The Heteroepitaxy of Thick β-Ga2O3 Film on Sapphire Substrate with a β-(AlxGa1−x)2O3 Intermediate Buffer Layer

A high aluminum (Al) content β-(AlxGa1−x)2O3 film was synthesized on c-plane sapphire substrate using the gallium (Ga) diffusion method. The obtained β-(AlxGa1−x)2O3 film had an average thickness of 750 nm and a surface roughness of 2.10 nm. Secondary ion mass spectrometry results indicated the homogenous distribution of Al components in the film. The Al compositions in the β-(AlxGa1−x)2O3 film, as estimated by X-ray diffraction, were close to those estimated by X-ray photoelectron spectroscopy, at ~62% and ~61.5%, respectively. The bandgap of the β-(AlxGa1−x)2O3 film, extracted from the O 1s core-level spectra, was approximately 6.0 ± 0.1 eV. After synthesizing the β-(AlxGa1−x)2O3 film, a thick β-Ga2O3 film was further deposited on sapphire substrate using carbothermal reduction and halide vapor phase epitaxy. The β-Ga2O3 thick film, grown on a sapphire substrate with a β-(AlxGa1−x)2O3 buffer layer, exhibited improved crystal orientation along the (-201) plane. Moreover, the scanning electron microscopy revealed that the surface quality of the β-Ga2O3 thick film on sapphire substrate with a β-(AlxGa1−x)2O3 intermediate buffer layer was significantly improved, with an obvious transition from grain island-like morphology to 2D continuous growth, and a reduction in surface roughness to less than 10 nm.


Introduction
β-Ga 2 O 3 has garnered significant attention in recent years due to its wide bandgap, high breakdown electrical field, and Baliga's figure of merit [1,2]. The synthesis of β-Ga 2 O 3 films on various substrates, such as Ga 2 O 3 [3,4], sapphire [5,6], and GaAs [7], has been widely reported. Among these substrates, homoepitaxial β-Ga 2 O 3 films exhibit a smooth surface without cracks or dislocations, making it the most favorable substrate currently available. However, the cost of β-Ga 2 O 3 substrate is relatively high, which limits its market application. Compared with β-Ga 2 O 3 substrate, sapphire is an economic and cost-effective substrate for heteroepitaxy. Several groups have attempted to synthesize β-Ga 2 O 3 on sapphire substrate using various methods, including pulsed layer deposition (PLD) [8], molecular beam epitaxy (MBE) [4,9], halide vapor phase epitaxy (HVPE) [3,10], carbothermal reduction [11], metal organic chemical vapor deposition (MOCVD) [12,13], and low-pressure chemical vapor deposition (LPCVD) [14]. PLD and MBE, allowing for precision controllability, are able to achieve high-quality β-Ga 2 O 3 films on sapphire. Nevertheless, their growth rate is not sufficient for fast growth of thick films. HVPE, carbothermal reduction, and MOCVD methods show competitive growth rates over 5 µm/h. Among these, carbothermal reduction is a promising technique for the growth of thick β-Ga 2 O 3 films, as it avoids the use of corrosive precursor gases. However, the crystal quality is significantly degraded under fast growth because of the lattice mismatch between the corundum structure of sapphire (α-Al 2 O 3 ) and the monoclinic structure of β-Ga 2 O 3 . One Materials 2023, 16, 2775 2 of 10 solution for heteroepitaxy on sapphire is to use a larger bandgap material as a buffer layer to mitigate the lattice mismatch-induced strain, as is commonly known for growth of GaN on AlN/sapphire template. In the case of β-Ga 2 O 3 , we employed the ((Al x Ga 1−x ) 2 [15].
Currently, molecular beam epitaxy (MBE) [16], pulsed laser deposition (PLD) [17], and metal-organic vapor deposition (MOCVD) [18] are common methods used to grow (Al x Ga 1−x ) 2 O 3 alloys. However, achieving a growth strategy for β-(Al x Ga 1−x ) 2 O 3 film remains a challenge. This obstacle is associated with synthesizing a stable high-Al-content β-(Al x Ga 1−x ) 2 O 3 by direct growth techniques, such as MOCVD, MBE, and PLD, due to the phase transformation from β to γ for (Al x Ga 1−x ) 2 O 3 when the Al content exceeds 30% [19,20]. Additionally, these methods are based on expensive vacuum equipment. Instead of direct epitaxial β-(Al x Ga 1−x ) 2 O 3 on sapphire, we utilized the gallium (Ga) diffusion method, which has previously been used to fabricate Ga-diffused waveguides in sapphire [21].
In this paper, the β-(Al x Ga 1−x ) 2 O 3 film was grown on c-plane sapphire substrate in a high-temperature tubular furnace by the gallium (Ga) diffusion method, serving as an intermediate buffer layer for the subsequent heteroepitaxial growth of the thick β-Ga 2 O 3 film. The obtained β-(Al x Ga 1−x ) 2 O 3 film displayed a high crystal quality, with a thickness of approximately 750 nm. The distribution of Al components in the film was homogenous, with an Al content of approximately 62%. After synthesizing the β-(Al x Ga 1−x ) 2 O 3 buffer layer, the β-Ga 2 O 3 thick film was further deposited on the β-(Al x Ga 1−x ) 2 O 3 /sapphire template using two methods: carbothermal reduction, reported recently by our group [11], and HVPE. Finally, we characterized the properties of the β-Ga 2 O 3 thick film on sapphire with and without the β-(Al x Ga 1−x ) 2 O 3 buffer layer. The results showed that the β-Ga 2 O 3 thick film, grown on a sapphire substrate with a β-(Al x Ga 1−x ) 2 O 3 buffer layer, improved crystal orientation and surface quality.

Experiments
The (Al x Ga 1−x ) 2 O 3 film was synthesized using a high-temperature tubular furnace, as illustrated in Figure 1. Ga2O3 powder with purity of 99.999% as the source material was put in a corundum crucible. The sapphire substrate was inserted into the corundum crucible, and the system was subjected to setting temperature of 1450 • C. The growth process took place in an anoxic atmosphere, where 1.5 slm argon (Ar) was maintained at the pressure of 3 × 10 4 Pa for 2 h. In a neutral gas atmosphere, Ga 2 O 3 powder underwent decomposition into volatile Ga 2 O(g), which further decomposed into gaseous Ga, as illustrated below: The Ga species diffused into α-Al 2 O 3 , resulting in the formation of a β-(Al x Ga 1−x ) 2 O 3 buffer layer according to the Al 2 O 3 -Ga 2 O 3 phase diagram.
The Ga 2 O 3 thick film was grown using the carbothermal reduction method in a homemade growing system as shown in Figure 2. During the growth process, 20 sccm of O 2 and 500 sccm of Ar were kept for 2 h at a pressure of 3 × 10 4 Pa. The growth condition for HVPE was as follows: the ratio of flow rate between HCl and O 2 , setting growth temperature, and pressure were 10/30, 1060 • C, and 5 × 10 4 Pa, respectively. The growth was kept for 2 h for HVPE. The Ga2O3 thick film was grown using the carbothermal reduction method in a homemade growing system as shown in Figure 2. During the growth process, 20 sccm of O2 and 500 sccm of Ar were kept for 2 h at a pressure of 3 × 10 4 Pa. The growth condition for HVPE was as follows: the ratio of flow rate between HCl and O2, setting growth temperature, and pressure were 10/30, 1060 °C, and 5 × 10 4 Pa, respectively. The growth was kept for 2 h for HVPE.

Results and Discussions
The thickness of the (AlxGa1−x)2O3 film was investigated using cross-sectional scanning electron microscopy (FEI Nova Nano SEM 450), as shown in Figure 3a, which presents a clear interface between (AlxGa1−x)2O3 and sapphire substrate. The β-(AlxGa1−x)2O3 film thickness was measured to be 750 nm, corresponding to a growth rate of 375 nm/h. The surface morphology and roughness of the (AlxGa1−x)2O3 film were characterized by top-view scanning electron microscopy and atomic force microscopy (AFM, Dimension Icon, Bruker, Germany). Figure 3b shows the equilateral triangular morphology of the (AlxGa1−x)2O3 film, which corresponds to the arrangement of oxygen atoms (equilateral triangles) on a c-plane sapphire substrate. When β-Ga2O3 is grown on a c-plane sapphire substrate, the oxygen atoms on the surface between the β-Ga2O3 (-201) plane and the c-plane sapphire are arranged in equilateral triangles, leading to (-201)oriented growth. The AFM image corroborates the equilateral triangular morphology and  The Ga2O3 thick film was grown using the carbothermal reduction method in a homemade growing system as shown in Figure 2. During the growth process, 20 sccm of O2 and 500 sccm of Ar were kept for 2 h at a pressure of 3 × 10 4 Pa. The growth condition for HVPE was as follows: the ratio of flow rate between HCl and O2, setting growth temperature, and pressure were 10/30, 1060 °C, and 5 × 10 4 Pa, respectively. The growth was kept for 2 h for HVPE.

Results and Discussions
The thickness of the (AlxGa1−x)2O3 film was investigated using cross-sectional scanning electron microscopy (FEI Nova Nano SEM 450), as shown in Figure 3a, which presents a clear interface between (AlxGa1−x)2O3 and sapphire substrate. The β-(AlxGa1−x)2O3 film thickness was measured to be 750 nm, corresponding to a growth rate of 375 nm/h. The surface morphology and roughness of the (AlxGa1−x)2O3 film were characterized by top-view scanning electron microscopy and atomic force microscopy (AFM, Dimension Icon, Bruker, Germany). Figure 3b shows the equilateral triangular morphology of the (AlxGa1−x)2O3 film, which corresponds to the arrangement of oxygen atoms (equilateral triangles) on a c-plane sapphire substrate. When β-Ga2O3 is grown on a c-plane sapphire substrate, the oxygen atoms on the surface between the β-Ga2O3 (-201) plane and the c-plane sapphire are arranged in equilateral triangles, leading to (-201)oriented growth. The AFM image corroborates the equilateral triangular morphology and corresponds to a root mean square roughness (RMS) of around 2.10 nm.

Results and Discussions
The thickness of the (Al x Ga 1−x ) 2 O 3 film was investigated using cross-sectional scanning electron microscopy (FEI Nova Nano SEM 450), as shown in Figure 3a, which presents a clear interface between (Al x Ga 1−x ) 2 O 3 and sapphire substrate. The β-(Al x Ga 1−x ) 2 O 3 film thickness was measured to be 750 nm, corresponding to a growth rate of 375 nm/h. The surface morphology and roughness of the (Al x Ga 1−x ) 2 O 3 film were characterized by top-view scanning electron microscopy and atomic force microscopy (AFM, Dimension Icon, Bruker, Germany). Figure 3b shows the equilateral triangular morphology of the (Al x Ga 1−x ) 2 O 3 film, which corresponds to the arrangement of oxygen atoms (equilateral triangles) on a c-plane sapphire substrate. When β-Ga 2 O 3 is grown on a c-plane sapphire substrate, the oxygen atoms on the surface between the β-Ga 2 O 3 (-201) plane and the c-plane sapphire are arranged in equilateral triangles, leading to (-201)-oriented growth. The AFM image corroborates the equilateral triangular morphology and corresponds to a root mean square roughness (RMS) of around 2.10 nm. Materials 2023, 16, x FOR PEER REVIEW 4 of 10 The crystalline orientation was characterized using high-resolution X-ray diffraction (HRXRD, Bruker D8 Advance).  Table 1, the diffraction peak position of β-(AlxGa1−x)2O3 film shifted to a higher diffraction angle. This phenomenon arose because the Ga 3+ ion was replaced by the smaller-radius Al 3+ ion, causing the lattice spacing to shrink and the diffraction peak to move to a higher angle. Considering the monoclinic structure, the Al compositions in β-(AlxGa1−x)2O3 film was determined using following expression [22,23].
where h = −4, k = 0, and l = 2. Based on the (-402) diffraction peak position and Equation (3), we obtained the Al compositions in films to be around 62%. The crystalline quality of β-(AlxGa1−x)2O3 film was characterized by ω rocking curve spectra. As shown in Figure 4b, the full width at half maximum (FWHM) of the (-201) plane was around 0.42°, indicating that the β-(AlxGa1−x)2O3 film had high crystalline quality, although the crystalline quality of β-(AlxGa1−x)2O3 films prepared by PLD, MBE, and MOVD deteriorated when the Al composition exceeded 30% [19,20]. The crystalline orientation was characterized using high-resolution X-ray diffraction (HRXRD, Bruker D8 Advance).  Table 1, the diffraction peak position of β-(Al x Ga 1−x ) 2 O 3 film shifted to a higher diffraction angle. This phenomenon arose because the Ga 3+ ion was replaced by the smaller-radius Al 3+ ion, causing the lattice spacing to shrink and the diffraction peak to move to a higher angle. Considering the monoclinic structure, the Al compositions in β-(Al x Ga 1−x ) 2 O 3 film was determined using following expression [22,23].
where h = −4, k = 0, and l = 2. Based on the (-402) diffraction peak position and Equation (3), we obtained the Al compositions in films to be around 62%. The crystalline quality of β-(Al x Ga 1−x ) 2 O 3 film was characterized by ω rocking curve spectra. As shown in Figure 4b, the full width at half maximum (FWHM) of the (-201) plane was around 0.42 • , indicating that the β-(Al x Ga 1−x ) 2 O 3 film had high crystalline quality, although the crystalline quality of β-(Al x Ga 1−x ) 2 O 3 films prepared by PLD, MBE, and MOVD deteriorated when the Al composition exceeded 30% [19,20].  The content of Al in β-(AlxGa1−x)2O3 films was further determined by X-ray photoelectron spectroscopy (XPS, K-alpha+). A wide survey spectrum clearly showed peaks of Al 2s and Al 2p for the β-(AlxGa1−x)2O3 film, along with β-Ga2O3 crystal as a reference, as shown in Figure 5a. The Al compositions in the film were estimated from the Al 2p and Ga 2p core-level peak areas, considering the sensitivity factors of the elements. Figure 5b,c show the Al 2p and Ga 2p core-level spectra for the film. The Al 2p peak in the β-(AlxGa1−x)2O3 film displayed a binding energy of 73.96 eV. This shift towards a lower binding energy compared with the Al 2p peak in the sapphire substrate (74.5 eV) can be attributed to the formation of Al-O-Ga bonds [24]. Based on the XPS results, the Al composition in the β-(AlxGa1−x)2O3 film was around 61.5%, which is consistent with the quantitative results obtained from the XRD analysis.   The content of Al in β-(Al x Ga 1−x ) 2 O 3 films was further determined by X-ray photoelectron spectroscopy (XPS, K-alpha+). A wide survey spectrum clearly showed peaks of Al 2s and Al 2p for the β-(Al x Ga 1−x ) 2 O 3 film, along with β-Ga 2 O 3 crystal as a reference, as shown in Figure 5a. The Al compositions in the film were estimated from the Al 2p and Ga 2p core-level peak areas, considering the sensitivity factors of the elements. Figure 5b,c show the Al 2p and Ga 2p core-level spectra for the film. The Al 2p peak in the β-(Al x Ga 1−x ) 2 O 3 film displayed a binding energy of 73.96 eV. This shift towards a lower binding energy compared with the Al 2p peak in the sapphire substrate (74.5 eV) can be attributed to the formation of Al-O-Ga bonds [24]. Based on the XPS results, the Al composition in the β-(Al x Ga 1−x ) 2 O 3 film was around 61.5%, which is consistent with the quantitative results obtained from the XRD analysis.   The content of Al in β-(AlxGa1−x)2O3 films was further determined by X-ray photoelectron spectroscopy (XPS, K-alpha+). A wide survey spectrum clearly showed peaks of Al 2s and Al 2p for the β-(AlxGa1−x)2O3 film, along with β-Ga2O3 crystal as a reference, as shown in Figure 5a. The Al compositions in the film were estimated from the Al 2p and Ga 2p core-level peak areas, considering the sensitivity factors of the elements. Figure 5b,c show the Al 2p and Ga 2p core-level spectra for the film. The Al 2p peak in the β-(AlxGa1−x)2O3 film displayed a binding energy of 73.96 eV. This shift towards a lower binding energy compared with the Al 2p peak in the sapphire substrate (74.5 eV) can be attributed to the formation of Al-O-Ga bonds [24]. Based on the XPS results, the Al composition in the β-(AlxGa1−x)2O3 film was around 61.5%, which is consistent with the quantitative results obtained from the XRD analysis.  The bandgap of β-(Al x Ga 1−x ) 2 O 3 film was determined by analyzing the O 1s corelevel spectra in XPS. This approach has been established by previous studies [25][26][27][28]. The bandgap energy can be derived from the difference between the core-level peak energy and the initial inelastic losses [29]. Figure 6 presents the O1s core-level spectra obtained from XPS analysis of the β-(Al x Ga 1−x ) 2 O 3 film. The O1s peak energy was 530.8 eV, while the initial inelastic losses were 536.8 eV. The bandgap of β-(Al x Ga 1−x ) 2 O 3 film extracted from the O1s spectra was approximately 6.0 ± 0.1 eV, which is consistent with the bandgap energy estimated by absorption spectra in previous literature [30]. The impurity in the β-(AlxGa1-x)2O3 film was investigated using time-of-flight secondary ion mass spectrometry (SIMS, IONTOF 5). Figure 7 shows the TOF-SIMS depth profile for the β-(Al x Ga 1−x ) 2 O 3 film, which revealed that the impurity in the film was negligible. Additionally, the β-(Al x Ga 1−x ) 2 O 3 film exhibited a homogenous distribution of Al. The bandgap of β-(AlxGa1−x)2O3 film was determined by analyzing the O 1s core-level spectra in XPS. This approach has been established by previous studies [25][26][27][28]. The bandgap energy can be derived from the difference between the core-level peak energy and the initial inelastic losses [29]. Figure 6 presents the O1s core-level spectra obtained from XPS analysis of the β-(AlxGa1−x)2O3 film. The O1s peak energy was 530.8 eV, while the initial inelastic losses were 536.8 eV. The bandgap of β-(AlxGa1−x)2O3 film extracted from the O1s spectra was approximately 6.0 ± 0.1 eV, which is consistent with the bandgap energy estimated by absorption spectra in previous literature [30]. The impurity in the β-(AlxGa1-x)2O3 film was investigated using time-of-flight secondary ion mass spectrometry (SIMS, IONTOF 5). Figure 7 shows the TOF-SIMS depth profile for the β-(AlxGa1−x)2O3 film, which revealed that the impurity in the film was negligible. Additionally, the β-(AlxGa1−x)2O3 film exhibited a homogenous distribution of Al.   The bandgap of β-(AlxGa1−x)2O3 film was determined by analyzing the O 1s core-level spectra in XPS. This approach has been established by previous studies [25][26][27][28]. The bandgap energy can be derived from the difference between the core-level peak energy and the initial inelastic losses [29]. Figure 6 presents the O1s core-level spectra obtained from XPS analysis of the β-(AlxGa1−x)2O3 film. The O1s peak energy was 530.8 eV, while the initial inelastic losses were 536.8 eV. The bandgap of β-(AlxGa1−x)2O3 film extracted from the O1s spectra was approximately 6.0 ± 0.1 eV, which is consistent with the bandgap energy estimated by absorption spectra in previous literature [30]. The impurity in the β-(AlxGa1-x)2O3 film was investigated using time-of-flight secondary ion mass spectrometry (SIMS, IONTOF 5). Figure 7 shows the TOF-SIMS depth profile for the β-(AlxGa1−x)2O3 film, which revealed that the impurity in the film was negligible. Additionally, the β-(AlxGa1−x)2O3 film exhibited a homogenous distribution of Al.   After preparing the buffer layer, the β-(AlxGa1−x)2O3/sapphire template was transferred to a fast epitaxial β-Ga2O3 thick film through carbothermal reduction and HVPE techniques, respectively. A reference sample of β-Ga2O3 on sapphire without a buffer layer was also grown under the same conditions for comparison. The growth rate for all samples was approximately 4~6 µm/h, depending on the film thickness measured by cross-sectional SEM image as shown in Supplementary Figure S1. Figure 9 presents the θ-2θ scan XRD characterization of β-Ga2O3 thick film on sapphire with and without a β-(AlxGa1−x)2O3 buffer layer, respectively. The samples grown on the buffer layer showed a clear appearance of dominant (-201) and high-order β-Ga2O3 diffraction peaks, as shown in Figure 9b,c for the growth carried out by carbothermal reduction and HVPE, respectively. However, the β-Ga2O3 on sapphire without a buffer layer revealed a competitive crystal orientation of (400), (002), (-403), and (-313) peaks marked in Figure  9a, except for the peak of (-201) orientation planes. This competitive crystal orientation is due to the lattice mismatch related anisotropic growth, which leads to the existence of rhombic prism faces, as marked in the rectangle area indicated in Figure 10a. The XRD results demonstrate that miscellaneous crystalline facets were strongly inhibited for β-Ga2O3 thick film on sapphire by means of a β-(AlxGa1−x)2O3 buffer layer, which implies a much-improved crystalline quality. The surface properties of β-Ga2O3 thick film were investigated by SEM. As shown in Figure 10a, the surface morphology of β-Ga2O3 on sapphire without a buffer layer displayed the pseudo hexagonal shape with an average grain size of 4 µm and welldefined boundaries. The details of coalesced β-Ga2O3 grain were composed of rhombic After preparing the buffer layer, the β-(Al x Ga 1−x ) 2 O 3 /sapphire template was transferred to a fast epitaxial β-Ga 2 O 3 thick film through carbothermal reduction and HVPE techniques, respectively. A reference sample of β-Ga 2 O 3 on sapphire without a buffer layer was also grown under the same conditions for comparison. The growth rate for all samples was approximately 4~6 µm/h, depending on the film thickness measured by cross-sectional SEM image as shown in Supplementary Figure S1. Figure 9 presents the θ-2θ scan XRD characterization of β-Ga 2 O 3 thick film on sapphire with and without a β-(Al x Ga 1−x ) 2 O 3 buffer layer, respectively. The samples grown on the buffer layer showed a clear appearance of dominant (-201) and high-order β-Ga 2 O 3 diffraction peaks, as shown in Figure 9b,c for the growth carried out by carbothermal reduction and HVPE, respectively. However, the β-Ga 2 O 3 on sapphire without a buffer layer revealed a competitive crystal orientation of (400), (002), (-403), and (-313) peaks marked in Figure 9a, except for the peak of (-201) orientation planes. This competitive crystal orientation is due to the lattice mismatch related anisotropic growth, which leads to the existence of rhombic prism faces, as marked in the rectangle area indicated in Figure 10a. The XRD results demonstrate that miscellaneous crystalline facets were strongly inhibited for β-Ga 2 O 3 thick film on sapphire by means of a β-(Al x Ga 1−x ) 2 O 3 buffer layer, which implies a much-improved crystalline quality. After preparing the buffer layer, the β-(AlxGa1−x)2O3/sapphire template was transferred to a fast epitaxial β-Ga2O3 thick film through carbothermal reduction and HVPE techniques, respectively. A reference sample of β-Ga2O3 on sapphire without a buffer layer was also grown under the same conditions for comparison. The growth rate for all samples was approximately 4~6 µm/h, depending on the film thickness measured by cross-sectional SEM image as shown in Supplementary Figure S1. Figure 9 presents the θ-2θ scan XRD characterization of β-Ga2O3 thick film on sapphire with and without a β-(AlxGa1−x)2O3 buffer layer, respectively. The samples grown on the buffer layer showed a clear appearance of dominant (-201) and high-order β-Ga2O3 diffraction peaks, as shown in Figure 9b,c for the growth carried out by carbothermal reduction and HVPE, respectively. However, the β-Ga2O3 on sapphire without a buffer layer revealed a competitive crystal orientation of (400), (002), (-403), and (-313) peaks marked in Figure  9a, except for the peak of (-201) orientation planes. This competitive crystal orientation is due to the lattice mismatch related anisotropic growth, which leads to the existence of rhombic prism faces, as marked in the rectangle area indicated in Figure 10a. The XRD results demonstrate that miscellaneous crystalline facets were strongly inhibited for β-Ga2O3 thick film on sapphire by means of a β-(AlxGa1−x)2O3 buffer layer, which implies a much-improved crystalline quality. The surface properties of β-Ga2O3 thick film were investigated by SEM. As shown in Figure 10a, the surface morphology of β-Ga2O3 on sapphire without a buffer layer displayed the pseudo hexagonal shape with an average grain size of 4 µm and welldefined boundaries. The details of coalesced β-Ga2O3 grain were composed of rhombic prism faces, which contribute to the sub-peak in XRD measurement as we mentioned above. Taking into account the β-Ga2O3 thick film on sapphire with β-(AlxGa1−x)2O3 buffer layer, the SEM image explicitly unveiled a significant improvement in surface quality through the obvious transition from grain island-like morphology to 2D continuous growth, as shown in Figure 10b,c, respectively. The surface roughness was finally identified by AFM measurements of 5 × 5 µm. The AFM results displayed highly correlated morphology with the high magnification SEM images for all three samples. The RMS surface roughness was 74 nm for the β-Ga2O3 thick film on sapphire without a buffer layer, as shown in Figure 11a. By contrast, the use of the β-(Al1−xGax)2O3 buffer layer resulted in a much smoother surface, as confirmed by the RMS values of 9 nm and 5 nm for the epitaxial film prepared by carbothermal reduction and HVPE, respectively, as shown in Figure 11b,c. Further effort would be to optimize the growth parameter to obtain an even smoother surface without additional pits. The surface properties of β-Ga 2 O 3 thick film were investigated by SEM. As shown in Figure 10a, the surface morphology of β-Ga 2 O 3 on sapphire without a buffer layer displayed the pseudo hexagonal shape with an average grain size of 4 µm and well-defined boundaries. The details of coalesced β-Ga 2 O 3 grain were composed of rhombic prism faces, which contribute to the sub-peak in XRD measurement as we mentioned above. Taking into account the β-Ga 2 O 3 thick film on sapphire with β-(Al x Ga 1−x ) 2 O 3 buffer layer, the SEM image explicitly unveiled a significant improvement in surface quality through the obvious transition from grain island-like morphology to 2D continuous growth, as shown in Figure 10b,c, respectively.
The surface roughness was finally identified by AFM measurements of 5 × 5 µm. The AFM results displayed highly correlated morphology with the high magnification SEM images for all three samples. The RMS surface roughness was 74 nm for the β-Ga 2 O 3 thick film on sapphire without a buffer layer, as shown in Figure 11a. By contrast, the use of the β-(Al 1−x Ga x ) 2 O 3 buffer layer resulted in a much smoother surface, as confirmed by the RMS values of 9 nm and 5 nm for the epitaxial film prepared by carbothermal reduction and HVPE, respectively, as shown in Figure 11b,c. Further effort would be to optimize the growth parameter to obtain an even smoother surface without additional pits. prism faces, which contribute to the sub-peak in XRD measurement as we mentioned above. Taking into account the β-Ga2O3 thick film on sapphire with β-(AlxGa1−x)2O3 buffer layer, the SEM image explicitly unveiled a significant improvement in surface quality through the obvious transition from grain island-like morphology to 2D continuous growth, as shown in Figure 10b,c, respectively. The surface roughness was finally identified by AFM measurements of 5 × 5 µm. The AFM results displayed highly correlated morphology with the high magnification SEM images for all three samples. The RMS surface roughness was 74 nm for the β-Ga2O3 thick film on sapphire without a buffer layer, as shown in Figure 11a. By contrast, the use of the β-(Al1−xGax)2O3 buffer layer resulted in a much smoother surface, as confirmed by the RMS values of 9 nm and 5 nm for the epitaxial film prepared by carbothermal reduction and HVPE, respectively, as shown in Figure 11b,c. Further effort would be to optimize the growth parameter to obtain an even smoother surface without additional pits.