# Fileset

[Rapid growth of a pit-free 011 -Ga2O3 homoepitaxial layer with very low residual Si and Cl concentrations by HCl-based halide vapor phase epitaxy-1.pdf](https://mdr.nims.go.jp/filesets/b5a5791e-9739-4bbf-8667-c0a41c1db68c/download)

## Creator

[Yuichi Oshima](https://orcid.org/0000-0001-8293-4891), [Takayoshi Oshima](https://orcid.org/0000-0001-8550-9735)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Rapid growth of a pit-free (011) β-Ga                    <sub>2</sub>                    O                    <sub>3</sub>                    homoepitaxial layer with very low residual Si and Cl concentrations by HCl-based halide vapor phase epitaxy](https://mdr.nims.go.jp/datasets/8169211e-55ce-406f-af72-c45c384d274f)

## Fulltext

Rapid growth of a pit-free (011) β-Ga2O3 homoepitaxial layer with very low residual Si and Cl concen  Science and Technology of Advanced Materials ISSN: 1468-6996 (Print) 1878-5514 (Online) Journal homepage: www.tandfonline.com/journals/tsta20 Rapid growth of a pit-free (011) β-Ga2O3 homoepitaxial layer with very low residual Si and Cl concentrations by HCl-based halide vapor phase epitaxy Yuichi Oshima \& Takayoshi Oshima To cite this article: Yuichi Oshima \& Takayoshi Oshima (2026) Rapid growth of a pit-free (011) β-Ga2O3 homoepitaxial layer with very low residual Si and Cl concentrations by HCl-based halide vapor phase epitaxy, Science and Technology of Advanced Materials, 27:1, 2751773, DOI: 10.1080/14686996.2026.2751773 To link to this article:  https://doi.org/10.1080/14686996.2026.2751773 © 2026 The Author(s). Published by National Institute for Materials Science in partnership with Taylor \& Francis Group. View supplementary material  Published online: 09 Nov 2026. Submit your article to this journal  Article views: 256 View related articles  View Crossmark data Full Terms \& Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=tsta20[https://www.tandfonline.com/journals/tsta20?src=pdf](https://www.tandfonline.com/journals/tsta20?src=pdf)[https://www.tandfonline.com/action/showCitFormats?doi=10.1080/14686996.2026.2751773](https://www.tandfonline.com/action/showCitFormats?doi=10.1080/14686996.2026.2751773)[https://doi.org/10.1080/14686996.2026.2751773](https://doi.org/10.1080/14686996.2026.2751773)[https://www.tandfonline.com/doi/suppl/10.1080/14686996.2026.2751773](https://www.tandfonline.com/doi/suppl/10.1080/14686996.2026.2751773)[https://www.tandfonline.com/doi/suppl/10.1080/14686996.2026.2751773](https://www.tandfonline.com/doi/suppl/10.1080/14686996.2026.2751773)[https://www.tandfonline.com/action/authorSubmission?journalCode=tsta20\&show=instructions\&src=pdf](https://www.tandfonline.com/action/authorSubmission?journalCode=tsta20&show=instructions&src=pdf)[https://www.tandfonline.com/action/authorSubmission?journalCode=tsta20\&show=instructions\&src=pdf](https://www.tandfonline.com/action/authorSubmission?journalCode=tsta20&show=instructions&src=pdf)[https://www.tandfonline.com/doi/mlt/10.1080/14686996.2026.2751773?src=pdf](https://www.tandfonline.com/doi/mlt/10.1080/14686996.2026.2751773?src=pdf)[https://www.tandfonline.com/doi/mlt/10.1080/14686996.2026.2751773?src=pdf](https://www.tandfonline.com/doi/mlt/10.1080/14686996.2026.2751773?src=pdf)[http://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773\&domain=pdf\&date\_stamp=09%20Nov%202026](http://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773&domain=pdf&date_stamp=09%20Nov%202026)[http://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773\&domain=pdf\&date\_stamp=09%20Nov%202026](http://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773&domain=pdf&date_stamp=09%20Nov%202026)[https://www.tandfonline.com/action/journalInformation?journalCode=tsta20](https://www.tandfonline.com/action/journalInformation?journalCode=tsta20)# Rapid growth of a pit-free (011) β-Ga2O3 homoepitaxial layer with very low  residual Si and Cl concentrations by HCl-based halide vapor phase epitaxyYuichi Oshima and Takayoshi OshimaResearch Center for Electronic and Optical Materials, National Institute for Materials Science, Tsukuba, Japan# ABSTRACTWe report the rapid growth of pit-free (011) β-Ga2O3 homoepitaxial layers with very low residual  \[Si\] and \[Cl\] by HCl-based halide vapor phase epitaxy (HCl-HVPE) in which GaCl was generated  from Ga and HCl. Along with the GaCl and O2 growth precursors, additional HCl was introduced  directly into the growth zone to suppress parasitic gas-phase reactions. The effects of O2 and  additional HCl partial pressures on residual impurity concentrations were systematically investi gated. The \[Si\] decreased sharply with increasing O2 partial pressure before reaching saturation,  whereas it increased monotonically with additional HCl partial pressure and fell below the  secondary ion mass spectrometry detection limit (1 × 1015 cm−3) in the absence of additional  HCl. Thermodynamic analysis of SiO2 etching by H2 and HCl qualitatively reproduced these trends  and suggested that residual Si originates predominantly from HCl-induced etching of quartz  components in the source zone in the reactor. Even without additional HCl, increasing the GaCl  supply enabled a growth rate of 13 μm/h, comparable to that previously achieved with additional  HCl, while maintaining \[Si\] below the detection limit and \[Cl\] below 2 × 1015 cm−3, without  particle-induced surface degradation. Furthermore, no pit formation was observed in the absence  of additional HCl, and the pits were found to originate from substrate pre-etching by additional  HCl before growth. These findings demonstrate that HCl-HVPE enables rapid homoepitaxy of pit-  free (011) β-Ga2O3 layers with very low residual impurity concentrations, offering a promising  route toward cost-effective vertical β-Ga2O3 power devices.# Impact StatementThis study identified the origins of residual Si contamination and pit formation in HCl-HVPE and  demonstrates rapid growth of pit-free (011) β-Ga2O3 homoepitaxial layers with very low  residual donor concentrations.ARTICLE HISTORY  Received 28 July 2026  Revised 4 October 2026  Accepted 15 October 2026KEYWORDS  β-Ga2O3; HVPE; impurity# Introductionβ-Ga2O3 is the most thermodynamically stable poly morph of Ga2O3 and possesses an ultra-wide bandgap  of 4.5–4.9 eV \[1,2\]. Owing to its wide bandgap, β-Ga2  O3 is expected to exhibit a high critical electric field  \[3\]. Furthermore, large-area, high-quality single-  crystal substrates grown by melt-growth techniques  are commercially available, enabling homoepitaxial  growth \[4–10\]. These advantages have established β-  Ga2O3 as one of the most promising semiconductor  materials for next-generation high-performance  power electronics. Indeed, high-performance proto type devices, including Schottky barrier diodesCONTACT Yuichi Oshima OSHIMA.Yuichi@nims.go.jp Research Center for Electronic and Optical Materials, National Institute for Materials Science  (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, JapanSupplemental data for this article can be accessed online at https://doi.org/10.1080/14686996.2026.2751773SCIENCE AND TECHNOLOGY OF ADVANCED MATERIALS  2026, VOL. 27, NO. 1, 2751773  https://doi.org/10.1080/14686996.2026.2751773Routledge LogoCrossmark Logo© 2026 The Author(s). Published by National Institute for Materials Science in partnership with Taylor \& Francis Group.  This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits  unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting  of the Accepted Manuscript in a repository by the author(s) or with their consent.[http://orcid.org/0000-0001-8293-4891](http://orcid.org/0000-0001-8293-4891)[http://orcid.org/0000-0001-8550-9735](http://orcid.org/0000-0001-8550-9735)[https://doi.org/10.1080/14686996.2026.2751773](https://doi.org/10.1080/14686996.2026.2751773)[http://www.tandfonline.com](http://www.tandfonline.com)[https://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773\&domain=pdf\&date\_stamp=2026-11-06](https://crossmark.crossref.org/dialog/?doi=10.1080/14686996.2026.2751773&domain=pdf&date_stamp=2026-11-06)(SBDs) \[11–14\] and metal–oxide–semiconductor  field-effect transistors (MOSFETs) \[15–19\], have  already been demonstrated.For high-voltage β-Ga2O3 power devices, the drift  layer is required to be both thick and lightly doped.  For example, realizing a device with a breakdown vol tage of approximately 10 kV requires a drift layer with  a thickness of 30–50 μm and a carrier concentration as  low as 5 × 1015 − 1 × 1016 cm−3 \[20\]. Therefore, an  epitaxial growth technique for high-voltage β-Ga2O3  power devices must provide both a high growth rate  for cost-effective fabrication and a residual carrier  concentration sufficiently lower than the required car rier concentration. Halide vapor phase epitaxy  (HVPE) is a promising epitaxial growth technique  potentially capable of meeting these requirements.HVPE of β-Ga2O3 can be broadly classified into  two types according to the method used to generate  the GaCl precursor. One approach employs the reac tion of metallic Ga with Cl2 gas to produce GaCl and is  hereafter referred to as Cl2-HVPE. A notable advan tage of Cl2-HVPE is its extremely low residual \[Si\],  which is typically below the secondary ion mass spec trometry (SIMS) detection limit (DL), for example,  \<5 × 1015 cm−3 \[21\]. Consequently, Cl2-HVPE has  been employed in most demonstrations of vertical β-  Ga2O3 power devices \[12–14,16,17,19\]. However,  increasing the precursor concentration to achieve  higher growth rates tends to promote parasitic gas-  phase reactions, resulting in particle formation \[22\].  As a result, commercially available (001) β-Ga2O3  epitaxial wafers are typically grown at a relatively low  growth rate of approximately 4 μm/h \[20\].The other type of HVPE employs the reaction of  metallic Ga with HCl gas to generate the GaCl pre cursor and is hereafter referred to as HCl-HVPE to  distinguish it from Cl2-HVPE according to the  chlorine source. A characteristic feature of HCl-  HVPE is the simultaneous generation of H2 as a by-  product during GaCl formation. In GaN grown by  HCl-HVPE, relatively high residual \[Si\] (typically on  the order of 1017 cm−3 \[23\]) have been widely  reported. This has been attributed to etching of the  quartz reactor components, which are extensively  used in HVPE reactors, by H-containing gases (H2,  HCl, and NH3), resulting in the generation of Si-  containing gaseous species. Indeed, it has been  reported that the residual \[Si\] can be reduced to  below 5 × 1014 cm−3 by eliminating quartz compo nents from the HVPE reactor \[23\]. By analogy with  these findings for GaN, HCl-HVPE has generally  been considered more susceptible than Cl2-HVPE  to Si contamination in β-Ga2O3 epilayers \[24\]. On  the other hand, HCl-HVPE generally enables higher  growth rates than Cl2-HVPE \[25–27\]. This is pre sumably because the H2 generated during growth,  which also acts as an etchant for β-Ga2O3,  suppresses particle formation caused by parasitic  reactions between the source gases. Furthermore,  the growth rate in HCl-HVPE can be further  increased by introducing additional HCl into the  growth zone \[26,28,29\]. For example, we routinely  achieve a high growth rate of 24 μm/h for (001) β-  Ga2O3 growth using HCl-HVPE with additional HCl  \[26,28\].To date, (001)-oriented substrates have been pre dominantly used for the epitaxial growth of β-Ga2O3  drift layers for vertical power devices. This is mainly  because the (001) plane is parallel to the \[010\] pulling  direction used in the edge-defined film-fed growth  (EFG) method, facilitating the fabrication of large-  area substrates, and because it generally allows higher  HVPE growth rates than other crystal orientations  \[30\]. However, a common drawback of (001) homo epitaxy by both types of HVPE is the relatively high  residual \[Cl\], with typical values around 1 × 1016 cm−3  \[20,26,31\]. One promising approach to overcoming  this issue is to employ the (011) orientation. For both  Cl2-HVPE and HCl-HVPE, the residual \[Cl\] in (011)  epilayers has been reported to be approximately one  order of magnitude lower than that in (001) epilayers  \[26,32\]. However, the growth rate on the (011) plane is  only about 60% of that on the (001) plane and is as low  as approximately 2 μm/h for Cl2-HVPE \[32\].  Nevertheless, several promising vertical β-Ga2O3  power devices based on (011) epilayers have been  demonstrated \[19,33,34\]. For example, Wakimoto  et al. reported a FinFET with a breakdown voltage  exceeding 10 kV incorporating an 85-μm-thick drift  layer grown by Cl2-HVPE \[19\].In contrast, HCl-HVPE with additional HCl has  been reported to achieve a much higher growth rate  of 14 μm/h than that reported for Cl2-HVPE, while  maintaining a low residual \[Cl\] of 1.7 × 1015 cm−3.  However, the residual \[Si\] remains relatively high at  2 × 1016 cm−3 \[26\]. Another issue is the formation of  pits on the epilayer surface at a density of approxi mately 103–104 cm−2, a phenomenon that has not been  reported for Cl2-HVPE \[26\]. These pits extend nearly  to the epilayer/substrate interface, suggesting that  their origin is associated with the epilayer/substrate  interface.If the residual \[Si\] and pit formation on the (011)  surface can be effectively suppressed, HCl-HVPE  could become a promising epitaxial growth technique  for β-Ga2O3 power devices by combining high growth  rates with the intrinsic advantage of low Cl incorpora tion. In this work, we investigate the origins of residual  Si incorporation and pit formation in HCl-HVPE.  Furthermore, we demonstrate rapid homoepitaxial  growth of pit-free (011) β-Ga2O3 with a residual \[Si\]  below the SIMS detection limit while maintaining the  low residual \[Cl\] characteristic of the (011)  orientation.Sci. Technol. Adv. Mater. 27 (2026) 2                                                                                                                                      Y. Oshima and T. Oshima# Experimental methods(011) β-Ga2O3 homoepitaxial growth was carried out by  atmospheric-pressure HVPE at 1030°C without inten tional doping on conductive Sn-doped (011) β-Ga2O3  single-crystal substrates grown by the EFG method  (Novel Crystal Technology, Inc.) using a laboratory-  built quartz reactor. A schematic illustration of the  HVPE reactor is shown in Figure 1. Not only the main  reactor tube but also all internal components were made  of quartz. High-purity O2 gas (\>99.99995%) was used as  the oxygen source. GaCl precursor gas was generated in  the source zone maintained at 610°C by reacting metallic  Ga (\>99.99999%) with source HCl gas (\>99.999%) in the  Ga boat. Under these conditions, the source HCl was  essentially completely consumed and converted into  GaCl and H2. In addition to the precursor gases, the  reactor was equipped with a supply line for additional  HCl gas to suppress parasitic reactions between the  source gases \[26,29\]. Unlike the source HCl, the addi tional HCl bypassed the Ga boat without reacting with  metallic Ga and was introduced directly into the growth  zone. N2 gas (dew point \< −110 °C) was used as the  carrier gas for all gas streams. The GaCl and additional  HCl gas streams were transported independently of the  O2 gas stream and mixed only in the growth zone, where  β-Ga2O3 growth took place. No intentional post-growth  annealing was performed.During the temperature ramp-up, O2 was supplied  at the same partial pressure as that used during the  growth step. After the reactor temperature had stabi lized, the GaCl supply was initiated to start epitaxial  growth. When additional HCl was used, its supply was  started 5 min before the introduction of GaCl.  Consequently, the β-Ga2O3 substrates were subjected  to a slight etching treatment prior to growth.The growth rate was determined from the substrate  weight change before and after epitaxial growth. The  surface morphology was characterized by Nomarski  optical microscopy using BX51 (OLYMPUS, Japan)  and scanning electron microscopy (SEM) using  SU8230 (Hitachi, Japan). Impurity concentrations  were measured by SIMS.To confirm that the residual carrier concentration  was sufficiently reduced, a SBD consisting of an unin tentionally doped homoepitaxial layer grown on  a conductive Sn-doped (011) β-Ga2O3 substrate was  fabricated. Its current–voltage (J–V) and capacitance–  voltage (C–V) characteristics were measured using a  semiconductor parameter analyzer B1500A (Keysight,  US). The C–V measurement was performed at an AC  frequency of 1 MHz, assuming a relative dielectric  constant of 10 for β-Ga2O3. For the cathode, Ti (30  nm)/Pt (20 nm)/Au (70 nm) contacts were deposited  by electron-beam (EB) evaporation, followed by rapid  thermal annealing (RTA) at 550°C for 1 min in N2 gas  flow. The ohmic behavior of the backside contact was  confirmed by the linear I–V characteristics measured  between adjacent cathode electrodes. The Schottky  anode consisted of a 50-nm-thick Au electrode with  an area of 2.3 × 10−3 cm2 was deposited by EB  evaporation.# Results and discussionFigure 2(a) shows the dependences of the residual  \[Si\] and \[Cl\], and growth rate on the supplied O2  partial pressure, P0(O2). Here, P0(X) denotes the  supplied partial pressure of gas species X. Note that  P0(HCl) hereafter refers to the partial pressure of the  additional HCl rather than that of the source HCl.  Throughout this experiment, both P0(GaCl) and P0  (HCl) were fixed at 25 Pa. The growth rate increased  with increasing P0(O2), gradually approaching  saturation. The residual \[Si\] was 1.5 × 1016 cm−3 at  the lowest P0(O2) investigated but decreased rapidly  to 3.2 × 1015 cm−3, approximately one-fifth of the  initial value, at P0(O2) = 0.625 kPa. A further increase  in P0(O2) produced little change in the residual \[Si\].  In contrast, the residual \[Cl\] exhibited a slight  increase with increasing P0(O2), while remaining as  low as 1.5 × 1015 cm−3 even at the highest P0(O2)  investigated. Considering that Cl is generally believed  to occupy O sites in β-Ga2O3 \[35\], one might expect  the Cl incorporation to decrease with increasing P0  (O2). However, the opposite trend was observed. This  result suggests that Cl incorporation is influenced by  mechanisms other than simple competition between  O and Cl for O lattice sites. The origin of this beha vior remains unclear at present and requires further  investigation.A schematic of an HVPE reactor for beta-Ga2O3 growth with source and growth zones. GaGaCl, HHClSubstrate1030 C 610 CSource zone Growth zoneOHCl OGa boatFigure 1.Schematic of the HVPE reactor used in this study. Carrier gas flows are omitted for clarity.Sci. Technol. Adv. Mater. 27 (2026) 3                                                                                                                                      Y. Oshima and T. OshimaFigure 2(b) shows the dependences of the residual  \[Si\] and \[Cl\], and the growth rate on P0(HCl). In this  experiment, P0(GaCl) and P0(O2) were fixed at 25 Pa  and 2.5 kPa, respectively. The growth rate initially  increased with increasing P0(HCl) and then gradually  decreased. This behavior is considered to result from  two competing effects. At low P0(HCl), the additional  HCl suppresses parasitic gas-phase reactions, thereby  improving the transport efficiency of the source  species to the substrate. At higher P0(HCl), how ever, the etching of β-Ga2O3 on the substrate  becomes dominant, resulting in a reduction in the  net growth rate. When no additional HCl was sup plied, the residual \[Si\] was below the SIMS detection  limit (1 × 1015 cm−3); therefore, this data point is not  included in Figure 2(b). The residual \[Si\] increased  monotonically with increasing P0(HCl) and reached  1.4 × 1016 cm−3 at P0(HCl) = 100 Pa. In contrast, the  residual \[Cl\] exhibited no clear dependence on P0  (HCl) and remained nearly constant in the range of  (1.3–1.9) × 1015 cm−3.As shown in Figure 2(a), the residual \[Si\] can be  substantially reduced by increasing P0(O2). This  reduction is most likely attributable to the suppression  of SiO2 reduction by the H-containing species, H2 and  HCl, in the presence of O2. A notable feature is that  the residual \[Si\] did not decrease continuously with  increasing P0(O2). Instead, it dropped sharply with  a relatively small increase in P0(O2) and then remained  nearly constant. This behavior suggests that even  a relatively low O2 partial pressure is sufficient to  effectively suppress the reduction of SiO2.The monotonic increase in the residual \[Si\] with  increasing P0(HCl), shown in Figure 2(b), further  suggests that HCl plays a significant role in etching  SiO2 and thereby enhances Si contamination. It  should be noted that, even in the absence of addi tional HCl, H2 generated during GaCl formation  was still present in the reactor at a partial pressure  of one-half that of P0(GaCl) (12.5 Pa). If H2 had an  etching effect on SiO2 comparable to or greater than  that of HCl under the present growth conditions,  a detectable concentration of Si would also be  expected even without additional HCl. However,  the residual \[Si\] remained below the SIMS detection  limit. These results therefore suggest that, under the  present growth conditions, the effective etching cap ability of H2 toward SiO2 was weaker than that  of HCl.To elucidate the experimentally observed behavior  of the residual \[Si\], the thermodynamic driving forces  for SiO2 etching by H2 and HCl were evaluated, and  their dependences on temperature and P0(O2) were  analyzed. The thermodynamic model and calculation  procedure are described below. Assuming that the  SiO2 etching rate is governed primarily by the equili brium generation of volatile Si-containing species, the  thermodynamic driving force for SiO2 etching is  expected to correlate with the total equilibrium partial  pressure of these species at the SiO2 surface.  Accordingly, the following quantity was adopted as  an indicator of the etching driving force:where P Sini Xið Þ is the equilibrium partial pressure of  a gaseous species Sini Xi.The equilibrium partial pressures of all gaseous  species were calculated using the thermodynamic  equilibrium software CatCalc (Research Institute of  Computational Thermodynamics, Inc.), which deter mines the equilibrium gas composition by minimizing  the Gibbs free energy of the system \[36\].For the analysis of SiO2 etching by H2, the following  gaseous species were included in the equilibrium cal culation: Si, Si2, Si3, SiH4, SiO, SiO2, Si(OH)4, SiH2O3,  H2, H2O, H2O2, O2, and O3, together with inert carrier  gas. For the analysis of SiO2 etching by HCl, the  following gaseous species were considered: Si, Si2,  Si3, SiH4, SiO, SiO2, Si(OH)4, SiH2O3, SiCl2, SiCl4,  SiH2Cl2, SiH3Cl, SiHCl3, H2, H2O, H2O2, O2, O3, Cl2,  Cl4, Cl2O, ClO2, HCl, HClO, and HClO4, together  with inert carrier gas.For the H2 etching calculations, P0(H2) was fixed at  12.5 Pa, corresponding to the H2 generated as a by-  product during GaCl formation from metallic Ga andTwo line graphs of residual silicon and chlorine and growth rate versus oxygen and hydrogen chloride. 0 25 50 75 10010151016101702468P0(HCl) \[Pa\]C once ntra tion  \[cm -3\]G row thr ate  \[m /h\]\[Si\]\[Cl\]P0(GaCl) = 25 Pa P0(O2) = 2.5 kPa0 1 2 310151016101702468P0(O2) \[kPa\]C once ntra tion  \[cm -3\]G row thr ate  \[m /h\]\[Si\]\[Cl\]P0(GaCl) = 25 Pa P0(HCl) = 25 Pa(a) (b)Figure 2.Dependences of the residual \[Si\] and \[Cl\], and growth rate on (a) P0(O2) and (b) P0(HCl).Sci. Technol. Adv. Mater. 27 (2026) 4                                                                                                                                      Y. Oshima and T. Oshimasource HCl under the growth conditions used for the  experiments shown in Figure 2(a,b), where P0(GaCl)  was 25 Pa. For the HCl etching calculations, P0(HCl)  was fixed at 25 Pa, corresponding to the experimental  conditions used in Figure 2(a). P0(O2), was varied  from 0 to 2.5 kPa to cover the experimental growth  conditions.Figure 3(a,b) show the calculated temperature  dependences of the equilibrium partial pressures of  individual volatile Si-containing species, Sini Xi, and  the etching driving force indicator, Ptot, for SiO2 etch ing by H2 (P0(H2) = 12.5 Pa) and HCl (P0(HCl) = 25  Pa), respectively, under P0(O2) = 0. The vertical  dashed lines indicate the source-zone temperature  (610 °C) and the growth-zone temperature (1030  °C). The equilibrium partial pressures of the Si-  containing species not shown are below the plotting  range.　The dominant volatile Si-containing species  generated by H2 and HCl etching are SiO and SiCl4,  respectively, and their equilibrium partial pressures  almost completely overlap with the corresponding  Ptot curves. A notable feature is that the temperature  dependence of the etching driving force is much  weaker for HCl than for H2. Consequently, although  the HCl-induced etching driving force is comparable  to that for H2 at the growth-zone temperature of  1030°C, it is much greater at the source-zone tempera ture of 610°C.Figure 4(a,b) show the calculated temperature  dependences of the etching driving force indicator,  Ptot, for SiO2 etching by H2 (P0(H2) = 12.5 Pa) and HCl  (P0(HCl) = 25 Pa), respectively, in the presence of O2.Two line graphs of equilibrium partial pressures versus temperature for hydrogen and hydrogen chloride etching. 600 800 1000 1200 10-1510-1410-1310-1210-1110-1010-910-810-710-610-510-410-310-210-1100 Par tial p ress ure  \[Pa\]Temperature \[ oC\]P (O2) = 0 Pa P (HCl) = 25 Pa600 800 1000 1200 10-1510-1410-1310-1210-1110-1010-910-810-710-610-510-410-310-210-1100P artia l pre ssur e\[P a\]Temperature \[ oC\]P (O ) = 0 Pa P (H ) = 12.5 Pa(a)PSiOSiO SiSiHSi(OH)SiSiOSiOSiClSiClSiHClSiH Cl(b)PFigure 3.Calculated temperature dependences of the equilibrium partial pressures of Si-containing gaseous species and the  etching driving force indicator, Ptot, for SiO2 etching by (a) H2 and (b) HCl.Two line graphs of etching driving force indicator versus temperature for varying oxygen supply. (a) (b)600 800 1000 1200 10-1510-1410-1310-1210-1110-1010-910-810-710-610-510-410-310-210-1100Temperature \[ oC\]P tot \[ Pa\]P0(O2) P0(H2) = 12.5 Pa1\.25 x 10-5 Pa 0 Pa1\.25 x 10-4 Pa 1.25 x 10-3 Pa 1.25 x 10-2 Pa 1.25 x 10-1 Pa 1.25 x 100 Pa1\.25 x 101 Pa600 800 1000 1200 10-15 10-14 10-13 10-12 10-11 10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100Temperature \[ oC\]P tot \[ Pa\]P0(HCl) = 25 Pa P0(O2)1\.25 x 101 Pa0 Pa 1.25 x 10-4 Pa 1.25 x 10-3 Pa 1.25 x 10-2 Pa 1.25 x 10-1 Pa 1.25 x 100 Pa1\.25 x 102 Pa 2.50 x 103 PaFigure 4.Calculated temperature dependences of the etching driving force indicator, Ptot, at various P0(O2): (a) H2 etching and (b)  HCl etching.Sci. Technol. Adv. Mater. 27 (2026) 5                                                                                                                                      Y. Oshima and T. OshimaFor both etchants, Ptot decreases markedly with O2  supply. In Figure 4(a), the Ptot curves for P0(O2)  ≥12.5 Pa almost completely overlap. It should also be  noted that the dominant volatile Si-containing species  depend on P0(O2). Under the standard experimental  condition of P0 (O2) = 2.5 kPa, Si(OH)4 becomes the  dominant Si-containing species for both H2 and HCl  etching, in contrast to the O2-free case. The calculated  temperature dependences of the equilibrium partial  pressures of individual Si-containing species and Ptot  for various P0(O2) values are presented in Figs. S1  and S2.Figure 5(a,b) show the calculated dependences of  Ptot on P0(O2) at 1030°C for SiO2 etching by H2 (P0  (H2) = 12.5 Pa) and HCl (P0(HCl) = 25 Pa), respec tively. In both cases, Ptot decreases sharply even with  a small O2 supply. For example, in the H2 etching  calculation, supplying O2 at only one-tenth of the H2  partial pressure (P0(O2) = 1.25 Pa) reduces Ptot to less  than one-thousandth of its value without O2. These  calculated results provide a qualitative explanation for  the experimentally observed sharp decrease in the resi dual \[Si\] with a relatively small increase in P0(O2),  followed by the appearance of a plateau, as shown in  Figure 2(a).Table 1 compares the calculated etching driving  forces for SiO2 by H2 and additional HCl in the source  zone and the growth zone of the HVPE reactor. In the  source zone, the etching driving force for additional  HCl is much greater than that for H2. This is because  the etching driving force of HCl exhibits a much  weaker temperature dependence than that of H2,  allowing HCl to maintain a high etching capability  even at the relatively low source-zone temperature of  610°C.In contrast, the etching driving forces in the growth  zone are much smaller than that for additional HCl in  the source zone for both etchants because of the strong  suppression by O2. These results therefore suggest that  SiO2 etching in the HVPE reactor is dominated by  additional HCl in the source zone. This mechanism  provides a qualitative explanation for the strong  dependence of the residual \[Si\] on P0(HCl) observed  in Figure 2(b).The experimental results described above  demonstrated that the residual \[Si\] in (011) β-Ga2  O3 homoepitaxial layers can be reduced below the  SIMS detection limit by eliminating the supply of  additional HCl. However, under these conditions,  the growth rate was limited to 4.6 μm/h for P0  (GaCl) = 50 Pa, which is considerably lower than  the reported value of 14 μm/h for HVPE with addi tional HCl \[26\]. Therefore, the growth rate was  increased by increasing P0(GaCl) while maintaining  P0(O2) at 2.5 kPa. Figure 6(a) shows the depen dences of the growth rate, the residual \[Si\] and \[Cl\]  on P0(GaCl). The labels (b) and (c) correspond to  the SEM images shown in Figure 6(b,c), respec tively. At P0(GaCl) = 100 Pa \[label (b)\], the growth  rate reached 13 μm/h, comparable to that reported  for HVPE with additional HCl, while no degrada tion of the surface morphology was observed, as  shown in Figure 6(b). When P0(GaCl) was further  increased to 200 Pa \[label (c)\], the growth rate  increased to 25 μm/h. However, numerous particles  were observed on the epilayer surface, as shown inTwo scatter plots of etching driving force indicator versus oxygen partial pressure, showing a sharp decrease then plateau. 10-610-510-410-310-210-1100 101 102 103 104 10-1010-910-810-710-610-510-410-310-2P0(O2) \[Pa\]P tot \[ Pa\]P0(H2) = 12.5 Pa, 1030 oC10-610-510-410-310-210-1100 101 102 103 104 10-1010-910-810-710-610-510-410-310-2P0(O2) \[Pa\]P tot \[ Pa\]P0(HCl) = 25 Pa, 1030 oC(a) (b)Figure 5.Calculated dependences of the etching driving force indicator, Ptot, on P0(O2) at 1030°C for (a) H2 etching and (b) HCl  etching.Table 1. Comparison of the etching driving force indicator, Ptot, for H2 and  HCl in the source and growth zones of the HVPE reactor.   Etchant Source zone (610 °C)  P0(O2) = 0 kPa Growth zone (1030 °C)  P0(O2) = 25 kPa  H2  P0(H2) = 12.5 Pa 8.8 × 10−14 Pa 4.1 × 10−14 Pa  HCl  P0(HCl) = 25 Pa 8.4 × 10−10 Pa 9.1 × 10−15 Pa Sci. Technol. Adv. Mater. 27 (2026) 6                                                                                                                                      Y. Oshima and T. OshimaFigure 6. These particles were most likely generated  by parasitic reactions between GaCl and O2 under  the high precursor concentration. To further  increase the growth rate while suppressing particle  formation, two approaches appear promising: sup plying a small amount of additional HCl while  maintaining the residual \[Si\] below the acceptable  level, or employing reduced-pressure growth.As also shown in Figure 6(a), no significant increase  in the residual \[Cl\] was observed even at the highest  growth rates, and the residual \[Si\] remained below the  SIMS detection limit (1 × 1015 cm−3) for all samples. In  contrast, for the (001) orientation, a significant  increase in the residual \[Cl\] has been reported for P0  (GaCl) increase \[26\]. This difference suggests that, in  addition to thermodynamic factors, crystal-  orientation-dependent kinetic effects also play an  important role in Cl incorporation. However, further  studies are required to clarify the underlying  mechanism.The possible relationship between the surface mor phology and residual impurity concentrations was also  examined by comparing epilayers grown under differ ent conditions. Although the surface morphology var ied with the growth conditions, no systematic  relationship was found between the surface morphol ogy and the residual \[Si\] or \[Cl\]. This suggests that the  residual impurity concentrations are not primarily  correlated with the morphology of the flat regions.To verify that the substantial reduction in both the  residual \[Si\] and \[Cl\] achieved by eliminating addi tional HCl also resulted in a low carrier concentration,  C–V and J–V measurements were performed on  a SBD. A schematic cross-sectional structure of the  SBD is shown in Figure 7(a), and an optical micro scope image of the Schottky anode is shown in  Figure 7(b). The homoepitaxial layer was grown  under the same conditions as those of the sample  shown in Figure 6(b), yielding \[Si\] \<1 × 1015 cm−3  and \[Cl\] = 2 × 1015 cm−3.A mixed line plot and two micrographs showing growth rate and residual chlorine versus gallium chloride. 0 50 100 150 2001015101610170102030C once ntra tion  \[cm -3\]G row thr ate  \[m /h\]P0(HCl) = 0 Pa P0(O2) = 2.5 kPa(a)(b)(c)\[Si\] \< DL (1x10 cm ) for all specimensFigure 6.(a) Dependences of the residual \[Cl\] and growth rate on P0(GaCl). The residual \[Si\] was below the SIMS detection limit  under all growth conditions. (b), (c) corresponding plan-view SEM images of the epilayer surfaces grown at P0(GaCl) = 100 and  200 Pa, respectively.Figure 7.(a) Cross-sectional structure of the SBD used for the J–V and C–V measurements. (b) Optical microscope image of the  Schottky anode. (c), (d) measured J–V and C–V characteristics, respectively.Sci. Technol. Adv. Mater. 27 (2026) 7                                                                                                                                      Y. Oshima and T. OshimaFigure 7 shows the J–V characteristics of the SBD.  The reverse leakage current was low, indicating good  barrier characteristics. From the slope of the forward  J–V curve above the turn-on voltage, the series resis tance was estimated to be approximately 360 kΩ,  indicating that the homoepitaxial layer was highly  resistive. Because of this high series resistance,  a substantial fraction of the applied voltage is expected  to drop across the epilayer. Therefore, the forward J–  V characteristics are not suitable for conventional  extraction of the ideality factor and Schottky barrier  height.Figure 7 shows the C–V characteristics of the SBD  at 1 MHz. The capacitance showed little dependence  on the applied voltage. The depletion width estimated  directly from the measured capacitance was approxi mately 3 μm, which is nearly equal to the epilayer  thickness, indicating that the epilayer was nearly fully  depleted. To examine whether the capacitance mea surement was affected by the high series resistance or  by insufficient response of slowly responding electri cally active states at 1 MHz, C–V measurements were  also performed at a lower frequency of 100 kHz.  Similar voltage-independent capacitance and  a depletion width close to the epilayer thickness were  obtained at 100 kHz, consistent with the results  obtained at 1 MHz. Assuming a built-in potential of  1 V, the ND – NA required for complete depletion of  the epilayer at this voltage was estimated to be at most  1.2 × 1014 cm−3.The estimated ND − NA was considerably lower than  the residual \[Cl\] of approximately 2 × 1015 cm−3. The  activation state of the residual Cl cannot be determined  from the present results, and this discrepancy may arise  from incomplete donor activation, compensation by  unintentional acceptors, or both.During the investigation of the dependence of the  residual impurity concentrations on P0(HCl), we  found that additional HCl plays a critical role in pit  formation on the epilayer surface. Figure 8(a) shows  a Nomarski optical micrograph of a homoepitaxial  layer grown for 1 h at a growth rate of 4.3 μm/h with  P0(GaCl) = P0(HCl) = 25 Pa and P0(O2) = 2.5 kPa.  Numerous pits are observed, with a density of  approximately 3.5 × 102 cm−2. In contrast, no pits  were observed in the homoepitaxial layer shown in  Figure 8(b), which was grown for 1 h under the same  precursor partial pressures without additional HCl at  a comparable growth rate of 4.6 μm/h. Figure 8(c)  shows a sample initially grown under the pit-free  conditions used for Figure 8(b). After the epilayer  thickness reached 2.3 μm, additional HCl was intro duced so that the growth conditions became identical  to those used for the sample shown in Figure 8(a), and  growth was continued until the total thickness reached  4.5 μm. Nevertheless, almost no pits were observed. It  should be noted that the residual \[Si\] in the upper part  of the epilayer in Figure 8(c) should be the same as  that in the epilayer shown in Figure 8(a), because they  were grown under the same additional-HCl condition.  Nevertheless, almost no new pits were formed in the  epilayer shown in Figure 8(c). This result suggests that  the increased Si incorporation itself is not the direct  cause of pit formation. These results strongly suggest  that pit formation is not caused by the presence of  additional HCl during epitaxial growth itself, but  rather by the pre-growth etching of the substrate sur face by additional HCl before the introduction of  GaCl.Figure 9(a) shows a Nomarski optical micrograph  of a (011) β-Ga2O3 substrate subjected only to HCl  pre-etching under the same conditions as those used  prior to the growth of the pit-containing sample  shown in Figure 8(a). To more clearly reveal the sur face morphology produced by etching, the etching  time was increased to 50 min, which is ten times  longer than that used in the growth experiments.  Numerous hillocks with a density of approximately  1.1 × 104 cm−2 were observed, as indicated by the  arrows.Figure 9(b) shows a plan-view SEM image of one of  these hillocks. An inclined facet plane, indicated by the  arrow, is clearly visible. The trace direction of the facet  on the (011) surface coincides with the intersection  direction between the (011) and (310) planes, i.e. the  \[133\] direction. Furthermore, the inclination angle of  the facet, determined geometrically from the apparent  change in facet width in tilted-view SEM observations,  was 50.9°, which agrees almost exactly with the theo retical angle of 51.0° between the (011) and  (310) planes. These observations indicate that the  facet is the (310) plane. The (310) plane is known toFigure 8.Nomarski optical micrographs of the surfaces of (011) β-Ga2O3 homoepitaxial layers grown (a) with additional HCl, (b)  without additional HCl, and (c) with additional HCl introduced after an initial growth period without additional HCl.Sci. Technol. Adv. Mater. 27 (2026) 8                                                                                                                                      Y. Oshima and T. Oshimabe one of the relatively stable low-surface-energy  planes of β-Ga2O3 \[37\], and the formation of similar  (310) facets has also been reported during selective  HCl etching of (010) and (001) β-Ga2O3 substrates  \[38,39\].Figure 9 shows a plan-view SEM image of a pit formed  on the homoepitaxial layer grown under additional-HCl  supply. Nearly all of the observed pits were accompanied  by the (310) facet, as indicated by the arrow. These results  strongly suggest that HCl pre-etching of the (011) sub strate produces stable (310) facets, which subsequently  act as the origin of pit formation. Owing to their high  stability, the growth rate on the (310) facets is presum ably lower than that on the surrounding (011) surface.  Consequently, the growth front above the facets lags  behind the surrounding epilayer during homoepitaxy,  while the (310) facets continue to develop, eventually  leading to pit formation.The dislocation density in EFG-grown β-Ga2O3 has  been reported to be on the order of 103 cm−2 \[20\],  which is reasonably close to the hillock density  observed after HCl pre-etching. This may indicate  that dislocations are involved in hillock formation.  On the other hand, the pit density on the homoepitax ial surface was only approximately 3.5 × 102 cm−2,  much lower than the hillock density. This suggests  that only a fraction of the hillocks develop into pits,  possibly because only those with particularly well-  developed (310) facets continue to evolve during sub sequent epitaxial growth.# ConclusionsIn summary, we experimentally clarified the effects of  P0(O2) and P0(HCl) on the residual \[Si\] and \[Cl\] in  (011) β-Ga2O3 homoepitaxial layers grown by HCl-  HVPE. the residual \[Si\] exhibited a characteristic  dependence on P0(O2), decreasing sharply with  a relatively small increase in P0(O2) and subsequently  remaining nearly constant. In addition, the residual  \[Si\] was reduced below the SIMS detection limit  (1 × 1015 cm−3) in the absence of additional HCl,  despite the presence of H2 generated during GaCl for mation, whereas it increased markedly with increasing  P0(HCl). In contrast, the residual \[Cl\] showed little  dependence on either P0(O2) or P0(HCl) and remained  below 2 × 1015 cm−3 throughout this study.To explain these experimental observations, the ther modynamic driving force for SiO2 etching was analyzed  using the equilibrium parameter Ptot. The calculations  revealed that the etching driving force of HCl is much  less temperature dependent than that of H2 under O2-  free conditions, allowing HCl to maintain a high etch ing capability even in the relatively low-temperature  source zone (610°C). In contrast, the addition of  a small amount of O2 dramatically suppressed the etch ing driving force for both etchants, predicting that SiO2  etching in the growth zone is much weaker than that by  additional HCl in the source zone. These predictions  consistently explain the experimentally observed strong  correlation between the residual \[Si\] and P0(HCl), the  absence of detectable Si without additional HCl despite  the presence of H2, and the characteristic dependence of  the residual \[Si\] on P0(O2).Furthermore, by eliminating additional HCl and  increasing P0(GaCl), a growth rate of 13 μm/h, compar able to that previously reported for HVPE with addi tional HCl, was achieved without particle formation or  increases in the residual \[Si\] and \[Cl\]. C–V measure ments of a SBD fabricated on such a homoepitaxial layer  confirmed its semi-insulating nature, confirming that the  residual carrier concentration was successfully reduced,  although carrier compensation by unintentionally incor porated acceptors, such as nitrogen, is suggested. We also  found that pit formation was dramatically suppressed in  the absence of additional HCl and showed that the pits  most likely originate from substrate pre-etching by addi tional HCl, which produces stable (310) facets before  epitaxial growth.Although HCl-HVPE of (011) β-Ga2O3 has been  recognized as a promising technique capable of much  higher growth rates than Cl2-HVPE, its practical  application has been hindered by high residual \[Si\]  and pit formation. The present study demonstrates  that both of these issues can be effectively overcome.  These findings establish HCl-HVPE as a practical  route for the high-rate growth of high-quality (011) β-  Ga2O3 homoepitaxial layers and are expected to con tribute significantly to the cost-effective manufactur ing and widespread deployment of high-performance  β-Ga2O3 power devices.Figure 9.(a) Nomarski optical micrograph and (b) plan-view SEM image of an HCl-etched (011) β-Ga2O3 substrate. (c) Plan-view  SEM image of a pit formed on a (011) β-Ga2O3 homoepitaxial layer during growth with additional HCl.Sci. Technol. Adv. Mater. 27 (2026) 9                                                                                                                                      Y. Oshima and T. Oshima# AcknowledgmentsThis work was supported by the ‘Advanced Research  Infrastructure for Materials and Nanotechnology in Japan  (ARIM)’ initiative of the Ministry of Education, Culture,  Sports, Science and Technology (MEXT) under Proposal  Numbers JPMXP1225NM5075 and JPMXP1226NM5276.# Disclosure statementNo potential conflict of interest was reported by the  author(s).# FundingThis study was based on the results obtained from a project,  JPNP22007, commissioned by the New Energy and  Industrial Technology Development Organization (NEDO).# ORCIDYuichi Oshima http://orcid.org/0000-0001-8293-4891Takayoshi Oshima http://orcid.org/0000-0001-8550-  9735# References\[1\] Orita M, Ohta H, Hirano M, et al. H. Deep-ultraviolet  transparent conductive β-Ga2O3 thin films. Appl  Phys Lett. 2000;77(25):4166. doi: 10.1063/1.1330559\[2\] Onuma T, Saito S, Sasaki K, et al. Valence band  ordering in β-Ga2O3 studied by polarized transmit tance and reflectance spectroscopy. Jpn J Appl Phys.  2015;54(11):112601. doi: 10.7567/JJAP.54.112601\[3\] Higashiwaki M, Sasaki K, Kuramata A, et al. Gallium  oxide (Ga2O3) metal-semiconductor field-effect tran sistors on single-crystal β-Ga2O3 (010) substrates.  Appl Phys Lett. 2012;100(1):013504. doi: 10.1063/1.  3674287\[4\] Aida H, Nishiguchi K, Takeda H, et al. Growth of β-  Ga2O3 single crystals by the edge-defined, film fed  growth method. Jpn J Appl Phys. 2008;47(11):8506.  doi: 10.1143/JJAP.47.8506\[5\] Kuramata A, Koshi K, Watanabe S, et al. High-quality  β-Ga2O3 single crystals grown by edge-defined film-  fed growth. Jpn J Appl Phys. 2016;55(12):1202A2. doi:  10.7567/JJAP.55.1202A2\[6\] Blevins JD, Stevens K, Lindsey A, et al. Development  of large diameter semi-insulating gallium oxide (Ga2  O3) substrates. IEEE Trans Semicond Manuf. 2019;32  (4):466. doi: 10.1109/TSM.2019.2944526\[7\] Galazka Z. Growth of bulk β-Ga2O3 single crystals by  the Czochralski method. J Appl Phys. 2022;131  (3):031103. doi: 10.1063/5.0076962\[8\] Hoshikawa K, Kobayashi T, Matsuki Y, et al. 2-inch  diameter (100) β-Ga2O3 crystal growth by the vertical  Bridgman technique in a resistance heating furnace in  ambient air. J Cryst Growth. 2020;545(May):125724.  doi: 10.1016/j.jcrysgro.2020.125724\[9\] Hoshikawa K, Kobayashi T, Ohba E, et al. 50 mm  diameter Sn-doped (001) β-Ga2O3 crystal growth  using the vertical bridgeman technique in ambient  air. J Cryst Growth. 2020;546(June):125778. doi: 10.  1016/j.jcrysgro.2020.125778\[10\] Ueda Y, Igarashi T, Koshi K, et al. Two-inch Fe-  doped β-Ga2O3 (010) substrates prepared using ver tical Bridgman method. Jpn J Appl Phys. 2023;62(SF):  SF1006. doi: 10.35848/1347-4065/acb55a\[11\] Higashiwaki M, Konishi K, Sasaki K, et al.  Temperature-dependent capacitance–voltage and  current–voltage characteristics of Pt/Ga2O3 (001)  schottky barrier diodes fabricated on n––Ga2O3 drift  layers grown by halide vapor phase epitaxy. Appl  Phys Lett. 2016;108(13):133503. doi: 10.1063/1.  4945267\[12\] Sasaki K, Wakimoto D, Thieu QT, et al. First demon stration of Ga2O3 trench MOS-Type schottky barrier  diodes. IEEE Electron Device Lett. 2017;38(6):783.  doi: 10.1109/LED.2017.2696986\[13\] Dong P, Zhang J, Yan Q, et al. 6 kV/3.4 mΩ·cm2  vertical β-Ga2O3 schottky barrier diode with BV2  /Ron,sp performance exceeding 1-D unipolar limit of  GaN and SiC. IEEE Electron Device Lett. 2022;43  (5):765. doi: 10.1109/LED.2022.3160366\[14\] Lin C-H, Ema K, Masuya S, et al. Uniformity  improvement of thickness and net donor concentra tion in halide vapor phase epitaxial β-Ga2O3 wafers  prepared on miscut angle substrates. Jpn J Appl Phys.  2023;62(SF):SF1005. doi: 10.35848/1347-4065/acb4fb\[15\] Higashiwaki M, Sasaki K, Kamimura T, et al.  Depletion-mode Ga2O3 metal-oxide-semiconductor  field-effect transistors on β-Ga2O3 (010) substrates  and temperature dependence of their device charac teristics. Appl Phys Lett. 2013;103(12):123511. doi:  10.1063/1.4821858\[16\] Li W, Nomoto K, Hu Z, et al. Single and multi-fin  normally-off Ga2O3 vertical transistors with  a breakdown voltage over 2.6 kV. In: 2019 IEEE  International Electron Devices Meeting (IEDM) vol  2019-Decem IEEE. p.12.4.1–.12.4.4. doi: 10.1109/  IEDM19573.2019.8993526\[17\] Wong MH, Goto K, Murakami H, et al. Current  aperture vertical β-Ga2O3 MOSFETs fabricated by  N- and Si-ion implantation doping. IEEE Electron  Device Lett. 2019;40(3):431. doi: 10.1109/LED.2018.  2884542\[18\] Peterson C, Alema F, Bhattacharyya A, et al. Kilovolt-  class β-Ga2O3 MOSFETs on 1-in. bulk substrates.  Appl Phys Lett. 2024;124(8):082104. doi: 10.1063/5.  0187989\[19\] Wakimoto D, Lin C-H, Ema K, et al. A multi-fin  normally-off β -Ga2O3 vertical transistor with  a breakdown voltage exceeding 10 kV. Appl Phys  Express. 2025;18(10):106502. doi: 10.35848/1882-  0786/ae0d2a\[20\] Sasaki K. Prospects for β-Ga2O3: now and into the  future. Appl Phys Express. 2024;17(9):090101. doi: 10.  35848/1882-0786/ad6b73\[21\] Goto K, Konishi K, Murakami H, et al. Halide vapor  phase epitaxy of Si doped β-Ga2O3 and its electrical  properties. Thin Solid Films. 2018;666  (September):182. doi: 10.1016/j.tsf.2018.09.006\[22\] Ema K, Sasaki K, Kuramata A, et al. Homo- and  hetero-epitaxial growth of β-gallium oxide via GaCl3  -O2-N2 system. J Cryst Growth. 2021;564  (March):126129. doi: 10.1016/j.jcrysgro.2021.126129\[23\] Fujikura H, Konno T, Yoshida T, et al. Hydride-vapor  -phase epitaxial growth of highly pure GaN layers  with smooth as-grown surfaces on freestanding GaN  substrates. Jpn J Appl Phys. 2017;56(8):085503. doi:  10.7567/JJAP.56.085503Sci. Technol. Adv. Mater. 27 (2026) 10                                                                                                                                    Y. Oshima and T. Oshima[https://doi.org/10.1063/1.1330559](https://doi.org/10.1063/1.1330559)[https://doi.org/10.7567/JJAP.54.112601](https://doi.org/10.7567/JJAP.54.112601)[https://doi.org/10.1063/1.3674287](https://doi.org/10.1063/1.3674287)[https://doi.org/10.1063/1.3674287](https://doi.org/10.1063/1.3674287)[https://doi.org/10.1143/JJAP.47.8506](https://doi.org/10.1143/JJAP.47.8506)[https://doi.org/10.7567/JJAP.55.1202A2](https://doi.org/10.7567/JJAP.55.1202A2)[https://doi.org/10.7567/JJAP.55.1202A2](https://doi.org/10.7567/JJAP.55.1202A2)[https://doi.org/10.1109/TSM.2019.2944526](https://doi.org/10.1109/TSM.2019.2944526)[https://doi.org/10.1063/5.0076962](https://doi.org/10.1063/5.0076962)[https://doi.org/10.1016/j.jcrysgro.2020.125724](https://doi.org/10.1016/j.jcrysgro.2020.125724)[https://doi.org/10.1016/j.jcrysgro.2020.125778](https://doi.org/10.1016/j.jcrysgro.2020.125778)[https://doi.org/10.1016/j.jcrysgro.2020.125778](https://doi.org/10.1016/j.jcrysgro.2020.125778)[https://doi.org/10.35848/1347-4065/acb55a](https://doi.org/10.35848/1347-4065/acb55a)[https://doi.org/10.1063/1.4945267](https://doi.org/10.1063/1.4945267)[https://doi.org/10.1063/1.4945267](https://doi.org/10.1063/1.4945267)[https://doi.org/10.1109/LED.2017.2696986](https://doi.org/10.1109/LED.2017.2696986)[https://doi.org/10.1109/LED.2022.3160366](https://doi.org/10.1109/LED.2022.3160366)[https://doi.org/10.35848/1347-4065/acb4fb](https://doi.org/10.35848/1347-4065/acb4fb)[https://doi.org/10.1063/1.4821858](https://doi.org/10.1063/1.4821858)[https://doi.org/10.1063/1.4821858](https://doi.org/10.1063/1.4821858)[https://doi.org/10.1109/IEDM19573.2019.8993526](https://doi.org/10.1109/IEDM19573.2019.8993526)[https://doi.org/10.1109/IEDM19573.2019.8993526](https://doi.org/10.1109/IEDM19573.2019.8993526)[https://doi.org/10.1109/LED.2018.2884542](https://doi.org/10.1109/LED.2018.2884542)[https://doi.org/10.1109/LED.2018.2884542](https://doi.org/10.1109/LED.2018.2884542)[https://doi.org/10.1063/5.0187989](https://doi.org/10.1063/5.0187989)[https://doi.org/10.1063/5.0187989](https://doi.org/10.1063/5.0187989)[https://doi.org/10.35848/1882-0786/ae0d2a](https://doi.org/10.35848/1882-0786/ae0d2a)[https://doi.org/10.35848/1882-0786/ae0d2a](https://doi.org/10.35848/1882-0786/ae0d2a)[https://doi.org/10.35848/1882-0786/ad6b73](https://doi.org/10.35848/1882-0786/ad6b73)[https://doi.org/10.35848/1882-0786/ad6b73](https://doi.org/10.35848/1882-0786/ad6b73)[https://doi.org/10.1016/j.tsf.2018.09.006](https://doi.org/10.1016/j.tsf.2018.09.006)[https://doi.org/10.1016/j.jcrysgro.2021.126129](https://doi.org/10.1016/j.jcrysgro.2021.126129)[https://doi.org/10.7567/JJAP.56.085503](https://doi.org/10.7567/JJAP.56.085503)[https://doi.org/10.7567/JJAP.56.085503](https://doi.org/10.7567/JJAP.56.085503)\[24\] Ahmadi E, Oshima Y. Materials issues and devices of  α- and β-Ga2O3. J Appl Phys. 2019;126(16):160901.  doi: 10.1063/1.5123213   \[25\] Oshima Y, Vίllora EG, Shimamura K. Quasi-  heteroepitaxial growth of β-Ga2O3 on off-angled sap phire (0001) substrates by halide vapor phase epitaxy.  J Cryst Growth. 2015;410:53–58. doi: 10.1016/j.jcrys  gro.2014.10.038   \[26\] Oshima Y, Oshima T. Rapid homoepitaxial growth of  (011) β-Ga2O3 by HCl-based halide vapor phase epi taxy. Sci Technol Adv Mater. 2025;26(1):2585551.  doi: 10.1080/14686996.2025.2585551   \[27\] Pechnikov AI, Stepanov SI, Chikiryaka AV, et al.  Thick α-Ga2O3 layers on sapphire substrates grown  by halide epitaxy. Semiconductors. 2019;53(6):780.  doi: 10.1134/S1063782619060150   \[28\] Oshima Y, Oshima T. Homoepitaxial growth of  (-102) β-Ga2O3 by halide vapor phase epitaxy.  Semicond Sci Technol. 2023;38(10):105003. doi: 10.  1088/1361-6641/acf241   \[29\] Oshima Y, Kawara K, Oshima T, et al. Rapid growth  of α-Ga2O3 by HCl-boosted halide vapor phase epi taxy and effect of precursor supply conditions on  crystal properties. Semicond Sci Technol. 2020;35  (5):55022. doi: 10.1088/1361-6641/ab7843   \[30\] Goto K, Murakami H, Kuramata A, et al. Effect of  substrate orientation on homoepitaxial growth of β -  Ga2O3 by halide vapor phase epitaxy. Appl Phys Lett.  2022;120(10):102102. doi: 10.1063/5.0087609   \[31\] Murakami H, Nomura K, Goto K, et al.  Homoepitaxial growth of β-Ga2O3 layers by halide  vapor phase epitaxy. Appl Phys Express. 2015;8  (1):015503. doi: 10.7567/APEX.8.015503   \[32\] Ema K, Lin C-H, Ueda Y, et al. Homo-epitaxial growth  on the β-Ga2O3 (011) substrate by HVPE. In: Prenented  in the 85th JSAP Autumn meeting 2024, presentation  number 20p-A22-4 (abstract page 16–085); \[33\] Hollar EJ, Farzana E. Over 3 kV and ultra-low leakage  vertical (011) β -Ga2O3 power diodes with engineered  schottky contact and high-permittivity dielectric field  plate. Appl Phys Lett. 2026;128(5):053503. doi: 10.  1063/5.0309746   \[34\] Wen J, Liu J, Li Q, et al. Over 3.5 kV (011) β-Ga2O3  schottky barrier diode and first demonstration of 4.7  kV nio-based (011) β-Ga2O3 heterojunction diode.  In: 2026 IEEE 38th International Symposium on  Power Semiconductor Devices and ICs (ISPSD).  IEEE. p. 317–320. doi: 10.1109/ISPSD64561.2026.  11553728   \[35\] Varley JB, Weber JR, Janotti A, et al. Oxygen vacan cies and donor impurities in β-Ga2O3. Appl Phys Lett.  2010;97(14):142106. doi: 10.1063/1.3499306   \[36\] Shobu K. CaTCalc: new thermodynamic equilibrium  calculation software. Calphad Comput Coupling  Phase Diag Thermochem. 2009;33(2):279. doi: 10.  1016/j.calphad.2008.09.015   \[37\] Yamaguchi H, Akito Kuramata TM. Slip system ana lysis and X-ray topographic study on β-Ga2O3.  Superlattices Microstruct. 2016;99:99. doi: 10.1016/j.  spmi.2016.04.030   \[38\] Oshima T, Oshima Y. Anisotropic non-plasma HCl  gas etching of a (010) β-Ga2O3substrate. Appl Phys  Express. 2023;16(6):66501. doi: 10.35848/1882-0786/  acdbb7   \[39\] Oshima T, Oshima Y. Plasma-free dry etching of  (001) β-Ga2O3 substrates by HCl gas. Appl Phys  Lett. 2023;122(16):162102. doi: 10.1063/5.0138736 Sci. Technol. Adv. Mater. 27 (2026) 11                                                                                                                                    Y. Oshima and T. Oshima[https://doi.org/10.1063/1.5123213](https://doi.org/10.1063/1.5123213)[https://doi.org/10.1016/j.jcrysgro.2014.10.038](https://doi.org/10.1016/j.jcrysgro.2014.10.038)[https://doi.org/10.1016/j.jcrysgro.2014.10.038](https://doi.org/10.1016/j.jcrysgro.2014.10.038)[https://doi.org/10.1080/14686996.2025.2585551](https://doi.org/10.1080/14686996.2025.2585551)[https://doi.org/10.1134/S1063782619060150](https://doi.org/10.1134/S1063782619060150)[https://doi.org/10.1088/1361-6641/acf241](https://doi.org/10.1088/1361-6641/acf241)[https://doi.org/10.1088/1361-6641/acf241](https://doi.org/10.1088/1361-6641/acf241)[https://doi.org/10.1088/1361-6641/ab7843](https://doi.org/10.1088/1361-6641/ab7843)[https://doi.org/10.1063/5.0087609](https://doi.org/10.1063/5.0087609)[https://doi.org/10.7567/APEX.8.015503](https://doi.org/10.7567/APEX.8.015503)[https://doi.org/10.1063/5.0309746](https://doi.org/10.1063/5.0309746)[https://doi.org/10.1063/5.0309746](https://doi.org/10.1063/5.0309746)[https://doi.org/10.1109/ISPSD64561.2026.11553728](https://doi.org/10.1109/ISPSD64561.2026.11553728)[https://doi.org/10.1109/ISPSD64561.2026.11553728](https://doi.org/10.1109/ISPSD64561.2026.11553728)[https://doi.org/10.1063/1.3499306](https://doi.org/10.1063/1.3499306)[https://doi.org/10.1016/j.calphad.2008.09.015](https://doi.org/10.1016/j.calphad.2008.09.015)[https://doi.org/10.1016/j.calphad.2008.09.015](https://doi.org/10.1016/j.calphad.2008.09.015)[https://doi.org/10.1016/j.spmi.2016.04.030](https://doi.org/10.1016/j.spmi.2016.04.030)[https://doi.org/10.1016/j.spmi.2016.04.030](https://doi.org/10.1016/j.spmi.2016.04.030)[https://doi.org/10.35848/1882-0786/acdbb7](https://doi.org/10.35848/1882-0786/acdbb7)[https://doi.org/10.35848/1882-0786/acdbb7](https://doi.org/10.35848/1882-0786/acdbb7)[https://doi.org/10.1063/5.0138736](https://doi.org/10.1063/5.0138736)- Abstract- Abstract- Introduction- Experimental methods- Results and discussion- Conclusions- Acknowledgments- Disclosure statement- Funding- ORCID- References