# Fileset

[042110_1_5.0186319.pdf](https://mdr.nims.go.jp/filesets/f429bcda-2231-4a5e-aa4a-b405c25d9b93/download)

## Creator

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

## Rights

This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in akayoshi Oshima et al., Appl. Phys. Lett. 124, 042110 (2024) and may be found at https://doi.org/10.1063/5.0186319.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Using selective-area growth and selective-area etching on (−102) β-Ga2O3 substrates to fabricate plasma-damage-free vertical fins and trenches](https://mdr.nims.go.jp/datasets/575c0527-225d-4519-abd6-14d9b31db681)

## Fulltext

Using selective-area growth and selective-area etching on (−102) β-Ga2O3 substrates to fabricate plasma-damage-free vertical fins and trenchesViewOnlineExportCitationCrossMarkRESEARCH ARTICLE |  JANUARY 25 2024Using selective-area growth and selective-area etching on(−102) β-Ga2O3 substrates to fabricate plasma-damage-freevertical fins and trenches Takayoshi Oshima   ; Yuichi Oshima Appl. Phys. Lett. 124, 042110 (2024)https://doi.org/10.1063/5.0186319 26 January 2024 00:02:07https://pubs.aip.org/aip/apl/article/124/4/042110/3132648/Using-selective-area-growth-and-selective-areahttps://pubs.aip.org/aip/apl/article/124/4/042110/3132648/Using-selective-area-growth-and-selective-area?pdfCoverIconEvent=citehttps://pubs.aip.org/aip/apl/article/124/4/042110/3132648/Using-selective-area-growth-and-selective-area?pdfCoverIconEvent=crossmarkjavascript:;https://orcid.org/0000-0001-8550-9735javascript:;https://orcid.org/0000-0001-8293-4891javascript:;https://doi.org/10.1063/5.0186319https://servedbyadbutler.com/redirect.spark?MID=176720&plid=2288780&setID=592934&channelID=0&CID=840268&banID=521619201&PID=0&textadID=0&tc=1&scheduleID=2209021&adSize=1640x440&data_keys=%7B%22%22%3A%22%22%7D&matches=%5B%22inurl%3A%5C%2Fapl%22%5D&mt=1706227327185546&spr=1&referrer=http%3A%2F%2Fpubs.aip.org%2Faip%2Fapl%2Farticle-pdf%2Fdoi%2F10.1063%2F5.0186319%2F18930418%2F042110_1_5.0186319.pdf&hc=67e17216768e53f6d831624589731d0d55fdbb9c&location=Using selective-area growth and selective-areaetching on (�102) b-Ga2O3 substrates to fabricateplasma-damage-free vertical fins and trenchesCite as: Appl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319Submitted: 5 November 2023 . Accepted: 11 January 2024 .Published Online: 25 January 2024Takayoshi Oshimaa) and Yuichi OshimaAFFILIATIONSResearch Center for Electronic and Optical Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0044, Japana)Author to whom correspondence should be addressed:OSHIMA.Takayoshi@nims.go.jpABSTRACTWe have demonstrated selective-area growth and selective-area etching on SiO2-masked (�102) b-Ga2O3 substrates using a HCl-basedhalide-vapor-phase epitaxy system that is capable of performing both growth and gas etching without plasma excitation. Since the surface ofthe (�102) substrate is perpendicular to the (100) plane, which has the lowest surface energy, we were able to use both methods to fabricateplasma-damage-free fins and trenches with (100)-faceted vertical sidewalls on windows striped along the [010] direction with high processingaccuracy. Furthermore, since the [010] window direction is aligned parallel to the majority of dislocations and line-shaped voids in the sub-strate—which extend along the [010] direction and could potentially act as leakage paths—such crystal defects are unlikely to appear on thesurfaces of the resulting fins and trenches. We believe that these selective-area growth/etching techniques can greatly accelerate research on,and the development of, b-Ga2O3-based vertical/lateral devices with fins or trenches.Published under an exclusive license by AIP Publishing. https://doi.org/10.1063/5.0186319The ultra-wide-bandgap semiconductor b-Ga2O3 has garneredconsiderable attention in recent years because it is expected to have acritical field strength as large as approximately 8MV cm�1,1 the widelycontrollable doping range of 1015–1020 cm�3,2 and high-quality wafersare available with the low defect densities of 103–105 cm�2.3 Theseproperties all indicate that this material has significant potential foruse in next-generation power semiconductors, following the footstepsof SiC and GaN. Because of such expectations, research on b-Ga2O3has advanced in various areas, including the growth of bulk crystals,homo- and hetero-epitaxy, impurity doping, device design and proc-essing, and related theoretical frameworks.4Processing technologies that can be used to shape various struc-tures on a wafer surface constitute one such area, which is indispens-able for device manufacturing. In the b-Ga2O3 research community,plasma-based dry etching that utilizes chlorine-based chemistries iswidely accepted as a proven process5 for obtaining various structures,including mesas, fins, and trenches. They are used to define deviceregions and fabricate fin channels, recessed-gate structures, trenchmetal-oxide-semiconductor (trench MOS) structures, edge termina-tions, and so on. Utilizing such processed structures, fin field-effecttransistors (FinFETs),6–8 gate-recessed MOSFETs,9 trench MOSSchottky barrier diodes (trench MOSSBDs),10,11 trench MOSFETs,12and mesa-terminated SBDs13,14 and PN diodes15 have demonstratedimproved device performance—for example, increased breakdownvoltages and/or normally off operations—confirming the benefit ofthese structural effects. However, reactive ions in the plasma causedamage to the processed surfaces, degrading electrical properties of thedevices such as the channel mobilities and transfer characteristics ofFinFETs,6,7 and the on-resistances of trench SBDs,10 gate-recessedMOSFETs,16 and mesa-terminated SBDs.13 In order to recover fromthis damage and restore proper device performance, post-treatmentslike wet chemical etching17–19 or annealing16,19,20 are necessary. Thus,plasma etching is actually a two-step process that requires a damage-recovery procedure. In contrast, non-plasma-based wet and dry etch-ing processes—such as metal-assisted chemical etching (MacEtch),21Ga-flux etching,22 and hydrogen-environment anisotropic thermaletching (HEATE)23—have emerged as alternatives to plasma-baseddry etching. These non-plasma techniques are quite promising, asevidenced by the facts that FinFETs formed using MacEtch shownear-zero hysteresis in their gate-transfer characteristics,24 thatGa-flux-etched recessed-gate MOSFETs have been produced with highdrain currents and high transconductances,25 and that trenchMOSSBDs have been formed with near-ideal forward characteristics.26In this context, we have recently demonstrated plasma-freeselective-area growth (SAG) and selective-area etching (SAE) of com-mercially available (010) and (001) b-Ga2O3 substrates using aAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-1Published under an exclusive license by AIP PublishingApplied Physics Letters ARTICLE pubs.aip.org/aip/apl 26 January 2024 00:02:07https://doi.org/10.1063/5.0186319https://doi.org/10.1063/5.0186319https://www.pubs.aip.org/action/showCitFormats?type=show&doi=10.1063/5.0186319http://crossmark.crossref.org/dialog/?doi=10.1063/5.0186319&domain=pdf&date_stamp=2024-01-25https://orcid.org/0000-0001-8550-9735https://orcid.org/0000-0001-8293-4891mailto:OSHIMA.Takayoshi@nims.go.jphttps://doi.org/10.1063/5.0186319pubs.aip.org/aip/apllaboratory-designed HCl-based halide-vapor-phase epitaxy (HVPE)system.27–30 This HVPE system includes an HCl etching-gas line inaddition to the growth-precursor lines that enable regulation of thebalance between growth and etching. This enables growth on windowswhile suppressing nucleation on amorphous masks, thereby achievingSAG with perfect selectively. Alternatively, SAE can be performed inwindow areas using the HCl etching gas alone. Thus, this system canalso be utilized as a halide-vapor-phase etching (HVPE) system.During both growth and etching, the sidewalls of the resulting struc-tures are dominated by flat (100) facets owing to their high chemicalstability, which can be attributed to the fact that they have the lowestsurface energy density.31,32 When the window direction is parallel tothe (100) plane, the sidewalls consist exclusively of (100) facets, facili-tating the production of high-aspect-ratio structures such as fins andtrenches, as shown in Fig. 1. In particular, vertical fins and trenchescan be created on (010) substrates due to the orthogonal relationshipbetween the (100) and (010) planes. Nevertheless, a (010)-orientedsubstrate is unsuitable for vertical power-device applications becauseof its crystal defects. Since most line-shaped voids (also called nano-pipes) and dislocations in the crystal exist along the [010] direction,33–35 it is very likely for these crystal defects to appear on the surface ofthe resulting structures and act as leakage paths, causing prematurebreakdown.36,37 Conversely, a (001) substrate can avoid such [010]-extended defect issues as the defects are parallel to the (001) plane andrarely surface on the substrate or the fins/trenches created thereon.However, the structures produced by the SAG or SAE on (001) sub-strates are inclined by 13.7� from the substrate normal because theface angle between the (100) and (001) planes is 103.7�. Such inclinedfins and trenches are not suitable for vertical power-device applica-tions, as the current and/or electric field is concentrated on either sideof the structures. It is, therefore, necessary to consider the substrate ori-entation for implementation of the SAG and SAE methods in process-ing vertical power devices.The (�102) plane is the optimal substrate orientation for usingSAG and SAE to create fins/trenches with flat and vertical sidewallsthat are suitable for vertical power-device applications. To producefins/trenches with (100)-faceted vertical sidewalls, the substrate surfacemust be perpendicular to the (100) plane due to the dominance of(100) facets in the structures formed using SAG or SAE. To minimizethe appearance of line-shaped voids and dislocations along the [010]direction on the surface, it is also necessary for the substrate to be ori-ented parallel to the [010] direction. To meet these two requirements,the substrate plane must be the (�102) plane, as shown in Fig. 1. Wenote that the (�102) plane can be set as the primary plane for edge-defined film-fed growth (EFG) because it is parallel to the [010] pullingdirection.38 This allows the fabrication of scalable wafers with highproductivity, which can be comparable to the productivity achieved for(001) and (�201) wafers. In addition, prior to the present study, we veri-fied that homoepitaxial growth is possible on (�102) substrates withoutmisoriented domains, which is similar to the cases of (010) and (001)substrates.39 Encouraged by these facts and expectations, we, therefore,performed experiments using SAG and SAE on (�102) substrates andinvestigated the resulting fin and trench structures.We prepared SiO2-masked (�102) b-Ga2O3 substrates as follows.10� 15mm2 substrates with the (�102) orientation were manufacturedby slicing, cutting, grinding, lapping, and chemically and mechanicallypolishing a Sn-doped ingot with (001) principal planes grown alongthe [010] direction using the EFG method (Novel Crystal Technology,Inc.). The (�102) substrates were monocrystalline, with highly crystal-line quality that was comparable to those of commercially available(001) and (010) substrates. This is evidenced by the x-ray diffraction(XRD) results summarized in Fig. 2. As expected, we observed only asingle peak in both the symmetric h–2h and skew-symmetric / scans[Figs. 2(a) and 2(b), respectively]. The full-width at half maximum val-ues of the x rocking curves measured in the symmetric and skew-symmetric geometries were as narrow as 20–30 arc sec [Fig. 2(c)].Prior to initiating the mask fabrication, we subjected the substrate sur-face to a meticulous cleaning process. This involved immersion inorganic solvents, specifically acetone and isopropyl alcohol, followedby treatment with HF and a H2SO4/H2O2 blend. This step was crucialto ensure the complete removal of any residues from the chemicalmechanical polishing slurry. Subsequently, the substrate surface wascoated with a 0.10-lm-thick SiO2 layer. This layer was applied usingplasma-assisted chemical vapor deposition, utilizing tetraethoxysilaneand O2 as the source precursors. We then opened square-, radial-line-,and stripe-patterned windows in the SiO2 layer using laser lithography,with buffered HF as the etchant. We removed the residual photoresistcompletely using organic solvents (acetone and isopropyl alcohol) andO2 plasma. The square window was 100� 100lm2 in size. The radialpattern consisted of 36 window lines in 10� steps starting from [010].The width of each window was 1.4lm. We prepared two types ofstripes: pattern A and pattern B. They were both parallel to the [010]direction, but they differed in the mask/window widths, which were2.6/1.4lm (pattern A) and 0.5/5.7lm (pattern B).We performed SAG and SAE using a laboratory-made HCl-basedHVPE system similar to that employed in our previous experimentson (010)- and (001)-oriented substrates.27–30 This system is equippedwith a horizontal quartz reactor, which is divided into an upstreamFIG. 1. (Two columns) Schematic crosssections of selective-area-grown andselective-area-etched structures on (010),(001), and (�102) substrates showing thedirectional relationships among the sub-strate planes, (100)-faceted sidewalls, and[010]-oriented propagation direction of themajority of voids and dislocations.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-2Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07pubs.aip.org/aip/aplsource zone and a downstream processing zone. In these zones, pre-cursor and/or etching gases are introduced to a vertically oriented sub-strate. This substrate is mounted on a rotating holder for uniformexposure, as depicted in Fig. S1. For SAG, GaCl precursor was synthe-sized in the source zone at 822 �C using the chemical reaction betweenGa metal (>99.99999% pure) and HCl gas (>99.999% pure). Underthe reaction conditions, the supplied HCl gas was completely con-sumed to generate an equimolar amount of GaCl gas. We then per-formed SAG by injecting the GaCl and O2 (>99.99995% pure)precursors, together with additional HCl etching gas, onto the surfaceof the SiO2-masked (�102) b-Ga2O3 substrate while heating the proc-essing zone at 1038 �C for 15min. It is necessary to supply the HCletching gas to suppress parasitic gas-phase reactions and nucleation onthe mask.27,40 For SAE, we performed the etching by introducing theHCl etching gas onto the SiO2-masked (�102) substrate while heatingthe processing zone at 1038 �C for 10min. For both SAG and SAE, thereactive gases were carried by purified N2 gas (dew point < �110 �C)under atmospheric pressure (approximately 100 kPa), and we main-tained the total gas flow through the reactor at 8 slm. The partial pres-sures of GaCl, O2, and HCl (for etching purposes) were set at 0.125,1.25, and 0.25 kPa in the case of SAG, and at 0, 0, and 62.5 Pa for SAE,respectively.The substrates, post-growth, and etching were examined to deter-mine surface compositions within 100� 100 lm2 window areas ofboth SAG and SAE substrates using Auger electron spectroscopy(AES). A bare (�102) substrate, subjected to the cleaning process, wasalso analyzed for comparison. We examined the grown and etchedstructures using scanning electron microscopy (SEM). Additionally, weobserved the cross sections of the structures using focused-ion-beamscanning electron microscopy (FIB–SEM), where we utilized carbondeposition as a protective layer against FIB milling. Furthermore, weviewed the atomic arrangement of the Ga atoms on the SAG b-Ga2O3in the vicinity of the mask edge from the [010] direction using high-angle annular dark-field scanning transmission electron microscopy(HAADF–STEM) with an accelerating voltage of 200 kV.We first assessed surface compositions following the SAG andSAE procedures. Figure 3 illustrates the typical AES spectra for theSAG, SAE, and the reference bare substrates. In all spectra, elements ofGa and O were present. C was also detected, likely due to hydrocarboncontaminants. Cl, however, was not detected within the analytical sen-sitivity of AES. The Ga/O atomic composition ratios for the SAG andSAE substrates, calculated using the relative sensitivity method, were0.724 and 0.721, aligning with the reference value of 0.724. This consis-tency suggests that the SAG and SAE processes do not alter the Ga/Osurface stoichiometry.We then characterized the in-plane anisotropies of the SAG andSAE structures on the (�102) substrates. Figures 4(a) and 4(b), respec-tively, show top-view SEM images of the SAG and SAE structures onthe radial-patterned windows. We determined the lengths of the lateralovergrowth and under-etching perpendicular to the windows fromthese images by measuring the distances between the window edgeand the end of the corresponding SAG fins and SAE trenches. Wethen plotted the acquired data in polar coordinates, as presented inFig. 4(c). The results are similar to those observed on (001) sub-strates.27,28 The lateral overgrowth and under etching were signifi-cantly reduced when the window directions were <010> because thecorresponding sidewalls were the chemically most stable (100) facets.This is reflected in the polar plots, as the deepest dips occur in the[201] and [�20�1] directions. Note that these directions are perpendicu-lar to the (100) plane. Slight deviations from the <010> directionresulted in remarkable enhancements of the lateral growth and etch-ing. For SAG, deviations of the window directions further from the<010> direction resulted in the deposition of polycrystals and/or theFIG. 3. (Single column) Auger electron spectra obtained from the 100� 100lm2window regions of the selective-area-grown and selective-area-etched samples.Additionally, it includes a spectrum from a cleaned bare (�102) substrate, serving asa reference.FIG. 2. (Single column) summary of the x-ray diffraction characterization of a(�102)-oriented b-Ga2O3 substrate using Cu Ka1 radiation. (a) Symmetric h–2h scanpattern. (b) Skew-symmetric / scan pattern at the tilt angle v� 59.8�. (c) The xrocking curves of the �204 and �401 diffraction peaks corresponding to (a) and (b),respectively.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-3Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07pubs.aip.org/aip/aplabsence of growth rather than in epitaxial growth of fins on the win-dows. Therefore, only the fins with (100) sidewalls survived, whilethose without were eliminated in the competition between growth andetching during SAG. For SAE, lateral etching was again suppressed asthe deviation of the window directions increased. This resulted in smalldips in the plots at the specific window directions where the sidewallsshould be {310} facets. Consequently, the window direction should be[010] for the best controllability of the cross-sectional shape becausethe lateral size deviation of the fins/trenches from the mask width canbe minimized.We then investigated the SAG and SAE structures on the stripedwindows along the [010] direction using SEM, as shown in Fig. 5.Here, we used pattern A striped windows with a narrow window widthto create both SAG fins and SAE trenches, while we used pattern Bstriped windows with a narrow mask width to fabricate SAE fins. In allcases, the overgrowth and under-etching lengths perpendicular to thewindows were much smaller than the corresponding grown height andetched depth, respectively. This means that the fins and trenches werefabricated nearly according to the window and mask widths [Figs.5(a), 5(f), and 5(k)]. In addition, vertical (100)-faceted sidewalls areclearly visible on both sides of the fins and trenches. In the SAG fins,the (100) sidewalls on the [�20�1] side were shorter compared to thoseon the [201] side. This is attributed to the formation of a downward-sloping facet at the top surface [Figs. 5(b)–5(e)], identified as the (�201)plane from subsequent cross-sectional analysis. Conversely, the (100)sidewalls on both sides were almost the same height in the SAEtrenches and fins [Figs. 5(g)–5(j) and 5(l)–5(o)]. The etched bottomsurfaces were relatively rough due to the presence of macro steps alongthe [010] direction. We also observed this roughness in the HCl-gas-etched bottom surface on the (001) substrates.28Next, we examined the cross-sectional profiles of the fabricatedSAG and SAE stripes along the [010] direction. Figures 6(a)–6(c) showtilted SEM images of the cross sections of the SAG and SAE stripescorresponding to those shown in Fig. 5. Figure 6(d) shows a schematiccross-sectional illustration with labeled facets for the SAG fins andSAE trenches on the pattern A masks. The SAG fins exhibited vertical(100)-faceted sidewalls and an inclined (�201)-faceted top surface. The(�201) facet’s emergence is likely due to its relatively low surface energydensity, second only to the (100) facet.32 The structural aspect ratio—which we define to be the (100)-sidewall height divided by the lateralovergrowth or under-etching length—was 2.7 ([�20�1] side)–6.7 ([201]side). It is important to note that there were small bumps on both sidesof the (100) sidewalls of the fins due to the formation of twin domainsduring the lateral overgrowth, which we discuss later. Conversely, theSAE trenches and fins consisted of vertical (100)-faceted sidewalls anda relatively rough bottom surface. Note that the (100) sidewall on the[�20�1] side was not perfectly vertical but slightly curved. Moreover, theunder-etching length on the [�20�1] side was a little bit longer than thatof the sidewall on the [201] side. The detailed reason for this left–rightasymmetry of the sidewalls is unclear at the moment, but it must berelated to the anisotropy of the monoclinic structure. Optimization ofthe etching parameters—such as the processing temperature and HClpartial pressure30—may flatten the (100) sidewalls on the [�20�1] sidebecause the (100) and (�1 00) planes that correspond to the opposingsidewalls are crystallographically equivalent. The structural aspect ratiowas 7.9 ([201] side)–11.2 ([�20�1] side), which is higher than the corre-sponding values for the SAG structures.FIG. 4. (Single column) top-view scanning electron microscope images of (a)selective-area-grown and (b) selective-area-etched b-Ga2O3 on radial-line windows.(c) Polar plots of the lateral-overgrowth length (empty squares) and under-etchinglength (empty circles) measured from (a) and (b), respectively. The symbols “�”and “c” indicate the cross product of the given directions and the [001] direction.The solid lines connecting the data in (c) are guides to the eye.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-4Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07pubs.aip.org/aip/aplFinally, we examined the atomic arrangements of the overgrownregion of the SAG b-Ga2O3 using STEM, as shown in Fig. 7. It ishighly probable for twinning to occur during lateral overgrowth on the(100) sidewalls because homoepitaxial growth on (100) b-Ga2O3 sub-strates without miscut angle forms twin lamellae due to the doublepositioning of the adatoms on the (100) plane.41 Figures 7(a) and 7(b)display HAADF–STEM images in the vicinity of the mask edge, inwhich the comparatively heavy Ga atoms are visible as white dots.We observed a single periodic arrangement of Ga atoms in the sub-strate region and in the SAG region above the window, which is con-sistent with the result from homoepitaxy on (�102) b-Ga2O3substrates.39 We found a twin domain in the lateral-overgrown regionabove the mask alongside a coherent boundary stemming from themask edge, as evidenced by the twin symmetry of the fast Fouriertransformed images of the two domain regions that cross the boundary[Figs. 7(c) and 7(d)]. These findings indicate that the twinning wasFIG. 5. (Two columns) top- and tilted-view scanning electron microscope images (at angles of 0� and 50� relative to the sample normal, respectively), of the selective-area-grown fins on the stripe mask of (a)–(e) pattern A and of the selective-area-etched trenches under the stripe masks of (f)–(j) pattern A and (k)–(o) pattern B.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-5Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07pubs.aip.org/aip/aplcaused by the same mechanism as that occurring during homoepitaxialgrowth on (100) substrates. The twin formation is anticipated not onlyfor the (�102) substrates but also on the SAG sidewalls for the (010)and (001) substrates, in line with this mechanism. Therefore, it is cru-cial to consider the presence of twin domains located proximate to the(100) sidewall surfaces when employing SAG structures for devices.In summary, we have explored the potential of SAG and SAEusing HVPE as one-step plasma-free processes for b-Ga2O3. By utiliz-ing (�102)-oriented substrates, which are perpendicular to the (100)plane and parallel to the [010] direction, we have been able to useSAG and SAE to fabricate high-aspect-ratio fins and trenches with(100)-faceted vertical sidewalls along the [010] direction. In additionto being free of plasma damage, these structures can also prevent line-shaped voids and dislocations that extend in the [010] direction fromappearing on their surfaces, making them suitable for both vertical andlateral device applications. However, we have also found that SAGFIG. 6. (Single column) (a)–(c) 54�-tilted-view scanning electron microscopeimages of the cross-sectional structures of selective-area-grown fins and selective-area-etched trenches. The fins in panel (a) correspond to Figs. 5(a)–5(e) (patternA). The trenches in panels (b) and (c) correspond to Figs. 5(f)–5(j) (pattern A) andFigs. 5(k)–5(o) (pattern B), respectively. (d) Schematic cross sections of (a) and (b).FIG. 7. (Single column) high-angle annular dark-field scanning transmission elec-tron microscope images around the mask edge of a selective-area-grown b-Ga2O3region taken at (a) low and (b) high magnifications. In (b), the small empty circlessurrounded by blue and red dashed parallelograms represent Ga atoms in unit cellsin the untwinned and twinned domains, respectively. The bottom panels show fastFourier transformed images of (c) the untwinned and (d) the twinned domainregions.Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-6Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07pubs.aip.org/aip/aplinevitably causes twinning in the lateral-overgrown region of the (100)sidewalls. This occurs because the lateral growth is virtually the sameas homoepitaxial growth on (100) substrates. Therefore, SAE is morepromising than SAG for practical device applications.See the supplementary material for the schematics illustrating thegas flows in the HVPE system (Fig. S1).The authors thank T. Harada of the National Institute forMaterials Science regarding XRD characterization and Y. Yao of theJapan Fine Ceramics Center for valuable discussions on dislocationsin b-Ga2O3. This research was partially funded by the TEPCOMemorial Foundation and supported by the Advanced ResearchInfrastructure for Materials and Nanotechnology in Japan (ARIM)of the Ministry of Education, Culture, Sports, Science andTechnology (MEXT), Proposal No. JPMXP1223NM5073.AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsTakayoshi Oshima: Conceptualization (lead); Data curation (lead);Methodology (equal); Writing – original draft (lead). Yuichi Oshima:Conceptualization (supporting); Methodology (equal); Writing –review & editing (lead).DATA AVAILABILITYThe data that support the findings of this study are available fromthe corresponding author upon reasonable request.REFERENCES1M. Higashiwaki, K. Sasaki, A. Kuramata, T. Masui, and S. Yamakoshi, Appl.Phys. Lett. 100, 013504 (2012).2J. Zhang, J. Shi, D.-C. Qi, L. Chen, and K. H. L. Zhang, APL Mater. 8, 020906(2020).3O. Ueda, M. Kasu, and H. Yamaguchi, Jpn. J. Appl. Phys., Part 1 61, 050101(2022).4A. J. Green, J. Speck, G. Xing, P. Moens, F. Allerstam, K. Gumaelius, T. Neyer,A. Arias-Purdue, V. Mehrotra, A. Kuramata, K. Sasaki, S. Watanabe, K. Koshi,J. Blevins, O. Bierwagen, S. Krishnamoorthy, K. Leedy, A. R. Arehart, A. T.Neal, S. Mou, S. A. Ringel, A. Kumar, A. Sharma, K. Ghosh, U. Singisetti, W.Li, K. Chabak, K. Liddy, A. Islam, S. Rajan, S. Graham, S. Choi, Z. Cheng, andM. Higashiwaki, APL Mater. 10, 029201 (2022).5R. Khanna, K. Bevlin, D. Geerpuram, J. Yang, F. Ren, and S. Pearton, GalliumOxide (Elsevier, 2019), pp. 263–285.6K. D. Chabak, N. Moser, A. J. Green, D. E. Walker, S. E. Tetlak, E. Heller, A.Crespo, R. Fitch, J. P. McCandless, K. Leedy, M. Baldini, G. Wagner, Z.Galazka, X. Li, and G. Jessen, Appl. Phys. Lett. 109, 213501 (2016).7Z. Hu, K. Nomoto, W. Li, Z. Zhang, N. Tanen, Q. T. Thieu, K. Sasaki, A.Kuramata, T. Nakamura, D. Jena, and H. G. Xing, Appl. Phys. Lett. 113,122103 (2018).8W. Li, K. Nomoto, Z. Hu, T. Nakamura, D. Jena, and H. G. Xing, 2019 IEEEInt. Electron Devices Meet. (IEEE, 2019).9K. D. Chabak, J. P. McCandless, N. A. Moser, A. J. Green, K. Mahalingam, A.Crespo, N. Hendricks, B. M. Howe, S. E. Tetlak, K. Leedy, R. C. Fitch, D.Wakimoto, K. Sasaki, A. Kuramata, and G. H. Jessen, IEEE Electron DeviceLett. 39, 67 (2018).10W. Li, K. Nomoto, Z. Hu, D. Jena, and H. G. Xing, IEEE Trans. ElectronDevices 68, 2420 (2021).11F. Otsuka, H. Miyamoto, A. Takatsuka, S. Kunori, K. Sasaki, and A. Kuramata,Appl. Phys. Express 15, 016501 (2022).12Y. Ma, X. Zhou, W. Tang, X. Zhang, G. Xu, L. Zhang, T. Chen, S. Dai, C. Bian,B. Li, Z. Zeng, and S. Long, IEEE Electron Device Lett. 44, 384 (2023).13S. Dhara, N. K. Kalarickal, A. Dheenan, C. Joishi, and S. Rajan, Appl. Phys. Lett.121, 203501 (2022).14H. Chen, H. Wang, and K. Sheng, IEEE Electron Device Lett. 44, 21 (2023).15K. Danno, M. Kado, T. Hara, T. Takasugi, H. Yamano, Y. Umetani, and T.Shoji, Jpn. J. Appl. Phys., Part 1 62, SF1007 (2023).16S. Sharma, L. Meng, A. F. M. A. U. Bhuiyan, Z. Feng, D. Eason, H. Zhao, andU. Singisetti, IEEE Electron Device Lett. 43, 2029 (2022).17Y. Zhang, A. Mauze, F. Alema, A. Osinsky, and J. S. Speck, Appl. Phys. Express12, 044005 (2019).18H.-K. Lee, H.-J. Yun, K.-H. Shim, H.-G. Park, T.-H. Jang, S.-N. Lee, and C.-J.Choi, Appl. Surf. Sci. 506, 144673 (2020).19Z. Wang, X. Yu, H. Gong, T. Hu, Y. Zhang, X. Ji, F. Ren, S. Gu, Y. Zheng, R.Zhang, A. Y. Kuznetsov, and J. Ye, J. Phys. Chem. Lett. 13, 7094 (2022).20J. Yang, F. Ren, R. Khanna, K. Bevlin, D. Geerpuram, L.-C. Tung, J. Lin, H.Jiang, J. Lee, E. Flitsiyan, L. Chernyak, S. J. Pearton, and A. Kuramata, J. Vac.Sci. Technol., B 35, 051201 (2017).21H. Huang, M. Kim, X. Zhan, K. Chabak, J. D. Kim, A. Kvit, D. Liu, Z. Ma, J.-M.Zuo, and X. Li, ACS Nano 13, 8784 (2019).22N. K. Kalarickal, A. Fiedler, S. Dhara, H.-L. Huang, A. F. M. A. U. Bhuiyan, M.W. Rahman, T. Kim, Z. Xia, Z. J. Eddine, A. Dheenan, M. Brenner, H. Zhao, J.Hwang, and S. Rajan, Appl. Phys. Lett. 119, 123503 (2021).23Y. Yamazaki, M. Tomoaki, A. Takeki, and A. Kikuchi, in 4th InternationalWorkshop on Gallium Oxide and Related Materials, 2022.24H.-C. Huang, Z. Ren, A. F. M. Anhar Uddin Bhuiyan, Z. Feng, Z. Yang, X. Luo,A. Q. Huang, A. Green, K. Chabak, H. Zhao, and X. Li, Appl. Phys. Lett. 121,052102 (2022).25N. K. Kalarickal, A. Dheenan, J. F. McGlone, S. Dhara, M. Brenner, S. A.Ringel, and S. Rajan, Appl. Phys. Lett. 122, 113506 (2023).26S. Dhara, N. K. Kalarickal, A. Dheenan, S. I. Rahman, C. Joishi, and S. Rajan,Appl. Phys. Lett. 123, 023503 (2023).27T. Oshima and Y. Oshima, Appl. Phys. Express 15, 075503 (2022).28T. Oshima and Y. Oshima, Appl. Phys. Lett. 122, 162102 (2023).29T. Oshima and Y. Oshima, Appl. Phys. Express 16, 066501 (2023).30Y. Oshima and T. Oshima, Jpn. J. Appl. Phys., Part 1 62, 080901 (2023).31S. Mu, M. Wang, H. Peelaers, and C. G. Van de Walle, APL Mater. 8, 091105(2020).32R. Schewski, K. Lion, A. Fiedler, C. Wouters, A. Popp, S. V. Levchenko, T.Schulz, M. Schmidbauer, S. Bin Anooz, R. Gr€uneberg, Z. Galazka, G. Wagner,K. Irmscher, M. Scheffler, C. Draxl, and M. Albrecht, APL Mater. 7, 022515(2019).33O. Ueda, N. Ikenaga, K. Koshi, K. Iizuka, A. Kuramata, K. Hanada, T.Moribayashi, S. Yamakoshi, and M. Kasu, Jpn. J. Appl. Phys.,Part 1 55, 1202BD(2016).34E. Ohba, T. Kobayashi, M. Kado, and K. Hoshikawa, Jpn. J. Appl. Phys., Part 155, 1202BF (2016).35Y. Yao, Y. Tsusaka, K. Sasaki, A. Kuramata, Y. Sugawara, and Y. Ishikawa,Appl. Phys. Lett. 121, 012105 (2022).36M. Kasu, T. Oshima, K. Hanada, T. Moribayashi, A. Hashiguchi, T. Oishi, K.Koshi, K. Sasaki, A. Kuramata, and O. Ueda, Jpn. J. Appl. Phys., Part 1 56,091101 (2017).37Y. Yao, D. Wakimoto, H. Miyamoto, K. Sasaki, A. Kuramata, K. Hirano, Y.Sugawara, and Y. Ishikawa, Scr. Mater. 226, 115216 (2023).38A. Kuramata, K. Koshi, S. Watanabe, Y. Yamaoka, T. Masui, and S. Yamakoshi,Jpn. J. Appl. Phys, Part 1 55, 1202A2 (2016).39Y. Oshima and T. Oshima, Semicond. Sci. Technol. 38, 105003 (2023).40Y. Oshima, K. Kawara, T. Oshima, M. Okigawa, and T. Shinohe, Semicond.Sci. Technol. 35, 055022 (2020).41R. Schewski, M. Baldini, K. Irmscher, A. Fiedler, T. Markurt, B. Neuschulz, T.Remmele, T. Schulz, G. Wagner, Z. Galazka, and M. Albrecht, J. Appl. Phys.120, 225308 (2016).Applied Physics Letters ARTICLE pubs.aip.org/aip/aplAppl. Phys. Lett. 124, 042110 (2024); doi: 10.1063/5.0186319 124, 042110-7Published under an exclusive license by AIP Publishing 26 January 2024 00:02:07https://doi.org/10.1063/1.3674287https://doi.org/10.1063/1.3674287https://doi.org/10.1063/1.5142999https://doi.org/10.35848/1347-4065/ac4b6bhttps://doi.org/10.1063/5.0060327https://doi.org/10.1016/B978-0-12-814521-0.00012-9https://doi.org/10.1016/B978-0-12-814521-0.00012-9https://doi.org/10.1063/1.4967931https://doi.org/10.1063/1.5038105https://doi.org/10.1109/IEDM19573.2019.8993526https://doi.org/10.1109/IEDM19573.2019.8993526https://doi.org/10.1109/LED.2017.2779867https://doi.org/10.1109/LED.2017.2779867https://doi.org/10.1109/TED.2021.3067856https://doi.org/10.1109/TED.2021.3067856https://doi.org/10.35848/1882-0786/ac4080https://doi.org/10.1109/LED.2023.3235777https://doi.org/10.1063/5.0123284https://doi.org/10.1109/LED.2022.3222878https://doi.org/10.35848/1347-4065/acb2d7https://doi.org/10.1109/LED.2022.3218749https://doi.org/10.7567/1882-0786/ab08adhttps://doi.org/10.1016/j.apsusc.2019.144673https://doi.org/10.1021/acs.jpclett.2c02167https://doi.org/10.1116/1.4986300https://doi.org/10.1116/1.4986300https://doi.org/10.1021/acsnano.9b01709https://doi.org/10.1063/5.0057203https://doi.org/10.1063/5.0096490https://doi.org/10.1063/5.0131996https://doi.org/10.1063/5.0151808https://doi.org/10.35848/1882-0786/ac75c8https://doi.org/10.1063/5.0138736https://doi.org/10.35848/1882-0786/acdbb7https://doi.org/10.35848/1347-4065/acee3bhttps://doi.org/10.1063/5.0019915https://doi.org/10.1063/1.5054943https://doi.org/10.7567/JJAP.55.1202BDhttps://doi.org/10.7567/JJAP.55.1202BFhttps://doi.org/10.1063/5.0098942https://doi.org/10.7567/JJAP.56.091101https://doi.org/10.1016/j.scriptamat.2022.115216https://doi.org/10.7567/JJAP.55.1202A2https://doi.org/10.1088/1361-6641/acf241https://doi.org/10.1088/1361-6641/ab7843https://doi.org/10.1088/1361-6641/ab7843https://doi.org/10.1063/1.4971957pubs.aip.org/aip/apl