# Fileset

[J_Appl_Phys_139_075302_(2026).pdf](https://mdr.nims.go.jp/filesets/5fdc39cd-c402-413e-8817-2b81dcad110d/download)

## Creator

[Yuichi Oshima](https://orcid.org/0000-0001-8293-4891), [Takayoshi Oshima](https://orcid.org/0000-0001-8550-9735), [Shiyu Xiao](https://orcid.org/0009-0007-0382-0396), Kazuto Murakami, Katsuhiro Imai, Takahiro Tomita

## Rights

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

## Other metadata

[Halide vapor phase epitaxy of a thick                    <i>c</i>                    -plane α-Ga2O3 film on a high-quality α-Cr2O3/sapphire template](https://mdr.nims.go.jp/datasets/488e7e22-e71e-4213-8a37-842621ba3bd8)

## Fulltext

Halide vapor phase epitaxy of a thick c-plane α-Ga2O3 film on a high-quality α-Cr2O3/sapphire templateViewOnlineExportCitationRESEARCH ARTICLE |  FEBRUARY 17 2026Halide vapor phase epitaxy of a thick c-plane α-Ga2O3 film ona high-quality α-Cr2O3/sapphire templateYuichi Oshima   ; Takayoshi Oshima  ; Shiyu Xiao  ; Kazuto Murakami; Katsuhiro Imai; Takahiro TomitaJ. Appl. Phys. 139, 075302 (2026)https://doi.org/10.1063/5.0319104Articles You May Be Interested InEvaluation of valence band offset and its non-commutativity at all oxide α-Cr2O3/β-Ga2O3 heterojunctionfrom photoelectron spectroscopyJ. Appl. Phys. (November 2021)High mobility electron gas with quasi-two-dimensional characteristics at the interface of Cr2O3/SrTiO3heterostructuresJ. Appl. Phys. (July 2023)Microstructure and transport properties of epitaxial topological insulator Bi2Se3 thin films grown on MgO(100), Cr2O3 (0001), and Al2O3 (0001) templatesJ. Appl. Phys. (September 2015) 17 February 2026 23:11:15https://pubs.aip.org/aip/jap/article/139/7/075302/3380272/Halide-vapor-phase-epitaxy-of-a-thick-c-planehttps://pubs.aip.org/aip/jap/article/139/7/075302/3380272/Halide-vapor-phase-epitaxy-of-a-thick-c-plane?pdfCoverIconEvent=citejavascript:;https://orcid.org/0000-0001-8293-4891javascript:;https://orcid.org/0000-0001-8550-9735javascript:;https://orcid.org/0009-0007-0382-0396javascript:;javascript:;javascript:;https://crossmark.crossref.org/dialog/?doi=10.1063/5.0319104&domain=pdf&date_stamp=2026-02-17https://doi.org/10.1063/5.0319104https://pubs.aip.org/aip/jap/article/130/17/175303/1063751/Evaluation-of-valence-band-offset-and-its-nonhttps://pubs.aip.org/aip/jap/article/134/3/035303/2902990/High-mobility-electron-gas-with-quasi-twohttps://pubs.aip.org/aip/jap/article/118/12/125309/142000/Microstructure-and-transport-properties-ofhttps://servedbyadbutler.com/redirect.spark?MID=188841&plid=3503912&setID=1044475&channelID=0&CID=1697306&banID=524364881&PID=0&textadID=0&tc=1&rnd=2298746690&scheduleID=3691594&adSize=1640x440&data_keys=%7B%22%22%3A%22%22%7D&mt=1771369875844248&spr=1&referrer=http%3A%2F%2Fpubs.aip.org%2Faip%2Fjap%2Farticle-pdf%2Fdoi%2F10.1063%2F5.0319104%2F20907907%2F075302_1_5.0319104.pdf&request_uuid=da7a8d0e-221a-48cc-b38f-4509e3172a45&hc=69ac2eeca8599833664798fe7c6ac4ac6a72d55a&location=Halide vapor phase epitaxy of a thick c-planeα-Ga2O3 film on a high-quality α-Cr2O3/sapphiretemplateCite as: J. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104View Online Export Citation CrossMarkSubmitted: 22 December 2025 · Accepted: 28 January 2026 ·Published Online: 17 February 2026Yuichi Oshima,1,a) Takayoshi Oshima,1 Shiyu Xiao,2 Kazuto Murakami,2 Katsuhiro Imai,2and Takahiro Tomita2AFFILIATIONS1Research Center for Electronic and Optical Materials, National Institute for Materials Science, 1-1 Namiki,Tsukuba 305-0044, Japan2NGK Insulators, Ltd, 2-56 Suda-cho, Mizuho, Nagoya 467-8530, Japana)Author to whom correspondence should be addressed: OSHIMA.Yuichi@nims.go.jpABSTRACTα-Ga2O3 is a promising ultra-wide-bandgap semiconductor for future power devices, and the use of α-Cr2O3 buffer layers represents aneffective approach to improve the crystalline quality of heteroepitaxial α-Ga2O3 films owing to the small lattice mismatch between the twomaterials. In this study, c-plane α-Ga2O3 films were grown using halide vapor phase epitaxy (HVPE) on high-quality α-Cr2O3/sapphiretemplates, and the dependence of crystalline quality on the film thickness was systematically investigated. HVPE growth was performedunder atmospheric pressure at 520 °C using GaCl and O2 as the precursors and at a growth rate of 14 μmh−1. The film thickness was variedfrom 0.24 to 21 μm by controlling the growth time. X-ray 2θ–ω scan and pole figure measurements helped confirm that the α-Ga2O3 epi-layers were phase-pure single-crystalline films. Thickness-dependent x-ray rocking curve measurements and reciprocal space mappingrevealed that lattice relaxation began at a thickness of approximately 0.47 μm or less and virtually completed at thicknesses of 11 μm orgreater. Cross-sectional scanning transmission electron microscopy results showed that dislocations were observed predominantly near thefilm surface and were absent at the α-Ga2O3/α-Cr2O3 interface. Etch-pit density measurements yielded a low dislocation density of5.6 × 107 cm−2 for the fully strained 0.24 μm-thick film. The almost fully relaxed 21 μm-thick film showed a higher dislocation density of3.9 × 108 cm−2. Nevertheless, this value was approximately one order of magnitude lower than that of an α-Ga2O3 film directly grown on ac-plane sapphire substrate under identical conditions.© 2026 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license(https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/5.0319104I. INTRODUCTIONCorundum-structured α-Ga2O3 is an ultra-wide-bandgapsemiconductor exhibiting the highest bandgap energy amongGa2O3 polymorphs (Eg = 5.3 eV).1 Owing to the large bandgap, ahigh critical electric field is expected, which is advantageous forpower device applications. Moreover, α-Ga2O3 can form solidsolutions without compositional limitations with other corundum-structured oxides, such as α-Al2O3, providing a high degree offreedom for band engineering.2 Although it is difficult to impartp-type conduction to α-Ga2O3, isomorphic p-type oxides, such asα-(Ir,Ga)2O3, can be utilized to form hetero-pn junctions.3–5 Thesecharacteristics make α-Ga2O3 a promising material for future high-performance power devices. In fact, several promising prototypedevices have been reported, including Schottky barrier diodes(SBDs) with a very low on-resistance of 0.1 mΩ cm2,6 ampere-classSBDs with a breakdown voltage (VB) of 1.7 kV,7 andmetal-oxide-semiconductor field-effect transistors with a VB of2.3 kV.8α-Ga2O3 can be grown using various epitaxial growth tech-niques, including mist chemical vapor deposition (mist CVD),1,3halide vapor phase epitaxy (HVPE),9,10 molecular beamepitaxy,11–13 and metal-organic vapor phase epitaxy.14 To date, sap-phire substrates have been predominantly used for growingJournal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-1© Author(s) 2026 17 February 2026 23:11:15https://doi.org/10.1063/5.0319104https://doi.org/10.1063/5.0319104https://pubs.aip.org/action/showCitFormats?type=show&doi=10.1063/5.0319104http://crossmark.crossref.org/dialog/?doi=10.1063/5.0319104&domain=pdf&date_stamp=2026-02-17https://orcid.org/0000-0001-8293-4891https://orcid.org/0000-0001-8550-9735https://orcid.org/0009-0007-0382-0396mailto:OSHIMA.Yuichi@nims.go.jphttps://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1063/5.0319104https://pubs.aip.org/aip/japα-Ga2O3. This is because sapphire shares the same corundumcrystal structure as α-Ga2O3 and, high-quality, large-area sapphiresubstrates are commercially available at a reasonable cost.However, due to the large lattice mismatch between sapphireand α-Ga2O3 (Δa/a =−4.5% and Δc/c =−3.3%), an α-Ga2O3 epi-layer grown on sapphire substrates without any specific counter-measures exhibits an extremely high dislocation density. Forexample, regarding α-Ga2O3 grown on c-plane sapphire substrates,Kaneko et al. reported that the dislocation density of a film grownby mist CVD was approximately 7 × 1010 cm−2 in the vicinity ofthe interface (∼100 nm).15 In addition, Oshima et al. reported adislocation density of approximately 3 × 1010 cm−2 in a 4-μm-thickHVPE-grown film.16 Such a high dislocation density can lead to adeterioration in device performance. For example, carrier scatteringby dislocations may reduce carrier mobility.17 According to the the-oretical calculations made by Takane et al., carrier scattering at acarrier concentration of 1 × 1016 cm−3 should be negligible whenthe dislocation density is 1 × 108 cm−2 or lower.17Several techniques have been reported to reduce the disloca-tion density in α-Ga2O3 epilayers, such as epitaxial lateral over-growth16,18,19 and low-temperature high-growth-rate techniques.20By employing these approaches, dislocation densities below1 × 108 cm−2 can be achieved.19,20 However, they requirelithography-based mask formation and/or relatively thick filmgrowth, posing challenges in terms of manufacturing cost.In recent years, a corundum-structured α-Cr2O3 buffer layerhas attracted considerable attention as a novel approach for reduc-ing the dislocation density in α-Ga2O3 epilayers. This is becausethe lattice mismatch between α-Cr2O3 and α-Ga2O3 (Δa/a =−0.45%and Δc/c = 1.23%) is significantly lower than that between sapphireand α-Ga2O3. Several studies have reported the epitaxial growth ofα-Ga2O3 on a sapphire substrate using an α-Cr2O3 film as the bufferlayer.21–23Stepanov et al. grew α-Ga2O3 using HVPE on a c-plane sapphiresubstrate on which a 150-nm-thick c-plane α-Cr2O3 layer had beendeposited by radio frequency magnetron sputtering. The full widthsat half maximum (FWHMs) of the x-ray rocking curves (XRCs) forthe 0006 (symmetric) and 10�18 (skew-symmetric) diffractions wereboth approximately 1400 arcsec. In contrast, for α-Ga2O3 growndirectly on sapphire, the corresponding values were 60 and2400 arcsec, respectively. From these XRC-FWHM values, the dislo-cation density was estimated to decrease from 2 × 1010 to5 × 109 cm−2, owing to the introduction of the α-Cr2O3 buffer layer.21However, this dislocation density is still high, which is likelyattributable to the insufficient crystalline quality of the α-Cr2O3buffer layer. The small lattice mismatch between α-Ga2O3 andα-Cr2O3 implies that α-Ga2O3 largely inherits its crystalline qualityfrom the underlying α-Cr2O3 layer. Therefore, the α-Cr2O3 bufferlayer must itself have a high crystalline quality to obtain high-quality α-Ga2O3 epilayers.Recently, a template comprising a high-quality, approximately100-μm-thick c-plane α-Cr2O3 layer formed on a c-plane sapphiresubstrate has been realized by a group at NGK Insulators, Ltd.Although the specific methods used to grow the α-Cr2O3 layerhave not been disclosed, the FWHMs of the XRCs for the 0006(symmetric) and 10�14 (skew-symmetric) diffractions were reportedto be approximately 100 and 120 arcsec, respectively.22Several studies have reported the growth of c-plane α-Ga2O3 onsuch high-quality α-Cr2O3 templates.22,23 Yamada et al. deposited anapproximately 0.7-μm-thick α-Ga2O3 layer using mist CVD; the dislo-cation density, estimated from the XRC-FWHM values, was4 × 107 cm−2, which is approximately three orders of magnitude lowerthan that of α-Ga2O3 grown directly on sapphire substrates.22 Takedaet al. reported the growth of a phase-pure α-Ga2O3 film using HVPE,for which the XRC-FWHMs of the 0006 and 10�14 diffractions were250 and 290 arcsec, respectively (the film thickness was not speci-fied).23 Based on these XRC-FWHM values, the dislocation density ofthis film was considered to be of the order of 108 cm−2.Clearly, the use of high-quality α-Cr2O3 templates appears tobe effective in improving the crystalline quality of α-Ga2O3.However, this effectiveness has, thus, far been demonstrated pri-marily for thin films with thicknesses of less than approximately1 μm. From the viewpoint of applications to α-Ga2O3-based verticalpower devices, it is essential to verify the effectiveness of such tem-plates for thick films suitable for use as drift layers. Therefore, inthis study, c-plane α-Ga2O3 layers with thicknesses of up to 21 μmwere grown at a high-growth rate using HVPE on high-qualityc-plane α-Cr2O3 templates, and the thickness dependence of theircrystalline quality was systematically investigated.II. EXPERIMENTALα-Ga2O3 was grown using a laboratory-made atmospheric-pressure HVPE reactor at a temperature of 520 °C. O2 (>99.999 95%purity) and GaCl were used as precursors. GaCl was synthesized insitu upstream in the reactor by the reaction between Ga metal(>99.999 99% purity) and HCl gas (>99.999% purity) at a temperatureof 570 °C. The partial pressures of O2 and GaCl were set to 1.25 kPaand 125 Pa, respectively. In addition to these precursors, HCl gas wasdirectly supplied to the growth zone at a partial pressure of 188 Pa tosuppress parasitic reactions.10 N2 (dew point <−110 °C) was used asthe carrier gas. Under these growth conditions, the growth rate wasapproximately 14 μmh−1. By varying the growth time, α-Ga2O3 layerswere grown with thicknesses (t) ranging from 0.24 to 21 μm.We used high-quality c-plane α-Cr2O3/sapphire templates(manufactured by NGK Insulators, Ltd) for the HVPE growth. Thethickness of the α-Cr2O3 layer was approximately 100 μm. TheXRC-FWHMs for the 0006 and 10�14 diffractions of the α-Cr2O3layer were typically less than approximately 100 and 200 arcsec,respectively. The templates were cleaned using hydrofluoric acid and asulfuric acid–hydrogen peroxide mixture prior to the HVPE growth.The phase purity and crystallographic orientation of thegrown epilayers were examined by x-ray 2θ–ω scan and pole figuremeasurements, respectively. The crystalline quality was evaluatedbased on the XRC-FWHMs for the 0006 (symmetric) and 10�14(skew-symmetric) diffractions, denoted by β(0006) and β(10�14),respectively. The parameter β(0006) ; βtilt is referred to as the tiltangle and represents the tilt spread of the c-plane. The twist angleβtwist, which characterizes the twist spread about the c axis, wascalculated using the following relationship:24,25β(10�14)2 ¼ (βtiltcos χ)2 þ (βtwistsin χ)2, (1)where χ is the inclination angle of the diffracting plane (10�14).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-2© Author(s) 2026 17 February 2026 23:11:15https://pubs.aip.org/aip/japBoth the templates and grown epiwafers exhibited convexwafer bowing, with relatively large curvatures ranging from −0.3 to−1.8 m−1, as shown in Fig. S1 in the supplementary material. Tominimize the influence of wafer curvature on the XRC-FWHMvalues, an optimized optical configuration was employed, andcurvature-induced broadening was corrected based on the mea-sured curvature of each sample. The values of β(hkml) used in thisstudy correspond to those obtained after curvature correction.Detailed descriptions of the optical setup and the curvature correc-tion procedure are provided in Appendix A.The strain state of the α-Ga2O3 films was evaluated by x-rayreciprocal space mapping (RSM) of the �20210 diffractions. For the�20210 diffractions of α-Ga2O3, the x-ray incidence angle to theepilayer surface was 74.399°, which corresponds to the sum of theBragg angle (42.492°) and the angle between the (0001) and (�1015)planes (31.907°).The dislocation behavior was investigated by cross-sectionalscanning transmission electron microscopy (STEM) observationsfrom the m axis direction. Dislocation types were identified usingthe g⋅b = 0 criterion under two-beam dark-field conditions withg = 0006 and 11�20. The dislocation density was estimated eitherfrom the XRC-FWHM values26 or by measuring the etch-pitdensity (EPD) using HCl gas etching.27 See Appendix B for thespecific procedure for EPD estimation.III. RESULTS AND DISCUSSIONFigure 1(a) shows the x-ray 2θ–ω scan profile of a2.4-μm-thick α-Ga2O3 epilayer measured in the out-of-plane con-figuration; Fig. 1(b) presents an enlarged view of the profile. Onlydiffraction peaks originating from the (0001) planes of α-Ga2O3and α-Cr2O3 can be observed. It should be noted that the theoreti-cal transmittance of a 100-μm-thick α-Cr2O3 film (linear absorp-tion coefficient of 8.989 × 102 cm−1 for Cu Kα1 radiation)28 is onlyabout 3.3 ppm, even when x rays are incident at the Bragg angle ofsapphire 00012 diffraction (θB = 45.36°). Therefore, the intensity ofx rays that reach the sapphire substrate is diffracted, and subse-quently transmits through the α-Cr2O3 film again to be detectedshould be virtually zero. This explains why no diffraction peaksfrom the sapphire substrate are observed in the 2θ–ω scan profilein Fig. 1. Figures 2(a) and 2(b) show the XRD pole figures for the10�14 diffractions of α-Ga2O3 and α-Cr2O3, respectively. In bothcases, diffraction peaks were observed only at the positionsexpected for single crystals with a corundum structure. Theseresults confirm that the epilayer grown was a phase-pure, single-crystalline α-Ga2O3 film.Figures 3(a) and 3(b) show the thickness dependences of βtiltand βtwist, respectively. The βtilt values of the α-Ga2O3 epilayerswere close to those of the underlying α-Cr2O3 layer and did notsignificantly increase with increasing film thickness. In contrast,βtwist remained comparable to that of the α-Cr2O3 layer up to athickness of t = 2.4 μm but began to increase markedly at approxi-mately t≈ 5.9 μm, followed by a substantial increase at larger thick-nesses. The slight reduction in βtwist observed at t = 21 μmcompared with that at t = 11 μm may be attributed to dislocationannihilation occurring during thick-film growth.20,29 It should benoted that a pronounced increase in βtwist of α-Cr2O3 was observedfor the sample with t = 21 μm. Although the origin of this behavioris not clear at present, this sample exhibited a much larger curva-ture than the other samples, as shown in Fig. S1 in thesupplementary material. Therefore, we speculate that a large stresswas applied to this sample, which may have resulted in the genera-tion of dislocations in the α-Cr2O3 layer. However, this possibilityhas not been verified at this stage, and further investigation isrequired.Figures 4(a)–4(f ) show the RSM measurement results. Att = 0.24 μm, the in-plane lattice constant of α-Ga2O3 coincides withthat of α-Cr2O3, indicating that the α-Ga2O3 epilayer was fullystrained. At t = 0.47 μm, the α-Ga2O3 peak remains located at thefully strained position; however, its tail appears to extend towardFIG. 1. (a) X-ray 2θ–ω scan profile of an α-Ga2O3 (t = 2.4 μm)/α-Cr2O3/sapphire epiwafer. (b) Enlarged view of (a).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-3© Author(s) 2026 17 February 2026 23:11:15https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://pubs.aip.org/aip/japthe fully relaxed position, suggesting the onset of lattice relaxation.At t = 2.4 and 5.9 μm, the α-Ga2O3 peak is still centered at the fullystrained position, while the changes in the peak tail become morepronounced. For t≥ 11 μm, the α-Ga2O3 peak is almost completelyrelaxed. The pronounced relaxation observed for t≥ 11 μm is con-sistent with the thickness dependence of βtwist.Based on our RSM measurements, lattice relaxation wasobserved when t≥ 0.47 μm. Therefore, the critical thickness islikely to lie between 0.24 and 0.47 μm. The theoretical criticalthickness of c-plane α-Ga2O3 grown on bulk α-Cr2O3 is estimatedto be 0.36 μm based on the model proposed by People and Bean.30This value is in good agreement with the experimental results. Inthis calculation, lattice relaxation was assumed to occur via theintroduction of edge dislocations with the Burgers vectors beingequal to the a axis lattice constant. Table I presents the parametersused in the calculation.Figure 5(a) shows a cross-sectional bright-field STEM imageof the sample with t = 5.9 μm, which is considered to be partiallyrelaxed. No dislocations were observed at the α-Ga2O3/α-Cr2O3interface, whereas dislocations were clearly observed near thesurface of the α-Ga2O3 layer. Typically, lattice relaxation in alattice-mismatched heteroepitaxial system proceeds via theFIG. 2. X-ray pole figures of the α-Ga2O3 (t = 2.4 μm)/α-Cr2O3/sapphire epiwafer measuring the 10�14 diffractions of (a) α-Ga2O3 and (b) α-Cr2O3.FIG. 3. (a) Tilt and (b) twist angles of the α-Ga2O3 film and α-Cr2O3 template layer estimated based on the XRC measurements as a function of the α-Ga2O3 filmthickness.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-4© Author(s) 2026 17 February 2026 23:11:15https://pubs.aip.org/aip/japnucleation of dislocation half-loops on the surface, which expandby gliding toward the interface and eventually form threading dis-locations and misfit dislocations therein. In Fig. 5(a), the fact thatdislocations are observed only near the surface and did not reachthe interface appears to be consistent with the RSM results, indicat-ing that this sample was at a relatively early stage of lattice relaxa-tion. Figures 5(b)–5(d) show two-beam dark-field TEM images ofthe dislocations observed in Fig. 5(a). Since the contrast of most ofthe dislocations disappeared under the g = 0006 condition, themajority of dislocations were inferred to be edge-type. This obser-vation is consistent with the results shown in Fig. 3, where theincrease in crystallographic orientation fluctuations with increasingfilm thickness is dominated by the twist component. It should benoted that the periodic fringe patterns and dark line-like contrastsobserved in Figs. 5(c) and 5(d) are attributed to equal-thicknessfringes and bend contours, respectively, and are not associated withcrystallographic defects.Figure 6 shows the thickness dependence of the dislocationdensity estimated either from the XRC measurements or from theFIG. 4. Reciprocal space mapping (RSM) of α-Ga2O3/α-Cr2O3/sapphire epiwafers with various α-Ga2O3 film thicknesses. Fully relaxed positions for α-Ga2O3 andα-Cr2O3 are indicated by × marks.TABLE I. Parameters used in calculating the critical thickness of c-plane α-Ga2O3grown on bulk α-Cr2O3.Parameter Value Sourcea axis length α-Ga2O3 0.498 26 nm PDF card No: 01-074-1610α-Cr2O3 0.496 00 nm PDF card No. 01-070-3765Poisson’s ratio 0.729 First-principles calculation31Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-5© Author(s) 2026 17 February 2026 23:11:15https://pubs.aip.org/aip/japEPD measurements. For comparison, the figure also shows the EPDof an α-Ga2O3 film with t = 21 μm grown directly on a c-plane sap-phire substrate under the same HVPE conditions. The edge and screwcomponents of the dislocation density, De and Ds, estimated from theXRC results were calculated using Eqs. (2) and (3), respectively,De ¼ β2twist(2π ln 2)b2e, (2)Ds ¼ β2tilt(2π ln 2)b2s, (3)where, be and bs denote the Burgers vectors of the edge and screwdislocations, respectively, for which be = 0.498 25 nm (the a axislattice constant) and bs = 1.343 31 nm (the c axis lattice constant)were used.26 The dislocation densities estimated from the XRCmeasurements shown in Fig. 6 represent the sum of the edge andscrew components.The dislocation density determined using the EPD methodincreased with increasing film thickness, similar to the trendobserved in the XRC-based results. However, in the thickness rangeof t≤ 5.9 μm, the EPD did not fully coincide with the lattice relaxa-tion behavior inferred from the XRC and RSM results. This dis-crepancy is likely attributable to the nonuniform distribution ofetch pits, as revealed by the Nomarski microscopy image in Fig. S3in the supplementary material, which presumably reflects spatialinhomogeneity in the crystalline quality of the α-Cr2O3 template. Itshould be noted that the dislocation density distribution in theα-Cr2O3 layer was not investigated in this study, because the dislo-cation density was too low to evaluate the distribution by TEMFIG. 5. (a) Bright-field cross-sectional STEM image of the α-Ga2O3 (t = 5.9 μm)/α-Cr2O3/sapphire epiwafer. (b)–(d) Dark-field TEM images of dislocations observed undertwo-beam conditions. Dislocations considered to have edge and screw components are labeled e and s, respectively.FIG. 6. Dislocation density of α-Ga2O3 epilayers estimated based on XRC mea-surements or by etch-pit counting as a function of the film thickness.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-6© Author(s) 2026 17 February 2026 23:11:15https://doi.org/10.60893/figshare.jap.c.8266762https://pubs.aip.org/aip/japobservations. To clarify the dislocation distribution in the α-Cr2O3layer in the future, it will be necessary to establish an etch-pit-basedmethod. Because the measurement area of the EPD (approximately5 × 13–10 × 26 μm2) was significantly lower than that probed by theXRC, the EPD was more susceptible to local variations in the dislo-cation density. Regardless of the estimation method, the dislocationdensity increased markedly for film thicknesses exceeding 5.9 μm.Nevertheless, even at large thicknesses, the dislocation densityremained approximately one order of magnitude lower than that ofan α-Ga2O3 film grown directly on a sapphire substrate.IV. CONCLUSIONSα-Ga2O3 epilayers with thicknesses of up to 21 μm weregrown using HVPE on high-quality c-plane α-Cr2O3/sapphire tem-plates, and the thickness dependence of their crystalline propertieswas systematically investigated. Although the HVPE growth wasperformed at a high-growth rate of 14 μmh−1, phase-pure, single-crystalline α-Ga2O3 films were successfully obtained, as confirmedby x-ray 2θ–ω scans and pole figure measurements. The tilt andtwist spreads of the α-Ga2O3 epilayers were evaluated based onXRC measurements of the 0006 and 10�14 diffractions. The tiltspread did not significantly increase with increasing film thickness.In contrast, the twist spread began to increase at approximatelyt = 5.9 μm and showed a pronounced increase for t≥ 11 μm. Thisrelaxation behavior was also consistently observed in the RSMresults. The dislocation density estimated by etch-pit counting wasapproximately 5.6 × 107 cm−2 for the thinnest, fully strained sample(t = 0.24 μm), which was comparable to that of the template layer.In contrast, the dislocation density increased to 3.9 × 108 cm−2 forthe virtually fully relaxed sample with t = 21 μm. Nevertheless, thisvalue remained substantially lower than that of the α-Ga2O3 filmgrown directly on a sapphire substrate under the same conditions(3 × 109 cm−2). To suppress carrier scattering by dislocations,further reduction in the dislocation density would be required.From this perspective, continued development of template layerswith even smaller lattice mismatches is highly desirable.SUPPLEMENTARY MATERIALSee the supplementary material for the epiwafer curvature,measured XRC-FWHMs before and after the curvature correction,and optical microscopy image exhibiting the inhomogeneousetch-pit distribution.AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsYuichi Oshima: Conceptualization (lead); Data curation (lead);Formal analysis (lead); Investigation (lead); Methodology (lead);Resources (lead); Visualization (lead); Writing – original draft(lead); Writing – review & editing (equal). Takayoshi Oshima:Conceptualization (supporting); Data curation (supporting);Formal analysis (supporting); Investigation (supporting);Methodology (supporting); Visualization (supporting); Writing –review & editing (equal). Shiyu Xiao: Conceptualization (support-ing); Data curation (supporting); Formal analysis (supporting);Investigation (supporting); Methodology (supporting); Resources(equal); Visualization (supporting); Writing – review & editing(equal). Kazuto Murakami: Conceptualization (supporting); Datacuration (supporting); Formal analysis (supporting); Investigation(supporting); Methodology (supporting); Resources (equal);Visualization (supporting); Writing – review & editing (equal).Katsuhiro Imai: Conceptualization (supporting); Data curation(supporting); Formal analysis (supporting); Investigation (support-ing); Methodology (supporting); Resources (equal); Visualization(supporting); Writing – review & editing (equal). TakahiroTomita: Conceptualization (supporting); Data curation (support-ing); Formal analysis (supporting); Investigation (supporting);Methodology (supporting); Resources (equal); Visualization(supporting); Writing – review & editing (equal).DATA AVAILABILITYThe data that support the findings of this study are availablewithin the article and its supplementary material.APPENDIX A: OPTICAL SETUP AND CURVATURECORRECTION PROCEDURE FOR XRC MEASUREMENTSThe α-Cr2O3/sapphire templates used in this study exhibited acurvature of approximately −0.3 m−1 prior to the epitaxial growthof α-Ga2O3, most likely due to the thermal stress arising from thedifference in thermal expansion coefficients. Upon epitaxial growthof α-Ga2O3, the curvature further increased with increasing filmthickness, reaching a value of −1.8 m−1 for the sample with a thick-ness of 21 μm. For samples with such a large curvature, the contri-bution of wafer bowing to the broadening of the XRC-FWHM mayno longer be negligible.To minimize the influence of epiwafer bowing on theXRC-FWHM values, the thickness of the incident x-ray beam wasreduced to h = 0.096 mm using an incident slit. Under this condi-tion, the length of the beam footprint in the direction perpendicu-lar to the rocking axis during the 0006 XRC measurement,FP(0006), was 0.28 mm, as given by FP(0006) = h/sinθ0006.A parallel-plate collimator (PPC) with a receiving angle of0.09° was placed in front of the x-ray detector. A slit was furtherinstalled to allow the x-ray beam to pass through only one of themultiple gaps of the PPC, with a gap width of 0.130 mm.Consequently, the effective beam footprint length on the sample inthe direction perpendicular to the rocking axis during the 10�14XRC measurement, FP(10�14), was limited to 0.45 mm, which canbe expressed as FP(10�14) ¼ gap/sinθ10�14.The peak broadening arising solely from wafer bowing,βr(hkml), can be expressed as βr(hkml) = FP(hkml)/R(hkml), whereR(hkml) is the radius of curvature for the (hkml) plane.24,25 Thecurvature radius R(hkml) was determined from the peak shifts ofthe 0006 and 10�14 XRCs measured while translating the samplealong the x axis.32 Figure S1 in the supplementary material showsthe relationship between the measured curvature, R(hkml)−1, andthe film thickness. Finally, the contribution of wafer bowing wasremoved using the following relationship to extract the intrinsicJournal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-7© Author(s) 2026 17 February 2026 23:11:15https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://pubs.aip.org/aip/japXRC--FWHM, β(hkml):24,25β(hkml)2 ¼ βm(hkml)2 � βr(hkml)2, (A1)where βm(hkml) is the experimentally measured XRC-FWHM forthe 0006 or 10�14 diffractions. Figures S2(a) and S2(b) in thesupplementary material show the thickness dependences ofβm(hkml); Figs. S2(c) and S2(d) in the supplementary materialpresent the corresponding thickness dependences of the curvature-corrected values β(hkml).APPENDIX B: PROCEDURE FOR EPD ESTIMATIONAs shown in Fig. S3 in the supplementary material, the EPDof the α-Ga2O3 epilayer was not necessarily uniform. EPD valueswere particularly high in the vicinity of regions corresponding tomacroscopic defects in the α-Cr2O3 layer, which were originallypresent at a density of approximately 103 cm−2 and are indicatedby the yellow circle in Fig. S3 in the supplementary material.Therefore, EPD in this study was estimated in regions located as faras possible from such macroscopic defects. In these regions, SEMobservations were carried out at four randomly selected locations,and the average EPD was calculated. The observation magnificationwas chosen according to the dislocation density, such that thenumber of etch pits in each field of view was approximately morethan ∼100; magnifications of 5k× (field of view: 10 × 26 μm2) and10k× (field of view: 5 × 13 μm2) were used as appropriate.REFERENCES1D. Shinohara and S. Fujita, Jpn. J. Appl. Phys. 47, 7311 (2008).2H. Ito, K. Kaneko, and S. Fujita, Jpn. J. Appl. Phys. 51, 100207 (2012).3K. Kaneko, S. Fujita, and T. Hitora, Jpn. J. Appl. Phys. 57, 117103 (2018).4S. Kan, S. Takemoto, K. Kaneko, I. Takahashi, M. Sugimoto, T. Shinohe, andS. Fujita, Appl. Phys. Lett. 113, 212104 (2018).5K. Kaneko, Y. Masuda, S. Kan, I. Takahashi, Y. Kato, T. Shinohe, and S. Fujita,Appl. Phys. Lett. 118, 102104 (2021).6M. Oda, R. Tokuda, H. Kambara, T. Tanikawa, T. Sasaki, and T. Hitora,Appl. Phys. Express 9, 021101 (2016).7T. Shinohe, in 2024 IEEE International Meeting for Future Electron Devices,Kansai (IEEE, 2024), pp. 1–4.8Y. J. Jeong, J. H. Park, M. J. Yeom, I. Kang, J. Y. Yang, H. Y. Kim, D. W. Jeon,and G. Yoo, Appl. Phys. Express 15, 074001 (2022).9Y. Oshima, E. G. Víllora, and K. Shimamura, Appl. Phys. Express 8, 055501(2015).10Y. Oshima, K. Kawara, T. Oshima, M. Okigawa, and T. Shinohe, Semicond.Sci. Technol. 35, 055022 (2020).11Z. Cheng, M. Hanke, P. Vogt, O. Bierwagen, and A. Trampert, Appl. Phys.Lett. 111, 162104 (2017).12R. Jinno, C. S. Chang, T. Onuma, Y. Cho, S.-T. Ho, D. Rowe, M. C. Cao,K. Lee, V. Protasenko, D. G. Schlom, D. A. Muller, H. G. Xing, and D. Jena,Sci. Adv. 7, eabd5891 (2021).13J. Steele, J. Chen, T. Burrell, N. A. Pieczulewski, D. Bhattacharya, K. Smith,K. Gann, M. O. Thompson, H. G. Xing, D. Jena, D. A. Muller, M. D. Williams,M. K. I. Senevirathna, and D. G. Schlom, APL Mater. 13, 101117 (2025).14A. F. M. A. U. Bhuiyan, Z. Feng, H.-L. Huang, L. Meng, J. Hwang, andH. Zhao, APL Mater. 9, 101109 (2021).15K. Kaneko, H. Kawanowa, H. Ito, and S. Fujita, Jpn. J. Appl. Phys. 51, 020201(2012).16Y. Oshima, K. Kawara, T. Shinohe, T. Hitora, M. Kasu, and S. Fujita,APL Mater. 7, 022503 (2019).17H. Takane, H. Izumi, H. Hojo, T. Wakamatsu, K. Tanaka, and K. Kaneko,J. Mater. Res. 38, 2645 (2023).18Y. Oshima, K. Kawara, T. Oshima, M. Okigawa, and T. Shinohe, Jpn. J. Appl.Phys. 59, 025512 (2020).19Y. Oshima and T. Shinohe, Appl. Phys. Lett. 126, 202104 (2025).20Y. Oshima, H. Ando, and T. Shinohe, Appl. Phys. Express 16, 065501 (2023).21S. I. Stepanov, V. I. Nikolaev, A. V. Almaev, A. I. Pechnikov, M. P. Scheglov,A. V. Chikiryaka, B. O. Kushnarev, and A. Y. Polyakov, Mater. Phys. Mech. 47,577 (2021).22K. Yamada, R. Iida, S. Xiao, K. Murakami, T. Onuma, T. Honda,M. Watanabe, and T. Yamaguchi, “Mist CVD growth and characterization ofα-Ga2O3 films on α-Cr2O3 templates,” in 72nd JSAP Spring Meeting, Noda,Japan (MDPI, 2025), p. 16p-Y1311-7.23K. Takeda, Y. Morishita, S. Xiao, K. Murakami, M. Watanabe, T. Tomita, andR. Miyagawa, “Epitaxial growth of Ga2O3 on Cr2O3 template by HVPE,” in 44thElectronic Materials Symposium (EMS-44), Nara, Japan (EMS, 2025), p. Fr1-12.24J. Q. Liu, J. F. Wang, Y. X. Qiu, X. Guo, K. Huang, Y. M. Zhang, X. J. Hu,Y. Xu, K. Xu, X. H. Huang, and H. Yang, Semicond. Sci. Technol. 24, 125007(2009).25M. A. Moram and M. E. Vickers, Rep. Prog. Phys. 72, 036502 (2009).26T. C. Ma, X. H. Chen, Y. Kuang, L. Li, J. Li, F. Kremer, F.-F. Ren, S. L. Gu,R. Zhang, Y. D. Zheng, H. H. Tan, C. Jagadish, and J. D. Ye, Appl. Phys. Lett.115, 182101 (2019).27Y. Oshima, S. Yagyu, and T. Shinohe, J. Cryst. Growth 576, 126387 (2021).28M. J. Berger, J. H. Hubbell, S. M. Seltzer, J. Chang, J. S. Coursey, R. Sukumar,D. S. Zucker, and K. Olsen, see https://www.nist.gov/pml/xcom-photon-cross-sections-database for “XCOM: Photon Cross SectionsDatabase version 1.5” (2010).29K. Fujito, S. Kubo, H. Nagaoka, T. Mochizuki, H. Namita, and S. Nagao,J. Cryst. Growth 311, 3011 (2009).30R. People and J. C. Bean, Appl. Phys. Lett. 47, 322 (1985).31T. Kawamura and T. Akiyama, Jpn. J. Appl. Phys. 61, 021005 (2022).32See https://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=english for “Rigakuapplication note B-XRD2031.”Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 075302 (2026); doi: 10.1063/5.0319104 139, 075302-8© Author(s) 2026 17 February 2026 23:11:15https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.60893/figshare.jap.c.8266762https://doi.org/10.1143/JJAP.47.7311https://doi.org/10.1143/JJAP.51.100207https://doi.org/10.7567/JJAP.57.117103https://doi.org/10.1063/1.5054054https://doi.org/10.1063/5.0027297https://doi.org/10.7567/APEX.9.021101https://doi.org/10.35848/1882-0786/ac7431https://doi.org/10.7567/APEX.8.055501https://doi.org/10.1088/1361-6641/ab7843https://doi.org/10.1088/1361-6641/ab7843https://doi.org/10.1063/1.4998804https://doi.org/10.1063/1.4998804https://doi.org/10.1126/sciadv.abd5891https://doi.org/10.1063/5.0299750https://doi.org/10.1063/5.0065087https://doi.org/10.1143/JJAP.51.020201https://doi.org/10.1063/1.5051058https://doi.org/10.1557/s43578-023-01015-8https://doi.org/10.35848/1347-4065/ab6fafhttps://doi.org/10.35848/1347-4065/ab6fafhttps://doi.org/10.1063/5.0269810https://doi.org/10.35848/1882-0786/acddcahttps://doi.org/10.1088/0268-1242/24/12/125007https://doi.org/10.1088/0034-4885/72/3/036502https://doi.org/10.1063/1.5120554https://doi.org/10.1016/j.jcrysgro.2021.126387https://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://www.nist.gov/pml/xcom-photon-cross-sections-databasehttps://doi.org/10.1016/j.jcrysgro.2009.01.046https://doi.org/10.1063/1.96206https://doi.org/10.35848/1347-4065/ac468fhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://rigaku.com/products/x-ray-diffraction-and-scattering/xrd/application-notes/b-xrd2031-curvature-single-crystal-substrate?setLang=englishhttps://pubs.aip.org/aip/jap