Growth of Altermagnetic α-MnTe Films: Substrate Variation and Surface Modification

Abstract: Manganese telluride (MnTe) in its hexagonal α-MnTe crystal structure has evolved as one of the altermagnet workhorse materials. The synthesis of MnTe thin films is highly relevant for both fundamental science and device applications. Here, we report on the epitaxial growth of MnTe thin films and heterostructures. The films are studied by X-ray and electron diffraction as well as soft X-ray angle-resolved photoemission spectroscopy. We demonstrate the ability to grow high-quality α-MnTe on various substrates, ranging from transparent band insulators, over topological insulators, metallic transition metal chalcogenides, to the van der Waals ferromagnet Fe$_3$GeTe$_2$. While insulating substrates are useful for transport experiments or optical spectroscopy, metallic topological surface states may trigger spintronic interface effects, such as spin-orbit torques. Metallic substrates, in general, are highly relevant to avoid charging at the insulating MnTe films in electron spectroscopy or microscopy methods. Lastly, ferromagnetic substrates will be of interest to control magnetization across the interface. In addition, we discuss the formation of superstructures on the MnTe(0001) surfaces, that emerge directly after growth, upon subsequent tellurium evaporation and after thermal treatment. This will be relevant in further studying the surface magnetic and electronic structure in α-MnTe.

• The paper establishes systematic growth protocols for $\alpha$-MnTe films on diverse substrates, including fluorides, topological insulators, and ferromagnets, enabling studies in spin electronics, spectroscopy, and surface magnetism.
• Interface bonding character, not lattice mismatch alone, governs the selection of the $\alpha$-phase, highlighting the importance of vdW interfaces.
• Reversible surface reconstruction on α-MnTe(0001) can be achieved through thermal treatments and Te dosing, offering promise for tailored magnetic and spintronic properties.

Overview and motivation

Altermagnetic $\alpha$-MnTe has become a central material for experimental studies of compensated collinear magnets with momentum-dependent spin splitting, evidenced by anomalous Hall effect measurements (Bey et al., 2024), X-ray magnetic circular dichroism (Krüger, 11 Apr 2025), and ARPES observations of Kramers-degeneracy lifting [(Tu et al., 2023)-style results reported in Krempaský et al., Nature 626, 517 (2024)]. While MBE growth on SrF$_2$(111) and InP(111) is well established, systematic studies of alternative substrates — needed for transport on insulators, spin–orbit torque heterostructures on topological metals, proximity coupling to ferromagnets, and charging-free spectroscopy on conductors — have been lacking. This work addresses that gap by establishing growth recipes for $\alpha$-MnTe on five substrates spanning insulating fluorides, a vdW topological insulator, topological platinum tellurides, and the vdW ferromagnet Fe$_3$GeTe$_2$, and by characterizing the surface reconstructions that form on the resulting $\alpha$-MnTe(0001) films.

Epitaxial growth across substrate classes

Films were grown by MBE from elemental effusion cells at an optimal flux ratio of BEP$_\mathrm{Te}$/BEP$_\mathrm{Mn}$ = 7, with thicknesses above 60 nm. Structural quality was assessed by ex-situ XRD and in-situ RHEED/LEED.

Substrate	Lattice mismatch	Properties	Optimal $T_\mathrm{sub}$
SrF$_2$ (111)	+1.2%	transparent insulator	440 °C
BaF$_2$0 (111)	−5.3%	transparent insulator	450 °C
Bi$_2$1Te$_2$2/BaF$_2$3 (0001)	−5.3%	vdW topological insulator	~300 °C
PtTe$_2$4/Pt$_2$5Te$_2$6 (0001)	+4.0%	vdW topological metal	~500 °C
Fe$_2$7GeTe$_2$8 (0001)	+4.0%	vdW ferromagnet, metallic	~410 °C

SrF$_2$9: the reference case. Growth on SrF$\alpha$0(111) yields pure hexagonal $\alpha$1-MnTe at all tested temperatures, with streaky RHEED patterns and sharp LEED spots consistent with layer-by-layer growth of large ordered domains. Substrate temperature affects only crystallographic quality, not the grown phase.

BaF$\alpha$2: phase selection by temperature and interface chemistry. On BaF$\alpha$3(111), which shares SrF$\alpha$4's chemical character but has a −5.3% mismatch, the outcome depends strongly on $\alpha$5. At 340 °C only the cubic $\alpha$6-MnTe (111) reflection appears; at 450 °C both $\alpha$7- and $\alpha$8-phase peaks are present. The authors interpret this as interface-driven nucleation of the better lattice-matched cubic phase near the substrate, with a temperature-driven transformation to the hexagonal phase in upper layers — supported by LEED observation of the $\alpha$9-MnTe(0001) $_3$0 reconstruction at the surface. Notably, inserting a few monolayers of vdW Bi$_3$1Te$_3$2 between BaF$_3$3 and MnTe eliminates the $_3$4-phase entirely: despite essentially unchanged lattice mismatch and even lower growth temperature (~300 °C), the film grows purely in the $_3$5-phase. This is a strong result: it shows that interface bonding character, not mismatch magnitude alone, controls phase selection — vdW-type interfaces tolerate large mismatch where ionic/covalent bonding does not.

Platinum tellurides: topological metallic substrates. On PtTe$_3$6(0001) (and equivalently Pt$_3$7Te$_3$8, whose surface converts to Te-rich PtTe$_3$9 under the Te-rich flux), epitaxial $_2$0-MnTe grows at all tested temperatures, with best quality at ~500 °C. RHEED indicates layer-by-layer growth; LEED spots are slightly broader than on the fluorides, indicating smaller mean island size — a limitation that carries over into slightly broadened ARPES linewidths.

Fe$_2$1GeTe$_2$2: altermagnet on a ferromagnet. Growth succeeds at ~410 °C with the sharpest LEED spots of any substrate in the study, indicating the largest terraces. Two implications follow directly. First, the combination of metallic conductivity (avoiding charging in electron spectroscopy/microscopy) and ferromagnetism (Curie temperature ~220–230 K) makes this system a platform for studying magnetic proximity effects on the altermagnetic order. Second, and unexpectedly, these films are the only ones lacking the $_2$3 surface reconstruction, discussed below.

Surface superstructures and their reversible control

All films except those on Fe$_2$4GeTe$_2$5 exhibit a $_2$6 reconstruction on $_2$7-MnTe(0001), making half-order diffraction spots a practical in-situ fingerprint of successful $_2$8-phase growth. Photon-energy-dependent XPS of the Te-4d region reveals, in addition to the bulk component ($_2$9), two surface doublets ($\alpha$0, $\alpha$1), proving at least two chemically distinct Te species within the $\alpha$2 unit cell. The authors favor a Te-deficient surface layer over a weakly bound Te adlayer interpretation, based on the reconstruction's thermal stability up to at least 500 °C; a possible atomic model is a kagome-like Te vacancy layer by analogy to MnAs(0001) [Ouerghi et al., PRB 74, 155412 (2006)], though this remains speculative.

The superstructure sequence is fully controllable and reversible via Te dosing and annealing:

Treatment	Resulting structure
As-grown / anneal ≥350 °C	$\alpha$3
Te deposition ~1 min at 80 °C	vacancy-healed $\alpha$4
Further Te deposition	$\alpha$5, then bulk-like Te adlayers
Anneal 250 °C	$\alpha$6
Anneal ~300 °C	metastable $\alpha$7
Anneal slightly above 300 °C	intermediate $\alpha$8

This demonstrates deterministic control over the surface atomic structure of an altermagnet — directly relevant given theoretical predictions that $\alpha$9-MnTe(0001) hosts metallic surface states capable of producing emergent anomalous Hall responses. The open questions here are concrete: the exact atomic arrangement of each superstructure is undetermined (a quantitative LEED-$_\mathrm{Te}$0 analysis is proposed but not performed), the $_\mathrm{Te}$1 termination is only marginally stable in a narrow temperature window, and the absence of $_\mathrm{Te}$2 on Fe$_\mathrm{Te}$3GeTe$_\mathrm{Te}$4 is unexplained. The authors' tentative explanation — larger terraces allow Te atoms to heal vacancies during growth — is consistent with, but not proven by, the LEED spot sharpness.

Electronic structure by soft X-ray ARPES

Bulk-sensitive SX-ARPES at the P04 beamline (DESY) confirms the expected electronic structure of the films. On SrF$_\mathrm{Te}$5, sharp valence bands are resolved along $_\mathrm{Te}$6–K, with the altermagnetic spin splitting at the valence band maximum clearly visible; the closed film on PtTe$_\mathrm{Te}$7 shows the same dispersion with modestly broadened linewidths, tracking its slightly inferior structural quality. Angle-integrated spectra show no Fermi edge and a valence band maximum ≈40 meV below $_\mathrm{Te}$8, consistent with p-doped semiconducting $_\mathrm{Te}$9-MnTe. For closed films thicker than 50 nm on metallic substrates, absence of a Fermi edge confirms full coverage and negligible substrate contribution.

By contrast, a poorer film exhibiting island-growth RHEED signatures exposes substrate areas, yielding a clear Fermi edge and Pt-4f core-level signal — demonstrating how morphological defects produce parasitic metallic signals that could confound spectroscopic studies of predicted MnTe surface states. The $_\mathrm{Mn}$0 reconstruction itself leaves the bulk band structure unaffected in these measurements, as expected for bulk states, though its influence on surface states remains untested.

Limitations and open questions

Several caveats bound the conclusions. The atomic structures of all observed superstructures remain unknown, resting on indirect evidence (thermal stability, analogy to isostructural systems). The mechanism suppressing $_\mathrm{Mn}$1 formation on Fe$_\mathrm{Mn}$2GeTe$_\mathrm{Mn}$3 is conjectural. No magnetic characterization of the films themselves (e.g., Néel order, domain structure, or proximity-induced modifications on Fe$_\mathrm{Mn}$4GeTe$_\mathrm{Mn}$5) is presented; the ferromagnetic-substrate functionality is asserted as a prospect rather than demonstrated. Film thicknesses above 60 nm preclude conclusions about ultrathin-limit behavior or interfacial strain effects analogous to those reported for GaAs substrates. Finally, whether the controlled surface terminations modify the predicted metallic surface states is left as an explicitly open experimental question.

Conclusion

This work extends the epitaxial parameter space of altermagnetic $_\mathrm{Mn}$6-MnTe from conventional fluoride substrates to vdW topological insulators, topological metals, and a vdW ferromagnet, demonstrating relaxed $_\mathrm{Mn}$7-phase growth across mismatches from −5.3% to +4.0% and identifying interface bonding — not mismatch magnitude — as the decisive factor in phase selection. Combined with the demonstrated reversible control of $_\mathrm{Mn}$8-MnTe(0001) surface superstructures and SX-ARPES confirmation of the altermagnetic band splitting, the study provides both reproducible growth recipes and a surface-preparation toolkit for forthcoming investigations of surface-state magnetism, spin–orbit torque devices, and magnetic proximity effects in altermagnetic heterostructures.