Wide bandgap single crystal Ga₂O₃ has been intensively researched due to its potential for realizing next generation ultra-wide band gap (E_g ~4.9 eV) electronic/optoelectronic applications such as high-power transistors and photodetectors operating in the solar blind wavelength range [1]. In addition, the availiablity of single crystal Ga₂O₃ substrates, manufactured by the conventional bulk crystal growth technique, has led to growing interest in developing epitaxial growth of Ga₂O₃ on either native Ga₂O₃ or foreign substrates (such as c- or m-plane sapphire substrate). Notably, single crystal Ga₂O₃ is known to exhibit different types of polymorphs, such as α-, β-, γ-, ε-, and κ- phases [2]. In addition to the most thermodynamically stable β phase Ga₂O₃, increasing attention has been given to different polymorphs, especially κ-Ga₂O₃, which is the second most stable form. This is because of its potential for next generation ferroelectric applications as it possesses large spontaneous polarization along its c- axis [3]. Many vacuum deposition techniques, such as pulsed laser deposition [4], molecular beam epitaxy [5], hydride vapor phase epitaxy [6], mist chemical vapor deposition (mist-CVD) [7,8], and metalorganic chemical vapor deposition (MOCVD) [9,10], have demonstrated epitaxial growth of single crystal κ-Ga₂O₃. Among those, MOCVD is the most economical and technologically matured crystal growth technique, and has been widely adopted in the lighting and microelectronics industries. In the MOCVD process, understanding the effects of different growth conditions is critical for achieving highly desired material properties, which include crystallinity and the electrical and optical properties of epitaxial thin films. Here, we present the epitaxial growth of κ-Ga₂O₃ by MOCVD with H₂O as an oxygen precursor. In particular, we investigate the impacts of key epitaxial growth parameters such as group III molar flow rate, H₂O vapor flow rate, and type of carrier gas, carried out by x-ray diffraction (XRD), atomic force microscopy (AFM), and photoluminescence (PL) measurements

The epitaxial growth of Ga₂O₃ on a c-plane sapphire substrate was performed using trimethyl-Ga (TMGa) and ultra-pure H₂O as Ga and O precursors, respectively, within a commercial AIX200/4 MOCVD reactor. The growth temperature was maintained at 690 °C and 50 mbar, while the TMGa and H₂O bubbler temperatures were set at 5 and 35 °C, respectively. The impact of growth temperature on the surface morphology and crystallinity was reported elsewhere [11], while 50 mbar is a typical growth pressure employed in low pressure MOCVD processes. During the growth, SiH4 was simultaneously supplied for Si doping. After the growth, cross-sectional field emission scanning electron microscopy (SEM) was used to measure the grown thickness. Atomic force microscopy was employed to further characterize the surface morphology. In addition, the structural characteristics of the Ga₂O₃ film were evaluated by high-resolution x-ray diffraction (HR-XRD). Electrical characteristics, including resistivity, mobility, and carrier concentration, were obtained by using the Van der Pauw Hall technique at room temperature. It must be noted that all of the as-grown κ-Ga₂O₃ samples in the present study were electrically resistive, although they were doped with Si. High conductivity was obtained only after post-growth annealing at 1,000 ℃, which in turn leads to a phase transition from κ to a thermodynamically stable β phase. A prior study showed that when the annealing temperature is higher than 900 °C, the metastable κ-Ga₂O₃ phase transforms to the most stable β phase, which was supported by XRD and scanning transmission electron microscopy results [11]. Optical characterizations were performed by PL measurement using an Ar ion laser with excitation wavelength of 244 nm.

3.1. Impact of group III and H₂O molar flow

The impacts of the TMGa flow rate on the growth rate and the structural properties were investigated by varying the TMGa flow rate from 3 to 7 sccm while maintaining the H₂O vapor flow rate at 1,600 sccm (8.95 × 10³ mol/min) at the growth temperature of 690 °C. For this study, only H₂ was used as a carrier gas.

As shown in Fig. 1, all samples exhibited similar multiple peaks located at 9.6, 19.4, 29.96, and 41.68°, which correspond to the calculated Bragg’s angles of the (002), (004), (006), and (008) planes of orthorhombic κ-phase Ga₂O₃, as summarized in Table I. This observation is in good agreement with a prior study, which showed the stabilized κ-phase of Ga₂O₃ grown on a c-plane sapphire wafer, verified by a TEM investigation [11]. These XRD peaks exhibited similar full width at half maximum (FWHM) values 600 arcsec at κ-Ga₂O₃ (004) reflection.

Figure 1. XRD spectra taken from samples grown with varying TMGa flow rates, where the VI/III molar flow ratio is also indicated in parentheses.

Table I. Summary of measured and calculated reflection angle from Ga₂O₃ grown on a c-plane sapphire substrate (lattice parameter of orthorhombic κ-Ga₂O₃: a=5.12 Å, b=8.78 Å, and c=9.4 Å)..

Phase	(h k l)	d-spacing [Å]	Calculated Bragg’s angle [°]	Measured peak position [°]
κ-Ga2O3 (Orthorhombic)	(002)	4.705	9.4	9.6
κ-Ga2O3 (Orthorhombic)	(004)	2.353	19.1	19.4
κ-Ga2O3 (Orthorhombic)	(006)	1.568	29.4	29.96
κ-Ga2O3 (Orthorhombic)	(008)	1.176	40.9	41.68

Figures 2(a)-(c) show AFM images of the samples grown with different TMGa flow rates. As the TMGa flow rate was increased from 3 (molar flow rate: 17.67 μmol/min) to 5 sccm, the root mean square (RMS) roughness value of the film improved from ~7.7 to 3.7 nm. On the other hand, when the TMGa flow rate was further increased from 5 to 7 sccm, the surface morphology remained similar with a RMS close to 4 nm. In addition, a monotonic, nearly linear increase in the growth rate was observed as the TMGa flow rate was increased from 3 to 7 sccm. Figure 2(d) summarizes these observations.

Figure 2. (a)–(c) 5 × 5 μm² AFM images showing the surface morphology of as-grown samples grown with varying TMGa flow rate, (d) summary of the growth rate and RMS roughness (Rq) as a function of TMGa flow rate, where the dashed lines are a guide for the eye.

Next, in order to investigate the impact of the H₂O vapor flow rate on the growth rate and subsequent structural/optical properties, the H₂O vapor flow rate was varied from 960 to 1,600 sccm, while maintaining the TMGa flow rate at 5 sccm.

Figures 3(a)-(c) show the surface morphology of the samples grown at varying H₂O vapor flow rates. When the H₂O vapor flow rate of 960 sccm was used, no (or partial) deposition was observed on most of the surface, suggesting that a 960 sccm H₂O vapor flow rate is not sufficient to form a uniform thin film over the entire area of the substrate. On the other hand, as the H₂O vapor flow rate was increased to either 1,280 or 1,600 sccm, relatively uniform deposition was observed with a similar growth rate, as summarized in Fig. 3(d). In contrast, the RMS roughness value of the film improved from ~5.13 to 3.7 nm as the H₂O vapor flow rate was increased from 1,280 to 1,600 sccm. No significant change in the growth rate (15–17 nm/min) was observed, indicating that the flow rate of TMGa is a dominant factor in determining the growth rate when a sufficient H₂O vapor flow rate is supplied to form a film.

Figure 3. (a) Top-view SEM showing no (or partial) deposition when a H₂O vapor flow rate of 960 sccm was employed, (b) and (c) 5 × 5 μm² AFM images showing the surface morphology of as-grown samples grown with varying H₂O vapor flow rate, and (d) summary of the growth rate and measured RMS roughness.

The samples grown with varying H₂O vapor flow rate exhibited multiple peaks located at 9.6, 19.4, 29.96, and 41.68°, as shown in Fig. 4, and these peaks are attributed to Bragg’s angles of (002), (004), (006), and (008) planes of orthorhombic κ-phase Ga₂O₃. Similar FWHM values ranging from 540 to 700 arcsec were obtained, given that the resolution in the measurement of the XRD spectra was 54 arcsec.

Figure 4. XRD spectra taken from samples grown with varying H₂O vapor flow rates.

3.2. Growth using N₂ carrier gas

The impacts of the carrier gas on the structural, optical, and electrical properties of κ-Ga₂O₃ grown on a sapphire substrate were investigated by changing the carrier gas from H₂ to N₂, while maintaining the other growth parameters such as the TMGa flow rate (5 sccm), H₂O vapor flow rate (1,600 sccm), and the growth temperature (690 °C). For this study, only H₂ was fed into the source line while N₂ was used for the carrier gas line, such that the N₂/H₂ ratio within the reactor during the growth was 1:1. The κ-Ga₂O₃ was grown for 30 min.

While a similar degree of surface roughness was observed from the materials grown by either N₂ or H₂ carrier gas, a noticeable difference (15 vs. 12 nm/min) in the growth rate was observed as shown in Fig. 5. After post-growth in situ annealing at 1,000 °C under N₂ within the MOCVD reactor, which was utilized to activate Si doping for n-type conductivity, evidence of a phase transition from κ to β phase was again observed, based on the peak positions corresponding to the β phase (310) and (620) peaks, as shown in Fig. 6.

Figure 5. (a, b) AFM images showing the surface morphology and (c, d) cross sectional SEM images showing the thickness and the corresponding growth rate (R_growth) for Ga₂O₃ grown for 30 min on a sapphire substrate.

Figure 6. XRD spectra for as-grown (grown with either N₂ or H₂ carrier gas) and annealed samples in situ under N₂ ambient for 1 min at 1,000 °C.

Figure 7 shows the PL spectra measured from as-grown and annealed samples at 1,000 °C. Both the as-grown sample and the annealed sample exhibited a peak position near 420 nm. On the other hand, the annealed sample also showed a peak near 370 nm. While finding the origin of these emissions is a subject of our ongoing study, a comparison between the PL spectrum of the Ga₂O₃ substrate, which is the most stable β-phase, and that of the annealed samples at 1,000 °C reveals analogous peak shapes, indicating that the annealed sample has gone through a phase transition from κ to β phase. A detailed structural analysis related to similar observations was also reported elsewhere [11].

Figure 7. PL spectra taken from as-grown and annealed samples (grown with H₂ or N₂ carrier gas, respectively).

Table II shows a summary of the Hall measurement results obtained after post-growth thermal annealing. As-grown samples were relatively resistive with a resistivity value of ~510 ohm-cm. On the other hand, the resistivity value of the annealed sample was reduced to 0.16 ohm-cm with n-type conductivity. Table II summarizes the Hall mobility and carrier concentration. A higher Hall mobility (22 cm²/V-s) was observed from the annealed sample grown with N₂ carrier gas, in comparison to that (8 cm²/V-s) of the annealed sample grown with H₂ carrier gas. The exact origin of this improvement is the subject of our ongoing study. Complex chemical interactions between metalorganic precursors, their associated by-products after thermal decomposition at the growth front and hydrogen carrier gas, can lead to the formation of higher density hydrocarbon radicals within the Ga₂O₃ epitaxial layer. The higher density hydrocarbon radicals can act as point defects, which accounts for the observed lower electron mobility within the Ga₂O₃ epitaxial layer grown with N₂ carrier gas [12].

Table II. Summary of Hall measurement..

	Resistivity [ohm-cm]	Hall mobility [cm2/V-s]	Carrier concentration [cm−3]
As-grown (N2 carrier gas)	5.1 × 102	N/A	N/A
in situ annealed (N2 carrier gas)	1.6 × 10−1	22	1.9 × 1018(n-type)
(ref) in situ annealed (H2 carrier gas)	5.8 × 10−1	8	1.4 × 1018(n-type)

In conclusion, we have carried out a systematic study on the effects of key epitaxial growth parameters such as group III molar flow rate, H₂O vapor flow rate, and types of carrier gas by x-ray diffraction, atomic force microscopy, and photoluminescence measurements. A higher growth rate and improved surface morphology were observed as the TMGa flow increases. On the other hand, it was observed that a certain amount of H₂O flow rate is necessary for uniform surface coverage of the grown film. Furthermore, we show that the growth using N₂ as a carrier gas, instead of using the conventional H₂ carrier gas, can lead to higher Hall electron mobility, which was measured after post-growth annealing.

This work was supported by the Technology Innovation Program (RS-2023-00235844, Development of nano-structured materials and devices for super steep subthreshold swing) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea) (1415187621).