Electronic Structures of Molecular Beam Epitaxially Grown SnSe₂ Thin Films on $\sqrt{\mathbf{3}}\mathbf{\times }\sqrt{\mathbf{3}}$-Sn Reconstructed Si(111) Surface

Abstract

SnSe₂, as a prominent member of the post-transition metal dichalcogenides, exhibits many intriguing physical phenomena and excellent thermoelectric properties, calling for both fundamental study and potential application in two-dimensional (2D) devices. In this article, we realized the molecular beam epitaxial growth of SnSe₂ films on a $\sqrt{3}\times \sqrt{3}$-Sn reconstructed Si(111) surface. The analysis of reflection high-energy electron diffraction reveals the in-plane lattice orientation as SnSe₂[110]//$\sqrt{3}$-Sn [112]//Si [110]. In addition, the flat morphology of SnSe₂ film was identified by scanning tunneling microscopy (STM), implying the relatively strong adsorption effect of $\sqrt{3}$-Sn/Si(111) substrate to the SnSe₂ adsorbates. Subsequently, the interfacial charge transfer was observed by X-ray photoemission spectroscopy. Afterwards, the direct characterization of electronic structures was obtained via angle-resolved photoemission spectroscopy. In addition to proving the presence of interfacial charge transfer again, a new relatively flat in-gap band was found in monolayer and few-layer SnSe₂, which disappeared in multi-layer SnSe₂. The interface strain-induced partial structural phase transition of thin SnSe₂ films is presumed to be the reason. Our results provide important information on the characterization and effective modulation of electronic structures of SnSe₂ grown on $\sqrt{3}$-Sn/Si(111), paving the way for the further study and application of SnSe₂ in 2D electronic devices.

1. Introduction

Post-transition metal dichalcogenides (PTMCs), composed of III-VA metal elements (Ga, In, Sn, Bi etc.) and chalcogen elements (S, Se, Te), are a new class of layered two-dimensional (2D) materials that have attracted extensive study recently. In addition to the highly anisotropic lattice structure and unique electronic structures, many outstanding electrochemical, thermoelectric, and optoelectronic properties were observed in PTMCs. Therefore, PTMCs have been applied in many fields like electrochemistry [1,2], electronics [3,4], optoelectronics [5,6] and gas sensing [7]. As a prominent member of PTMCs, SnSe₂ is an intrinsic layered semiconductor, exhibiting considerable bandgap and excellent thermoelectric properties like ultra-low thermal conductivity and high carrier mobility [8,9]. Together with a high abundance of Sn, SnSe₂ shows potential applications in optoelectronics [10,11,12,13], thermoelectric devices [14], lithium-ion batteries [15] and phase change memory [16,17].

In addition, intriguing physical properties like charge density wave (CDW), superconductivity and interfacial polarons were observed in SnSe₂ from previous works [18,19,20]. The interface properties of monolayer (ML) SnSe₂ could be effectively tuned by choosing different substrates. For examples, a recent study showed that ML and bilayer SnSe₂ films grown on Si(111) undergo a commensurate CDW transition at T_C ~78 K driven by Fermi surface nesting between symmetry inequivalent electron pockets, forming a corresponding (2 × 2) superlattice [20]. In addition, the formation of interfacial polarons induced by charge accumulation and electron–phonon coupling between SrTiO₃ and SnSe₂ was evidenced [19]. Furthermore, the superconductivity in SnSe₂, which can be modulated by cation intercalation [21] and dielectric gating techniques [22], was reported to be enhanced in the epitaxial SnSe₂ on graphitized SiC(0001) substrate [18]. However, there are only a few studies on SnSe₂ film grown on Si(111), with the study on direct characterization of band structures still lacking in particular. In this paper, we synthesize SnSe₂ films on Si(111) by molecular beam epitaxy (MBE), and employ in situ reflection high energy electron diffraction (RHEED), scanning tunneling microscopy (STM), X-ray photoemission spectroscopy (XPS) and angle-resolved photoemission spectroscopy (ARPES) to characterize the atomic and electronic structures of them. We found that the SnSe₂ shows layer-by-layer growth mode on $\sqrt{3}$-Sn/Si(111) substrate with a high quality of flatness. The interfacial charge transfer effect is observed in the SnSe₂ film grown on $\sqrt{3}$-Sn/Si(111) substrate. Moreover, an emerging band above the original valence band of bulk SnSe₂ is found in the SnSe₂ thin film, resulting in the reduction of the indirect band gap. This emerging band is suggested to be induced by the interfacial strain-driven structural phase transition of the grown SnSe₂. Our results provide important information for further applications of SnSe₂ film in 2D electronic devices.

2. Materials and Methods

All of the experiments were performed in a combined MBE-STM-ARPES ultra-high vacuum (UHV) system with a base pressure of ~2 × 10⁻¹⁰ mbar. First, to obtain a clean Si(111)-(7 × 7) reconstructed surface, a Si(111) substrate (n-doped by P) was initially degassed at $650 °\mathrm{C}$, followed by flash-annealing to ~$1250 °\mathrm{C}$ for 20 cycles. To passivate the dangling bonds of the Si(111)-(7 × 7) surface and enhance the surface diffusion for the growth of SnSe₂, we evaporated $1/3$ ML of Sn on the Si(111)-(7 × 7) surface at $550 °\mathrm{C}$, leading to a well-ordered $\sqrt{3}\times \sqrt{3} R30°$-Sn recontruction surface on the Si(111) ($\sqrt{3}$-Sn/Si(111)) surface. Next, the SnSe₂ films were grown by co-depositing high-purity Sn (99.99%) and Se (99.99%) from standard Knudsen cells separately onto the $\sqrt{3}$-Sn/Si(111) substrate at $200 °\mathrm{C}$. The temperatures of the Sn and Se sources were maintained at $920 °\mathrm{C}$ and $130 °\mathrm{C}$, respectively, keeping the flux ratio at Sn:Se ~1:20, and leading the growth rate of SnSe₂ films as about ~15 min/layer. During the film growth, an in situ RHEED was used for real time monitoring and surface structural characterization. The electron beam energy of RHEED was kept at 18 keV. After growth, the surface morphology was further characterized by an in situ room temperature STM (RT-STM), whose set point was kept at V_b = 1.0 V and I_t = 100 pA. Subsequently, to investigate the electronic structures, the in situ low temperature (~10 K) XPS and ARPES were performed by a shared DA30 analyzer (Scienta Omicron AB, Uppsala, Sweden). The monochromatic X-ray of XPS and the ultraviolet light source of ARPES were generated from an Al electrode excitation source (Al Kα, 1486.7 eV) and a Helium lamp with a monochromator (He I, 21.218 eV), respectively.

3. Results and Discussion

The lattice structure of the $\sqrt{3}$-Sn/Si(111) surface is illustrated in Figure 1a, showing an in-plane 30° rotation and $\sqrt{3}\times \sqrt{3}$ reconstruction of a Sn layer relative to a Si(111) surface. Each Sn atom is back-bonded to three Si atoms beneath, with a remaining dangling bond at the surface. Figure 1b presents the lattice structure of T-phase SnSe₂ (T-SnSe₂). As a typical MX₂ of PTMCs, single layer T-SnSe₂ consists of two layers of chalcogenides and a layer of transition metal sandwiched between them, presenting a Se-Sn-Se sandwich-like structure. Within the layer, covalent bonding is achieved through Se-Sn, while interlayer stacking is achieved through Se-Se van der Waals interactions, presenting the lattice constants as $\mathit{a}=3.81 \AA$ and $\mathit{c}=6.14 \AA$ [23].

Figure 1c–e present the RHEED patterns of an annealed Si(111)-(7 × 7) surface, a $\sqrt{3}$-Sn/Si(111) surface and an ML SnSe₂ film along the Si$\left[1\overline{1}0\right]$ direction, respectively. Corresponding RHEED patterns along the Si$\left[\overline{1}\overline{1}2\right]$ direction are also obtained by 30° rotating the sample in-plane, as shown in Figure 1f–h. From Figure 1c,f, we can see the clearly (7 × 7) reconstructed patterns of the annealed Si surface, which disappear and are displaced by a new pair of (1 × 1) diffraction stripes after the $\sqrt{3}$-Sn passivation, as shown in Figure 1d,g. Subsequently, from Figure 1e,h, we can see the new (1 × 1) diffraction stripes from the epitaxial SnSe₂ film, which are very clear and sharp, implying the high-quality crystallization.

By performing Lorentz fittings on the corresponding RHEED patterns, we can accurately obtain the (1 × 1) main diffraction stripe spacings of both substrates and the SnSe₂ films (see Figure S1 of the Supplementary Materials for details), on which we can conduct more detailed analysis and calculations. The orientation of the $\sqrt{3}$-Sn buffer layer is in-plane rotated by 30° relative to the Si(111), with the calculated lattice constant as ${\mathit{a}}_{\sqrt{3}-\mathrm{Sn}}=6.66\pm 0.05 Å$, approximately equal to $\sqrt{3}$ times ${\mathit{a}}_{\mathrm{Si}\left(111\right)}=3.84 Å$. This result is consistent with the previous reports [24,25]. Subsequently, the epitaxial SnSe₂ film is rotated by 30° again relative to the $\sqrt{3}$-Sn/Si surface, resulting in the same in-plane orientation with the Si(111) substrate. Furthermore, the in-plane lattice constant of ML SnSe₂ is also obtained as ${\mathit{a}}_{{\mathrm{SnSe}}_2}=3.83\pm 0.05 Å$, which is in agreement with the previous report [23], and shows no obvious interfacial strain effect considering the $\pm 0.05 Å$ errors in RHEED analysis.

After growth, we utilized the in situ STM to investigate the surface morphology of ML SnSe₂ grown on $\sqrt{3}$-Sn/Si(111), as shown in Figure 1i. Different from three-dimensional (3D) island growth mode on graphene substrates [26], the SnSe₂ grown on $\sqrt{3}$-Sn/Si(111) tends to be layer-by-layer growth mode, which provides the better film flatness. The different growth modes of SnSe₂ on two substrates may be attributed to the larger local density of states of $\sqrt{3}$-Sn/Si(111) than that of graphene substrate, leading to an increased interface adsorption ability [27,28]. The height of ML SnSe₂ is ~0.62 Å [Figure 1j], which is consistent with the previous report [23].

Having determined the lattice structure, we then investigated the chemical states of the SnSe₂ film via in situ XPS. The full XPS spectra of Si(111), $\sqrt{3}$-Sn/Si(111) and SnSe₂ are shown in the Supplementary Materials Figure S2. Figure 2a presents the Si 2p core-level spectra of the Si(111)-(7 × 7) surface, $\sqrt{3}$-Sn/Si(111) and ML SnSe₂ film. For the Si(111)-(7 × 7) surface, the core levels of Si 2p_1/2 and Si 2p_3/2 orbitals are located at 99.7 eV and 99.1 eV, in accordance with the previous literature [29]. The peaks red-shift by about 0.1 eV compared to the core levels of Si 2p_1/2 (99.6 eV) and Si 2p_3/2 (99.0 eV) orbitals of the $\sqrt{3}$-Sn/Si(111), indicating a charge transfer effect from Sn to Si upon the surface Sn-Si bond formation, which is reasonable since the electronegativity of Si is stronger than that of Sn. In addition, after the growth of ML SnSe₂, the core levels of Si 2p_1/2 (99.8 eV) and Si 2p_3/2 (99.2 eV) again shift toward higher binding energies, which suggests interfacial charge transfer from the $\sqrt{3}$-Sn/Si(111) substrate to ML SnSe₂.

Subsequently, in Figure 2b, the core levels of Sn 3d_3/2 (493.3 eV) and 3d_5/2 (484.9 eV) orbitals of $\sqrt{3}$-Sn/Si(111) exhibit obvious blue-shift compared to the core levels of Sn 3d_3/2 (494.6 eV) and 3d_5/2 (486.2 eV) orbitals of ML SnSe₂. This shift can be explained by the much stronger bond polarity of Sn-Se than that of Sn-Si, which leads to the weakened screening to the Coulomb potential created by the positive core of Sn, and finally resulting in the higher binding energies of the remaining electrons. Moreover, in Figure 2c the core levels of Se 3d orbitals of ML SnSe₂ split into two sets, Se₁ 3d_3/2 (54.6 eV) and Se₁ 3d_5/2 (53.7 eV), and Se₂ 3d_3/2 (55.0 eV) and Se₂ 3d_5/2 (54.2 eV), respectively. The splitting of Se 3d orbitals implies the presence of two nonequivalent chemical environments of Se atoms in ML SnSe₂. This can be attributed to the fact that the bottom layer of Se atoms is adjacent to the √3-Sn surface, while the top layer of Se atoms is located at the top surface of SnSe₂.

To further investigate the interface effect, we characterized the electronic structures of the SnSe₂/$\sqrt{3}$-Sn/Si(111) system via in situ ARPES. Figure 3 shows the electronic structures of the Si(111)-(7 × 7) and $\sqrt{3}$-Sn/Si(111) surface. The constant energy mappings near the Fermi level of the Si(111)-(7 × 7) and $\sqrt{3}$-Sn/Si(111) surface clearly display the six-fold symmetry and the relative 30° rotation with respect to each other [Figure 3b,c], in accordance with the corresponding schematic of the Brillouin zone (BZ) illustrated in Figure 3a. The band structure of Si(111)-(7 × 7) surface is shown in Figure 3d, with the intensity enhanced zoom-in spectra plotted at the bottom panel of Figure 3e for more detailed observation. Three surface states marked as S₁, S₂ and S₃ can be clearly observed, in agreement with the previous literature [30]. The S₃ is induced by the back bonds of the (7 × 7) reconstruction, showing the obvious band dispersion with its maximum located at about −1.8 eV. The nearly flat S₂, induced by the rest-atoms, is located at about −0.9 eV, while the S₁, induced by the adatoms of the surface reconstruction, lies near the Fermi level.

The band structure of the $\sqrt{3}$-Sn/Si(111) surface and its zoom-in spectra with enhanced intensity are plotted in Figure 3f and the top panel of Figure 3e, respectively. The surface states S₂’ is located at −1~−2 eV, and the minimum of surface states S₁’ is located at −0.7 eV. The S₂’ and S₁’ are attributed to the three back bonds with Si and the one remaining dangling bond of each Sn adatom, respectively [31,32]. Our results indicate the well processed substrate, on which the SnSe₂ films can be grown with high-quality flatness and crystallization.

Finally, we investigated the electronic structures of SnSe₂ films grown on the $\sqrt{3}$-Sn/Si(111) surface, as shown in Figure 4. Recent first-principles calculations and an ARPES study of SnSe₂ showed that bulk SnSe₂ possesses an indirect band gap of ~1.07 eV, whose conduction band minimum (CBM) and valence band maximum (VBM) are located along the M-L and Γ-M (K) directions, respectively [33,34]. In contrast, ML SnSe₂ film exhibits an indirect band gap of ~1.69 eV, with CBM and VBM located at the M and Γ points, respectively [34]. The increased band gap of ML SnSe₂ is attributed to the quantum confinement of electrons in quasi-2D samples and the reduction of the screening. The CBM of both ML and bulk SnSe₂ are higher than the Fermi level, making it difficult to be characterized directly via ARPES in pristine SnSe₂.

Figure 4a is the constant energy mapping taken near the Fermi level of ML SnSe₂, presenting clear six-fold symmetry, with electron pockets centered at the M points of BZ. It is quite similar to that of the potassium(K)-doped SnSe₂ single crystals, suggesting the charge transfer from the $\sqrt{3}$-Sn/Si(111) substrate to the grown SnSe₂ film. In addition, the CBM of different layers of SnSe₂ are obtained from energy distribution curve (EDC) fittings, as shown in Figure S3. It can be clearly seen that all the CBM features appear near the Fermi level at the M point, shifting from −199 meV in ML SnSe₂ to −112 meV in few-layer (FL) SnSe₂ (the thickness of FL SnSe₂ is estimated to be ~3 ML), and finally to the edge of the Fermi level in multi-layer (Multi-L, the estimated thickness is about ~6 ML) SnSe₂ [Figure 4i–k]. The upward shift in CBM with the increase in layers again confirms the interfacial charge transfer scenario.

Furthermore, the ARPES spectra of ML, FL and Mult-L SnSe₂ along the Γ-M direction are plotted in Figure 4b–d, respectively. The VBM are all located at the Γ point. For more detailed observation, the zoom-in ARPES spectra near the VBM around the Γ point and corresponding EDCs along the red lines are illustrated in Figure 4e–h. Intriguingly, we note that a new band α above the original VBM of bulk SnSe₂ emerges in the ML and FL SnSe₂ films, which disappears in the Mult-L SnSe₂. The emerging α band can be better resolved in the second-derivative spectra, as shown in Figure S4. From the EDC fittings in Figure 4h, the maximum of the α band is located at −1.05 eV and −0.97 eV in ML and FL SnSe₂, respectively, leading to the reduced band gaps as 0.85 eV in both of them. The position, as well as the relatively flat dispersion of the α band, is quite similar to the top valence band of H-phase SnSe₂ (H-SnSe₂), which was characterized and marked as ‘7/8’ band in a previous study [35]. Moreover, a recent first-principle calculation showed a reduced band gap of H-SnSe₂ compared to T-SnSe₂, in accordance with our ARPES results [36]. Therefore, we believe that the newly emerged energy band α would be attributed to the presence of a small amount of H-SnSe₂ formed in ML and FL SnSe₂. Indeed, a pressure-induced structural phase transition to H-SnSe₂ has been reported before [37]. Therefore, we assume that the interfacial strain from $\sqrt{3}$-Sn/Si(111) substrate would be the main reason for the emergence of H-SnSe₂, which vanishes gradually as the film becomes thicker. Similar interfacial-induced structural phase transitions in other 2D materials were also observed in previous studies. For example, the lattice structure of WSe₂ grown on SrTiO₃ substrate was 1T’-phase [38], while the mixture of 1T’- and 1H-phase WSe₂ was found on graphene [39,40]. However, this interfacial strain effect is hard to be directly detected by RHEED due to the large relative error (~1.3%) in RHEED analysis, while the lattice constants of H-SnSe₂ and T-SnSe₂ are quite similar. Furthermore, the indistinct band structures of ML and FL SnSe₂, contrary to the clear band structures in Multi-L SnSe₂, also provide evidence for our assumption, since the band structures of both T-SnSe₂ and H-SnSe₂ can be detected by ARPES and eventually overlap together in the spectra, as shown in Figure 4b,c.

Our results indicate the potential application prospects of SnSe₂/$\sqrt{3}$-Sn/Si(111) in 2D devices. First of all, the Si is a traditional semiconductor with mature large-scale device fabrication processes, which is beneficial for the practical application of SnSe₂ grown on Si substrate. Actually, there have been several studies on SnSe₂/Si-based 3D devices in the past, such as temperature-dependent phase change memory [17] and self-powered broadband photodetectors which can be enhanced by Rhenium doping [12,13]. In our study, the $\sqrt{3}$-Sn passivated Si(111) substrate can not only effectively prevent the formation of Sn and Se clusters on the surface of Si, but also provide relatively strong adsorption effects, which allows us to achieve the growth of SnSe₂/$\sqrt{3}$-Sn/Si(111) thin films with better crystallization and film flatness. Moreover, the interfacial charge transfer gives rise to the n-doped SnSe₂ film, allowing us to build atomically thin SnSe₂-based p-n junction devices. Furthermore, our results also show that the band gap of thin SnSe₂ film grown on $\sqrt{3}$-Sn/Si(111) substrate can be more effectively modulated by changing the number of layers because of the appearing α band in ML and FL SnSe₂. Therefore, SnSe₂/$\sqrt{3}$-Sn/Si(111) would appeal to higher performance tunable, broad-spectrum photodetection, showing its possible application in 2D optoelectronics. In addition, the possible partial lattice distortion may lead to changes in resistivity, indicating its prospect in phase change memory application. However, to determine the specific performance of the SnSe₂/$\sqrt{3}$-Sn/Si(111) system, electric transport measurements are still needed to further explore its application prospects.

4. Conclusions

In summary, we have realized the MBE growth of SnSe₂ films on $\sqrt{3}$-Sn passivated Si(111) substrate. The analysis of RHEED patterns demonstrates the successful growth and the same in-plane lattice orientation of the grown SnSe₂ with Si(111) substrate, whereas the Sn buffer layer conducts a $\left(\sqrt{3}\times \sqrt{3}\right)$ reconstruction with in-plane rotation by 30° relative to both Si(111) and SnSe₂. In addition, the layer-by-layer growth mode of SnSe₂ identified by STM indicates the relatively stronger adsorption effect of $\sqrt{3}$-Sn/Si(111) on SnSe₂ adsorbates than the commonly used graphene substrate. Subsequently, from the XPS measurements, we observed the interfacial charge transfer from substrate to SnSe₂, which was further confirmed in the ARPES measurements. Moreover, a new valence band α located above the original VBM of bulk SnSe₂ emerges in the ML and FL SnSe₂ films, which disappears in the Multi-L SnSe₂ film. We attribute it to the partial T→H structural phase transition of thin SnSe₂ films grown on $\sqrt{3}$-Sn/Si(111) substrate, probably driven by the interfacial strain. The direct characterizations of electronic structures of SnSe₂ films, as well as the layer-dependent structural transition, along with the effective band modulation provide significant and important information for the further applications of SnSe₂ films in 2D electronic devices, such as optoelectronics and phase change memory.