Structural and surface characterizations of 2D β-In₂Se₃/3D β-Ga₂O₃ heterostructures grown on c-Sapphire substrates by molecular beam epitaxy

Abstract

Integrating two-dimensional (2D) layered materials with wide bandgap β-Ga₂O₃ has unveiled impressive opportunities for exploring novel physics and device concepts. This study presents the epitaxial growth of 2D β-In₂Se₃/3D β-Ga₂O₃ heterostructures on c-Sapphire substrates by plasma-assisted molecular beam epitaxy. Firstly, we employed a temperature-dependent two-step growth process to deposit Ga₂O₃ and obtained a phase-pure $(\overline{2 }01)$ β-Ga₂O₃ film on c-Sapphire. Interestingly, the in-situ reflective high-energy electron diffraction (RHEED) patterns observed from this heterostructure revealed the in-plane ‘b’ lattice constant of β-Ga₂O₃ ~ 3.038Å. In the next stage, for the first time, 2D In₂Se₃ layers were epitaxially realized on 3D β-Ga₂O₃ under varying substrate temperatures (T_sub) and Se/In flux ratios (R_VI/III). The deposited layers exhibited (00l) oriented β-In₂Se₃ on $(\overline{2 }01)$ β-Ga₂O₃/c-Sapphire with the epitaxial relationship of $[11\overline{2 }0]$ β-In₂Se₃ || [010] β-Ga₂O₃ and $[10\overline{1 }0]$ β-In₂Se₃ || [102] β-Ga₂O₃ as observed from the RHEED patterns. Also, the in-plane ‘a’ lattice constant of β-In₂Se₃ was determined to be ~ 4.027Å. The single-phase β-In₂Se₃ layers with improved structural and surface quality were achieved at a T_sub ~ 280 °C and R_VI/III ~ 18. The microstructural and detailed elemental analysis further confirmed the epitaxy of 2D layered β-In₂Se₃ on 3D β-Ga₂O₃, a consequence of the quasi-van der Waals epitaxy. Furthermore, the β-Ga₂O₃ with an optical bandgap (E_g) of ~ 5.04 eV (deep ultraviolet) when integrated with 2D β-In₂Se₃, E_g ~ 1.43eV (near infra-red) can reveal potential applications in the optoelectronic field.

Introduction

Since the advent of two-dimensional (2D) layered van der Waals (vdWs) materials, Indium Selenide (In_xSe_y), one of the prominent candidates in this family, has been widely explored in the scientific community to investigate its novel and impeccable properties^{1,2,3,4,5,6,7,8}. It belongs to a complex system that crystallizes into different stoichiometric ratios (stacking configurations), e.g., InSe (β, γ), In₂Se₃ (α, β, γ and δ), In₃Se_4, etc., under various deposition conditions and growth techniques^4,7,8. Among these, the β-In₂Se₃ with its rhombohedral crystal structure has a primitive unit cell (a = b = 4.00 Å, and c = 28.33 Å) comprising three monolayers that are stacked vertically and repeatedly by weak vdW forces, with an in-plane covalently bonded atomic sequence of “Se − In − Se − In − Se”². The β-In₂Se₃ is renowned for its exceptional chemical stability and remarkable optical activity at room temperature (RT) and further exhibits strong 2D quantum confinement effects with its absorption edge in the near infra-red (IR) spectral range (~ 1.43 eV)^1,2. In addition, recent studies demonstrated that the phase-engineering of In₂Se₃ from α to β through thermal annealing has resulted in ultrahigh responsivity and detectivity of 8.8 × 10⁴ A/W and 2.9 × 10¹³ Jones, respectively⁶.

On the other hand, recently, there has been significant research interest in exploring the integration of 2D layered materials with wide bandgap (WB) semiconductors, particularly Gallium Oxide (Ga₂O₃)^{9,10,11,12,13}. Being a fourth-generation semiconducting material, Ga₂O₃, one of the group-III metal sesquioxide exhibits various polymorphs: α, β, γ, δ, and ε¹⁴. Among which the monoclinic β-phase (a = 12.23 Å, b = 3.04 Å, c = 5.80 Å, and β = 103.71°) with its direct bandgap (E_g) ~ 4.9 eV is considered to be the thermodynamically stable structure^14,15. Due to its ultrawide E_g, high breakdown electric field of ~ 8 MV/cm, and robust chemical/thermal stability, it has exhibited tremendous progress in high-power electronics and deep ultraviolet (UV) optoelectronic devices^16,17,18. Integrating this material with 2D layered materials can unveil novel opportunities in device physics. For instance, Wang et al. demonstrated a solar-blind photodetector with p-GaSe/n-Ga₂O₃ vdWs heterostructure that showed a high responsivity of 51.9 A/W and a pronounced specific detectivity up to 10¹⁴ Jones, resulting from the efficient separation of charge carriers across the pn junctions¹¹. An ambipolar p-TMD (p-MoTe₂ or p-WSe₂)/n-Ga₂O₃ junction field effect transistor (JFET) was reported by Choi et al., with two different types of channels in a single device architecture with their respective charge carriers¹². Despite the challenge in realizing the enhancement mode (e-mode) operation of the Ga₂O₃ device due to lack of p-type doping, Yang et al. fabricated a β-Ga₂O₃ FET with ferroelectric α-In₂Se₃ wrapped-gate that changed from depletion- to e-mode operation by effectively controlling the threshold voltage¹³. These findings collectively highlight the significance and potential of 2D material/β-Ga₂O₃ heterostructures for future device applications.

Nevertheless, the integration of these 2D materials with WB-Ga₂O₃ from the previous works was constrained to the ex-situ techniques, particularly by exfoliation or transfer methods of either the 2D vdW layers or the underlying Ga₂O₃ layers from the bulk substrates. Although the results are encouraging, such methods offer limited control over film thickness, may be prone to contamination and defects, and, most importantly, accessible with reduced scalability, therefore limiting their usage in large-area applications. Owing to these challenges, the utilization of molecular beam epitaxy (MBE) emerges as a proven growth technique to fabricate these heterostructures in situ with its ultra-high vacuum (UHV) conditions, high pure elements, thickness controllability and further yielding single crystalline materials with reduced defects.

For the first time in this study, the mixed-dimensional 2D β-In₂Se₃/3D β-Ga₂O₃ heterostructures were realized in situ using plasma-assisted molecular beam epitaxy (PA-MBE) on c-Sapphire. To achieve high-quality heteroepitaxial films, careful optimization of the initial β-Ga₂O₃ growth process is essential. A strategic approach involves the introduction of a low-temperature (LT) buffer (nucleation) layer, which proves effective in two key aspects. Firstly, the LT buffer layer serves as a sacrificial template by incorporating and localizing threading dislocations (TDs) that arise due to the lattice mismatch concerning the substrate^19,20. Secondly, it provides a homo-surface, circumventing lattice constraints²¹ and facilitating a smoother transition for high-temperature (HT) film growth. Consequently, we used a two-stepped β-Ga₂O₃ film grown under LT and HT conditions, commonly used for the heteroepitaxy on a Sapphire substrate, to improve the crystal quality effectively^19,21. Amidst the daunting challenge of the inherent and uneven surface of 3D Ga₂O₃, we successfully achieved the epitaxy of 2D In₂Se₃, thanks to our vigilant in-situ reflective high-energy electron diffraction (RHEED) tool for providing the information about the structural changes, in-plane lattice constants, and epitaxial relationships of the grown films. Besides the rich phases of In₂Se_3, we achieved the dominant phase 2D β-In₂Se₃ on 3D β-Ga₂O_3, which was confirmed by X-ray Diffraction (XRD) and Raman Spectroscopy. Furthermore, the surface morphological changes of the grown layers were studied carefully using Atomic Force Microscope (AFM) measurements. The microstructural and detailed elemental analysis across the heterostructures grown on c-Sapphire was thoroughly investigated by (Scanning) Transmission Electron Microscopy-(S)TEM measurements. The results presented in this study establish a fundamental understanding of the epitaxy of 2D In₂Se₃/3D Ga₂O₃ heterostructures, which is crucial for its commercialization in large-area applications.

Results and discussion

Figure 1a,b shows the in-situ RHEED patterns of the c-Sapphire substrate (before growth). Soon after the growth of LT-Ga₂O₃ film (substrate temperature, T_sub ~ 450 °C), the transition in the RHEED patterns occurred along both the azimuthal directions (repeated for every 60° rotation), as shown in Fig. 1c,d, indicating the change in crystal structure from rhombohedral (α) c-Sapphire to monoclinic (β) Ga₂O₃. The in-plane epitaxial relationship observed from RHEED patterns revealed that the Ga₂O₃ was aligned along [010] β-Ga₂O₃ || $[10\overline{1 }0]$ c-Sapphire and [102] β-Ga₂O₃ || $[11\overline{2 }0]$ c-Sapphire, and the respective growth directional views were showed in Fig. S1. This preferential alignment of β-Ga₂O₃ is attributed to its similar oxygen atomic arrangement compared to c-Sapphire, and the effort to minimize lattice strain to establish a coherent epitaxial relationship between them. This is attributed to the fact that the lattice points within the growth directional planes, (0001) Sapphire and $(\overline{2 }01)$ β-Ga₂O_3, maintain closer repeated interatomic distances (rectangle with white circled corners), as shown in Fig. 1g,h. Further exerting an in-plane lattice mismatch of -3.2% and -10.7% (minus indicates compressively strained β-Ga₂O₃) in their respective directions. This mismatch might arise due to the slight distortion of regular hexagon redistribution of oxygen atoms with a single O–O distance: 4.76 Å in (0001) Sapphire to two O–O distances: 4.96 Å and 5.15 Å in $(\overline{2 }01)$ β-Ga₂O₃¹⁵, owing to the change in ionic radii of the Al³⁺ (0.54 Å) to Ga³⁺ (0.62 Å)²². Moreover, this initially grown LT-Ga₂O₃ can serve as a nucleation film between the c-Sapphire and the HT-Ga₂O₃ (T_sub ~ 700 °C) film by minimizing this lattice and in-plane thermal expansion (α) mismatches between (0001)-Sapphire (α_s ~ 5 × 10^–6 K⁻¹)²³ and $(\overline{2 }01)$ β-Ga₂O₃ (α_g ~ 7.8 × 10^–6 K⁻¹)²⁴. Corroborating this, the evolution of RHEED patterns from the LT β-Ga₂O₃ film grown on c-Sapphire exhibited a significant improvement in the crystalline quality from starting to finishing growth at T_sub ~ 450 °C, as shown in Fig S2. Furthermore, Fig. 1e,f shows sharper and streakier RHEED patterns observed from the HT-grown Ga₂O₃ film. This could result from the lattice mismatch compensation and uniform nucleation in the LT-Ga₂O₃ film, providing a decent surface for the homoepitaxy at HT. More detailed information on the temporal evolution of RHEED patterns from the two-stepped (HT/LT) β-Ga₂O₃ film indicating the improvement in crystalline quality is discussed in the supplementary information.

Figure 1

In-situ RHEED patterns of (a,b) c-Sapphire substrate (before growth), (c,d) LT β-Ga₂O_3, and (e,f) HT/LT β-Ga₂O₃ films at ~ 5 min of their respective growths. The in-plane crystallographic views of (g) (0001) Sapphire and (h) $(\overline{2 }01)$ β-Ga₂O₃ along the growth direction were visualized using the ball and stick model by VESTA Software⁴².

Notably, the information about the 'b' lattice constant of the β-Ga₂O₃ crystal structure can be obtained from the RHEED pattern along [102] β-Ga₂O₃. In this regard, we extracted the RHEED intensity profiles along respective directions, as shown in Fig. 2a. The reciprocal lattice spacing of ${b}_{{G}_{[102]}}^{*}$= 2.068 Å⁻¹ extracted from known lattice distance of c-Sapphire along ${b}_{{S}_{[11\overline{2 }0]}}$= 8.241 Å (${\sqrt{3}a}_{S}$)²⁵ in real space, has yielded the ‘b’ lattice constant of β-Ga₂O₃, ${b}_{{G}_{\left[102\right]}}$= 2π/${b}_{{G}_{[102]}}^{*}$ Å ~ 3.038 Å, which matches well with the theoretical value^15,24, which could indicate the fully relaxed β-Ga₂O₃ film grown on c-Sapphire. This was determined quantitatively by the translation of streak spacing in the reciprocal lattice by the number of pixels achieved from the RHEED patterns²⁶. The additional diffraction streaks observed in this direction (indicated by blue arrows) also maintained a similar streak spacing. This coexistence of patterns along [102] β-Ga₂O₃ might arise from the octahedral and tetrahedral planes of Ga atoms within the $(\overline{2 }01)$ β-Ga₂O₃; however, further understanding is required to confirm its origin. The surface morphology of the as-grown LT-Ga₂O₃ and two-stepped Ga₂O₃ films grown on c-Sapphire is shown in Fig. 2b,c from the AFM scans. At LT, the surface of the film exhibited small granular morphology with dense grain boundaries owing to a root mean square (RMS) of ~ 0.56 nm. This is because the adatoms at LT will not have sufficient energy to transfer and nucleate with adjacent atoms on the surface, thus resulting in high nucleation sites, as seen in Fig. 2b. On the other hand, this decreased mobility of surface species can promote uniform dispersion of nuclei that can effectively cover the substrate¹⁹ and provide a homo-surface for the HT film growth. At HT conditions, Fig. 2c, the surface of the film was covered with large grains with reduced grain boundaries and exhibited a rougher surface (RMS ~ 5.83 nm). This could be ascribed to the greater likelihood of an adatom encountering an existing island formed during the ripening stage and promoting further growth primarily due to an increased adatom diffusion coefficient at HT.

Figure 2

(a) RHEED intensity profiles of $[11\overline{2 }0]$ c-Sapphire and [102] β-Ga₂O₃ (HT/LT) after growth, with the in-plane ‘b’ lattice constant evaluated to be ~ 3.038 Å. The inset shows the corresponding patterns, and the profiles are extracted from respectively. 5 × 5 µm² AFM scans of (b) LT-Ga₂O₃ and (c) HT/LT-Ga₂O₃ exhibiting a smaller and larger granular morphology, respectively.

Figure 3a shows XRD 2θ-scans of the Ga₂O₃ films grown on c-Sapphire. The LT nucleation film exhibited distinct diffraction peaks at ~ 18.9° and ~ 38.3°, corresponding to the $(\overline{2 }01)$ and $(\overline{4 }02)$ diffraction peaks of β-Ga₂O₃, respectively^15,17 (ICDD Card No. 01–082-3838). Moreover, the ratio calculated between these $(\overline{2 }01)$ and $(\overline{4 }02)$ diffraction peaks was found to be ~ 1.89 (ideal value ~ 2.2)²⁷, suggesting the dominant β-Ga₂O₃ when grown at a low T_sub ~ 450 °C, in contrast to the observation of secondary phases by Oshima et al.²⁷ Also, the d-spacing measured between the $(\overline{2 }01)$ planes of LT β-Ga₂O₃ is determined to be ~ 0.468 nm, with the thickness of film ~ 32 nm as shown by TEM images in the supplementary information, Fig. S3. Further, depositing the HT Ga₂O₃ layers, the $(\overline{2 }01)$ family of 2θ diffraction peaks persisted by preserving the single oriented β-phase. The employment of LT nucleation film presents a key advantage in minimizing the likelihood of defects propagating into the HT film due to lattice mismatch^19,20. Additional information on the crystalline quality of Ga₂O₃ without and with LT nucleation film is shown in Fig. S4, suggesting the improved Ga₂O₃ film quality with incorporating LT nucleation film. The Raman spectra of the Ga₂O₃ films grown on c-Sapphire exhibited the phonon modes corresponding to β-phase, as shown in Fig. 3b. These peaks are segregated into three categories: the lower frequency peaks located ~ 147.3 cm⁻¹(B_g²), ~ 170.5 cm⁻¹(A_g²), and ~ 202.0 cm⁻¹(A_g³) are attributed to libration and translation of octahedral-tetrahedral chains, the mid-frequency peaks located ~ 350.3 cm⁻¹(A_g⁵), and ~ 483.7 cm⁻¹(A_g⁷/B_g⁴) are attributed to deformation of GaO₆ octahedra, and the higher frequency peak located ~ 656.7 cm⁻¹(A_g⁹/B_g⁵) relates to the stretching and bending of GaO₄ tetrahedra^28,29,30. The more pronounced vibrational modes were observed from the two-stepped film with the FWHM of A_g³ mode ~ 5.5 cm⁻¹. Figure 3c shows the transmittance spectra of the HT/LT β-Ga₂O₃ film at RT, and the spectrum exhibited an average transmittance of ~ 94% with clear interference fringes in the visible region. The direct bandgap of a semiconductor can be determined from the UV–visible spectra by using the Tauc relation ${\left(\alpha h\nu \right)}^{2}\propto {(h\nu -E}_{g}$)³¹, where α is the absorption coefficient, hν is the incident photon energy, and E_g is the optical bandgap. An abrupt decrement in the wavelength was observed at the absorption edge around 250 nm, indicating the presence of an optical bandgap. The E_g of the β-Ga₂O₃ film is estimated by extrapolating the intercept of the energy axis at α = 0 with an approximate value ~ 5.04 eV (deep UV region), similar to previously reported values^30,32,33. By virtue of the successful epitaxy of two-stepped β-Ga₂O₃ film on c-Sapphire, in the following crucial stage, we focused on depositing and investigating the 2D-In₂Se₃ films on 3D β-Ga₂O₃ using MBE.

Figure 3

(a) XRD 2θ-scans and (b) Raman Spectra of LT β-Ga₂O₃ (green) and HT/LT β-Ga₂O₃ (red) grown on a c-Sapphire substrate (black). (c) Transmittance spectrum of HT/LT β-Ga₂O₃. Inset shows the plot of (αhν)² versus photon energy, where α and hν represent the absorption coefficient and photon energy, respectively. The optical bandgap of ~ 5.04 eV was estimated by extrapolating α to 0.

The epitaxy of the chalcogenide material in a typical solid source UHV-MBE system is employed at a high chalcogen-to-metal flux ratio due to increased volatility and lower sticking coefficient of chalcogen atoms at the growth surface³⁴. Owing to this, we maintained Se-rich conditions for the In₂Se₃ epitaxy in this work by setting the Se/In flux ratio (R_VI/III) larger than ~ 15⁸. Figure 4a,b shows the RHEED patterns of the as-grown HT/LT β-Ga₂O₃/c-Sapphire before the growth of In₂Se₃. When the growth was maintained at T_sub of 480 °C (R_VI/III ~ 28), the RHEED patterns remained similar to β-Ga₂O₃ for the whole growth, as seen in Fig. 4c,d. This indicates that the epitaxy of In₂Se₃ layers didn’t occur at these conditions, which might be caused by the kinetic limitations (rate of adsorption and desorption of adatoms) or nucleation barriers³⁵ that may not be favorable at 480 °C, leading to hindered growth. Further decreasing the T_sub ~ 330 °C, the transition in the RHEED patterns was observed along both the azimuthal directions, as shown in Fig. 4e,f, indicating the change in the crystal structure from monoclinic β-Ga₂O₃ to rhombohedral β-In₂Se₃ structure. In addition, the streak spacing ratio between the a-a/m-m planes was measured to be ~ $\sqrt{3}$, representing the six-fold symmetry of In₂Se₃ layers. Despite the 3D surface morphology of the β-Ga₂O₃ film, the In₂Se₃ layers were successfully grown on it. This could be a consequence of quasi-vdWs epitaxy being independent of the surface lattice conditions of the underlying layer³⁶. The In₂Se₃ layers grown on β-Ga₂O₃ film followed the in-plane epitaxial relationship of $[11\overline{2 }0]$ β-In₂Se₃ || [010] β-Ga₂O₃ and $[10\overline{1 }0]$ β-In₂Se₃ || [102] β-Ga₂O₃. The spotty pattern observed in this condition may have originated from the 3D growth of In₂Se₃ layers. Gradually, the RHEED patterns became streakier upon further reducing the T_sub ~ 280 °C, indicating improved lateral growth, as shown in Fig. 4g,h. Following this, the In₂Se₃ layers were grown with varied R_VI/III of 38 and 18 at a T_sub of 280 °C. The RHEED patterns became broader with slight spots for the sample grown at an increased flux ratio of 38, Fig. 4i,j, suggesting the declined surface quality. Among the whole series, a clear and streakier pattern was observed for the entire epitaxy when grown at R_VI/III ~ 18 (T_sub ~ 280 °C), suggesting the improved surface of the In₂Se₃ layers, as shown in Fig. 4k,l. Furthermore, the in-plane reciprocal streak spacing along $[10\overline{1 }0]$ β-In₂Se₃ with the respective [102] β-Ga₂O₃ has yielded the real space ‘a’ lattice constant of β-In₂Se₃, ${b}_{{I}_{\left[10\overline{1 }0\right]}}$= 2π/${b}_{{I}_{[10\overline{1 }0]}}^{*}$ Å ~ 4.027 Å as shown Fig. 4m. The surface morphologies of In₂Se₃ layers grown on 3D β-Ga₂O₃ films at varying epitaxial conditions are shown in Fig. 5 from the 5 × 5 µm² AFM scans. At a T_sub of 330 °C (R_VI/III ~ 28), we can observe a high density of smaller triangles (~ 250 nm) with a pronounced vertical stacking, resulting in a 3D surface morphology of In₂Se₃ (RMS ~ 13.70 nm) which is in correspondence with the observation of spotty RHEED patterns. Further reducing the T_sub ~ 280 °C, the density of triangular domains is reduced by exhibiting an improved lateral growth (RMS ~ 4.89 nm). However, increasing the R_VI/III to ~ 38 by maintaining the T_sub ~ 280 °C resulted in an increment in the density of triangular domains with reduced size. It might be caused by the excess Se atoms occupying the surface sites, causing limited surface diffusion³⁷ and further promoting vertical growth, as evidenced by the enhanced RMS ~ 7.09 nm. In contrast, a smoother surface, comprising 0° and 180°-oriented triangles with improved lateral size ~ 450 nm, was observed when the R_VI/III was reduced to ~ 18 (RMS ~ 3.94 nm). The step profile analysis reveals the thickness of the monolayer measured to be ~ 0.95 nm, as shown in Fig. S5, which matches well with other reports from the literature⁴. Therefore, we claim that both the T_sub and R_VI/III play vital roles in controlling the nucleation density and surface quality of the In₂Se₃ layers grown on 3D β-Ga₂O₃/c-Sapphire.

Figure 4

In-situ RHEED patterns of (a,b) as-grown HT/LT β-Ga₂O₃/c-Sapphire, with the epitaxy of In₂Se₃ layers grown on it under varied R_VI/III/T_sub conditions, (c,d) 28/480 °C, (e,f) 28/330 °C, where m and a are denoted as the diffraction planes from hexagonal crystal symmetry and along $[11\overline{2 }0]$ and $[10\overline{1 }0]$ azimuth rotations, (g,h) 28/280 °C, (i,j) 38/280 °C and (k,l) 18/280 °C. (m) RHEED intensity profiles of [102] HT/LT β-Ga₂O₃ and $[10\overline{1 }0]$ β-In₂Se₃ (18/280 °C) after growth, with the in-plane lattice constant evaluated to be ~ 4.027 Å. The inset shows the corresponding patterns, and the profiles are extracted from respectively.

Figure 5

5 × 5 µm² AFM scans of β-In₂Se₃ layers grown at R_VI/III/T_sub of (a) 28/330 °C, (b) 28/280 °C, (c) 38/280 °C and (d) 18/280 °C epitaxial conditions on HT/LT β-Ga₂O₃/c-Sapphire heterostructures.

On the other hand, it has been a challenging issue to identify and differentiate the commonly obtained crystal phases of In₂Se₃, among which the rhombohedral crystal structures of α- and β-In₂Se₃ share similar but different space groups (R3m and R$\overline{3 }$m)^5,38. The primary difference between these two structures lies in the location of In atoms at tetrahedral and octahedral sites covered by the Se packing in α- and β-In₂Se_3, respectively³⁸. Subsequently, insisting on a demanding characterization method and prudent analysis to distinguish the respective crystal phases. Figure 6a shows the XRD 2θ-scans of In₂Se₃ layers grown on two-stepped β-Ga₂O₃/c-Sapphire at different epitaxial conditions. The sample grown at T_sub ~ 480 °C (R_VI/III ~ 28) exhibits only the diffraction peaks related to β-Ga₂O₃, indicating the absence of In₂Se₃ epitaxy which agrees well with the observed RHEED patterns. On the other hand, at all the other epitaxial conditions, apart from the diffraction peaks of β-Ga₂O₃, we can observe two distinct additional peaks, ~ 9.4° and ~ 28.5° corresponding to (003) and (009) diffraction planes of rhombohedral β-In₂Se₃ crystal structure (ICDD Card No. 35–1056, space group R$\overline{3 }$m). Furthermore, no additional peaks are present in the 2θ-scans, confirming the epitaxial growth of single phase β-In₂Se₃ layers on 3D two-stepped Ga₂O₃ films. However, the peak overlapping at ~ 18.9° and ~ 38.3° originated from the diffraction signals of (006) and (0012) planes of β-In₂Se₃ and $(\overline{2 }01)$ and $(\overline{4 }02)$ planes of β-Ga₂O₃, respectively, make it difficult to validate the pristine properties of β-Ga₂O₃ underneath layers after β-In₂Se₃ deposition. Hence, selected 2θ-XRD peak analysis, as well as a direct growth of β-In₂Se₃ on c-sapphire, have been performed to support this validation, as shown in Fig. S6(a,b). Among all the samples, the lowest FWHM of (003) and (009) 2θ-diffraction peaks were observed to be ~ 0.29° and ~ 0.35° for the sample grown at T_sub ~ 280 °C and R_VI/III ~ 18.

Figure 6

(a) XRD 2θ-scans and (b) Raman Spectra of β-In₂Se₃ layers grown at different R_VI/III/T_sub epitaxial conditions on HT/LT β-Ga₂O₃/c-Sapphire heterostructure along with (c) active Raman vibrational modes of β-In₂Se₃.

Moreover, detailed information is further essential to classify the grown layers. As mentioned earlier, Raman spectroscopy is a robust and non-destructive technique used to characterize the samples with different phases based on the molecular fingerprints obtained from various active phonon modes¹. Figure 6b shows the Raman spectra of In₂Se₃ layers grown at different epitaxial conditions on β-Ga₂O₃ films. Here as well, the epitaxy performed at T_sub ~ 480 °C (R_VI/III ~ 28) shows only peaks related to β-Ga₂O₃ film. All the other samples exhibited three clear peaks observed ~ 110.4 cm⁻¹, ~ 176.3 cm⁻¹, and ~ 206.9 cm⁻¹ attributed to A₁(LO + TO), A₁(TO), and A₁(LO) phonon modes, respectively, as shown in Fig. 6c, which are characteristics of β-In₂Se₃, that are similar to the previously reported results^1,2,39. A similar peak overlapping was observed between the pronounced A_g³ mode from the β-Ga₂O₃ and the A₁(LO) mode of β-In₂Se_3, as shown in Fig. S6(c,d). The active vibrational modes exhibited by β-In₂Se₃ with regard to the similarly structured α-In₂Se₃ are validated upon comparing the typical Raman peaks of α-In₂Se₃ as summarized by Liu et al.³⁹. However, the sample grown at T_sub of 280 °C under the R_VI/III ~ 38 exhibited two additional Raman modes ~ 151.3 cm⁻¹ and ~ 252.6 cm⁻¹, along with the respective β-In₂Se₃ modes. These peaks are characteristics of γ-In₂Se₃, with the former phonon mode corresponding to the zone center vibration and the latter to the excess contribution of Se atoms’ linkage to the Se-Se bond due to the high Se flux used in this series^40,41. This indicates that, at these epitaxial conditions, the growth leads to the co-existence of β-In₂Se₃ and γ-In₂Se₃ with the dominance in the former phase. The existence of additional 3D γ-In₂Se₃ may cause predominantly vertical growth, resulting in a rougher surface, as evident from the surface morphology characterization mentioned above.

Furthermore, Fig. 7a shows the STEM high-angle annular dark-field (HAADF) cross-sectional view of the β-In₂Se₃/β-Ga₂O₃ heterostructure grown on c-Sapphire. The thicknesses of the LT-, HT β-Ga₂O_3, and β-In₂Se₃ films are determined to be ~ 32 nm, ~ 120 nm, and ~ 28 nm, respectively, with the corresponding growth rates of ~ 0.53, ~ 1.0, and ~ 0.47 nm/min. Figure 7b,c provides a detailed visualization of the interfaces between the LT β-Ga₂O₃/c-Sapphire and β-In₂Se₃/HT β-Ga₂O₃ heterostructures. Regardless of the relatively rough surface of the two-stepped β-Ga₂O₃ film that may result in a non-abrupt 2D/3D interface, the layered structure of β-In₂Se₃ is clearly observed, attributed to the quasi-van der Waals epitaxy. The detailed elemental mappings of the entire β-In₂Se₃/β-Ga₂O₃ heterostructure grown on c-Sapphire are shown in Fig. 7d–i, which reveal an abrupt transition and uniform distribution of respective elements within the specific layers. These results evidently support the objective of the present work on realizing the epitaxial growth of β-In₂Se₃ on β-Ga₂O₃, which shows the potential scope to study the mixed dimensional heterostructures for future device applications using MBE technique.

Figure 7

(a) Low magnification STEM-HAADF cross-sectional view of the β-In₂Se₃/β-Ga₂O₃ heterostructure grown on c-Sapphire; (b,c) represent the high-magnification TEM images of the interfaces between β-In₂Se₃/HT β-Ga₂O₃ and LT β-Ga₂O₃/c-Sapphire, respectively. (d) Low magnification STEM-HAADF cross-sectional view of β-In₂Se₃/β-Ga₂O₃/c-Sapphire heterostructure grown on c-Sapphire showing the corresponding elemental mappings (e–i) of Al, Ga, O, In, and Se atoms in their respective films.

So far, we have successfully achieved the epitaxy of the single phase 2D β-In₂Se₃ layers on 3D two-stepped β-Ga₂O₃/c-Sapphire. In the present series, the better structural and surface quality of β-In₂Se₃ layers was obtained when grown at R_VI/III and T_sub of ~ 18 and ~ 280 °C, respectively. Finally, such a mixed dimensional (2D β-In₂Se₃/3D β-Ga₂O₃) heterostructure can avail the benefits offered by both materials, specifically in the optoelectronic field, with its absorption edges extending from Near-IR (~ 1.43 eV)¹ to deep UV regions (~ 5.04 eV), and can be used as a dual-band photodetector. Also, the epi-grown In₂Se₃ and Ga₂O₃, being intrinsically n-type, can form a heterostructure exhibiting an nN isotype heterojunction, forming 2DEG upon bandgap engineering. Hence, studying the band alignment of this heterostructure can unveil new opportunities in the (opto-) power electronic field.

In conclusion, we successfully realized 2D β-In₂Se₃/3D β-Ga₂O₃ heterostructures on c-Sapphire substrates using the PA-MBE technique. A two-stepped Ga₂O₃ growth was employed to improve the crystalline quality of the film, as indicated by the XRD 2θ-scans and Raman Spectra. For the first time, the in-plane ‘b’ lattice constant of β-Ga₂O₃ (~ 3.038Å) grown on c-Sapphire was determined using in-situ RHEED patterns. In the next stage, the 2D β-In₂Se₃ layers were successfully grown on 3D β-Ga₂O₃ films resulting from quasi-vdWs epitaxy. The In₂Se₃ layers followed an in-plane epitaxial relationship of $[11\overline{2 }0]$ β-In₂Se₃ || [010] β-Ga₂O₃ and $[10\overline{1 }0]$ β-In₂Se₃ || [102] β-Ga₂O₃ with the in-plane lattice constant of β-In₂Se₃ determined to be ~ 4.027Å. The single phase β-In₂Se₃ layers with improved structural and surface quality were achieved when growth was maintained at R_VI/III ~ 18 and T_sub ~ 280 °C on β-Ga₂O₃/c-Sapphire. The (S)TEM microstructural and detailed elemental analysis has clearly indicated the successful realization of 2D β-In₂Se₃/3D β-Ga₂O₃ heterostructure on c-Sapphire, completely in-situ using PA-MBE. Such an epitaxial realization of 2D layers on 3D films can enhance the potential of mixed-dimensional heterostructures by increasing the scalability and reducing the possibility of contamination compared to other transfer methods. The realized β-In₂Se₃/β-Ga₂O₃ heterostructure with its optical bandgap energies (E_g) ~ 1.43 eV (Near-IR)¹ and ~ 5.04 eV (Deep UV), respectively, has potential applications in the field of optoelectronics.

Experimental methods

The epitaxy of Ga₂O₃ and In₂Se₃ thin films was performed by the SVT associates PA-MBE system at a background pressure of ~ 2 × 10^–10 torr, using the Knudsen cells with high purity Gallium (7N), Indium (6N), and Selenium (6N) sources. The active oxygen species for the Ga₂O₃ growth was supplied by a Radio-frequency (RF) Plasma source. Firstly, a two-stepped Ga₂O₃ film was grown on the c-Sapphire substrate at LT and HT conditions. The Ga cell beam equivalent pressure (BEP) and T_sub for the epitaxy under LT and HT conditions were 2 × 10^–8 torr and 450 °C, 6 × 10^–8 torr and 700 °C, respectively. The oxygen plasma source was maintained at an RF power of 300W with a flow rate of 1.0 sccm for the two-stepped growth. The growth times of LT- and HT-Ga₂O₃ films were one and two hours, respectively. After the epitaxy of two-stepped Ga₂O₃ thin film, a series of In₂Se₃ layers were grown at different T_sub (480°–280 °C) and at varying Se/In BEP flux ratios (R_VI/III) (18–38) by maintaining a constant R_VI/III of 28 and T_sub of 280 °C respectively, for 1 h. The Se and In BEPs used in the present series range between 5.0–6.75 × 10^–7 and 1.8–2.85 × 10^–8 torr, respectively. The in-situ surface reconstructions of the films during the growth were monitored by RHEED operated at an electron beam energy of 12 keV. The surface morphology of the as-grown films was investigated using the atomic force microscope (AFM, Bruker Dimension Icon). The crystal quality and phase characterization of the films were determined by X-ray diffraction (XRD, Bruker New D8 Discover) 2θ-scans using Cu-Kα radiation (λ = 1.54056 Å) and Raman Spectrum using a LabRam iHR550 HORIBA spectrometer under 532 nm laser excitation. The microstructural and interfacial analysis at atomic resolutions was determined using the (Scanning) Transmission Electron Microscope (S)TEM (FEI Talos F200X, ThermoFisher Scientific), operated at 200 kV. The optical transmittance spectra were obtained using a JASCO V-780 UV–Vis-NIR Spectrophotometer. The in-plane and out-of-plane crystallographic views of β-Ga₂O₃ on c-Sapphire along the growth direction were visualized using the ball and stick model by VESTA Software version 3.5.7.