# Fileset

[024303_1_5.0303598.pdf](https://mdr.nims.go.jp/filesets/69992250-1f69-46e8-8472-69d31a23243b/download)

## Creator

[Akihiro Ohtake](https://orcid.org/0000-0002-3519-4613), [Jun Nara](https://orcid.org/0000-0002-0486-2981), [Yoshiki Sakuma](https://orcid.org/0000-0001-6804-7217)

## Rights

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

## Other metadata

[Epitaxial configuration of unidirectionally aligned MoS2 monolayer on sapphire](https://mdr.nims.go.jp/datasets/c7a13968-c1ba-498d-a3d0-9a87b9843d82)

## Fulltext

Epitaxial configuration of unidirectionally aligned MoS2 monolayer on sapphireViewOnlineExportCitationRESEARCH ARTICLE |  JANUARY 12 2026Epitaxial configuration of unidirectionally aligned MoS2monolayer on sapphireAkihiro Ohtake   ; Jun Nara  ; Yoshiki Sakuma J. Appl. Phys. 139, 024303 (2026)https://doi.org/10.1063/5.0303598Articles You May Be Interested InDiffraction studies of WS2 crystallographic ordering during laser MBE growth on Al2O3(0001)AIP Advances (March 2025)Determination of helicities in unidirectional assemblies of graphitic or graphiticlike tubular structuresAppl. Phys. Lett. (October 2008)Unidirectional self-assembling of SiGe Stranski-Krastanow islands on Si(113)Appl. Phys. Lett. (May 2005) 14 January 2026 01:48:17https://pubs.aip.org/aip/jap/article/139/2/024303/3377230/Epitaxial-configuration-of-unidirectionallyhttps://pubs.aip.org/aip/jap/article/139/2/024303/3377230/Epitaxial-configuration-of-unidirectionally?pdfCoverIconEvent=citejavascript:;https://orcid.org/0000-0002-3519-4613javascript:;https://orcid.org/0000-0002-0486-2981javascript:;https://orcid.org/0000-0001-6804-7217https://crossmark.crossref.org/dialog/?doi=10.1063/5.0303598&domain=pdf&date_stamp=2026-01-12https://doi.org/10.1063/5.0303598https://pubs.aip.org/aip/adv/article/15/3/035057/3341106/Diffraction-studies-of-WS2-crystallographichttps://pubs.aip.org/aip/apl/article/93/14/141903/928463/Determination-of-helicities-in-unidirectionalhttps://pubs.aip.org/aip/apl/article/86/22/223109/128189/Unidirectional-self-assembling-of-SiGe-Stranskihttps://servedbyadbutler.com/redirect.spark?MID=188841&plid=3318315&setID=1044475&channelID=0&CID=1578722&banID=524059805&PID=0&textadID=0&tc=1&rnd=2882221249&scheduleID=3474289&adSize=1640x440&data_keys=%7B%22%22%3A%22%22%7D&metadata=%5B%5D&mt=1768355297082655&spr=1&referrer=http%3A%2F%2Fpubs.aip.org%2Faip%2Fjap%2Farticle-pdf%2Fdoi%2F10.1063%2F5.0303598%2F20868602%2F024303_1_5.0303598.pdf&request_uuid=8284ad71-05dd-4d07-8797-2ee9e9ededc6&hc=689078011207ae7a47cd87a9f6144f2980021d98&location=Epitaxial configuration of unidirectionally alignedMoS2 monolayer on sapphireCite as: J. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598View Online Export Citation CrossMarkSubmitted: 23 September 2025 · Accepted: 17 December 2025 ·Published Online: 12 January 2026Akihiro Ohtake,a) Jun Nara, and Yoshiki SakumaAFFILIATIONSNational Institute for Materials Science (NIMS), Tsukuba 305-0044, Japana)Author to whom correspondence should be addressed: OHTAKE.Akihiro@nims.go.jpABSTRACTHighly oriented MoS2 monolayer (ML) on the sapphire(0001) substrate was grown by metal-organic chemical vapor deposition. Theepitaxial configuration of ML-MoS2/sapphire has been studied using low-energy electron diffraction (LEED) combined with first-principlescalculations. LEED analysis based on the dynamical diffraction theory revealed that the MoS2 ML is grown with the epitaxial relationship ofMoS2 [1120] // Al2O3 [1120] and MoS2 [1100] // Al2O3 [1100] and that the formation of antiparallel (inversion) domains is effectivelysuppressed. The observed epitaxial relationship is insensitive to the direction of surface steps on the sapphire substrate.© 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/10.1063/5.0303598I. INTRODUCTIONMonolayer (ML) transition-metal dichalcogenides (TMDCs)have attracted considerable research interest, because of theirunique electrical and optical properties. In particular, ML-MoS2has been considered one of the most promising 2D materials forsubstantial potential applications in electronic and optoelectronicdevices. The synthesis of ML-MoS2 on sapphire, scalable andindustry-compatible substrates, is crucial from an industrial per-spective and is suitable for a wide range of advanced electronic andoptoelectronic applications, particularly those requiring transpar-ency and excellent electrical insulation.Numerous studies have reported the successful epitaxialgrowth of wafer-scale MoS2 ML on the sapphire substrate using avariety of growth methods such as powder-source chemical vapordeposition (CVD)1–7 and metal-organic CVD (MOCVD).8–11However, the epitaxial ML-MoS2 film on sapphire substrates oftencontains both parallel (0�, 120�, or 240� domain) [Fig. 1(a)] andantiparallel (60�, 180�, or 300� domain) [Fig. 1(b)] domains.1,2,6,8The coexistence of inverted domains potentially induces opticalscattering and reduces carrier mobilities at domain boundaries,which severely degrade the electrical and optoelectronic propertiesof the MoS2 film. Thus, unidirectional alignment of MoS2 on sap-phire is highly preferred for achieving high-quality, wafer-scalesingle-crystalline films that are essential for advanced electronicand optoelectronic applications.Several attempts have been made recently to realize unidirec-tional alignment of MoS2 on sapphire.3–5,7,9,11 Determination ofthe epitaxial relationship is crucial to confirm whether the unidi-rectional alignment is achieved. Epitaxial relationship between MoS2and sapphire has extensively studied using a variety of experimentaltechniques such as optical microscopy,1–3,5–7 atomic-force micros-copy,4,9 x-ray diffraction (XRD),6,8–10 electron diffraction,3–5,8,11 andcross-sectional transmission electron microscopy (TEM).3–5,7,11Epitaxially grown MoS2 islands often exhibit triangular shapes andshow preferential alignment of their edges along certain crystallo-graphic directions of the sapphire substrate. While such macro-scopic alignment could be observed in optical microscopy andatomic-force microscopy images, it is usually not sufficient to con-clusively determine the precise epitaxial orientation. In addition,the orientation of MoS2 triangular islands enclosed byMo-terminated edges is rotated by 180� with respect to those withS-terminated edges,12 which makes it difficult to determinewhether the ML-MoS2 films were grown on the sapphire substratein a parallel or antiparallel configurations.Cross-sectional TEM directly provides information aboutatomic structures at the MoS2=Al2O3 interface,3–5,7,11 whichenables us to clearly distinguish parallel and antiparallel configura-tions. However, TEM data may not be fully representative of thewhole sample, because atomic structures in a localized area areimaged in TEM. While XRD has been widely used to study theJournal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-1© Author(s) 2026 14 January 2026 01:48:17https://doi.org/10.1063/5.0303598https://doi.org/10.1063/5.0303598https://pubs.aip.org/action/showCitFormats?type=show&doi=10.1063/5.0303598http://crossmark.crossref.org/dialog/?doi=10.1063/5.0303598&domain=pdf&date_stamp=2026-01-12https://orcid.org/0000-0002-3519-4613https://orcid.org/0000-0002-0486-2981https://orcid.org/0000-0001-6804-7217mailto:OHTAKE.Akihiro@nims.go.jphttps://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1063/5.0303598https://pubs.aip.org/aip/japepitaxial relationship between MoS2 and sapphire,6,8–11 parallel andantiparallel configurations could not be distinguished by conven-tional XRD measurements.This paper reports a combined experimental and theoreticalstudy on the detailed epitaxial orientation of ML-MoS2 on thec-plane sapphire substrate. From low-energy electron diffraction(LEED) analysis based on dynamical diffraction theory, we foundthat the MoS2 ML on sapphire is unidirectionally aligned in theparallel configuration, i.e., [1100]MoS2 // [1100]sapphire [Figs. 1(a)and 1(c)] over the whole sample. Theoretical calculations revealedthat the stability of unidirectionally oriented MoS2 ML criticallydepends on the atomic registry between the MoS2 and the sapphirelattices.II. EXPERIMENTALThe samples were prepared using the MOCVD reactor.11 The2 in. c-sapphire substrates with a miscut angle of �0:2� along them-axis direction (c/m = �0:2�) were procured by Orbray Co. Ltd.and were cut into 20 � 20 mm2 pieces for the growth experiments.Prior to the growth, the substrates were annealed in a mufflefurnace at 1150 �C for 1 h to produce a step-and-terrace structurewith ML-height steps. The ML-MoS2 film was grown at a substratetemperature of 975 �C using molybdenum oxychloride (MoO2Cl2)and hydrogen sulfide (H2S) as the molybdenum and sulfur precur-sors, respectively. The complete surface coverage was attained by60 min. The uniformity of the MoS2 coverage was confirmed byRaman and AFM measurements.11 Representative AFM imagesand Raman spectra were shown in Figs. S1 and S2, respectively, inthe supplementary material.The samples were analyzed by LEED (OCI BDL600IR) underultrahigh vacuum (UHV) conditions at 1� 10�8 Pa. Before theLEED measurements, the samples were degassed at 150 �C for30 min in UHV. The LEED patterns were acquired at room temper-ature with a 1 eV step in the energy range of 30–380 eV. The LEEDintensity–voltage (I–V) curves for five nonequivalent beams wereextracted from LEED patterns with the background being sub-tracted. The total cumulative energy range (ΔE) was approximately1455 eV. The samples were also characterized by reflection high-energy electron diffraction (RHEED) (Staib EK-35-R) with an elec-tron beam energy of 20 keV.III. CALCULATIONSTheoretical calculations were performed by using the PHASE/0code,13 which is based on DFT14 and pseudo-potential schemes15,16with plane-wave basis sets. For the exchange-correlation term, thePBE form was used.17 For the van der Waals interactions, theDFT-D2 method was applied.18 The cutoff energies for the wave-functions and charge density were 56 and 506 Ry, respectively. Thenumber of k points sampled in the Brillouin zone was more than 5� 5 per the surface unit cell of c-plane sapphire. All the models wereoptimized to meet the force criterion of 0.02 eV/Å. Sapphire slabemployed here consists of six Al2O3 layers and the thickness of avacuum region is 0.9 nm. The top (bottom) layer of a sapphire slabis terminated with only one Al atom per surface unit, to make itselectronic states semiconducting. The calculated lattice constantfor the hexagonal MoS2 monolayer is 0.318 nm, while that for sap-phire is 0.4798 nm. In this study, the change in the lattice mismatchof MoS2 grown on sapphire is quite essential to see the stability ofMoS2/sapphire heterostructures. Thus, the lattice constant forsapphire is set to 0.4789 nm to follow the experimental ratio of1.506 for MoS2 and sapphire. This treatment is justified, becausethere is no chemical bond between MoS2 monolayer andsapphire slab, and the lattice constant of sapphire slab is notaffected by MoS2. Adsorption energy, Ead is defined asEad ¼ (EMoS2 þ Esapphire � EMoS2=sapphire)=S, where EMoS2 , Esapphire,and EMoS2=sapphire are the calculated total energies for MoS2 mono-layer, sapphire, and MoS2/sapphire, respectively. S is the area of asuperstructure. In this definition, the larger Ead is, the more stablethe system is.IV. RESULTS AND DISCUSSIONFigures 2(a) and 2(b) show RHEED patterns of the sapphire(0001) substrate taken from the [1120] and [1100] directions,respectively. The asymmetric features of the RHEED patternobserved along the [1120] direction indicate the threefold symme-try of the sapphire (0001) surface, which is more clearly seen inLEED patterns [Fig. 2(d)], in which the intensity of the 1 0 spot isFIG. 1. Schematic drawings of MoS2(0001) [(a) and (b)] and Al2O3(0001)(c) lattices. The MoS2 lattices in (a) and (b) are in parallel and antiparallel con-figurations, respectively, with the Al2O3 lattice.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-2© Author(s) 2026 14 January 2026 01:48:17https://doi.org/10.60893/figshare.jap.c.8207132https://pubs.aip.org/aip/japhigher than that of the 0 1 spot. The RHEED patterns of 1ML-MoS2 film grown on the sapphire substrate are shown inFigs. 2(e) and 2(f). The spacings of streaks [a�(f ) and b�(f ) inFigs. 2(e) and 2(f )] are 1.5 times larger than those of sapphire[a�(s) and b�(s) in Figs. 2(a) and 2(b)] in both [1120] and [1100]directions, being consistent with lattice constants of sapphire(0.476 nm) and MoS2 (0.316 nm). Thus, it turns out that thecrystalline MoS2(0001) films were grown on the sapphire(0001)substrate with the epitaxial relationship of MoS2 [1100] // sapphire[1100]. However, from the results in Fig. 2, it is difficult to deter-mine if [1100] directions of MoS2 and sapphire are in a parallel[Fig. 1(a)] or antiparallel [Fig. 1(b)] configuration with the sapphiresubstrate [Fig. 1(c)].Figures 3(a)–3(c) show LEED patterns of 1 ML-MoS2 takenfrom the three locations across the sample. It is clearly seen thatthe spot intensities in LEED patterns measured at all of the loca-tions drastically change with the incident electron beam energy:whereas LEED patterns taken at 190 and 240 eV show threefoldsymmetry, the relative intensities of spots A and B are reversed. Incontrast, a sixfold symmetry is observed at 204 eV. We confirmedthat almost identical LEED patterns were observed at eight mea-surement points on the sample.To identify whether the ML-MoS2 films were grown on thesapphire substrate in a parallel or antiparallel configuration, weperformed LEED I–V curve analysis on the basis of dynamicaldiffraction theory. LEED I–V curves were calculated usingSATLEED package provided by Barbieri and Van Hove.19,20 Thepresent calculation used 10 phase shifts for the description of theelectron-crystal interaction. The inner potential V0 þ iVim was setto be independent of energy: the real part V0 was initially set tobe 10 eV and adjusted during the fitting process and theimaginary part Vim was set to be �4 eV. The isotropic thermalvibrational amplitudes represented by Debye temperatures werealso optimized to obtain good agreement with the experimentalI–V curves. The resultant Debye temperatures are 800 K forboth Mo and S atoms. To quantify the agreement between mea-sured and calculated I–V curves, we use Pendry’s reliabilityfactor (RP).21Figure 3(d) shows measured LEED I–V curves together withthe calculated ones from the MoS2 ML. This calculations assumedthat the ML-MoS2(0001) lattice is perfectly aligned to the sapphire(0001) lattice with the epitaxial relationship of MoS2 [1100] // sap-phire [1100], as shown in Figs. 1(a) and 1(c) (parallel configura-tion). As a first step of the analysis, for simplicity, I–V curves werecalculated from the isolated ML-MoS2(0001)-(1�1) model usingatomic coordinates fixed at bulk values. The analysis yields theoverall R factor of RP = 0.29, showing a good agreement with theLEED experiments: as indicated by vertical dashed lines inFig. 3(d), the intensity of the 1 0 (0 1) spot is higher than that ofthe 0 1 (1 0) spot at 190 eV (240 eV) for both measured and calcu-lated I–V curves. On the other hand, as shown in Fig. 4(c), whenthe antiparallel configuration [Figs. 1(b) and 1(c)] was considered,the calculation could not reproduce the experimental I–V curves(RP = 1.00), suggesting that the formation of antiparallel domains(inversion domains) could be effectively suppressed. These resultsclearly show that LEED spots A, B, C, D, and E could be indexedas 1 0, 0 1, 1 1, 2 0, and 0 2 reflections, respectively, and that theMoS2 grows on the sapphire substrate with its [1100] direction par-allel to the sapphire [1100] direction (parallel configuration). Inaddition, when I–V curves were calculated for bilayer MoS2[Fig. 4(d)] and bulk MoS2 [Fig. 4(e)], they resulted in larger Rfactors (RP = 0.47 for 2 ML-MoS2 and 0.48 for bulk MoS2),FIG. 2. RHEED patterns of the sap-phire(0001) substrate (a) and (b) and 1ML-MoS2(0001) (e) and (f ) takenalong the [1120] and [1100] directionsof sapphire. (c) and (g) show sche-matic drawing of reciprocal lattices ofsapphire and MoS2, respectively. LEEDpatterns taken from sapphire (d) andMoS2 (h) with an incident electronbeam energy of 120 eV.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-3© Author(s) 2026 14 January 2026 01:48:17https://pubs.aip.org/aip/japindicating that the MoS2 film mainly consists of ML-MoS2domains. This is in good agreement with AFM observations.11Since, as mentioned earlier, the in-plane lattice constant ofsapphire is approximately 1.5 times larger than that of MoS2, theMoS2 (3 � 3) supercell aligned in the same direction of the Al2O3(2 � 2) cell reduces the effective lattice mismatch from �33:6% to�0:42%. Possible atomic configurations of MoS2 (3 � 3)/Al2O3 (2� 2) superstructure are shown in Figs. 5(a)–5(c). In the models Aand C, as indicated with dashed circles, Mo and S atoms arelocated just above the surface Al atoms, respectively, whereassurface Al atoms are located beneath the center of Mo-S hexagonin the model B. To study the relative stability of these models, weperformed theoretical calculations. The present calculationsassumed the Al2O3 surface terminated with a single Al layer[Fig. 5(d)], which has been reported most energetically stable22,23and has been confirmed by experiments.24,25 After the structuraloptimization, the surface Al atoms in models A and B are displaceddownward by a large amount of 0.07–0.08 nm from their initialposition (horizontal dashed line). On the other hand, in the modelC, a slight upward displacement (�0.02 nm) is observed. Theatomic displacements of Mo and S atoms from their positions inoptimized ML-MoS2 are much less than 0.01 nm in all models.The energetic stability of MoS2/sapphire strongly depends onthe lateral registry between MoS2 and Al2O3 lattices: the models Aand B have 61 meV and 66meV higher adsorption energies per(3 � 3) unit cell, respectively, than the model C. We performedLEED analysis using the optimized atomic coordinates for MoS2(3 � 3)/Al2O3 (2 � 2) superstructures. As compared with the anal-ysis using unrelaxed MoS2(1 � 1) model (RP = 0.29), the modelsA and B show improved R factors of RP = 0.24 and RP = 0.23,respectively, whereas no significant improvement was found forthe model C (RP = 0.28). The statistical error limit, var(R)¼ Rminffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi(8Vim=ΔE)p,21 was estimated to be 0.034, indicating a sig-nificant difference between models A and B and model C. Thus,we conclude that the models A and B are more likely for theMoS2=Al2O3 interface structure.The present experiments show that the epitaxial orientation ofMoS2 on sapphire is uniquely determined with the MoS2 [1120]and [1100] directions being aligned with the Al2O3 [1120] and[1100] directions, respectively (parallel configuration). On the otherhand, our DFT calculations show that the difference in the adsorp-tion energy between the parallel and antiparallel domains is negligi-bly small (energy difference � 0.1 meV/MoS2(1 � 1)).11 Thus, thepresent experimental results could not be explained by the ener-getic stability. Another possible explanation for the onset of theparallel configuration is based on the presence of surface steps onthe sapphire substrates. Early studies have shown that the presenceof step edges on the substrate enables the unidirectional alignmentof MoS2 domains:3 triangular MoS2 islands are nucleated withtheir edges being parallel to the surface steps on the sapphire sub-strate, thus facilitating the unidirectional alignment of MoS2domains. Specifically, the use of sapphire substrate with the miscutorientation towards the [1120] axis promotes the unidirectionalalignment of 90�-rotated MoS2 domains on sapphire.3To study whether surface steps play a crucial role in determin-ing the epitaxial orientation, we prepared sapphire substrates withdifferent miscut orientations, as shown in Fig. 6(a). The miscutdirections and angles of the sapphire substrate were evaluatedby XRD rocking-curve measurements (see Fig. S3 in thesupplementary material). Figure 6(b) compares LEED I–V curvesmeasured from the ML-MoS2 grown on the three substrates; almostidentical results were obtained irrespective of the direction of stepson the sapphire substrate. All of the measured curves (A-C) are ingood agreement with the calculated ones for the MoS2/sapphiremodels shown in Figs. 5(a) and 5(b); RP is in the range of0.22�0.24. These results clearly show that the direction of surfaceFIG. 3. LEED patterns taken from three locations across the 1 ML-MoS2/sap-phire sample with incident electron beam energies of 190 (a), 204 (b), and240 eV (c). (d) Measured (red) and calculated (blue) LEED I–V curves.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-4© Author(s) 2026 14 January 2026 01:48:17https://doi.org/10.60893/figshare.jap.c.8207132https://pubs.aip.org/aip/japFIG. 4. (a) LEED I–V curves mea-sured from the 1 ML-MoS2 sample.Calculated I–V curves for 1 ML-MoS2on sapphire in parallel (b) and antipar-allel (c) configurations. Curves (d) and(e) were calculated from 2 ML MoS2and bulk MoS2, respectively.FIG. 5. Schematics of MoS2(3 � 3)/Al2O3(2 � 2) superstructures with different lateral registries between MoS2 and Al2O3 (a–c) and the Al2O3 surface (d). Horizontaldashed line indicates the vertical position of the surface Al atom of the Al2O3 substrate in (d).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-5© Author(s) 2026 14 January 2026 01:48:17https://pubs.aip.org/aip/japsteps on the sapphire substrate is less likely to cause the changes inthe epitaxial orientation of MoS2.The observed epitaxial relationship (0� domain) is in goodagreement with those reported in Refs. 1, 8–11. and is the orthogo-nal configuration of those (90� domains) reported in Refs. 2–7. Inaddition, as shown in Figs. 6(a)-C and 6(b)-C, the 90� domainswere not formed even when the sapphire substrate with the miscutorientation towards the [1120] direction was used, in marked con-trast with the result in Ref. 3. We performed DFT calculations forthe 0� and 90� domains to check if the formation of the 0� domainis governed by energetic stability, and found that the 0� domain ismore stable by 0.164 eV/MoS2(1 � 1). When the MoS2 lattice isrotated by 90�, MoS2 (5 � 5) is nearly lattice matched to Al2O3(2ffiffiffi3p � 2ffiffiffi3p)-R30�, resulting in a lattice mismatch of �4:18% (seeFig. S4 in the supplementary material). Thus, one may ascribe thehigher stability of the 0� domain to smaller lattice mismatch(�0:42%). However, in actual systems, the formation of highlystrained MoS2 lattices is unlikely, because of weak van der Waalsinteraction between MoS2 and Al2O3. Indeed, under the first-orderapproximation, the 90� domain becomes most stable in an incom-mensurate situation.1 Thus, no definitive answer is available for therelative stability of the 0� and 90� domains on the basis of the DFTcalculations.In reviewing previous papers on the epitaxial relationship forMoS2/sapphire, we found that the epitaxial orientation stronglydepends on the growth method: the epitaxial relationship of MoS2[1120] // sapphire [1120] (0� domain) has been commonlyobserved for the MoS2 ML grown by MOCVD,8–11 while thegrowth using powder-source CVD results in the formation of 90�domains,2–7 except for the result in Ref. 1. Thus, while not conclu-sive, it is possible that the two growth methods provide differentgrowth environments, e.g., surface reconstructions, chemical adsor-bates, and possible existence of interface layers, giving rise to thedifferent epitaxial orientations of MoS2.V. CONCLUSIONSWe have studied the epitaxial relationship of MOCVD-grownMoS2 on sapphire. LEED I–V curve analysis clearly shows that theMoS2 ML is unidirectionally aligned with the sapphire substratewith the MoS2 [1120] and [1100] directions being aligned with theAl2O3 [1120] and [1100] directions, respectively. On the basis ofDFT calculations, we proposed the possible interface structures forMoS2/sapphire, which agree well with the LEED experiments. Theobserved epitaxial relationship is hardly affected by the presence ofsurface steps on the sapphire substrate.SUPPLEMENTARY MATERIALSee the supplementary material for AFM images (Fig. S1) andRaman spectra (Fig. S2) for ML-MoS2 on sapphire, XRD data forsapphire substrate with different off-directions (Fig. S3), and theatomic configurations of MoS2 on sapphire for 0� and 90� domains(Fig. S4).ACKNOWLEDGMENTSWe would like to thank T. Hiroto for XRD measurements.Theoretical calculations were performed by using the NumericalMaterials Simulator of NIMS and the Earth Simulator (ES) ofJAMSTEC. This work was partly supported by JST-CREST (GrantNo. JPMJCR24A3) and JSPS KAKENHI (Grant No. JP23K04592).AUTHOR DECLARATIONSConflict of InterestThe authors have no conflicts to disclose.Author ContributionsAkihiro Ohtake: Conceptualization (equal); Data curation (lead);Formal analysis (lead); Investigation (equal); Methodology (lead);Writing – original draft (lead); Writing – review & editing (lead).Jun Nara: Conceptualization (equal); Data curation (lead); Formalanalysis (lead); Investigation (equal). Yoshiki Sakuma:Conceptualization (lead); Funding acquisition (lead); Investigation(equal); Project administration (lead); Supervision (lead).FIG. 6. (a) Schematic illustrations of sapphire substrates with miscut angles of�0.2� along the [1100] (A), [1100] (B), and [1120] (C) directions. The miscutangles determined by the XRD measurements are c/m =�0:2485� (A),c/m =þ0:199� (B), and c/a =þ0:1586� (C). (b) LEED I–V curves of ML-MoS2grown on sapphire substrates with different off-directions.Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-6© Author(s) 2026 14 January 2026 01:48:17https://doi.org/10.60893/figshare.jap.c.8207132https://doi.org/10.60893/figshare.jap.c.8207132https://pubs.aip.org/aip/japDATA AVAILABILITYThe data that support the findings of this study are availablefrom the corresponding author upon reasonable request.REFERENCES1D. Dumcenco et al., “Large-area epitaxial monolayer MoS2,” ACS Nano 9,4611–4620 (2015).2A. Aljarb, Z. Cao, H.-L. Tang, J.-K. Huang, M. Li, W. Hu, L. Cavallo, andL.-J. Li, “Substrate lattice-guided seed formation controls the orientation of 2Dtransition-metal dichalcogenides,” ACS Nano 11, 9215–9222 (2017).3T. Li et al., “Epitaxial growth of wafer-scale molybdenum disulfide semiconduc-tor single crystals on sapphire,” Nat. Nanotechnol. 16, 1201–1207 (2021).4J. H. Fu et al., “Oriented lateral growth of two-dimensional materials onc-plane sapphire,” Nat. Nanotechnol. 18, 1289–1294 (2023).5L. Li et al., “Epitaxy of wafer-scale single-crystal MoS2 monolayer via bufferlayer control,” Nat. Commun. 15, 1825 (2024).6Y. Hua et al., “Eight in. wafer-scale epitaxial monolayer MoS2,” Adv. Mater. 36,2402855 (2024).7H. Chen, C. Ji, Y. Chen, H. Hou, W. Li, J. Shen, C. Cao, H. Zhu, H. Li, andW. Kong, “Interfacial atomic mechanisms of single-crystalline MoS2 epitaxy onsapphire,” Adv. Mater. 37, 2414317 (2025).8X. Xiang et al., “Monolayer MoS2 on sapphire: An azimuthal reflection high-energy electron diffraction perspective,” 2D Mater. 8, 025003 (2021).9I. Kandybka, B. Groven, H. Medina Silva, S. Sergeant, A. Nalin Mehta,S. Koylan, Y. Shi, S. Banerjee, P. Morin, and A. Delabie, “Chemical vapor deposi-tion of a single-crystalline MoS2 monolayer through anisotropic 2D crystalgrowth on stepped sapphire surface,” ACS Nano 18, 3173–3186 (2024).10C. Chen et al., “Effect of growth temperature on the microstructure and prop-erties of epitaxial MoS2 monolayers grown by metalorganic chemical vapor dep-osition,” J. Vac. Sci. Technol., A 42, 022201 (2024).11Y. Sakuma et al., “Self-aligned and self-limiting van der Waals epitaxy ofmonolayer MoS2 to annihilate grain boundaries for scalable 2D electronics,”(unpublished), available at https://doi.org/10.21203/rs.3.rs-7386422/v1 (2025).12J. Dong, D. Ding, C. Jin, Y. Liu, and F. Ding, “Edge reconstruction-dependentgrowth kinetics of MoS2,” ACS Nano 17, 127–136 (2023).13T. Yamasaki, A. Kuroda, T. Kato, J. Nara, J. Koga, T. Uda, K. Minami, andT. Ohno, “Multi-axis decomposition of density functional program for strongscaling up to 82 944 nodes on the K computer: Compactly folded 3D-FFT com-municators in the 6D torus network,” Comput. Phys. Commun. 244, 264–276(2019).14W. Kohn and L. J. Sham, “Self-consistent equations including exchange andcorrelation effects,” Phys. Rev. 140, A1133–A1138 (1965).15N. Troullier and J. L. Martins, “Efficient pseudopotentials for plane-wave cal-culations,” Phys. Rev. B 43, 1993–2006 (1991).16D. Vanderbilt, “Soft self-consistent pseudopotentials in a generalized eigen-value formalism,” Phys. Rev. B 41, 7892–7895 (1990).17J. P. Perdew, K. Burke, and M. Ernzerhof, “Generalized gradient approxima-tion made simple,” Phys. Rev. Lett. 77, 3865–3868 (1996).18S. Grimme, “Semiempirical GGA-type density functional constructed with along-range dispersion correction,” J. Comput. Chem. 27, 1787–1799 (2006).19P. J. Rous, J. B. Pendry, D. K. Saldin, K. Heinz, K. Müller, and N. Bickel,“Tensor LEED: A technique for high-speed surface-structure determination,”Phys. Rev. Lett. 57, 2951–2954 (1986).20P. J. Rous and J. B. Pendry, “The theory of tensor LEED,” Surf. Sci. 219,355–372 (1989).21J. B. Pendry, “Reliability factors for LEED calculations,” J. Phys. C 13, 937–944(1980).22T. J. Godin and J. P. LaFemina, “Atomic and electronic structure of thecorundum (α-alumina) (0001) surface,” Phys. Rev. B 49, 7691–7696(1994).23X.-G. Wang, A. Chaka, and M. Scheffler, “Effect of the environment onα-Al2O3 (0001) surface structures,” Phys. Rev. Lett. 84, 3650–3653 (2000).24J. Ahn and J. W. Rabalais, “Composition and structure of the Al2O3{0001}-(1� 1) surface,” Surf. Sci. 388, 121–131 (1997).25E. A. Soares, M. A. Van Hove, C. F. Walters, and F. F. McCarty, “Structure ofthe α-Al2O3 (0001) surface from low-energy electron diffraction: Al-terminationand evidence for anomalously large thermal vibration,” Phys. Rev. B 65, 195405(2002).Journal ofApplied PhysicsARTICLE pubs.aip.org/aip/japJ. Appl. Phys. 139, 024303 (2026); doi: 10.1063/5.0303598 139, 024303-7© Author(s) 2026 14 January 2026 01:48:17https://doi.org/10.1021/acsnano.5b01281https://doi.org/10.1021/acsnano.7b04323https://doi.org/10.1038/s41565-021-00963-8https://doi.org/10.1038/s41565-023-01445-9https://doi.org/10.1038/s41467-024-46170-6https://doi.org/10.1002/adma.202402855https://doi.org/10.1002/adma.202414317https://doi.org/10.1088/2053-1583/abce08https://doi.org/10.1021/acsnano.3c09364https://doi.org/10.1116/6.0003296https://doi.org/10.21203/rs.3.rs-7386422/v1https://doi.org/10.1021/acsnano.2c05397https://doi.org/10.1016/j.cpc.2019.04.008https://doi.org/10.1103/PhysRev.140.A1133https://doi.org/10.1103/PhysRevB.43.1993https://doi.org/10.1103/PhysRevB.41.7892https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1002/jcc.20495https://doi.org/10.1103/PhysRevLett.57.2951https://doi.org/10.1016/0039-6028(89)90513-Xhttps://doi.org/10.1088/0022-3719/13/5/024https://doi.org/10.1103/PhysRevB.49.7691https://doi.org/10.1103/PhysRevLett.84.3650https://doi.org/10.1016/S0039-6028(97)00383-Xhttps://doi.org/10.1103/PhysRevB.65.195405https://pubs.aip.org/aip/jap