Epitaxial growth of 2D manganese dibromide thin films on Au(111) with buffering overlayers

Abstract

As promising candidates for two-dimensional (2D) magnetic semiconductors, layered transition metal halides (TMHs) have attracted increasing interest due to their diverse insulating bandgaps and magnetic properties. Here, we report the epitaxial growth of atomically thin MnBr₂ films on Au(111) by molecular beam epitaxy (MBE). Scanning tunneling microscopy (STM) reveals the initial formation of different buffering layers with distinct atomic structures on the Au(111) surface, followed by the subsequent growth of MnBr₂ layers. X-ray photoelectron spectroscopy (XPS) confirms that Mn atoms are in different coordination environments in the buffering layers (MnBr_x, x < 2) from MnBr₂ crystals. The electronic properties of MnBr₂ thin films with different thicknesses are investigated by scanning tunneling spectroscopy (STS) and density functional theory (DFT) calculations, showing that the electronic band structures are nearly independent of the thickness and the Au(111) substrate. This study provides a better insight into the growth behavior and interfacial properties of 2D layered TMHs on metal substrates.

---

Introduction

Two-dimensional (2D) materials, such as semi-metallic graphene and semiconducting MoS₂, have been booming in the last two decades as promising candidates for next-generation electronic devices.^1–4 Notably, the integration of a semiconducting bandgap with long-range magnetic ordering in a 2D crystal is coveted, as it will enable the manipulation of both charge and spin for exotic magnetoelectric properties.^5–7 Layered transition metal halides (TMHs), including trihalides (MH₃), dihalides (MH₂), and other stoichiometries, represent a prototypical group of van der Waals magnetic semiconductors.^8–10 Intensive theoretical calculations have predicted diverse spin textures in 2D TMHs, such as ferromagnetism (FM), antiferromagnetism (AFM), helical magnetic structure, and skyrmions.^8,11–14 Recently, many of them have been experimentally verified, including CrI₃,¹⁵ CrBr₃,¹⁶ FeCl₂,¹⁷ FeBr₂,¹⁸ NiCl₂,¹⁹ NiI₂,²⁰ CoCl₂,²¹ MnI₂,²² and so on.

Manganese dihalides have been predicted to exhibit antiferromagnetic direct exchange with extremely large magnetic moments (e.g., ∼4.5–5.0μ_B per Mn atom) as Mn ions are in the half-filled high-spin d⁵ state.^14,22,23 Manganese dihalides have been predicted to exhibit AFM direct exchange with extremely large magnetic moments (e.g., ∼4.5–5.0μ_B per Mn atom) as Mn ions are in the half-filled high-spin d⁵ state.^14,22–25 In monolayer MnBr₂, first-principles calculations suggest a noncollinear AFM ground state with flexible orientations due to its small in-plane magnetic anisotropy energy (MAE).^26,27 Intriguingly, in an In₂Se₃/MnBr₂/In₂Se₃ heterostructure, the sandwiched MnBr₂ monolayer can be reversibly switched between the AFM and FM states by controlling the ferroelectric polarization direction of α-In₂Se₃.²⁸ Moreover, it is possible to realize an FM Mn₃Br₈ monolayer by introducing 1/4 Mn vacancies into MnBr₂.²⁹ Another theoretical study suggests spontaneous valley polarization and the anomalous valley Hall effect in MnBr.³⁰ Experimentally, the antiferromagnetism in MnBr₂ bulk was investigated by neutron diffraction in 1958,^31,32 but the realization of monolayer MnBr₂ was reported only recently on graphene/Ir(111).³³ Intensive experimental investigations are required for the validation of these Mn-based halides with remarkable potential in spintronic devices.

In this study, we report the successful fabrication of atomically thin MnBr₂ films on Au(111) by molecular beam epitaxy (MBE). Systematic investigations of the growth behaviors were carried out by scanning tunnelling microscopy (STM) combined with X-ray photoelectron spectroscopy (XPS). Initially, different buffering layers (MnBr_x, x < 2) form on the Au(111) surface with different atomic densities and configurations. Atomically thin MnBr₂ layers subsequently grow upon the buffering layers, where the interference from the metal substrate could be largely reduced. The electronic properties are further investigated by scanning tunnelling spectroscopy (STS) combined with first-principles calculations. This study represents a pioneer experimental exploration of 2D Mn-based halides due to their potential in device applications.

Results and discussion

Fig. 1a shows the optimized structure of bulk MnBr₂ by first-principles calculations (Methods). It exhibits a layered structure similar to that of MoS₂, where each monolayer contains Br–Mn–Br triple atomic layers in an energetically favorable trigonal (1T) configuration.^23,29 The lattice parameters are a = b = 3.85 Å and c = 6.24 Å, consistent with previously reported values.³⁴ More lattice parameters of the atomic models are summarized in Table S1 in the SI. As the thickness is reduced to a monolayer (ML) and a bilayer (BL), the in-plane lattice parameter remains nearly constant (3.84–3.85 Å), while the out-of-plane lattice parameter c increases to 6.46 Å in BL MnBr₂. Fig. 1b–d show the spin-polarized densities of states (DOSs) of ML, BL, and bulk MnBr₂ calculated based on both the Perdew–Burke–Ernzerhof (PBE) and Heyd–Scuseria–Ernzerhof 2006 (HSE06) methods. It is well known that the PBE method severely underestimates the bandgap of semiconductors and insulators, and an additional Hubbard U correction could achieve a more accurate description of the electronic structure,²⁶ while the HSE06 method is more accurate by introducing the hybrid Hartree–Fock (HF) density function.^35,36 All these three methods predict that MnBr₂ possesses a semiconducting bandgap independent of the layer number,³⁷ with values of 1.2 eV (PBE), 2.1 eV (PBE+U) and 3.6 eV (HSE06), respectively. The highly asymmetric distributions of the spin-up and spin-down DOSs could be attributed to the extremely large magnetic moments of Mn atoms.^23,37

Fig. 1 Atomic model and DOS of MnBr₂. (a) Top and side views of bulk MnBr₂ optimized with DFT. (b and c) Spin-up and spin-down DOSs of ML, BL, and bulk MnBr₂ calculated using the (b) PBE, (c) PBE+U (U = 2 eV), and (d) HSE06 methods, respectively. The dashed lines indicate the conduction and valence band edges.

To fabricate MnBr₂ thin films, we used MnBr₂ powder (Thermo Scientific, 99.9%) as the precursor and Au(111) as the substrate. It is found that the substrate temperature and atomic density/coverage are two key factors determining the atomic structures of the as-fabricated compounds (Fig. S1 and S2). After optimizing the growth conditions, atomically flat thin films were obtained by holding the substrate at 100 °C–180 °C during the deposition, followed by post-annealing at 100 °C–220 °C. The compounds were completely desorbed from the Au(111) surface when the annealing temperature was higher than 250 °C.

Initially, buffering layers (MnBr_x) with distinct atomic structures were observed on Au(111) before the formation of 1T-MnBr₂. Three typical examples, namely buffering layers I, II and III, are demonstrated in Fig. 2. Buffering layer I (BuffL-I) was most frequently observed when the substrate was held at ∼100 °C during deposition, while buffering layers II (BuffL-II) and III (BuffL-III) became dominant when the substrate temperature increased to 180 °C. The heights of these buffering layers are ∼1.3–1.6 Å as revealed by the line profiles shown in the inset of Fig. 2a and further illustrated in Fig. S4–S7. The slight variations are due to the scanning bias and tip conditions. This value is much smaller than that reported for other MH₂ monolayers grown on metal substrates, which are above 3.0 Å.^17,38,39 The BuffL-I layer is composed of distorted triangular and hexagonal rings, where the atoms are loose-packed with short-range ordering (Fig. 2c). The BuffL-II region becomes more ordered with close-packed hexagonal atomic structures (Fig. 2d–f), where atom vacancies exist. Additionally, irregular triangle patterns with dimensions of several nanometers are imposed on the herringbone superstructures (Fig. 2f), which might be attributed to a quasiperiodic moiré pattern.³⁸ While in the BuffL-III layer, the atomic structure is highly ordered and almost defectless (Fig. 2g–i). Concomitantly, an incommensurate moiré superstructure with a wave-shape modulation (Fig. S4) is observed, indicating the high flexibility of the MnBr_x buffering layers. As summarized in Table S2, the surface atomic density is only ∼5.6 atoms per nm² in Buff-I, which increases to ∼6.5 atoms per nm² in BuffL-II and ∼6.6 atoms per nm² in BuffL-III. The lattice constants of BuffL-II and BuffL-III are almost the same, which are 4.0 ± 0.2 Å.

Fig. 2 Different buffering layers formed in the sub-monolayer regions: (a–c) BuffL-I, (d–f) BuffL-II, and (g–i) BuffL-III. In the left column, the large-scale images show the coexistence of the buffering layers and the bare Au(111) surface, where herringbone superstructures are visible all over the surface. In the middle column, the enlarged STM images reveal that the length (L, along yellow lines) and width (W, along green arrows) vary with the buffering layers. In the right column, the highly resolved STM images demonstrate their different atomic configurations, where incommensurate moiré patterns are observable in BuffL-II and III. (a) V = 0.4 V and I = 400 pA; (b) V = −0.4 V and I = 500 pA; (c) V = 1.0 V and I = 500 pA; (d–f) V = −1.0 V and I = 100 pA; (g) V = 1.0 V and I = 100 pA; (h) V = −0.4 V and I = 400 pA; and (i) V = −0.05 V and I = 400 pA.

From the large-scale STM images, herringbone-like superstructures analogous to clean Au(111) (Fig. S1) are visible atop all these buffering layers. This indicates that the reconstructed superstructure of the underneath Au(111) surface^40,41 is preserved upon the formation of buffering layers due to the relatively weak interfacial interactions between them. Interestingly, it is noted that the herringbone superstructure observed atop BuffL-III is extremely long along the Au(111) highest symmetric directions, e.g., [1̄1̄2] in Fig. 2g. The longest one can extend over 300 nm (Fig. S4), which is much longer than the average length (L) obtained on clean Au(111) (below 50 nm). Moreover, the width (W) along the [11̄0] direction (red arrows) is as large as 7.2 nm for BuffL-III, as revealed by the line profile shown in Fig. 2h. This is also larger than that for clean Au(111), whose width is 6.6 nm corresponding to the (23 × √3) reconstructed structure.^40,41 Considering the Au(111) lattice constant of 2.88 Å (a_Au), the periodicity of 7.2 nm corresponds to 25a_Au. Such reconstruction of the Au(111) herringbone also occurs upon the formation of BuffL-I and BuffL-II. In the former (Fig. 2b), L is over 50 nm, and W remains the same as that of Au(111); meanwhile, in the latter (Fig. 2e), L can reach over 100 nm and W becomes the same as that of Buff-III. Such reconstructions might be induced by the different interfacial strain between the different MnBr_x adlayers and the Au(111) substrate. We further analyze the lattice orientations of the buffering layers with respect to the substrate. As shown in Fig. 2c, f and i, the red arrows represent the [11̄0] direction of Au(111), and the black arrows represent the buffering layers. The intersection angles are found to be different, with α₁ = 50° ± 2°, α₂ = 69° ± 2°, and α₃ = 29° ± 2° for BuffL-I, II and III respectively.

Similarly, buffering layers have been previously observed in other MBE-grown MH₂ crystals, which distinctly differ from Br adlayers.^38,42 For example, a nonstoichiometric Cr–I buffering layer has been observed in the MBE-fabricated CrI₂/NbSe₂ heterostructure,³⁸ and also a NiBr_x overlayer has been observed for NiBr₂ films grown on Au(111).⁴² Such buffering layers could be composed of partially dehalogenated MH₂ compounds (or molecules) formed by two possible processes: (1) MH₂ molecules could partially dehalogenate during thermal evaporation (e.g., the evaporation temperature of the MnBr₂ source is as high as 270 °C–320 °C in this study) and (2) the noble metal substrates could promote dehalogenation due to their catalytic activity. Unfortunately, precise determination of the atomic structures of such buffering layers is very difficult due to their amorphous and nonstoichiometric structure. The chemical stoichiometry of the MnBr_x buffering layers will be further investigated by XPS as discussed below.

As the coverage increases, monolayer and multilayer MnBr₂ films form sequentially on the buffering layers. In Fig. 3a, the overall MnBr₂ coverage is ∼0.7 layer (L), with ML, BL, and buffering layer regions observed at the same time on the Au(111) surface (with Br adatoms). While in Fig. 3b, the MnBr₂ thickness is ∼2.2 L, and no herringbone is visible. The line profile shown in Fig. 3d corresponds to the white line in Fig. 3a. It reveals that the height of a single MnBr₂ layer is ∼ 3.3–3.6 Å for both ML and BL MnBr₂ (see also Fig. S6 and S7), which is consistent with the layer thicknesses of FeCl₂¹⁷ and CrI₂³⁸ previously determined by STM studies. This value is largely distinct from that of ∼1.6 Å for the buffering layers. The height difference between the ML-MnBr₂ and Au(111) regions of ∼ 5.1 Å is the combined thickness of a single MnBr₂ layer and a buffering layer (Fig. S7). Fig. 3c shows a typical image recorded in the BL MnBr₂ regions to show the atomic structure, which is the same as the ML region (Fig. S7). The unit cell is highlighted by a blue rhombus, corresponding to the hexagonal pattern shown in the inset obtained by fast Fourier transform (FFT). Due to the absence of herringbone superstructures and moiré modulation, only six sharp spots are visible in this FFT pattern. In contrast, the FFT patterns obtained from buffering layers exhibit complex features (Fig. S5). The lattice constant is determined to be 3.9 ± 0.1 Å, which is close to 3.85 Å obtained from DFT optimization (Fig. 1a and Table S2).²⁹

Fig. 3 The growth of MnBr₂ layers on the MnBr_x buffering layers. (a and b) Large-scale morphological images revealing the formation of 0.7 L and 2.2 L MnBr₂ atop the buffering layers, respectively. (c) An atomically resolved STM image recorded in the BL region, where the inset shows a hexagonal pattern obtained by FFT. (d) The line profile corresponds to the blue line in panel (a), revealing the different heights of the MnBr₂ layers and MnBr_x buffering layers. (e) A high-resolution STM image recorded at the edge of ML MnBr₂ and Buff-I. (f) dI/dV spectra of ML, BL, and TL MnBr₂ (setting points: V = −2.5 V, I = 500 pA, f = 877 Hz, and V_r.m.s. = 20 mV). (a–c) V = 1.0 V and I = 100 pA; (e) V = −1.0 V and I = 200 pA.

Obviously, the MnBr₂ layers coexist with different buffering layers, as revealed by high-resolution STM images recorded in the edge regions, as shown in Fig. 3e and Fig. S6. Interestingly, we find that the intersection angle β between the MnBr₂ (blue line) and Au(111) (red arrow) lattices is fixed at 62° ± 2° over the whole sample, which does not change with the adjacent buffering layers. To realize such epitaxial growth behavior, one possible mechanism is that the underneath buffering layers undergo reconstructions to align with the top MnBr₂ layers. Another possibility is that the buffering MnBr_x layers directly converse into the MnBr₂ layers, which could be excluded by further XPS analyses discussed later.

Fig. 3f shows STS spectra obtained for the MnBr₂ films with various thicknesses, which all manifest a typical ‘U’ shape for semiconductors. There is no visible signature from Au(111) in the gaps even for ML MnBr₂ (black curve), suggesting that the buffering layer largely reduces the electronic interference from the metal substrate. As the bias voltages were applied to the sample, the conduction band minima (CBM) are determined to be 0.16 ± 0.03 eV for all the thicknesses. The valence band maximum (VBM) is −1.92 ± 0.05 eV for the ML region, which slightly shifts to −2.0 ± 0.05 eV for the TL. Thus, the bandgaps are nearly thickness independent, consistent with theoretical calculations. The value of ∼2.1–2.2 eV is comparable to the 2.1 eV predicted by the PBE+U (U = 2 eV) method, but much smaller than the 3.6 eV predicted by the HSE06 method. The difference between these two methods might be due to the screening effect from the metallic substrate.⁴³ The thickness-independent behavior has been observed in CrI₂⁴⁴ and ReS₂,⁴⁵ in contrast to most other 2D semiconductors, e.g., MoS₂⁴³ and WSe₂.⁴⁶ One possible reason is that the interlayer space of these materials is relatively larger, and thus adjacent layers are largely decoupled in terms of electronic structure.^44,45

In the following, XPS measurements were carried out to determine the chemical components, stoichiometries, and thickness of the MnBr_x compounds. Fig. 4a–c show the core level spectra recorded in Au 4f, Br 3d and Mn 2p regions with different coverages, namely clean Au(111), 0.6 L BuffL, 0.9 L BuffL, 1.1 L MnBr₂ on BuffL, and 2.1 L MnBr₂ on BuffL (confirmed by STM). As the coverage increases, the intensities of the Au 4f peaks decrease gradually while the Br 3d and Mn 2p peaks increase linearly. In Fig. 4a, the binding energies (BEs) of Au 4f_5/2 and 4f_7/2 for the clean Au(111) substrate are 87.7 eV and 84.0 eV, respectively, which remain constant after the growth of the MnBr_x compounds. This indicates that no Au alloy or strong chemical bond forms at the interface, which is consistent with the STM observations. From Fig. 4b, it is found that the Br 3d_3/2 and 3d_5/2 peaks appear at 69.6 eV and 68.6 eV, respectively, in 0.6 L and 0.9 L BuffL, which shift by 0.5 eV to higher BE in the thicker films with the formation of 1.1 L and 2.1 L MnBr₂ on BuffL. The shift could be attributed to their different coordination environments or interfacial interactions in thicker films.

Fig. 4 XPS measurements of samples with different coverages: clean Au(111), 0.6 L BuffL, 0.9 BuffL, 1.1 L MnBr₂ on BuffL, and 2.1 L MnBr₂ on BuffL. (a) BEs of Au 4f core levels remain constant, while their intensities gradually decrease. (b) From the BuffL to MnBr₂ layers, Br 3d peaks shift slightly to the higher BE side by 0.5 eV. A small amount of Br residual remains on the clean Au(111) surface after sputtering. (c) Four main peaks from Mn 2p_3/2 and 2p_1/2 and their shape-up satellites can be identified with the overlap of the Au 4p_1/2 peak.

The signals from Mn 2p core levels are much more complicated, as shown in Fig. 4c, where the Au 4p_1/2 peak that appears at 642.6 eV strongly overlaps with the Mn 2p_3/2 peaks in the MnBr_x compounds. Four main peaks from Mn 2p_3/2 and 2p_1/2 and their shake-up satellites^47,48 can be clearly identified from the thick films, as the signals from Au 4p_1/2 are suppressed. Careful analyses were carried out to fit the multiplet splitting of the Mn peaks, which arises from the coupling between unpaired electrons in transition metals.⁴⁰ The fitting parameters and full XPS survey can be found in Table S3. In 0.6 L and 0.9 L BuffL, both the Mn 2p_3/2 multiplets are composed of three main components (blue) with the highest one centered at 641.2 eV. While in the thick films, the BuffL multiplets remain, and new multiplets assigned to MnBr₂ arise. In contrast, the MnBr₂ multiplets contain four main components with the highest one located at 641.5 eV. These binding energies are very close to that reported for Mn²⁺.^47,48 However, their different components strongly suggest that the Mn atoms are in different coordination environments, e.g., different valence and/or atomic configurations. We employed quantitative analysis to determine the stoichiometries. As summarized in Table S4, the areas (A) of the peaks increase linearly with the coverages, and the stoichiometric ratio between Mn and Br elements can be determined by comparing their normalised intensity (I_normalised). Based on the method provided by the Manual,⁴⁹ it yields a Br : Mn ratio (x) of 1.83 ± 0.07 for 0.6 L BuffL, which gradually increases to 1.99 ± 0.04 for 2.1 L MnBr₂ on BuffL (Table S4). This suggests that the stoichiometry of the top layers is very close to the ideal value of 2. As discussed in the SI (section V), we confirm that the BuffL signals (blue peaks in Fig. 4b and c) persist upon the subsequent formation of MnBr₂ layers, excluding the direct growth of MnBr₂ on Au(111). As the stoichiometry of BuffL and thus the element valence only slightly differs from MnBr₂, the changes in the core levels could be mainly due to their different atomic configurations, consistent with STM observations.

As shown in Fig. 5a, a (3 × 3) MnBr₂ supercell was placed on a (4 × 4) Au(111) supercell for the calculations (see the Methods section). Fig. 5b and c show the differential charge density Δρ(z) (black curves) and the cumulative charge transfer ΔQ(z) (red curves) of the heterostructures with ML and BL MnBr₂, respectively. In the ΔQ(z) curve, a distinct peak observed at the interface represents the total amount of charges (ΔQ_T) transferred between the substrate and the adlayer. This shows that the electrons transfer from MnBr₂ to Au(111) for both cases, which are 0.19e and 0.18e per unit, respectively. These could be understood as the calculated work function (WF) of Au(111) is 5.53 eV, which is higher than that of 4.90 eV of MnBr₂ (Fig. 5f and Fig. S9). The side views of the differential charge density isosurfaces are given in the insets, where the yellow (cyan) represents electron accumulation (depletion). It is interesting to find that in the BL case, the charge redistributions only occur in the 1^st layer, which mainly accumulate at the MnBr₂/Au(111) interfaces (Fig. 5c). By contrast, the distribution is more dispersive in the ML MnBr₂ (Fig. 5b). Bader charge analysis also reveals that the Au(111) substrates obtain electrons from the ML and BL MnBr₂, which are 0.068e and 0.061e per unit, respectively.

Fig. 5 Calculated charge redistributions and PDOSs of ML and BL MnBr₂ on Au(111). (a) Top view of the atomic model. (b and c) The plots of the plane-averaged Δρ(z) (black) and ΔQ(z) (red) for the ML and BL MnBr₂ on Au(111), respectively. The insets show the side views of the differential charge density isosurfaces (isovalue: ±0.0002 e Å⁻³), where the cyan (yellow) represents electron accumulation (depletion). (d) PDOSs of the whole ML-MnBr₂/Au(111) system and the Au(111) substrate. (e) PDOS of the MnBr₂ layer extracted from (d). (f) The calculated WF of ML-MnBr₂/Au(111). PDOS of the BL-MnBr₂/Au(111) system: (g) the total PDOS of Au(111), (h) the 1^st MnBr₂ layer, and (i) the 2^nd MnBr₂ layer.

The PDOSs of the ML and BL MnBr₂ with the Au(111) substrate are shown in Fig. 5d and 5g, respectively. For both cases, the total PDOS is spin polarized (light blue and red), while the substrate is not spin polarized (black and grey). The PDOSs of each MnBr₂ layer are shown in Fig. 5e for the ML MnBr₂/Au(111) system and Fig. 5h and i for the 1^st and 2^nd layers of the BL MnBr₂/Au(111) system, respectively. Obviously, all the MnBr₂ layers exhibit prominent spin polarization, which only changes slightly after introducing the substrate. Furthermore, only small amounts of DOS appear in the gap of the 1^st MnBr₂ layer (Fig. 5e and h), while the 2nd layer preserves the characteristics of the free-standing ones (Fig. 5i). Moreover, our PBE-derived results also show the same trend (Fig. S10). These observations are confirmed by the spin-polarized band structures of the ML-MnBr₂/Au(111) system with little hybridization (Fig. S11). Therefore, this indicates that both the electronic and magnetic properties of the MnBr₂ layers could be well preserved by using Au(111) as a substrate, and the formation of buffering layers could further reduce the interference from the substrate.

Conclusions

In summary, we systematically investigate the growth behaviors of manganese bromide thin films on Au(111) by using STM combined with XPS. Initially, MnBr_x (x < 2) buffering layers with various atomic densities and configurations are observed in sub-monolayer regions. Subsequently, MnBr₂ layers are grown atop the buffering layers, exhibiting a nearly constant bandgap, namely ∼2.1–2.2 eV for the ML to the TL. Such thickness-independent behavior is confirmed by DFT calculations. We also find that the distribution of the spin-polarized density of states of the MnBr₂ monolayer is well preserved on the Au(111) substrate. This suggests that the impact of the Au(111) substrate is relatively weak, which could be further reduced by introducing a spacer, e.g., a MnBr₂ monolayer or buffering layers. Leveraging the predicted large magnetic moment and antiferromagnetic properties of MnBr₂, this study could inspire more research interest in 2D Mn-based halides for high-sensitivity, high-efficiency spintronic devices.

Methods

Calculations based on density functional theory (DFT) were performed using the Vienna ab initio Simulation Package (VASP).⁵⁰ The exchange correlation interaction was described by the Perdew–Burke–Ernzerhof (PBE) functional,⁵¹ and the interaction between valence electrons and ion nuclei is described by the projector-augmented wave potential (PAW) method.⁵² The Grimme's D3 correction was used to consider the van der Waals (vdW) interactions. The cutoff value for the plane wave basis set was set to 400 eV, and to eliminate the physical interactions caused by periodic boundary conditions, the vacuum spaces were larger than 15 Å. For the projected density of states (PDOS) calculations, the k-point meshes were set as follows: 18 × 18 × 1 for monolayer and bilayer systems and 18 × 18 × 14 for the bulk system. A Γ-centered k-mesh of 7 × 7 × 1 was used for geometry optimization and electronic structure calculations in the Brillouin zone (BZ). The Heyd–Scuseria–Ernzerhof hybrid functional (HSE06)⁵³ is also used to obtain a more accurate value of band gaps. The criteria for energy and atom force convergence are set to 1 × 10⁻⁶ eV and 0.01 eV Å⁻¹, respectively.

An ultra-high vacuum (UHV) STM (Createc) system interfaced with a custom-designed MBE (Fermi) chamber was used to prepare the samples. The background pressures in the LT-STM and MBE chambers were below 1 × 10⁻¹⁰ mbar and 3 × 10⁻¹⁰ mbar, respectively. A clean Au(111) surface was prepared by repeated Ar⁺ bombardment (1.0 kV, 1 × 10⁻⁶ mbar) followed by annealing at 400 °C. MnBr₂ powder (Thermo Scientific, 99.9%) was evaporated from a Knudsen cell at a deposition rate of 0.01–0.03 ML per minute. During the growth, the pressure was better than 2.5 × 10⁻⁸ mbar, and the Au(111) substrate was held at room temperature (RT) to 180 °C. In situ STM measurements were carried out using a Nanonis controller in constant current mode (sample bias) at 78 K. For dI/dV spectra, the tunneling current was obtained using a lock-in amplifier, with a modulation of 877 Hz and 20 mV.

X-ray photoemission spectroscopy (XPS) was performed at room temperature using a Thermo Fisher EscaLab Xi+ system integrated with an Al-K_α radiation source (hν = 1486.61 eV). The XPS peaks were fitted using the Avantage program, utilizing a Shirley background with the GL(30) line shape. The energy resolution was 0.8 eV.

Author contributions

Z. Z. W.: conceptualization, investigation, formal analysis, visualization, writing – original draft, and writing – review & editing; M. Y. S. and Y. D. W.: STM measurements and analysis; Z. P. J. and C. H.: STM measurements; Y. D. W. and Q. X. W.: XPS measurements and analysis; Y. J. Z. DFT calculations, supervision, and writing – review & editing; Y. L. H.: supervision, resources, writing – review & editing and project administration.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. Supplementary Information: Tables S1–S4, additional DFT calculations, and more experimental details. See DOI: https://doi.org/10.1039/d5nr03723a.

Acknowledgements

Y. L. H. acknowledges financial support from the National Natural Science Foundation of China (Grant No. 12350610236), the Natural Science Foundation of Fujian Province (2022J06035), and the start-up funding from the International Joint Institute of Tianjin University, Fuzhou. Q. W. acknowledges financial support from the National Natural Science Foundation of China (Grant No. 62304186).

References

1. K. S. Kim, J. Kwon, H. Ryu, C. Kim, H. Kim, E. K. Lee, D. Lee, S. Seo, N. M. Han, J. M. Suh, J. Kim, M. K. Song, S. Lee, M. Seol and J. Kim, Nat. Nanotechnol., 2024, 19, 895–906.
2. C. Liu, H. Chen, S. Wang, Q. Liu, Y. G. Jiang, D. W. Zhang, M. Liu and P. Zhou, Nat. Nanotechnol., 2020, 15, 545–557.
3. S. Bertolazzi, P. Bondavalli, S. Roche, T. San, S.-Y. Choi, L. Colombo, F. Bonaccorso and P. Samorì, Adv. Mater., 2019, 31, 1806663.
4. S. Das, A. Sebastian, E. Pop, C. J. McClellan, A. D. Franklin, T. Grasser, T. Knobloch, Y. Illarionov, A. V. Penumatcha, J. Appenzeller, Z. Chen, W. Zhu, I. Asselberghs, L.-J. Li, U. E. Avci, N. Bhat, T. D. Anthopoulos and R. Singh, Nat. Electron., 2021, 4, 786–799.
5. D. Kong, C. Zhu, C. Zhao, J. Liu, P. Wang, X. Huang, S. Zheng, D. Zheng, R. Liu and J. Zhou, Chem. Soc. Rev., 2024, 53, 11228–11250.
6. Z. Qu, C. Huang and E. Kan, Chin. Sci. Bull., 2023, 68, 3385–3396.
7. T. Dietl, A. Bonanni and H. Ohno, J. Semicond., 2019, 40, 080301.
8. X. Jiang, Q. Liu, J. Xing, N. Liu, Y. Guo, Z. Liu and J. Zhao, Appl. Phys. Rev., 2021, 8, 030401.
9. L. Xu, Z. Wang and C. Gao, Jpn. J. Appl. Phys., 2024, 63, 050805.
10. Y. Yao, X. Zhan, M. G. Sendeku, P. Yu, F. T. Dajan, C. Zhu, N. Li, J. Wang, F. Wang, Z. Wang and J. He, Nanotechnology, 2021, 32, 472001.
11. Q. Sun and N. Kioussis, Phys. Rev. B, 2018, 97, 094408.
12. S. Tomar, B. Ghosh, S. Mardanya, P. Rastogi, B. S. Bhadoria, Y. S. Chauhan, A. Agarwal and S. Bhowmick, J. Magn. Magn. Mater., 2019, 489, 165384.
13. X. Li, C. Xu, B. Liu, X. Li, L. Bellaiche and H. Xiang, Phys. Rev. Lett., 2023, 131, 036701.
14. A. S. Botana and M. R. Norman, Phys. Rev. Mater., 2019, 3, 044001.
15. B. Huang, G. Clark, E. Navarro-Moratalla, D. R. Klein, R. Cheng, K. L. Seyler, D. Zhong, E. Schmidgall, M. A. McGuire, D. H. Cobden, W. Yao, D. Xiao, P. Jarillo-Herrero and X. Xu, Nature, 2017, 546, 270–273.
16. W. Chen, Z. Sun, Z. Wang, L. Gu, X. Xu, S. Wu and C. Gao, Science, 2019, 366, 983–987.
17. X. Zhou, T. Jiang, Y. Tao, Y. Ji, J. Wang, T. Lai and D. Zhong, ACS Nano, 2024, 18, 10912–10920.
18. F. Xiang, N. Bisht, B. Da, M. S. Mohammed, C. Neiss, A. Görling and S. Maier, ACS Nano, 2024, 18, 18870–18879.
19. A. Aguirre, A. P. Solé, D. S. Polo, C. González-Orellana, A. Thakur, J. Ortuzar, O. Stesovych, M. Kumar, M. Peña–Díaz and A. Weber, Adv. Mater., 2024, 36, 2402723.
20. M. Amini, A. O. Fumega, H. González-Herrero, V. Vaňo, S. Kezilebieke, J. L. Lado and P. Liljeroth, Adv. Mater., 2024, 36, 2311342.
21. S. Kerschbaumer, S. E. Hadjadj, A. Aguirre-Baños, D. Longo, A. P. Solé, O. Stetsovych, A. E. Candia, P. Angulo-Portugal, D. Caldevilla and F. Choueikani, ACS Nano, 2025, 19(22), 20693–20701.
22. X. Cai, Z. Xu, S.-H. Ji, N. Li and X. Chen, Chin. Phys. B, 2021, 30, 028102.
23. V. V. Kulish and W. Huang, J. Mater. Chem. C, 2017, 5, 8734–8741.
24. G. H. Al-Hazmi, A. Zaman, N. Akhter, S. H. Alrefaee, P. Ahmad, T. Orlova, A. Nurmuhammedov, V. Tirth, A. Algahtani and N. Hadia, Int. J. Quantum Chem., 2025, 125, e70031.
25. J. Luo, G. Xiang, Y. Tang, K. Ou and X. Chen, J. Appl. Phys., 2020, 128, 113901.
26. X. Li, Z. Zhang and H. Zhang, Nanoscale Adv., 2020, 2, 495–501.
27. T. Guo, X. Xu, H. Zhang, Y. Xie, H. Yang, X. Rui, Y. Sun, X. Yao, B. Wang and X. Zhang, Appl. Phys. Lett., 2024, 124, 062404.
28. X. Zhang, B. Yang, X. Guo, Y. Liu, X. Han and Y. Yan, Appl. Surf. Sci., 2023, 614, 156201.
29. Y. Hu, S. Jin, Z. F. Luo, H. H. Zeng, J. H. Wang and X. L. Fan, Nanoscale Res. Lett., 2021, 16, 72.
30. Y. Wang, H. Sun, C. Wu, W. Zhang, S.-D. Guo, Y. She and P. Li, Phys. Rev. B, 2025, 111, 085432.
31. L. Zhuo, Y. Li, X. Wang, J. Lu, L. Qiao, D. Chen, J. Zheng, X. Xiong, X. Peng and J. Li, Mater. Res. Express, 2019, 6, 085046.
32. Y. Li, D. Chen, X. Dong, L. Qiao, Y. He, X. Xiong, J. Li, X. Peng, J. Zheng and X. Wang, J. Phys.: Condens. Matter, 2020, 32, 335803.
33. A. Safeer, O. Güleryüz, N. Atodiresei, W. Jolie, T. Michely and J. Fischer, arXiv, 2025, preprint, arXiv:2508.19694, https://arxiv.org/abs/2508.19694.
34. E. O. Wollan, W. C. Koehler and M. K. Wilkinson, Phys. Rev., 1958, 110, 638–646.
35. Y. Huang, T. Ma, Q. Y. Chen, C. Cao and Y. He, Sci. Rep., 2015, 5, 12750.
36. R. Pawar and A. A. Sangolkar, Comput. Theor. Chem., 2021, 1205, 113445.
37. G. C. C. de Jesus, R. D. C. Vasconcelos, L. B. do Vale, R. F. de Menezes, K. G. Sprenger and R. Gargano, Langmuir, 2025, 41, 5877–5883.
38. P. Li, J. Zhang, D. Zhu, C. Q. Chen, E. Yi, B. Shen, Y. Hou, Z. Yan, D. X. Yao, D. Guo and D. Zhong, Nano Lett., 2024, 24, 9468–9476.
39. Y. Wang, X. Zhao, L. Yao, H. Liu, P. Cheng, Y. Zhang, B. Feng, F. Ma, J. Zhao, J. Sun, K. Wu and L. Chen, Nat. Commun., 2024, 15, 10916.
40. C. Woll, S. Chiang, R. J. Wilson and P. H. Lippel, Phys. Rev. B: Condens. Matter, 1989, 39, 7988–7991.
41. J. V. Barth, H. Brune, G. Ertl and R. J. Behm, Phys. Rev. B: Condens. Matter, 1990, 42, 9307–9318.
42. D. Bikaljevic, C. Gonzalez-Orellana, M. Pena-Diaz, D. Steiner, J. Dreiser, P. Gargiani, M. Foerster, M. A. Nino, L. Aballe, S. Ruiz-Gomez, N. Friedrich, J. Hieulle, L. Jingcheng, M. Ilyn, C. Rogero and J. I. Pascual, ACS Nano, 2021, 15, 14985–14995.
43. Y. L. Huang, Y. Chen, W. Zhang, S. Y. Quek, C. H. Chen, L. J. Li, W. T. Hsu, W. H. Chang, Y. J. Zheng, W. Chen and A. T. Wee, Nat. Commun., 2015, 6, 6298.
44. L. Peng, J. Zhao, M. Cai, G.-Y. Hua, Z.-Y. Liu, H.-N. Xia, Y. Yuan, W.-H. Zhang, G. Xu, L.-X. Zhao, Z.-W. Zhu, T. Xiang and Y.-S. Fu, Phys. Rev. Res., 2020, 2, 023264.
45. S. Tongay, H. Sahin, C. Ko, A. Luce, W. Fan, K. Liu, J. Zhou, Y.-S. Huang, C.-H. Ho, J. Yan, D. F. Ogletree, S. Aloni, J. Ji, S. Li, J. Li, F. M. Peeters and J. Wu, Nat. Commun., 2014, 5, 3252.
46. W. Zhao, Z. Ghorannevis, L. Chu, M. Toh, C. Kloc, P.-H. Tan and G. Eda, ACS Nano, 2013, 7, 791–797.
47. E. S. Ilton, J. E. Post, P. J. Heaney, F. T. Ling and S. N. Kerisit, Appl. Surf. Sci., 2016, 366, 475–485.
48. M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Appl. Surf. Sci., 2011, 257, 2717–2730.
49. Thermo Fisher Scientific, Avantage Data Analysis Manual, Version 6, Thermo Fisher Scientific, Waltham, MA, 2023.
50. G. Kresse and J. Hafner, Phys. Rev. B:Condens. Matter Mater. Phys., 1993, 47, 558–561.
51. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
52. P. E. Blöchl, Phys. Rev. B:Condens. Matter Mater. Phys., 1994, 50, 17953–17979.
53. J. Heyd, G. E. Scuseria and M. Ernzerhof, J. Chem. Phys., 2003, 118, 8207–8215.

---