**1. Introduction**

As an ultra-wide-bandgap semiconductor with the obvious advantages of stable physical chemistry, low dielectric constant, and high mechanical strength, gallium oxide (Ga2O3) is attracting increasing attention as a new promising competitor to III-nitrides and SiC for various applications in high-voltage and high-power electronics and ultraviolet optoelectronics [1]. Compared with other oxides such as ZnO (3.24 eV) and In2O3 (3.6 eV), Ga2O3 has a larger bandgap energy of approximately 5 eV, which means a shorter absorption cuto ff wavelength and a much higher power application. There are a total of five di fferent polytypes (<sup>α</sup>, β, ε, δ, and γ) for Ga2O3. Until now, the most studied polytype was β-Ga2O3 because it is easy to obtain bulk and film β-Ga2O3 materials by the conventional crystal growth or epitaxial growth techniques, such as edge-defined film-fed growth, float-zone method, Czochralski method, molecular beam epitaxy, and metal organic chemical vapor deposition. β-Ga2O3 has a bandgap of 4.8 eV and high Baliga's figures of merit (FOM) of 3000, which is obviously superior to GaN and SiC. However, β-Ga2O3 is not the best candidate in various phases for the power application considering the bandgap. Compared to β phase, corundum-structured α-Ga2O3, another important phase for Ga2O3, has a wider bandgap of around 5.3 eV which can result in a larger Baliga's FOM, in theory. Thus, α-Ga2O3 has grea<sup>t</sup> potential for application in power devices. A wider bandgap

of about 5.3 eV means that the absorption cutoff wavelength can be shorter than 240 nm and then α-Ga2O3 is more suitable for ultraviolet optoelectronics. Although α-Ga2O3 has grea<sup>t</sup> application potential, the research of α-Ga2O3 still lags far behind β-Ga2O3 in large part because it is more difficult to obtain the high-quality α-Ga2O3 material than β-Ga2O3.

α-Ga2O3 is a metastable phase and the bulk material still cannot be obtained. There is no commercial α-Ga2O3 bulk substrate to date, so its homo-epitaxial growth is still difficult. Fortunately, the heterogeneous epitaxy provides an efficient way to obtain the α-Ga2O3 material. Corundum-structured gallium oxide belongs to the space group of R-3c with the lattice parameters a = b = 4.98 Å, c = 13.43 Å, α = β = 90◦, and γ = 120◦ [2], and the lattice mismatches between α-Ga2O3 and α-Al2O3 (sapphire) are only 4.81% and 3.54% in the a- and c-axis directions. The same crystal structure and small lattice mismatch make it easy to grow the α-Ga2O3 material on the α-Al2O3 sapphire substrate. More importantly, the same crystal structure means that it is attractive from the viewpoint of fabricating alloys for α-Ga2O3 with other corundum-structured materials, such as α-Al2O3, Fe2O3, and Cr2O3 for bandgap and material engineering [3], which is another advantage of α- Ga2O3.

Recent studies have shown that the growth of crystalline α-Ga2O3 on an inexpensive sapphire substrate is an efficient way to obtain the α-Ga2O3 material. The key technology for the growth of α-Ga2O3 is ultrasonic mist chemical vapor deposition (mist-CVD) method [4–6]. In the growth of a metal oxide, water solutions of safe and inexpensive chemicals containing the metal, for example, acetate or acetylacetonate, have been used as the source. By atomizing the source solution ultrasonically, it turns into mist particles, which are then transferred by a carrier gas to a reaction chamber. In this way, metal elements are supplied without the use of organometallic sources. They react with an oxygen source, which may be water or oxygen gas. This offers sufficient overpressure of oxygen with respect to the metal source and prevents the formation of oxygen vacancies. Therefore, mist-CVD method is suitable for epitaxial α-Ga2O3 on sapphire substrate and can reduce the material cost [7,8]. Based on the grown α-Ga2O3, various applications of Metal Epitaxial-Semiconductor Field Effect Transistor [9], Schottky barrier diodes [10], and solar-blind photodetectors [11] have been demonstrated. For example, a high performance Schottky diode with a breakdown voltage over 1 kV and a small specific on-resistance of 2.5 <sup>m</sup>Ω·cm<sup>2</sup> has been achieved and a normally-off MOSFET has been shown based on α-Ga2O3 material grown by a mist-CVD system [9,10]. However, the present device performance is greatly inferior to the β-Ga2O3 counterparts and the main reason is still the poor α-Ga2O3 film quality. Thus, more attention is urgently required to improve the quality of α-Ga2O3 now and in the future. Recently, highly crystalline α-Ga2O3 thin films have been successfully grown at atmospheric pressure by mist-CVD on c-sapphire substrates, whose temperatures of 400–500 ◦C are reasonably low and the optimal growth conditions of solution concentration, growth temperature, carrier gas velocity, and film thickness have also been investigated [12,13]. However, there is still no systematical study about how the different carrier gas affects the film quality.

In this paper, we systematically investigate the influence of different carrier gases (O2, N2, and air) on the film quality for the growth of α-Ga2O3 on c-plane sapphire substrates by using the mist-CVD method. It is demonstrated that the crystallization quality will be different when the gallium source is carried by different gases. When N2 and O2 are used as the carrier gases, α-Ga2O3 achieves a relative smooth surface. When O2 is the carrier gas, α-Ga2O3 achieves the smallest half-height width. The oxygen element in the carrier gas may be an important reason to prevent the generation of oxygen vacancies, thus influencing the quality of the thin films. The results provide constructive perspectives for the material quality improvement.

#### **2. Materials and Methods**

In the present experiment, we used gallium acetylacetonate as the gallium source, which was dissolved in deionized water. A small quantity of hydrochloric acid was added to dissolve gallium acetylacetonate completely. The concentration of the solution was adjusted to 0.05 M. By atomizing the source solution ultrasonically, it turned into mist particles (diameter of ~3 μm at an ultrasonic

frequency of 2.4 MHz), which was then carried by air, N2, and O2, respectively, to the heated reaction chamber. In the chamber, the (0001) sapphire substrate was placed on a sample holder that was kept at 400 ◦C, because under this condition a high quality α-Ga2O3 can be obtained, as shown in Figure S1 in the supplemental information. The growth time was kept at 1h and the rate of the carrier gas was set to be 6 L/min.

The structural properties of Ga2O3 films were investigated using a range of complementary techniques. X-ray diffraction (XRD) patterns were obtained from an X-ray diffractometer (D8 Advance, Bruker, Karlsruhe, Germany). The transmittance was measured by a dual-beam 950 UV-VIS spectrometer. XPS measurements were performed by the Escalab 250Xi (Waltham, MA, USA) with a source of monochromatic Al-Ka (1486.6 eV). The film morphologies were characterized by a field emission scanning electron microscope (SEM JSM-7800F, Tokyo, Japan), atomic force microscopy (AFM) (Agilent 5500, Palo Alto, Santa Clara, CA, USA), and high-resolution transmission electron microscopy (TEM) (Tecnai G2 F20 S-Twin, Hillsboro, OR, USA). The Raman spectra were measured using a confocal Jobin Yvon LavRam HR800 micro-Raman spectrometer (Edison, NJ, USA) with a charge-coupled device (CCD) detector.
