://doi.org/10.1103/PhysRevB.86.224408
16. M. Schlipf, M. Betzinger, M. Lezaic, C. Friedrich, and S. Bluegel, "Structural, electronic, and magnetic properties of the europium chalcogenides: A hybrid-functional DFT study", Physical Review B 88, 094433 (2013). DOI: 10.1103/PhysRevB.88.094433, https://doi.org/10.1103/PhysRevB.88.094433
17. L. Petit, Z. Szotek, M. Lueders, and A. Svane, "Rare-earth pnictides and chalcogenides from first-principles", Journal of Physics: Condensed Matter 28, 223001 (2016). DOI: 10.1088/0953-8984/28/22/223001, https://doi.org/10.1088/0953-8984/28/22/223001
18. A. V. Fedorov, G. Poelchen, S. V. Eremeev, S. Schulz, A. Generalov, C. Polley, C. Laubschat, K. Kliemt, N. Kaya, C. Krellner, E. V. Chulkov, K. Kummer, D. Yu. Usachov, A. Ernst, and D. V. Vyalikh, "Insight into the Temperature Evolution of Electronic Structure and Mechanism of Exchange Interaction in EuS", The Journal of Physical Chemistry Letters 12, 8328 (2021). DOI: 10.1021/acs.jpclett.1c02274, https://doi.org/10.1021/acs.jpclett.1c02274
19. T. Heider, T. Gerber, O. Koksal, M. Eschbach, E. Mlynczak, P. Loemker, P. Gospodaric, M. Gehlmann, M. Ploetzing, R. Pentcheva, L. Plucinski, C. M. Schneider, and M. Mueller, "Temperature-dependent spin-resolved electronic structure of EuO thin films", Physical Review B 106, 054424 (2022). DOI: 10.1103/PhysRevB.106.054424, https://doi.org/10.1103/PhysRevB.106.054424 (arXiv:1809.00631)