EuIn2As2: Magnetic Topology & Electronic Structure
- EuIn2As2 is a layered rare-earth Zintl compound characterized by an inverted band structure and complex magnetic phases including broken-helical order.
- Experimental studies using neutron diffraction, ARPES, and STM/STS reveal symmetry-sensitive spin arrangements and surface state reconstructions key to its axion-insulating behavior.
- Fermi-level tuning via doping and thin-film synthesis are critical strategies to control carrier density and realize the topological boundary phenomena in EuIn2As2.
Searching arXiv for papers on EuIn2As2 to ground the article in the latest literature. EuInAs is a layered rare-earth Zintl compound that crystallizes in the hexagonal space group and has become a central material in the study of magnetic topology because its low-energy band structure is inverted while its magnetic order is symmetry sensitive. In the early topological literature it was proposed as an antiferromagnetic axion-insulator and higher-order-topological-insulator candidate; later diffraction work showed that the zero-field magnetic ground state is not the originally assumed simple collinear A-type antiferromagnet but a lower-symmetry helical or broken-helical order, which changes the symmetry protection of its boundary states rather than eliminating the topological problem itself (Regmi et al., 2019, Riberolles et al., 2020). Across this literature, EuInAs is best understood as a material in which crystal symmetry, magnetic symmetry, carrier density, and surface termination all materially affect whether the experimentally relevant regime is a bulk-insulating axion phase, a topological-crystalline axion phase with selected gapless surfaces, or a hole-doped metallic state with strong magnetic-reconstruction effects (Gong et al., 2022, Yan et al., 2024).
1. Crystal structure and electronic constituents
EuInAs crystallizes in the hexagonal structure (No. 194), with alternating Eu layers and InAs layers stacked along the crystallographic 0-axis (Regmi et al., 2019, Riberolles et al., 2020). Single-crystal neutron work reported lattice parameters 1 Å and 2 Å, while scanning-tunneling work described the material as an in-plane hexagonal layered compound with 3 Å and 4 Å; both descriptions are consistent with a layered crystal that cleaves along the 5 plane (Riberolles et al., 2020, Gong et al., 2022). The cleaved surface is not always an ideal bulk truncation: STM identified partially Eu-terminated reconstructed surfaces, including stripe and atomic reconstructions, with terraces separated by multiples of 6 (Gong et al., 2022).
The local moments are carried by Eu7. Several studies describe the Eu ion as having 8, 9, and a local moment of about 0, while the Eu 1 states lie well below the Fermi level at about 2 eV binding energy (Sato et al., 2020, Liebman-Pelaez et al., 15 Jan 2025). This separation is important because the near-3 states are instead dominated by In 4 and As 5 orbitals, so the Eu 6 sector primarily supplies magnetism rather than itinerant carriers (Sato et al., 2020). In the band-inversion problem central to the topological classification, the relevant inversion is between In 7 and As 8 states at 9 (Regmi et al., 2019).
In practice, bulk crystals and films are typically hole doped rather than ideally insulating. Soft-x-ray ARPES resolved a small three-dimensional hole pocket centered at bulk 0, with an estimated hole density of roughly 1, consistent with Hall data giving 2 at 3 K (Sato et al., 2020). This persistent hole doping is a recurrent limitation in efforts to realize the bulk axion-insulating regime directly.
2. Magnetic order and its revision from collinear to broken-helical descriptions
The magnetic description of EuIn4As5 evolved substantially. Early first-principles and ARPES work treated the ordered state below 6 K in terms of two nearly degenerate collinear antiferromagnetic configurations: AFM-B with in-plane Eu moments and AFM-C with out-of-plane moments, with an energy difference of less than 7 meV (Regmi et al., 2019). Within that framework, both configurations were topologically nontrivial, but the spin orientation controlled whether the system was discussed as coexisting with topological crystalline insulating behavior or with higher-order topology and hinge modes (Regmi et al., 2019).
Later neutron diffraction overturned the assumption that the zero-field ground state is a simple A-type collinear antiferromagnet. Instead, two magnetic transitions were found on cooling, at 8 K and 9 K (Riberolles et al., 2020). Between these temperatures the system realizes a pure 0-helix phase, while below 1 it enters the lower-symmetry broken-helix phase. The diffraction data identified propagation vectors 2 with 3, close to 4, and 5, which immediately rules out pure A-type order as the zero-field ground state (Riberolles et al., 2020). In the broken helix, the Eu moments lie in the 6-plane and stack helically along 7; at 8 K the ordered moment was reported as about 9, while the saturation moment from magnetization is 0 (Riberolles et al., 2020).
The later dynamical literature recast the low-temperature state as a multi-1 “broken helix” built from 2 and 3, with the additional 4 modulation appearing below 5 K (Liebman-Pelaez et al., 15 Jan 2025). In this formulation the ordered state behaves as a nearly-6 easy-plane magnet whose low-energy orientational degree of freedom can be described by an in-plane nematic director 7 (Liebman-Pelaez et al., 15 Jan 2025). This viewpoint is particularly useful for understanding the pseudo-Goldstone mode and the role of strain.
A persistent issue in the EuIn8As9 literature is therefore not whether magnetism matters to topology, but which magnetic symmetry is actually realized under a given condition. The historical tension is between early collinear models used for topological classification and later experimental evidence for non-collinear coplanar helix and broken-helix phases. That tension is not merely terminological: it changes the surviving symmetry operations and thereby the expected location and nature of gapless or gapped boundary states (Regmi et al., 2019, Riberolles et al., 2020).
3. Band inversion, axion topology, and symmetry-dependent boundary physics
The basic topological mechanism in EuIn0As1 is an inversion-driven, spin-orbit-coupled band structure. DFT+2 calculations found a pronounced band inversion at 3 between In 4 and As 5 states, with an inverted gap of approximately 6 eV in the non-SOC calculation; after SOC is included, all crossings are gapped and valence and conduction bands are cleanly separated (Regmi et al., 2019). In the early collinear-antiferromagnetic classification, the inversion parity data over the eight TRIMs gave 7 for both AFM-B and AFM-C, which was interpreted as the indicator of an axion-insulator state (Regmi et al., 2019). In the standard field-theoretic language used in later ARPES analysis, the relevant response is
8
with 9 for the nontrivial axion phase when protected by inversion or suitable magnetic crystalline symmetry (Sato et al., 2020).
Within the early collinear picture, the moment direction controlled the additional topological structure. The in-plane configuration was linked to coexistence with a topological crystalline insulator phase, whereas the out-of-plane configuration was linked to higher-order topology with hinge modes (Sato et al., 2020). This is the origin of the frequent description of EuIn0As1 as a higher-order-topological-insulator candidate. The key caveat is that those papers did not image hinge states directly; the HOTI assignment was symmetry based and indirect (Regmi et al., 2019).
The helical revision did not remove the axion classification, but it changed its symmetry basis. Neutron diffraction and symmetry analysis argued that the experimentally realized helical and broken-helical phases are adiabatically connected to an inversion-symmetric A-type reference state with 2, while preserving the bulk gap, so that 3 survives in the actual low-symmetry state (Riberolles et al., 2020). In this description EuIn4As5 is a magnetic topological-crystalline axion insulator protected not by 6 or 7 separately, both of which are broken, but by the antiunitary crystalline symmetry 8 (Riberolles et al., 2020).
That symmetry has immediate boundary consequences. For the broken-helix ground state, 9 and 0 surfaces were predicted to host gapless 1-protected Dirac cones, whereas surfaces such as 2 were predicted to have gapped Dirac cones and half-integer QAH-type conductivity 3 (Riberolles et al., 2020). Because 4 is broken, these surface Dirac cones are not pinned to TRIM points; the work described them as “unpinned” Dirac cones whose crossings can move away from high-symmetry momenta (Riberolles et al., 2020). This feature distinguishes the helical-phase topological-crystalline description from the earlier collinear HOTI language, even though both are rooted in the same inverted bulk band structure.
A recurring practical limitation is that the ideal topological response requires the chemical potential to lie in the bulk gap. Multiple studies emphasized that measured crystals are slightly or strongly hole doped, so the topological classification is often better viewed as the topology of the underlying magnetic band structure than as a directly realized insulating transport phase (Gong et al., 2022, Singh et al., 6 Aug 2025).
4. Spectroscopic evidence and the status of surface states
The spectroscopic case for topology in EuIn5As6 is cumulative and deliberately cautious. Temperature-dependent VUV ARPES measured across the magnetic transition found a clear near-7 reconstruction between the paramagnetic state at 8 K and the antiferromagnetic state at 9 K (Regmi et al., 2019). In the AFM phase an inner circular feature appears near the zone center, most clearly at 0 meV below 1, and along 2-3-4 two bands split very close to the Fermi level at low temperature, whereas only a single near-5 peak is observed in the paramagnetic phase (Regmi et al., 2019). The authors interpreted this as a magnetic-transition-driven reorganization consistent with a magnetic topological state, but they did not claim direct observation of a complete topological surface Dirac cone, a magnetic gap pinned at 6, or hinge modes (Regmi et al., 2019).
A more bulk-sensitive ARPES study using both soft-x-ray and VUV photons refined that picture. In the paramagnetic phase it identified a small three-dimensional bulk hole pocket at 7 together with a much larger, 8-independent, heavily hole-doped surface state 9, which was interpreted as a trivial termination-induced state rather than the desired topological Dirac surface state (Sato et al., 2020). On cooling below 00, the low-energy bulk spectrum developed a near-01 “M”-shaped band 02 within about 03 eV of 04, while the trivial surface band remained essentially unchanged (Sato et al., 2020). That behavior was taken as qualitative evidence for the predicted SOC-gapped band inversion at 05, but the authors were explicit that the experiment did not directly prove the axion state and that the actual topological Dirac cone might lie mostly above 06 (Sato et al., 2020).
STM/STS added a complementary real-space surface perspective. On reconstructed, partially Eu-terminated surfaces, spectroscopy found a spin-orbit-induced bulk gap of about 07 meV located only a few meV above the Fermi energy, together with in-gap surface-state features on the atomic surface (Gong et al., 2022). Temperature-dependent spectra were interpreted as showing a partial surface-state gap of about 08 meV below the antiferromagnetic transition, which decreases with increasing temperature but remains finite above 09 (Gong et al., 2022). The modeling required a strongly anisotropic, nodal gap function to reproduce the density of states, so the observed gap was described as partial rather than a simple isotropic massive Dirac gap (Gong et al., 2022).
Taken together, these measurements support four propositions. EuIn10As11 has an inverted low-energy electronic structure; magnetic order reconstructs the states nearest 12; surface states exist within the bulk gap; and the low-temperature surface spectrum is at least partially gapped (Regmi et al., 2019, Sato et al., 2020, Gong et al., 2022). They do not yet provide direct spectroscopic detection of hinge conduction, a fully resolved magnetic Dirac gap at the ideal chemical potential, or quantized axion electrodynamics.
5. Transport, carrier tuning, and thin-film realization
Transport studies repeatedly show that EuIn13As14 is usually not in the ideal insulating limit. Hall data in both bulk crystals and films consistently indicate hole-type carriers of order 15 (Sato et al., 2020, Karim et al., 2023). This has motivated a major line of work on Fermi-level tuning. In Ca16Eu17In18As19, isovalent Ca substitution from 20 to 21 shrinks both lattice constants, decreases the hole carrier density, lowers 22, and preserves the susceptibility maximum at the Néel transition together with the topological Hall effect observed in pristine EuIn23As24 (Yan et al., 2024). Because no ZFC/FC splitting was observed, the doped series was argued not to enter a spin-glass regime, in contrast to other substitution strategies (Yan et al., 2024). The interpretation is that Ca acts mainly as chemical pressure, shifting the Fermi energy while keeping the essential AFM structure, so that further Ca substitution may realize the axion-insulating state (Yan et al., 2024).
Thin-film synthesis addressed the same problem from a device perspective. Molecular-beam epitaxy on 25 sapphire stabilized 26-axis-oriented EuIn27As28 films when the substrate temperature reached 29C or above, suppressing competing zincblende phases (Karim et al., 2023). The films reproduced the bulk-like magnetic phenomenology, including in-plane easy-axis behavior, a spin-flop feature near 30 T at 31 K, and saturation around 32, but they remained 33-type with 34 and mobility 35 (Karim et al., 2023). Their magnetoresistance is negative up to 36 T, highly anisotropic below about 37 T, and nearly isotropic at higher field; the work presented these films primarily as a platform for future gating and magneto-optical studies rather than as a direct observation of axion transport (Karim et al., 2023).
Field-dependent transport in the broken-helix regime has also become a subject in its own right. A later study found a field-induced metamagnetic transition with large hysteresis in magnetoresistance and a particularly sharp upturn when the field is tilted by 38 from the 39-axis (Singh et al., 6 Aug 2025). Combining magnetization, Hall, and theory, that work argued that the field converts low-resistivity antiferromagnetic domain walls into high-resistivity domain walls by reducing the interaction area of As-40-derived Fermi-surface sheets across the wall (Singh et al., 6 Aug 2025). This places EuIn41As42 among the comparatively rare topological-magnet candidates in which domain-wall transport is itself a leading experimental variable.
6. Spin-space symmetry, exchange-driven response, and collective dynamics
Recent work has broadened the significance of EuIn43As44 beyond static topological classification. A spin-space-symmetry analysis of the helical and broken-helical phases showed that both phases support an out-of-plane odd-wave order in momentum space, characterized by
45
with a single unpolarized nodal plane at 46 (Pari et al., 2024). Only the broken-helical phase admits an additional in-plane 47-wave order, which appears in one in-plane spin component and is protected by four nodal planes (Pari et al., 2024). In DFT without SOC, the helical phase showed 48-only spin splittings up to 49 meV, whereas the broken-helical phase showed in-plane splittings up to 50 meV and reduced odd-wave 51 splitting up to 52 meV (Pari et al., 2024).
Within the same framework, EuIn53As54 was predicted to exhibit a non-relativistic linear Edelstein effect generated by magnetic exchange alone rather than SOC. Without SOC, the only surviving response tensor element in both helical phases is the out-of-plane intraband component, so an electric field along 55 induces a spin density along 56 (Pari et al., 2024). The computed magnitude of this response differs strongly between the two phases: over 57 eV, the dominant 58 in the helical phase is about five times larger than in the broken-helical phase (Pari et al., 2024). This was proposed as a transport-based diagnostic of the debated magnetic transition and as a way to distinguish these non-collinear phases from the amplitude-modulated collinear phases proposed by Donoway et al. (Pari et al., 2024).
The low-frequency spin dynamics of the broken helix reveal a similarly symmetry-controlled structure. Ultrafast optical polarimetry identified a long-lived Goldstone-like mode near 59 GHz and a more strongly damped optical mode near 60 GHz at 61 K (Liebman-Pelaez et al., 15 Jan 2025). In the field-dominated regime, where the in-plane field overcomes strain pinning, the Goldstone mode corresponds to nearly uniform spin precession and obeys 62, which the authors derived as the lowest symmetry-allowed order for the 63-symmetric broken helix (Liebman-Pelaez et al., 15 Jan 2025). When strain dominates, the zero-field mode is already gapped and follows
64
The same study reported local spin-flop fields 65 T and 66 T, a helix-to-fan field 67 T, and saturation near 68 T (Liebman-Pelaez et al., 15 Jan 2025). These results make clear that local strain is not a minor perturbation: it selects the realized magnetic symmetry and therefore can affect the topological protection inherited from the magnetic order.
The broader implication is that EuIn69As70 is not only a candidate axion material with an inverted band structure. It is also a non-collinear easy-plane magnet in which exchange-only spin textures, pseudo-Goldstone dynamics, and strain-selective symmetry breaking are experimentally accessible (Pari et al., 2024, Liebman-Pelaez et al., 15 Jan 2025). That combination explains why the material remains important even when the chemical potential is not yet ideally placed for a bulk-insulating axion response.