# Fileset [c-BGO-growth111017(SST)_2ndSubmission_Final.pdf](https://mdr.nims.go.jp/filesets/0ddb7479-ac62-4334-ae23-bab0d9f5dcae/download) ## Creator [Yuichi Oshima](https://orcid.org/0000-0001-8293-4891), [Elaheh Ahmadi](https://orcid.org/0000-0002-8330-9366), Stephen Kaun, Feng Wu, James S Speck ## Rights © 2017 IOP Publishing Ltd
This is an author-created, un-copyedited version of an article accepted for publication/published in Semiconductor Science and Technology. IOP Publishing Ltd is not responsible for any errors or omissions in this version of the manuscript or any version derived from it. The Version of Record is available online at https://doi.org/10.1088/1361-6641/aa9c4d.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/) ## Other metadata [Growth and etching characteristics of (001) β- Ga2O3 by plasma-assisted molecular beam epitaxy](https://mdr.nims.go.jp/datasets/8aa2e172-638d-4607-af4d-67df1403fa12) ## Fulltext 1 Growth and etching characteristics of (001) -Ga2O3 by plasma-assisted molecular beam epitaxy Yuichi Oshima, Elaheh Ahmadi*, Stephen Kaun, Feng Wu, and James S. Speck Materials Department, University of California, Santa Barbara, CA, 93106 USA *E-mail: elaheh@ece.ucsb.edu We investigated the homoepitaxial growth and etching characteristics of (001) -Ga2O3 by plasma-assisted molecular beam epitaxy. The growth rate of -Ga2O3 increased with increasing Ga-flux, reaching a clear plateau of 56 nm/h, and then decreased at higher Ga-flux. The growth rate decreased from 56 nm/h to 42 nm/h when the substrate temperature was increased from 750 to 800 °C. The growth rate was negative (net etching) when only Ga-flux was supplied. The etching rate proportionally increased with increasing the Ga-flux, reaching 84 nm/h. The etching was enhanced at higher temperatures. It was found that Ga-etching of (001) -Ga2O3 substrates prior to the homoepitaxial growth markedly improved the surface roughness of the film. 2 1. Introduction -Ga2O3 is attracting remarkable attention for its great potential to realize high-performance power devices due to its large band gap (Eg ~ 4.7 eV) [1] and availability of melt-grown high-quality single crystal substrates [2-5]. (010) -Ga2O3 substrates are used in most cases to fabricate -Ga2O3 power devices such as Schottky barrier diodes (SBD) [6], and field effect transistors (FET) [7]. Currently, the edge-defined film-fed growth (EFG) method is used to produce commercial -Ga2O3 single crystal substrates [3,4]. This growth method gives board-shaped crystal with a thickness of up to a few centimeters. In melt growth techniques of -Ga2O3, the crystal pulling (growth direction) needs to be toward the unique b-axis (i.e., [010] direction) so that the cleavage planes (100) and (001) are parallel to the growth direction to prevent twinning and formation of small angle grain boundaries. The area of EFG-grown (010) -Ga2O3 wafers, so far, has been limited to around 1015 mm2. Therefore, it is difficult to produce large-area (010) -Ga2O3 wafers with a diameter of 6 inches or more, which will be required for the mass production of -Ga2O3 power devices. The Czochralski (Cz) method is also under investigation [5]. The challenges for Czochralski growth of -Ga2O3 include difficulties in damage of the iridium crucibles, which is strongly enhanced with increasing the melt volume [8]. In contrast, there is no fundamental limitation to produce large-area -Ga2O3 wafers with a principal crystal plane in the [010] zone (e.g., planes (h0l)) by the EFG method. In practice, EFG-grown 6-inch (2̅01) bulk crystal has already been demonstrated by the Tamura Corporation. Unfortunately, (2̅01) homoepitaxial layers suffered from a high 3 density of stacking faults [9]. The (100) plane is also in the [010] zone. However, the MBE growth rate on the (100) plane is extremely small because of the preferential formation of volatile suboxide Ga2O, and (100) is not suitable for industrial use [6,10]. We therefore need to explore other crystal planes to enable high-quality epilayers which can be grown at reasonable growth rates. In this work, we focus on growth on (001) oriented -Ga2O3 substrates, which are also in the [010] zone and thus are scalable. In practice, Tamura Corporation has successfully grown large-size (001) -Ga2O3 board-like bulk crystals by EFG method and high-quality 2-inch (001) -Ga2O3 wafers are commercially available. In addition, Konishi et al. demonstrated (001) -Ga2O3 SBDs with breakdown voltage of over 1 kV using homoepitaxial film grown by halide vapor phase epitaxy [12]. Hence, (001) is promising for power device applications. To fabricate FETs or high electron mobility transistors (HEMTs) using (001) -Ga2O3, we need to develop thin film growth technologies for (001) -Ga2O3. Molecular beam epitaxy (MBE), which enables precise thickness control and abrupt interfaces, is advantageous for such purpose. The present work investigates positive and negative growth (i.e. etching) characteristics of (001) -Ga2O3 by MBE. 2. Experimental -Ga2O3 was grown on EFG-grown (001) -Ga2O3 single crystal substrates by plasma-assisted MBE (PA-MBE) at substrate temperatures Ts = 700 ~ 800 °C. Each -Ga2O3 substrate was indium-bonded on a Si backing wafer. Ts was measured by a thermocouple located near a heater behind the backing wafer. Metallic aluminum (melting temperature 4 660°C) on a -Ga2O3 substrate melted at Ts = 645°C. The Ga flux was supplied using conventional K-cells. Beam equivalent pressure (BEP) of the Ga-flux 𝑃𝐺𝑎0 was 1.710-8 ~ 9.510-8 Torr. Plasma-activated oxygen, presumably atomic oxygen, was supplied from a Veeco Uni-bulb oxygen RF-plasma source. The oxygen foreline pressure and RF plasma power were 60 Torr and 200 W, respectively. The Ga-flux, oxygen plasma supply, and Ts were kept constant throughout the growth. The film thickness was calculated from thickness fringe spacing in high resolution x-ray diffraction (HRXRD) symmetrical -2 scans measured in a triple-axis configuration using Cu K1 radiation. A very thin layer of -(AlxGa1-x)2O3 was grown for 1.5 min with Al-flux ratios (the ratio of Al flux to the total group III flux) of 0.021 ~ 0.137 in prior to the growth of -Ga2O3 in order to obtain clear thickness fringes in -2 scans. In the case of (010) -(AlxGa1-x)2O3 layers, the Al content x tends to be higher than the Al-flux ratio [13]. In this work, the Al content of the interlayers was not measured, but the tendency should be similar. Typical -2 scan profile is shown in Fig. 1. The surface morphology was observed by atomic force microscopy (AFM). Figure 1. XRD -2 scan profile of a (001) -Ga2O3 homoepitaxial wafer. 31.2 31.4 31.6 31.8 32 32.2Intnsity [arb. units]2 [deg]002 5 3. Results and discussion Figure 2 shows the dependence of the growth rate of (001) -Ga2O3 on 𝑃𝐺𝑎0 at Ts = 750 °C. Our results on (100) [10] and (010) [11] planes are also shown for comparison. The growth rate on (001) first increased with increasing 𝑃𝐺𝑎0 , reaching clear plateau, and then decreased at higher 𝑃𝐺𝑎0 . It has been reported that the dependence of growth rate on 𝑃𝐺𝑎0 for (2̅01) -Ga2O3 homoepitaxial growth shows similar behavior, and growth rate decrease in Ga-rich condition can be attributed to the formation of volatile suboxide Ga2O [14]. Thermodynamic analysis of PA-MBE of Ga2O3 tells us that the formation of Ga2O is favored in the Ga-rich regime, and the growth rate should decrease with increasing 𝑃𝐺𝑎0 [15]. However, the thermodynamic analysis does not predict the plateau. The existence of the plateau indicates that Ga2O starts to form even in O-rich conditions. Vogt et al. pointed out that the (2̅01) -Ga2O3 surface could act as a catalyst for Ga2O formation [14]. It is likely that a similar phenomenon is taking place also on (001) -Ga2O3. Such catalytic effect could be different depending on the crystal plane and result in different growth rates. The growth rate at the plateau was approximately 56 nm/h, which is about 1/4 of that on (010), and about twice of that on (100) under similar growth conditions. Using ozone MBE, Sasaki et al. investigated the growth rate of -Ga2O3 for various crystal planes, and the trend shown here agrees with their results [6]. When oxygen plasma was not supplied, the growth rate was negative. The etching rate was estimated by comparing the homoepitaxial film thickness before and after the 30 min etching. The film thickness was measured by the method described in the experimental part. It was confirmed that no measurable etching took place even at Ts = 850°C when neither Ga-flux nor oxygen plasma was supplied. The etching rate proportionally 6 increased with increasing 𝑃𝐺𝑎0 , reaching 84 nm/h. This phenomenon can be utilized as a clean etching technique. This “Ga-etching technique” uses no contaminant, and can be carried out in MBE chambers. Therefore, it is possible to reduce surface contamination of -Ga2O3 substrates just before the MBE growth. CMP-related damage could also be removed. The Ga-etching can be applied not only to (001) but also to other crystal planes, such as (010). We have applied this technique to decrease the Si accumulation on (010) -Ga2O3 surface, and successfully demonstrated a modulation-doped FET [16]. Figure 2. Growth rate of -Ga2O3 as functions of 𝑃𝐺𝑎0 . Foreline oxygen pressures were 60 Torr for (010) and (001) samples, and 50 Torr for (100) samples. RF plasma power was 200W. Broken lines are guides for eyes. Figure 3 shows the dependence of the growth rate of (001) -Ga2O3 on substrate temperature Ts. The growth rate decreased with increasing Ts. Figure 4 shows the dependence of the etching rate of (001) -Ga2O3 on substrate temperature Ts. The etching rate increased with increasing Ts. The thermodynamic calculation does not predict 0 2 4 6 8 10-100-50050100150200Growth rate [nm/h] (010) (001) (100) 𝑃𝐺𝑎0 [10-8 Torr] (001) (w/o oxygen plasma) 7 significant increase of Ga2O formation with increasing Ts in the investigated temperature range [15]. It would be possible that the catalytic effect of Ga2O3 surface is enhanced at higher temperatures and results in reduced growth rate. Figure 3. Growth rate of (001) -Ga2O3 as a function of substrate temperature Ts. Figure 4. Etching rate of (001) -Ga2O3 as a function of substrate temperature Ts. 740 760 780 800 8200102030405060Ts [°C]Growth rate [nm/h]𝑃𝐺𝑎0 = 2.9 × 10−8 Torr 650 700 750 800 850 90004080120Etching rate [nm/h]Temperature [°C]𝑃𝐺𝑎0 = 5.9 × 10−8 Torr 8 Figures 5 (a), (b) and (c), (d) show AFM images of a virgin (001) -Ga2O3 substrate and Ga-etched surface of the same substrate respectively. The rms roughness of the virgin substrate was 0.2 nm, and the miscut was 0.03° in an azimuthal direction 63° clockwise from [010] to [100]. The etching was carried out with 𝑃𝐺𝑎0 = 5.910-8 Torr at Ts = 750 °C for 30 min. Under these conditions, the etched depth should be 29 nm. The rms roughness did not show significant increase by the Ga-etching although the step-terrace structure became less clear. Figure 5. AFM images of a (001) -Ga2O3 substrate. (a),(b): virgin substrates, (c),(d): After Ga-etching (-29 nm) RMS=0.2 nm (a) RMS=0.2 nm (b) (c) (d) [010] [100] 9 Figures 6 (a) and (b) show AFM images of homoepitaxial (001) -Ga2O3 layers grown on a virgin (001) -Ga2O3 substrate and Ga-etched one, respectively. The Ga-etched substrate was the same one shown in Figs. 5 (c) and (d). The films were grown with 𝑃𝐺𝑎0 = 2.910-8 Torr at Ts = 750 °C (56 nm/h) for 2 hrs. The film grown on a virgin substrate tends to be bumpy. On the other hand, the Ga-etched substrate gave very smooth film surface with rms roughness of ~0.2 nm. We speculate that the Ga-etching could remove surface contamination and/or polish-related damage, and resulted in the smooth surface. Stripe-like morphological features along [010] were observed on the surface. The direction of the stripes is different from the miscut direction. This is probably because the formation of such morphology should be attributed to the (100) faceting, which is also observed on HVPE-grown (001) homoepitaxial films [17]. Fig. 6 AFM images of (001) -Ga2O3 films grown on (a) virgin substrate, (b) Ga-etched substrate 4. Summary In summary, growth and etching characteristics of (001) -Ga2O3 was investigated by PA-MBE. The Ga-flux dependence of the growth rate exhibited clear plateau, probably due to the preferential formation of volatile suboxide Ga2O. The growth rate was negative (a) (b) [010] 10 when oxygen plasma was not supplied, and the etching rate increased proportionally to the Ga-flux. The rms roughness did not show significant increase after the Ga-etching. The rms roughness of homoepitaxial layer was markedly improved by the Ga-etching prior to the growth. Acknowledgements This work was supported by the Air Force Office of Scientific Research (AFOSR, Program Manager Dr. Ali Sayir) through grant # FA9550-14-1-0112. Additional support for J.S.S. was provided by the MRSEC Program of the U.S. National Science Foundation under Award No. DMR-1121053. References [1] Tippins H H 1965 Optical Absorption and Photoconductivity in the Band Edge of β-Ga2O3 Phys. Rev. 140 A316 [2] Víllora E G, Shimamura K, Yoshikawa Y, Aoki K, Ichinose N 2004 Large-size β-Ga2O3 single crystals and wafers J. Cryst. Growth 270 420 [3] Aida H, Nishiguchi K, Takeda H, Aota N, Sunakawa K, Yaguchi Y 2008 Growth of β-Ga2O3 Single Crystals by the Edge-Defined, Film Fed Growth Method J. Appl. Phys. 47 8506 [4] Kuramata A, Koshi K, Watanabe S, Yamaoka Y, Masui T, and Yamakoshi S 2016 High-quality β-Ga2O3 single crystals grown by edge-defined film-fed growth Jpn. J. Appl. Phys. 55 1202A2. [5] Galazka Z, Irmscher K, Uecker R, Bertram R, Pietsch M, Kwasniewski A, Naumann 11 M, Schulz T, Schewski R, Klimm D, and Bickermann M 2014 On the bulk β-Ga2O3 single crystals grown by the Czochralski method J. Cryst. Growth 404 184 [6] Sasaki K, Kuramata A, Masui T, Víllora E G, Shimamura K, and Yamakoshi S 2012 Device-Quality β-Ga2O3 Epitaxial Films Fabricated by Ozone Molecular Beam Epitaxy Appl. Phys. Express 5 035502 [7] Higashiwaki M, Sasaki K, Kamimura T, Wong M H, Krishnamurthy D, Kuramata A, Masui T, Yamakoshi S 2013 Depletion-mode Ga2O3 metal-oxide-semiconductor field-effect transistors on β-Ga2O3 (010) substrates and temperature dependence of their device characteristics Appl. Phys. Lett. 103 123511 [8] Galazka Z, Uecker R, Klimm D, Irmscher K, Naumann M, Pietsch M, Kwasniewski A, Bertram R, Ganschow S, and Bickermann M 2017 Scaling-Up of Bulk β-Ga2O3 Single Crystals by the Czochralski Method ECS J. Sol. Stat. Sci. Technol. 6 Q3007 [9] Sasaki K 2016 Study of homoepitaxial growth and n-type doping of -Ga2O3 Dr. Thesis, Faculty of engineering, Kyoto university [10] Tsai M Y, Bierwagen O, White M E, and Speck J S 2010 -Ga2O3 growth by plasma-assisted molecular beam epitaxy J. Vac. Sci. Technol. A 28 354 [11] Ahmadi E, Koksaldi O S, Kaun S K, Oshima Y, Short D B, Mishra U K, Speck, J S, Ge doping of β-Ga2O3 films grown by plasma-assisted molecular beam epitaxy, Applied Physics Express 10, 041102 (2017) [12] Konishi K, Goto K, Murakami H, Kumagai Y, Kuramata A, Yamakoshi S, and Higashiwaki M 1kV Vertical Ga2O3 Field-Plated Schottky Barrier Diodes Appl. Phys. Lett. 110, 103506 (2017) [13] Oshima Y, Ahmadi E, Badescu S C, Wu F, Speck J. S 2016 Composition determination of -(AlxGa1-x)2O3 layers coherently grown on (010) -Ga2O3 substrates by 12 high-resolution X-ray diffraction Appl. Phys. Express 9 061102 [14] Vogt P and Bierwagen O 2016 Reaction kinetics and growth window for plasma-assisted molecular beam epitaxy of Ga2O3: Incorporation of Ga vs. Ga2O desorption Appl. Phys. Lett. 108 072101 [15] Sawada Y, Weda N, Konishi K, Kumagai Y 2017“Thermodynamic Analysis on Molecular Beam Epitaxy of Ga2O3 The 78th JSAP Autumn meeting 7p-C17-9 [16] Ahmadi H, Koksaldi O S, Zheng X, Mates T, Oshima Y, Mishra U K, and Speck J S 2017 Demonstration of -(AlxGa1-x)2O3/-Ga2O3 modulation doped field-effect transistors with Ge as dopant grown via plasma-assisted molecular beam epitaxy Appl. Phys. Express 10 071101 [17] Murakami H, Nomura K, Goto K, Sasaki K, Kawara K, Thieu Q T, Togashi R, Kumagai Y, Higashiwaki M, Kuramata A, Yamakoshi S, Monemar B, and Koukitu A 2015 Homoepitaxial growth of -Ga2O3 layers by halide vapor phase epitaxy Appl. Phys. Express 8 015503 Figure captions Fig. 1 XRD -2 scan profile of a (001) -Ga2O3 homoepitaxial wafer. Fig. 2 Growth rate of -Ga2O3 as functions of 𝑃𝐺𝑎0 . Foreline oxygen pressures were 60 Torr for (010) and (001) samples, and 50 Torr for (100) samples. RF plasma power was 200W. Broken lines are guides for eyes. Fig. 3 Growth rate of (001) -Ga2O3 as a function of substrate temperature Ts. Fig. 4 Etching rate of (001) -Ga2O3 as a function of substrate temperature Ts. Fig. 5 AFM images of (001) -Ga2O3 substrates. (a),(b): virgin substrates, (c),(d): After 13 Ga-etching (-29 nm) Fig. 6 AFM images of (001) -Ga2O3 films grown on (a) virgin substrate, (b) Ga-etched substrate