# Fileset

[Advanced Science - 2025 - Yang - Highly Oriented Epitaxial Hexagonal Boron Nitride Multilayers on.pdf](https://mdr.nims.go.jp/filesets/c376af6c-10c7-45ff-b4e5-11fe56559bcb/download)

## Creator

[Xu Yang](https://orcid.org/0000-0001-8195-5850), Markus Pristovsek, Shugo Nitta, Yoshio Honda, [Akihiro Ohtake](https://orcid.org/0000-0002-3519-4613), [Yoshiki Sakuma](https://orcid.org/0000-0001-6804-7217), [Takanobu Hiroto](https://orcid.org/0000-0002-6176-5782), Takayuki Ishida, Michio Ikezawa, Qixin Guo, Hiroshi Amano

## Rights

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

## Other metadata

[Highly Oriented Epitaxial Hexagonal Boron Nitride Multilayers on High‐Temperature‐Resistant Single‐Crystal Aluminum Nitride (0001)](https://mdr.nims.go.jp/datasets/1194ce42-f7ca-4414-9ab3-087a037e368b)

## Fulltext

Highly Oriented Epitaxial Hexagonal Boron Nitride Multilayers on High‐Temperature‐Resistant Single‐Crystal Aluminum Nitride (0001)RESEARCH ARTICLEwww.advancedscience.comHighly Oriented Epitaxial Hexagonal Boron NitrideMultilayers on High-Temperature-Resistant Single-CrystalAluminum Nitride (0001)Xu Yang,* Markus Pristovsek, Shugo Nitta, Yoshio Honda, Akihiro Ohtake,Yoshiki Sakuma, Takanobu Hiroto, Takayuki Ishida, Michio Ikezawa, Qixin Guo,and Hiroshi AmanoThe epitaxy of high-quality hexagonal boron nitride (hBN) multilayers ondielectric wafers is essential for hBN applications but remains challenging.Herein, highly-oriented hBN multilayers grown on single-crystal aluminumnitride (AlN)—AlN on sapphire and bulk AlN substrates—via metalorganicvapor phase epitaxy and high-temperature annealing is reported. HexagonalAlN (0001) not only provides a crystallographically commensurate base forhBN epitaxy but is thermally stable for hBN annealing up to 1800 °C, enablingthe first instance of large-area multilayer hBN with both superior out-of-planeand in-plane alignments grown directly on dielectrics using a fullyindustry-compatible approach. Elevated temperatures also reduce carbon andallow control over the separation of related single photon emission centers inhBN. These centers exhibit a record-narrow wavelength distribution(578 ± 5 nm) with small zero-phonon linewidths down to 1.44 meV, indicatingthe high uniformity of the achieved multilayer hBN films. This work paves anindustry-compatible way toward producing highly-oriented homogeneoushBN multilayers on dielectrics, promising for future device and integrationapplications.1. IntroductionLayered hexagonal boron nitride (hBN) is a 2D material witha very wide bandgap and deep well-isolated states, which hasX. Yang, M. Pristovsek, S. Nitta, Y. Honda, H. AmanoInstitute of Materials and Systems for SustainabilityNagoya UniversityNagoya 464–8601, JapanE-mail: x.yang@nagoya-u.jpA.Ohtake, Y. SakumaResearchCenter for Electronic andOpticalMaterialsNational Institute forMaterials Science1-1Namiki, Tsukuba, Ibaraki 305-0044, JapanThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/advs.202509354© 2025 The Author(s). Advanced Science published by Wiley-VCHGmbH. This is an open access article under the terms of the CreativeCommons Attribution License, which permits use, distribution andreproduction in any medium, provided the original work is properly cited.DOI: 10.1002/advs.202509354been used for next-generation advancedtechnologies, such as 2D electronics, room-temperature single-photon emission, andneutron detection.[1–7] Compared withmonolayer hBN, multilayer hBN servesmore effectively as a substrate and dielec-tric layer in fabricating 2D semiconductortransistors and favors stabilizing quantumemission in hBN. Since only small hBNflakes can be mechanically exfoliated frombulk crystals, research has focused onepitaxial growth of hBN using large-areasubstrates. Chemical vapor deposition(CVD) of 2D layered BN on metals hasbeen extensively studied. However, thecatalytic action only works for the firstlayer and limits hBN films to one or a fewmonolayers at most.[8–13] Moreover, devicefabrication and characterization often de-mand nonmetal substrates. In such cases,an extra transfer process is required thatinevitably introduces contamination anddamage. Consequently, direct growth ofhigh-quality, large-area hBN multilayers onmicrofabrication-compatible dielectric substrates such as sap-phire remains highly desirable for broader applications andintegration.[14] Unfortunately, dielectric wafers, such as sap-phire and Si, are catalytically inactive for hBN growth. HighT. HirotoResearch Network and Facility Service DivisionNational Institute for Materials Science1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, JapanT. Ishida, M. IkezawaInstitute of Pure and Applied SciencesUniversity of Tsukuba1-1-1 Tennoudai, Tsukuba, Ibaraki 305–8571, JapanQ. GuoDepartment of Electrical and Electronic EngineeringSaga UniversityHonjo-1, Saga 840–8502, JapanAdv. Sci. 2025, 12, e09354 e09354 (1 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbHhttp://www.advancedscience.commailto:x.yang@nagoya-u.jphttps://doi.org/10.1002/advs.202509354http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadvs.202509354&domain=pdf&date_stamp=2025-09-29www.advancedsciencenews.com www.advancedscience.comtemperatures (>1500 °C) have been suggested to improve theepitaxial quality of hBN.[15] However, silicon has a melting pointof ≈1410 °C, while sapphire begins to decompose above 1400–1500 °C.[16] Though direct synthesis of hBN on sapphire abovethese temperatures via epitaxy and annealing is possible, sap-phire degradation has led to uncontrollable step-bunching andincreased roughness on the surface.[17,18] Furthermore, a signif-icant increase in oxygen impurities, stemming from the ther-mally degraded sapphire substrate, has been clearly observed inhBN grown at 1500 °C.[19] These limitations greatly hinder theuse of high-temperature synthesis process that hBN favors.[18–20]Aluminum nitride (AlN) is a wurtzite III-nitride material with ahexagonal (0001) surface resembling that of hBN. AlN bulk crys-tals are synthesized via physical vapor transport above 2000 °C.Even on sapphire, a sufficiently thick AlN layer (at least ≈50 nm)is stable up to 1700 °C,[21,22] stabilizing the underlying sapphireand mitigating the up-diffusion of oxygen from sapphire. Free-standing AlN substrates are expected to remain stable at evenhigher temperatures. However, prior limited attempts to growlayered BN on relevantly thick AlN (0001) layers on foreign sub-strates have shown inferior lattice alignments.[23–25]Herein, we report the successful synthesis of highly-orientedepitaxial hBN multilayers on single-crystal AlN (0001) surfaces,either relevantly thick AlN on sapphire or bulk AlN substrates, viaa fully industry-compatible approach, i.e., high-temperature met-alorganic vapor phase epitaxy (MOVPE) and subsequent anneal-ing at higher temperatures up to 1800 °C. This developed pro-cess, which leverages high temperature (up to 1800 °C) and high-temperature-resistant AlN as a substrate in conjunction with op-timized MOVPE growth, results in significant defect annihila-tion, superior lattice alignment, and relevant quantum emissionin hBN. This scalable and well-controllable approach for directlygrowing highly-oriented hBN multilayer films on dielectric sub-strates paves the way toward emerging hBN-based electronic andphotonic applications.2. Results and Discussion2.1. MOVPE of hBN Multilayers on AlN and Post-GrowthAnnealing ControlHereafter, we will call the single-crystal AlN epilayer on c-planesapphire as “AlN template” to distinguish it from the single-crystal bulk “AlN substrate”. Figure 1a shows the prepared typ-ical smooth AlN surface with an average surface roughness (Ra)of 0.2 ± 0.1 nm prior to hBN epitaxy. On these flat surfaces,hBN was grown by MOVPE using a pulsed-mode method withtriethylborane (TEB) and ammonia (NH3) as B and N precur-sors (Figure 1b). Note that the optimized growth temperature of1380 °C is limited by our MOVPE setup (maximum: ≈1400 °C),but it is still lower than the temperature favored by hBN as epi-taxially grown on dielectrics without catalyst assistance.[15] Af-ter MOVPE growth, the as-grown layers were thus annealed athigher temperatures up to 1800 °C in N2 with a face-to-face con-figuration, as illustrated in Figure 1c (details in Experimental Sec-tion). Besides precise control of the initial MOVPE growth, com-bining high-temperature annealing with thermally robust AlN asthe substrate is key to obtaining highly-oriented hBN grains.We found that elevating the growth temperature with reducednucleation rate is critical to achieve the initial epitaxial hBNmul-tilayers on AlN (0001) surfaces byMOVPE. Reduced growth tem-peratures and increased TEB flow rates both led to degraded hBNgrowth, which potentially limits the quality of the resulting hBNcrystal film even after subsequent high-temperature annealing.Using optimized growth conditions, we first grew well-defined2D hBN layers on AlN, where the layer thickness can be modu-lated by adjusting the number of pulsed cycles during MOVPEgrowth. High-resolution transmission electron microscopy (HR-TEM) confirmed the c-axis-aligned layered hBN on the AlN tem-plates with thicknesses of ≈1.5, 12, and 29 nm (Figure 1d–f). Thebuckling seen in the initial ≈1.5 nm of BN above the AlN surfaceis expected and has been observed before,[26] which is due to theformation of covalent bonds between the initial hBN nuclei andAlN. These covalent bonds follow the bond orientation in AlNand thus are not immediately flat.Both wrinkled and nearly wrinkle-free multilayer hBN filmswere achieved (Figure 1g,h). Wrinkle formation is commonlyascribed to the thermal expansion mismatch between 2D lay-ers (e.g., hBN and graphene) and the underlying substrates.[27,28]Thick hBN films on sapphire tend to exhibit pronounced wrin-kling due to increased compressive strain,[29] leading to largersurface roughness. In this work, the 12 nm thick hBN exhib-ited clear wrinkles with height variations of ≈1–5 nm and a Raof 1.41 nm (Figure 1h,i). In contrast, the thinner 1.5–2.0 nmhBN showed minimal surface wrinkling with low height varia-tions (0.5–1.5 nm) and a small Ra of 0.68 nm (Figure 1g–i). Thesurface roughness in the few-layer hBN films is also affected bythe initial buckling growth. Hence, smoother hBN films on AlNcould be obtained by reducing the initial buckling. Lowering thegrowth temperature and increasing the TEB flux assisted in re-ducing buckling but compromised the interfacial crystallinity.X-ray diffraction (XRD) measures the crystallographic proper-ties of hBN averaged over large areas but is limited by the lowscattering cross-sections of boron and nitrogen atoms. Further-more, the symmetric 0002 reflection requires at least 4 monolay-ers, i.e., ≈1.3 nm hBN. Thus, it is difficult to detect the ultrathin1.5 nm thick layered hBN by XRD. However, the XRD 2𝜃-𝜔mea-surement of the 12 nm thick hBN clearly showed the 0002 reflec-tion at 2𝜃 = 26.6° (Figure 2a), indicating c-oriented hBN growthon the AlN template by MOVPE. The in-plane 𝜑 measurementof the hBN 10-10 reflection displayed six peaks appearing at 60°intervals (Figure 2b), confirming the hexagonal symmetry andhigh rotational alignment of the hBNmultilayers. Due to the rel-atively low growth temperature, the reflections exhibited broadpeaks with large full-width at half maximum (FWHM) values. Af-ter annealing the hBN at 1700 °C, the 0002 reflection narrowedand increased in amplitude, and even the hBN 0004 reflectionat ≈54.8° appeared (Figure 2a), indicating enhanced c-axis align-ment. Similarly, the hBN 10-10 reflection in the in-plane 𝜑mea-surement narrowed from ≈7.1° for as-grown film to ≈1.8° afterannealing at 1700 °C (Figure 2b), both exceeding the XRD in-strumental resolution of ≈0.2° (for the in-plane configuration).The only comparable data in literature is the in-plane misalign-ment, which gives the rotational misorientation of mostly single-crystal monolayer hBN films relative to metal-based substrates.Our XRD FWHMs still compare well with the reported misalign-ment of state-of-the-art single-crystal hBN monolayer or trilayerAdv. Sci. 2025, 12, e09354 e09354 (2 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 1. Synthesis of epitaxial hBN multilayers on single-crystal AlN (0001) by MOVPE and high-temperature (HT)-annealing. a–c) Schematic illustra-tions depicting the synthesis process of multilayer hBN epilayers on atomically flat single-crystal AlN-based substrates. The AFM image in (a) showsa representative smooth AlN surface with a step-like feature. d–f) Cross-sectional TEM images of ≈1.5-, 12- and 29-nm-thick 2D layered hBN filmsgrown on AlN templates, respectively. BN at the AlN/hBN interface features dome-like buckling growth. Scale bar: 5 nm. g,h) AFM images of (g) a nearlynon-wrinkled thin hBN multilayer film (≈1.5 nm) and (h) a wrinkled thick hBN multilayer film (12 nm). i) Height profiles of the ≈1.5 nm (red) and 12nm (blue) hBN multilayer films on AlN templates, measured along the white dotted lines in (g) and (h).films grown on metals, such as Cu, Ni, and Au,[11–13,30,31] and aremuch better than that of previously reported CVD-grown hBNmultilayers on sapphire (Figure 2c),[32] even though our in-planeXRD FWHM is additionally broadened by size effects and themisorientation of individual layers within the multilayer hBN.Sharp and streaky RHEED patterns observed at every 60° con-firmed the high in-plane orientation with hexagonal symmetry(Figure 2d). The stripe spacing along the [10–10] and [11–20] az-imuths in the RHEED patterns yielded an in-plane constant of≈2.55 Å, close to the reported a-lattice constant of hBN.[33,34] Fur-thermore, the coincidence of the hBN peak positions with thoseof AlN in the in-plane XRD 𝜑-scans suggested that the epitaxialrelationship is [10-10]hBN//[10-10]AlN (Figure S1, Supporting In-formation). It has been reported that the epitaxy of layered BNon Ni can be guided by surface steps on the Ni.[35] To examinethe effect of AlN surface morphology, AlN templates with step-flow and step-bunched surfaces were prepared for hBN epitaxy.After growth, the resulting hBN films exhibited similar surfacewrinkles and identical epitaxial alignment (Figure S2, Support-ing Information). These results suggest that the epitaxial growthof hBN on AlN is primarily governed by the lattice symmetry ofAlN and minimally influenced by its surface steps.The microstructure of the 12 nm hBN annealed at 1700 °C onthe AlN template was directly examined by cross-sectional TEM.The TEM cross-section in Figure 2e revealed a well-defined c-oriented layered structure that was stacked parallel to the AlNsurface with an interlayer spacing of≈3.34 Å. It is consistent withthe reported value of 3.33 Å for bulk hBN.[36] Compared with theas-grown hBN multilayer (Figure 1e), the annealed hBN multi-layer had amore clearly defined layered structure and fewer struc-tural imperfections. There was no significant interface reactionbetween AlN and hBN, as neither the dome-like buckling nor thesurface wrinkling changed after annealing (Figure 2e,f), and thesurface showed uniform contrast (Figure 2g). Thus, hBN doesnot react with AlN up to 1700 °C. The electron energy loss spec-trum (EELS) acquired from the well-ordered hBN cross-sectionshown in Figure 2e exhibited the 𝜋* and 𝜎* energy losses for bothboron (≈190 eV) and nitrogen (≈400 eV) (Figure 2h), confirmingthe sp2 bonding configuration of hBN.[37] No signals for carbon,oxygen, and aluminumwere observed in the EELS spectrum afterannealing at 1700 °C. Furthermore, no boron was detected in theunderlying AlN (Figure S3, Supporting Information). These ob-servations further confirm the stability of the hBN on AlN. Addi-tionally, high-temperature annealing reduced impurities in hBNAdv. Sci. 2025, 12, e09354 e09354 (3 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 2. Structural and spectroscopic characterization of MOVPE-grown and annealed multilayer hBN on AlN templates. a) Symmetric XRD 2𝜃-𝜔measurements of as-grown and 1700 °C annealed hBN, showing reflections from hBN 0002, hBN 0004 (after annealing) and AlN 0002 and sapphire0006 from the AlN template. b) Corresponding in-plane XRD ϕ measurements of hBN {10-10} reflections for the hBN shown in (a). c) Comparison ofthe in-plane misorientation of mono- and tri-layers on metals and CVD-grown hBN multilayers on sapphire from the literature with the 10-10 in-planeFWHM of our hBN (as-grown and annealed hBN on AlN templates (1700 °C) and substrates (1800 °C)). d) RHEED patterns of the 1700 °C annealedhBN along the [11–20] and [10–10] directions. Yellow arrows indicate the positions of integer-order reflections. e) Cross-sectional TEM image of the hBNfilm shown in (d). Scale bar: 5 nm. Inset: Photograph of the annealed 12 nm hBN on the AlN template. f) AFM (2 × 2 μm2) and g) optical Nomarskimicroscope image of the 1700 °C annealed hBN in (e); Scale bar, 10 μm. h) EELS spectrum of the TEM cross section in (e). i,j) XPS core-levels of (i) O1s and (j) C 1s for as-grown and annealed hBN. k) Room-temperature PL spectra of as-grown and annealed hBN.as well. As observed by XPS, both the O 1s and C 1s core-levelintensities decreased significantly after annealing at 1650 °C andabove (Figure 2i,j). Consistent with the EELS results, no Al peakswere found in the XPS survey spectrum for hBN on the AlN tem-plate after annealing at 1700 °C (Figure S4, Supporting Informa-tion). This contrasts sharply with hBN grown on sapphire, wherehBN was damaged with the formation of AlN-based compoundson the surface after annealing at 1650 °C (Figure S5, SupportingInformation). As a more indirect indication, the intensity of de-fect luminescence decreased strongly after annealing at 1650 and1700 °C (Figure 2k). Particularly, the emission ≈2.1 eV has beenassociated with carbon-related defects in hBN.[38]Recent studies have identified carbon as a key constituent ofsingle photon emission (SPE) centers in hBN,[39,40] though theprecise structural origin remains debatable. Compared with as-grown hBN, the significant reduction in carbon levels after an-nealing enables us to resolve individual quantum emitters in themultilayer hBN even on an AlN template. Higher annealing tem-peratures result in fewer emission centers, as directly observedin the PL mapping measurements (Figure S6, Supporting Infor-Adv. Sci. 2025, 12, e09354 e09354 (4 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 3. Single photon emission from annealed multilayer hBN. a) Confocal PL map showing normalized emission intensity integrated over 550–650nm. An isolated single photon emitter, denoted as Emitter A and indicated by the white arrow, is clearly identified. Scale bar: 1 μm. b) PL spectrum(red) of Emitter A and its Gaussian-fitted line (gray). c) Second-order autocorrelation function g2(𝜏) for Emitter A (open circles) fitted with a two-levelmodel (red line), showing clear antibunching at zero-time delay (g2(0) <0.5). d) Luminescence stability measurement of Emitter A over 70 min, showingno bleaching or blinking. e) Statistical distribution of FWHM and wavelength localization of ZPL measured at ≈5K (bin size = 2 nm). The dashed linerepresents the spectrometer resolution. f) Comparison of the ZPL spectral distribution window of hBN emitters obtained in this work with those ofpreviously reported hBN emitters. All FWHM data were acquired at cryogenic temperatures.mation). This trend aligns with the XPS C 1s core-level data inFigure 2j, which exhibits a clear decrease in C 1s peak intensitywith increasing annealing temperature. These findings indicatethat the decrease in SPE centers is likely linked to the reducedcarbon content in our hBN layers.In this study, no observable PL emission appears within theSPE wavelength range for ultrathin hBN films (4 nm or less),as shown in Figure S7a (Supporting Information). In contrast,thicker hBN films (7 nm or more) exhibit clearly detectableemissions (Figure S7b, Supporting Information).However, whenthe thickness becomes too large (e.g., 12 nm), the emitter den-sity increases excessively, making it challenging to identify iso-lated single-photon emitters (Figure S7c, Supporting Informa-tion). Confocal PL mapping was thus conducted to measurethe emission intensity in the 550–650 nm range from a 7-nm-thick hBN annealed at 1700 °C, where we found a single iso-lated emitter within a 6 × 6 μm2 area (indicated by the whitearrow in Figure 3a). Its PL spectrum exhibited a zero-phononline (ZPL) at 575 nm and a phonon sideband (PSB) 158 meVbelow (Figure 3b), in agreement with previous studies.[3,41] Sin-gle photon emission was confirmed by measuring the second-order autocorrelation function, g(2)(𝜏), using a Hanbury Brownand Twiss setup at room temperature (Figure 3c). The achievedg(2)(0) ≈ 0.4 indicates the quantum nature of this emitter. Time-resolved PL measurements revealed stable fluorescence with noblinking or bleaching over 70 min for the hBN single photonemitter (Figure 3d). This is partly due to less contamination andthe low defect density surrounding the emitter.We measured ≈60 emitters from the annealed hBN multi-layer at room and cryogenic temperatures under an excitationAdv. Sci. 2025, 12, e09354 e09354 (5 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.compower of 100 μW. In both cases, ≈80% of the emitters exhib-ited ZPL wavelengths at ≈578 ± 5 nm (Figure 3e,f; Figure S8a,Supporting Information). This represents a considerably nar-rower spectral window compared to commonly reported visi-ble SPE from hBN,[39,42–49] highlighting the superior homogene-ity of the multilayer hBN achieved in this work. As represen-tatively shown in Figure S8b (Supporting Information), theseemitters also showed a small ZPL width with a mean of 4.9meV and a minimum of 1.44 meV measured at ≈5K. The nar-row FWHMs achieved in this study are overall comparable tothose reported for visible SPE from hBN grown on catalytic met-als and from hBN bulk crystals,[41,42,47,49,50] although ZPL widthcan be influenced by the excitation power used during mea-surement, which may vary across studies. ZPL broadening isalso related to defects, i.e., charge traps, in hBN. The observedsmall FWHM implies a low defect density surrounding theemitters.To get insights into carbon-related defect candidates that maycontribute to the SPE in our hBN, we compare our experimen-tal findings with previous experimental and theoretical stud-ies on carbon-related defects for hBN SPE. Several defects havebeen theoretically proposed to account for single-photon emis-sion ≈2 eV, including CBVN,[51] VBCN,[39] and carbon trimerssuch as C2CN and C2CB.[40,52] Among these, CBVN and VBCNare associated with high formation energies, several eV higherthan other carbon defects like dimers and trimers.[52] In con-trast, C2CN and C2CB trimers are energetically favorable andshow good agreement with the experimentally observed prop-erties of SPE ≈2 eV. Accordingly, we align the optical proper-ties of our SPE, including the ZPL energy (2.145 eV), PSB (158meV), Huang–Rhys (HR) factor (1.27), and Debye–Waller (DW)factor (0.28), with the reported theoretical predications.[52,53] TheC2CN trimer shows excellent agreement with our results acrossall these parameters, including a close match in the PL lineshape. Therefore, we identify the C2CN trimer as a likely candi-date to explain the observed results. However, our identificationis qualitative, and further microscopic investigations are neces-sary to unambiguously determine the atomic structure of theseemitters.Though the presence of defects in hBN that can act as SPE cen-ters, our high-temperature annealed hBN still exhibited decentelectrical properties. After transferring the hBN films from AlNtemplates onto Pt-coated quartz substrates using a wet process,Cr/Au top electrodes were deposited to form a metal/hBN/metalsandwich-type structure. The dielectric strength of hBN alongthe c-axis was evaluated using these device structures. The volt-age at which a sharp increase in leakage current occurs, corre-sponding to hard breakdown, was defined as the dielectric break-down voltage (VBD). For 18 nm hBN, the Au/Cr/hBN/Pt devicesexhibited a maximum VBD exceeding 20 V (Figure S9, Support-ing Information). The average breakdown field (EBD) was esti-mated to be 10.4 MV cm−1, comparable to that of hBN bulk crys-tals and superior to that of many previously reported CVD-grownhBN layers, indicating the high electrical quality of our multi-layer hBNfilms. Device-to-device variations in I–V characteristicswere observed, which are most likely caused by wrinkles, tears,and defects in hBN that form during the non-optimized transferprocess.2.2. Mechanism of Defect Healing in hBN via High TemperatureAnnealing2.2.1. Vacancy Migration-Mediated Defect HealingTo elucidate the microscopic processes behind the defect healingduring high-temperature annealing, we focused on the existenceof defects in hBN. Commonly known defects include vacancy andimpurity-related defects (e.g., VB, VN, CB, CN, OB, and ON).[54]Recently, self-healing induced by defect migration has been ob-served in graphene and conventional 3D materials.[55,56] In thecontext of hBN, the high annealing temperatures used in thisstudy are likely to promote defect migration and annihilation inhBN, healing imperfections in grains and grain boundaries. Asreported,[54] an annealing temperature of 1700 °C can overcomethe migration barriers of boron and nitrogen vacancies, allowingthem to move through the hBN lattice and potentially be annihi-lated by other defects or adatoms. In addition, hBN growth canpotentially form grain boundaries, which inevitably introducevarious line defects incorporating different types of B–N rings,as previously reported.[57] High-temperature annealing enlargesthe grain size of hBN, resulting in fewer grain boundaries, as ev-idenced by the enhanced crystallinity observed after annealing.Consequently, it is reasonable to expect a reduction in line defectswithin the hBN during the high-temperature annealing process.By plotting the FWHMs of the hBN 0002 and 10-10 reflections asa function of annealing temperature in Figure 4a, we extractedan activation energy of ≈4.0 eV for both. This is close to the ac-tivation energies calculated by density functional theory for themigration of VB (2.33–3.3 eV) and BN divacancy (VB-VN) migra-tion (3.0–4.5 eV).[54,58] This suggests that VB and VB-VN defectsmigrate (and annihilate at grain boundaries) during annealing.Additionally, the bond strengths of heteroatoms in hBN, such asB─C and C─N, are weaker than the B─N bond in hBN.[59] Thus,it is likely that heteroatoms like C and O move even faster thanvacancies during high-temperature annealing. Noteworthily, an-nealing at 1600 °C resulted in only a slight improvement relativeto as-grown hBN, whereas the FWHM strongly decreased afterannealing at 1650 and 1700 °C (Figure 4a). This trend mirrorsthe reduction in carbon observed in XPS in Figure 2j and the de-crease in PL defect signals in Figure 2k. Thus, it appears that car-bon migration may be the limiting process during annealing upto 1700 °C. This also underscores the need for sufficiently hightemperatures to effectively heal defects and remove impurities inhBN.The annihilation of VN and the removal of carbon may alsobe facilitated by active nitrogen generated by N2 dissociation dur-ing high-temperature annealing. While molecular nitrogen hasa high dissociation energy (≈9.76 eV),[60,61] thermal dissociationeffectively produces atomic nitrogen above 1600–1650 °C, as cal-culated (Experimental Section) and plotted in Figure 4b. The re-sulting active nitrogen can adsorb on and diffuse into the hBN,occupying VN sites and reactingwith heteroatoms to form volatilemolecules (e.g., CN) that then desorb from the hBN. This ex-plains the XRD and XPS observations, which revealed significantcrystallinity improvement along with a markedly reduced carbonin hBN after annealing at 1700 °C. Indeed, near-edge X-ray ab-sorption fine structure (NEXAFS) measurements have indicatedAdv. Sci. 2025, 12, e09354 e09354 (6 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 4. Mechanism of defect annihilation in hBN during high-temperature annealing. a) FWHM values from 𝜔-scan of hBN (0002) and 𝜑-scan ofhBN (10-10) as a function of temperature. b) Active nitrogen generated from thermal dissociation of N2 as a function of temperature at 90 kPa. Inset:Schematic of N2 decomposition into active nitrogen. c) B 1s NEXAFS spectra of as-grown (black) and 1700 °C annealed (green) hBN. d) Schematicillustration depicting the evolution in grains and grain boundaries of hBN before and after high-temperature annealing. Misalignment in hBN grains: 𝜃x(x = 0, 1; 𝜃1 < 𝜃0).the removal of the nitrogen vacancy in hBN after annealing at1700 °C (Figure 4c), evidenced by a sharper peak at 192 eV in B(1s) photoabsorption spectra and the disappearance of the peak at≈194 eV associated with VN defects.[62] Consequently, in additionto vacancymigration-mediated defect healing, the active nitrogenfrom thermal decomposition of N2 may also directly influence va-cancies as well as carbon and oxygen impurities present inweaklybound grain boundaries and unstable defects. These together ul-timately enhance lattice orientation and material purity by rotat-ing and enlarging the hBN grains, as depicted in Figure 4d.2.2.2. Control of Oxygen-Diffusion-Induced Damage fromUnderlying SapphireAnother crucial limitation arises with AlN templates from theunderlying sapphire substrate. HBN (≈12 nm) grown directlyon sapphire showed severe surface damage after annealing at1650 °C (Figure S5, Supporting Information). To further deter-mine whether the deterioration starts at the sapphire/BN inter-face or on the BN surface, we also annealed a thicker hBN film(≈30 nm) grown on sapphire at 1650 °C. The surface of the 30 nmhBN film on sapphire exhibited reduced damage after anneal-ing, preserving initial surface wrinkling across most areas andforming only a few holes (Figure S10, Supporting Information).However, the XPS O 1s peak was more intense, and the Al 2pand Al 2s peaks became detectable after annealing (Figure S10,Supporting Information). These results indicate that the damageoriginates from the bottom, where Al and O from the decom-posed sapphire diffuse into BN layer and react on the surfaceand at defects within the BN. Indeed, significant up-diffusionof Al and O impurities from the sapphire substrate into the as-grown hBN has been observed during MOVPE growth at 1380and 1500 °C,[19,63] even though the direct growth of hBN on sap-Adv. Sci. 2025, 12, e09354 e09354 (7 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 5. Oxygen-diffusion-induced degradation in annealed hBN on the AlN/sapphire template and its elimination on bulk AlN. a) Calculated diffusionlength of oxygen in AlN as a function of annealing temperature and time. The dashed line indicates the typical thickness (≈430 nm) of the AlN epilayerin AlN templates used. Inset: Schematic of hBN degradation on the AlN template caused by oxygen diffusion during annealing. b,c) AFM images (5 × 5μm2) of multilayer hBN after 1800 °C annealing for 20 min on (b) bulk AlN substrate and (c) AlN template. Surface damage and disappearance ofwrinkles are indicated by the dashed circle. d) XRD 2𝜃–𝜔 scans of the hBNmultilayers shown in (b) and (c). e) In-plane 𝜑-scan of the hBN {10-10} planefor the sample on bulk AlN shown in (b).phire at higher temperatures is possible.[17,18] In contrast, hBN onAlN templates withstood annealing up to 1700 °C—more than atleast 50 °C higher than the limit observed for hBN on sapphireunder our experimental conditions. Thus, a sufficiently thick AlNepitaxial layer pre-formed on the sapphire substrate effectivelyprevents oxygen diffusion-induced damage to hBN. The findingsabove also confirm that hBN itself remains stable at these ele-vated temperatures.Based on the temperature-dependent diffusion coeffi-cient (D) for oxygen in AlN,[64] we estimated the diffusionlength (𝜆) of oxygen in AlN using the Einstein equation𝜆 = 2√D × t with t being the annealing time. Figure 5a presentsthe calculated 𝜆 as a function of temperature and time, wherethe dashed line indicates the typical thickness (≈430 nm) ofthe AlN epilayer pre-formed on sapphire in this study. After1700 °C annealing for 20 min, the oxygen diffusion length issmaller than 430 nm, indicating that oxygen cannot penetratethrough the AlN layer to reach the overlying hBN. Since thermaldegradation is exponential, however, higher temperatures greatlyaccelerate the oxidation of the AlN layer, enabling more oxygento reach the hBN/AlN interface and then react with hBN. After1800 °C annealing for 20 min, the AlN 0002 reflection at 2𝜃= 36° exhibited significant broadening, while a new reflectionat 2𝜃 ≈ 57° assigned to AlON was observed (Figure 5d). Thesurface wrinkling disappeared, and the crystallinity degraded(Figure 5c,d). The XPS O 1s core-level spectrum also showeda pronounced increase in peak intensity, indicating substantialoxygen diffusion from the AlN/sapphire template into the over-lying hBN layer during the 1800 °C annealing process (FigureS11, Supporting Information). Therefore, oxygen diffusion fromthe decomposing sapphire through the AlN during annealingultimately sets an upper limit on AlN templates, while tempera-tures above ≈1600–1650 °C are needed for N2 dissociation andlikely also for vacancy migration.[54] This issue can be overcomeby growing hBN directly on more stable oxygen-free AlN bulksubstrates. Notably, even after annealing at 1800 °C, the hBNsurface maintained its characteristic wrinkling without damage(Figure 5b). Instead, its crystallinity was further improved rel-ative to that annealed at 1700 °C. The X-ray rocking curve ofthe hBN (0002) reflection for the sample annealed at 1800 °Cexhibited a narrower FWHM of ≈0.6° (Figure S12, SupportingInformation), which is smaller than that of hBN annealed at1700 °C and lower temperatures (Figure 4a). Furthermore,Figure 2c shows that the in-plane misalignment of hBN grownon the AlN bulk substrate and annealed at 1800 °C is lesspronounced than that of hBN on the AlN template and annealedat 1700 °C. These results highlight the enhanced crystal qualityof hBN achieved by higher-temperature annealing. With furtherprogress in nitride development, the thermally robust AlN bulksubstrate may become an interesting dielectric platform forhigh-quality wafer-scale hBN epitaxy.It is important to note that this work is fundamentally differ-ent from previous studies involving MOVPE and MBE growthof hBN on nitridated sapphire formed by intentional or unin-tentional surface nitridation.[17,65] In those cases, ultrathin layersof nominal AlxOyNz or AlN (000-1) potentially form on the sap-phire surface,[66] which differmarkedly in chemical composition,atomic configuration, and surface properties from the single-crystalline AlN (0001) investigated in this work. Moreover, theformation of only a few monolayers of AlxOyNz or AlN on sap-Adv. Sci. 2025, 12, e09354 e09354 (8 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comphire is insufficient to prevent the upward diffusion of oxygenand aluminum impurities from the degraded sapphire substrateinto the hBN at elevated temperatures.Additionally, very recent studies have demonstrated the for-mation of wafer-scale single-crystal hBN monolayers on insulat-ing substrates using a non-epitaxial “stamped-like” method.[67,68]In this method, single-crystalline hBN monolayers are first epi-taxially grown on both sides of a Cu foil that is loosely at-tached to the insulator surface. After growth, the Cu foil is re-moved by melting or etching, leaving the as-grown hBN layersattached to the underlying insulator. Unlike direct epitaxy on in-sulating substrates, this approach offers broader substrate com-patibility. Since the initial monolayer hBN is epitaxially grownon catalytic Cu foils, the resulting hBN typically exhibits bet-ter crystallinity and larger grain size relative to hBN grown di-rectly on catalyst-free insulating substrates. However, the non-epitaxial method still relies on the epitaxy of hBN layers on cat-alytic metals. Amajor limitation lies in thickness control becausecatalyst-assisted growth on the metal surface is generally effec-tive only for forming a single layer or, at most, a few mono-layers. Indeed, the two referenced studies only demonstratedthe successful preparation of monolayer hBN. Thus, achievinghigh-quality multilayer hBN on insulating substrates using thismethod remains potentially challenging. In contrast, the ap-proach developed in this work enables flexible control over hBNthickness, while also being industry-compatible and readily scal-able. Overall, the reported non-epitaxial technique and the ap-proach developed in this study can be considered complemen-tary strategies for forming hBN layers on insulating substrates.The most suitable strategy will depend on specific application re-quirements, such as layer thickness, grain size, interface quality,etc.3. ConclusionIn summary, we reported the successful synthesis of large-area highly-oriented multilayer hBN on single-crystal AlN (0001)via high-temperature MOVPE and subsequent annealing upto 1800 °C. High-temperature pulsed-mode MOVPE, whichovercomes large migration barriers of boron species on thegrowing surface and suppresses unfavorable gas-phase reac-tions between sources, ensures the initial epitaxy of hBN onAlN. Ultrahigh-temperature annealing reduces defects via de-fect migration and guides the prior MOVPE-grown hBN grainsinto better alignment. Combining the high-temperature pro-cess with thermally stable AlN as a substrate is key to ob-taining the epitaxial hBN multilayers on dielectric substrateswith both superior out-of-plane and in-plane alignments. Ele-vated temperatures also reduce carbon levels so that individualquantum emitters in hBN multilayers can be observed. Theseemitters exhibited stable room-temperature luminescence, anarrow spectral localization at ≈578 ± 5 nm, and small ZPLwidths down to 1.44 meV. Furthermore, the achieved hBNpresented a dielectric breakdown strength of 10.4 MV cm−1,which is comparable to that of exfoliated hBN from bulk crys-tals. This industry-compatible approach for scalable, high-qualitymultilayer hBN fabrication on dielectric substrates is expectedto accelerate the application of 2D photonics and electron-ics.4. Experimental SectionPreparation of High-Quality AlN-Based Substrate with Atomically Flat Sur-face: Single-crystal AlN templates (high-crystallinity AlN epitaxial layeron c-plane sapphire) were prepared by sputtering followed by high-temperature annealing.[22] The typical thickness of the AlN epilayer is≈430 nm with threading dislocation densities (TDs) of ≈108 cm−2. Ad-ditionally, commercial bulk AlN substrates (HexTech, Inc.) with TDs be-low 104 cm−2 were used for hBN epitaxy. Prior to growth, both tem-plates and substrates of AlN were annealed at 1350 °C in H2 for 3 min toachieve a clean, atomically flat step-like surface, as representatively shownin Figure 1a. Typically, AlN-based substrates of ≈1 × 1 cm2 were used forthe MOVPE growth and subsequent annealing of hBN.MOVPE of hBN and Post-Growth Annealing: Multilayer hBN wasgrown on templates and substrates of AlN byMOVPE using TEB and NH3as B and N precursors, respectively. As reported in our prior studies,[26,69]TEB andNH3 sources were alternately injected into the reactor in a pulsed-mode growth process. Compared with simultaneous TEB andNH3 supply,pulsed-mode MOVPE growth helps suppress adverse parasitic reactionsbetween TEB and NH3 in the vapor phase, resulting in hBN growth withimproved quality. The pulsed duration for each cycle was 2 s for TEB and1 s for NH3 without interruption. The growth was carried out at 1280–1380 °C in an H2 atmosphere at ≈4 kPa. The flow rate of TEB ranged from15 to 30 μmol min−1, while the NH3 flux was accordingly modulated tomaintain a nominal V/III ratio of 3000. The optimized TEB flow rate andtemperature for hBN growth on AlN were 15 μmol min−1 and 1380 °C,respectively. For comparison, hBNmultilayers were also directly grown onc-plane sapphire substrates using the same method. After pulsed-modeMOVPE, the as-grown hBN samples were annealed in a repurposed phys-ical vapor transport system with a face-to-face configuration (Figure 1c),which minimizes surface damage and allows the vapor pressure of onelayer to stabilize the other.[70] Annealing was typically conducted in N2 at90 kPa for 20 min at temperatures between 1500 and 1800 °C. All temper-atures in this study were recorded by thermocouple readings.Characterization: Symmetric XRD measurements (2𝜃-𝜔 and 𝜔 scans)were conducted on a PANalytical X’pert five-axis high-resolution systemusing an open detector and a Cu K𝛼1 source (𝜆 = 1.5406 Å, 45 kV, 40 mA).In-plane XRD experiments (𝜑 scans) were performed using Rigaku Smart-Lab with a Cu K𝛼 source (𝜆 = 1.5418 Å, 45 kV, 200 mA).Surface morphology was studied by tapping mode atomic force mi-croscopy (NanoNavi, SII NanoTech). The crystalline quality and mi-crostructure of hBN were examined by transmission electron microscopy(TEM, H-9500, Hitachi) operated at 200 kV. TEM lamellas were made us-ing a dual-beam focused ion beam (DB-FIB, NB5000) milling via the lift-out method, with protective deposits of platinum, tungsten, and carbonemployed during Ga ion milling. Chemical composition and fine structurewere investigated by EELS using a JEOL instrument (JEMARM200F) oper-ated at 200 kVwith a≈2 nm focal spot, which is sufficiently localized to dis-tinguish the signal from the 2D hBN multilayers. RHEED measurementswere conducted using an electron beam energy of 20 keV. As reported inthe previous study,[71] a well-defined MoSe2 monolayer epitaxially grownon a GaAs (111) B substrate was used as the reference sample for RHEEDcalibration. NEXAFSmeasurements were carried out at the beamline BL12at Saga Light Source in Japan using a synchrotron source with an excita-tion photon energy ranging from 60 to 800 eV. Chemical and bonding sig-natures were analyzed by XPS (Thermo Scientific) using a monochromaticAl K𝛼 source (h𝜈 = 1486.7 eV). All XPS data in this study were recordedat a 90° take-off angle using the C1s core-level at 248.8 eV for calibration.FTIR (FTIR-6100, JASCO) of BN was acquired in a reflection mode witha wavenumber resolution of 4 cm−1. All spectra were normalized to thereflectance spectrum of an aluminum mirror.Photoluminescence (PL) measurements were carried out using acontinuous-wave 532 nm laser for excitation. PLmapping and spectra wereacquired with a lab-built confocal micro-PL system at room and cryogenictemperatures (detection area: ≈0.8 or ≈1.6 μm), which allowed raster-scanned PL mapping using a motorized XY stage. Spectra were recordedon a JASCO CT-25 spectrometer equipped with a CCD (Andor DU970PC-UVB). For low-temperature measurements, samples were cooled in a liq-Adv. Sci. 2025, 12, e09354 e09354 (9 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comuid helium continuous-flow cryostat (JANIS ST-500) with a window lead-ing to a lab-built confocal microscope setup. To investigate the quantumnature of emitters, the second-order photon correlation function g(2)(𝜏)was recorded at room temperature via a Hanbury-Brown and Twiss setupconsisting of a fiber-coupled 50/50 beam splitter connected to two single-photon counting modules. The detected events were input to a time-correlated single-photon counting device (TimeHarp 260 P, PicoQuant;Base resolution: 25 ps) for monitoring the second order correlation func-tion and temporal stability of the emitter luminescence. A relatively broad100 nm bandpass filter covering the ZPL and the PSB was used for time-correlated single-photon counting measurements. The second-order au-tocorrelation data were fitted using a two-level model with the followingequation:g(2) (𝜏) = A(1 − k × exp(−||t − t0||𝜏)(1)where A is the uncorrelated amplitude, k is a pre-factor related to single-to-background ratio, t0 is the zero-delay offset, and 𝜏 is the excited statelifetime. Note that neither background subtraction nor additional spectralfiltering was carried out for the autocorrelation analysis. The overall timeresolution of themeasurement system,≈430 ps, also affects themeasuredsingle-photon purity. Therefore, the actual quantum emission purity couldbe higher than the measured one.Theoretical Calculations—N2 Dissociation at High Temperature: Nitro-gen molecules indeed could provide relevant active nitrogen through ther-mal dissociation at high temperatures above ≈1600–1650 °C. Assuminghead-on collisions, at least two nitrogen molecules with an energy greaterthan half of the dissociation energy Ed = 9.76 eV/2 = 4.88 eV are needed.The fraction of molecules (Nd) with an energy above Ed can be obtainedfrom the Maxwell–Boltzmann distribution in its energy formulation by in-tegration from Ed to ∞, as noted using the following equation:Nd = 1 − ∫Ed0 2√E𝜋 (kBT)exp(− EkBT)dE= 1 − [erf(√EdkBT)− 2√𝜋√EdkBTexp(−EdkBT)] (2)where kB is the Boltzmann constant, T is the temperature. The error func-tion can be replaced by 1, since erf (√EdkBT) ≈ 1 for Ed = 4.88 eV and thetemperature span considered in this study. The number of molecules ineach given value comes from the ideal gas equation:N =pkBT(3)The absolute number of active nitrogen (Nnitro), arising from N2 dis-sociation, can thus be derived using Equation (4) by twice the product ofEquation (2) and the total number of molecules at a given pressure, 90 kPaused in high-temperature annealing in this study.Nnitro = 2 × Nd =4√𝜋√EdkBTexp(−EdkBT)×pkBT(4)Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsThis work was supported by the CIRFE Young Researchers Support Projectand JSPS KAKENHI Grant No. JPJSJRP 20221603. M.I. acknowledges sup-port from JSPS KAKENHI Grant No. JP23K03272. X.Y. thanks E. Li at theUniversity of Tsukuba for helpful discussions and support; Y. Furusawaand X. Li for assistance with related experiments.Conflict of InterestThe authors declare no conflict of interestData Availability StatementThe data that support the findings of this study are available from the cor-responding author upon reasonable request.Keywordsepitaxy and annealing, hexagonal boron nitride, high temperature, multi-layer, single-crystal AlNReceived: May 23, 2025Revised: August 19, 2025Published online: September 29, 2025[1] C. R. Dean, A. F. Young, I. Meric, C. Lee, L. Wang, S. Sorgenfrei, K.Watanabe, T. Taniguchi, P. Kim, K. L. Shepard, J. Hone, Nat. Nan-otechnol. 2010, 5, 722.[2] S. Fukamachi, P. Solís-Fernández, K. Kawahara, D. Tanaka, T. Otake,Y.-C. Lin, K. Suenaga, H. Ago, Nat. Electron. 2023, 6, 126.[3] T. T. Tran, K. Bray, M. J. Ford, M. Toth, I. Aharonovich,Nat. Nanotech-nol. 2016, 11, 37.[4] Y. Sasama, T. Kageura, M. Imura, K. Watanabe, T. Taniguchi, T.Uchihashi, Y. Takahide, Nat. Electron. 2022, 5, 37.[5] Y. Kobayashi, K. Kumakura, T. Akasaka, T. Makimoto, Nature 2012,484, 223.[6] Z. Liu, Y. Gong, W. Zhou, L. Ma, J. Yu, J. C. Idrobo, J. Jung, A. H.MacDonald, R. Vajtai, J. Lou, P. M. Ajayan, Nat. Commun. 2013, 4,2541.[7] J. Li, R. Dahal, S. Majety, J. Y. Lin, H. X. Jiang, Nucl. Instrum. MethodsPhys. Res. A 2011, 654, 417.[8] L. Wang, X. Xu, L. Zhang, R. Qiao, M.Wu, Z. Wang, S. Zhang, J. Liang,Z. Zhang, Z. Zhang, W. Chen, X. Xie, J. Zong, Y. Shan, Y. Guo, M.Willinger, H. Wu, Q. Li, W. Wang, P. Gao, S. Wu, Y. Zhang, Y. Jiang, D.Yu, E. Wang, X. Bai, Z.-J. Wang, F. Ding, K. Liu, Nature 2019, 570, 91.[9] L. Song, L. Ci, H. Lu, P. B. Sorokin, C. Jin, J. Ni, A. G. Kvashnin, D. G.Kvashnin, J. Lou, B. I. Yakobson, P. M. Ajayan, Nano Lett. 2010, 10,3209.[10] G. Lu, T. Wu, Q. Yuan, H. Wang, H. Wang, F. Ding, X. Xie, M. Jiang,Nat. Commun. 2015, 6, 6160.[11] J. S. Lee, S. H. Choi, S. J. Yun, Y. I. Kim, S. Boandoh, J.-H. Park, B.G. Shin, H. Ko, S. H. Lee, Y.-M. Kim, Y. H. Lee, K. K. Kim, S. M. Kim,Science 2018, 362, 817.[12] T.-A. Chen, C.-P. Chuu, C.-C. Tseng, C.-K. Wen, H.-S. P. Wong, S. Pan,R. Li, T.-A. Chao, W.-C. Chueh, Y. Zhang, Q. Fu, B. I. Yakobson, W.-H.Chang, L.-J. Li, Nature 2020, 579, 219.[13] K. Y. Ma, L. Zhang, S. Jin, Y. Wang, S. I. Yoon, H. Hwang, J. Oh, D. S.Jeong, M. Wang, S. Chatterjee, G. Kim, A.-R. Jang, J. Yang, S. Ryu, H.Y. Jeong, R. S. Ruoff, M. Chhowalla, F. Ding, H. S. Shin, Nature 2022,606, 88.[14] A.-R. Jang, S. Hong, C. Hyun, S. I. Yoon, G. Kim, H. Y. Jeong, T. J.Shin, S. O. Park, K. Wong, S. K. Kwak, N. Park, K. Yu, E. Choi, A.Mishchenko, F. Withers, K. S. Novoselov, H. Lim, H. S. Shin, NanoLett. 2016, 16, 3360.Adv. Sci. 2025, 12, e09354 e09354 (10 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.com[15] K. Nakamura, J. Electrochem. Soc. 1986, 133, 1120.[16] K. Akiyama, T. Araki, H. Murakami, Y. Kumagai, A. Koukitu, Phys. Sta-tus Solidi (C) 2007, 4, 2297.[17] R. Page, J. Casamento, Y. Cho, S. Rouvimov, H. G. Xing, D. Jena, Phys.Rev. Mater. 2019, 3, 064001.[18] F. Liu, X. Rong, Y. Yu, T. Wang, B. W. Sheng, J. Q. Wei, S. F. Liu, J.J. Yang, F. Bertram, F. J. Xu, X. L. Yang, Z. H. Zhang, Z. X. Qin, Y. T.Zhang, B. Shen, X. Q. Wang, Appl. Phys. Lett. 2020, 116, 142104.[19] S. J. Grenadier, A. Maity, J. Li, J. Y. Lin, H. X. Jiang, Appl. Phys. Lett.2018, 112, 162103.[20] K. Ahmed, R. Dahal, A. Weltz, J.-Q. Lu, Y. Danon, I. B. Bhat, Appl. Phys.Lett. 2016, 109, 113501.[21] S. Liu, T. Lu, Y. Yuan, J. Cao, W. Wan, T. Li, T. Wang, Z. Chen, X. Sun,X. Wang, Adv. Funct. Mater. 2025, 2503580, 2503580.[22] H. Miyake, C.-H. Lin, K. Tokoro, K. Hiramatsu, J. Cryst. Growth 2016,456, 155.[23] M. Chubarov, H. Pedersen, H. Högberg, J. Jensen, A. Henry, Cryst.Growth Des. 2012, 12, 3215.[24] N. Coudurier, R. Boichot, F. Mercier, R. Reboud, S. Lay, E. Blanquet,M. Pons, Phys. Procedia 2013, 46, 102.[25] N. Coudurier, M. Chubarov, R. Boichot, F. Mercier, E. Blanquet, R.Reboud, S. Lay, A. Crisci, S. Coindeau, T. Encinas, M. Pons, Cryst.Res. Technol. 2016, 51, 231.[26] X. Yang, S. Nitta, M. Pristovsek, Y. Liu, Y. Liao, M. Kushimoto, Y.Honda, H. Amano, 2D Mater. 2019, 7, 015004.[27] S. Deng, V. Berry,Mater. Today 2016, 19, 197.[28] X. Li, S. Sundaram, Y. El Gmili, T. Ayari, R. Puybaret, G. Patriarche, P.L. Voss, J. P. Salvestrini, A. Ougazzaden, Cryst. Growth Des. 2016, 16,3409.[29] K. Bera, D. Chugh, A. Patra, H. H. Tan, C. Jagadish, A. Roy, Solid StateCommun. 2020, 310, 113847.[30] J. Li, A. Samad, Y. Yuan, Q. Wang, M. N. Hedhili, M. Lanza, U.Schwingenschlögl, I. Abate, D. Akinwande, Z. Liu, B. Tian, X. Zhang,Nat. Commun. 2024, 15, 8589.[31] Y. Wang, C. Zhao, X. Gao, L. Zheng, J. Qian, X. Gao, J. Li, J. Tang, C.Tan, J. Wang, X. Zhu, J. Guo, Z. Liu, F. Ding, H. Peng, Nat. Mater.2024, 23, 1495.[32] S. H. Lee, H. Jeong, O. F. N. Okello, S. Xiao, S. Moon, D. Y. Kim, G.-Y.Kim, J.-I. Lo, Y.-C. Peng, B.-M. Cheng, H. Miyake, S.-Y. Choi, J. K. Kim,Sci. Rep. 2019, 9, 10590.[33] R. M. Ribeiro, N. M. R. Peres, Phys. Rev. B 2011, 83, 235312.[34] W.-Q. Han, L. Wu, Y. Zhu, K. Watanabe, T. Taniguchi, Appl. Phys. Lett.2008, 93, 223103.[35] L. Wang, J. Qi, W. Wei, M. Wu, Z. Zhang, X. Li, H. Sun, Q. Guo, M.Cao, Q. Wang, C. Zhao, Y. Sheng, Z. Liu, C. Liu, M. Wu, Z. Xu, W.Wang, H. Hong, P. Gao, M. Wu, Z.-J. Wang, X. Xu, E. Wang, F. Ding,X. Zheng, K. Liu, X. Bai, Nature 2024, 629, 74.[36] R. S. Pease, Acta Crystallogr. 1952, 5, 356.[37] N. G. Chopra, R. J. Luyken, K. Cherrey, V. H. Crespi, M. L. Cohen, S.G. Louie, A. Zettl, Science 1995, 269, 966.[38] W. Liu, N.-J. Guo, S. Yu, Y. Meng, Z.-P. Li, Y.-Z. Yang, Z.-A. Wang, X.-D.Zeng, L.-K. Xie, Q. Li, J.-F. Wang, J.-S. Xu, Y.-T. Wang, J.-S. Tang, C.-F.Li, G.-C. Guo,Mater. Quantum. Technol. 2022, 2, 032002.[39] N. Mendelson, D. Chugh, J. R. Reimers, T. S. Cheng, A. Gottscholl, H.Long, C. J. Mellor, A. Zettl, V. Dyakonov, P. H. Beton, S. V. Novikov,C. Jagadish, H. H. Tan, M. J. Ford, M. Toth, C. Bradac, I. Aharonovich,Nat. Mater. 2021, 20, 321.[40] C. Jara, T. Rauch, S. Botti, M. A. L. Marques, A. Norambuena, R. Coto,J. E. Castellanos-Águila, J. R. Maze, F. Munoz, J. Phys. Chem. A 2021,125, 1325.[41] X. Li, G. D. Shepard, A. Cupo, N. Camporeale, K. Shayan, Y. Luo, V.Meunier, S. Strauf, ACS Nano 2017, 11, 6652.[42] Z.-Q. Xu, C. Elbadawi, T. T. Tran, M. Kianinia, X. Li, D. Liu, T. B.Hoffman, M. Nguyen, S. Kim, J. H. Edgar, X. Wu, L. Song, S. Ali, M.Ford, M. Toth, I. Aharonovich, Nanoscale 2018, 10, 7957.[43] N. Mendelson, Z.-Q. Xu, T. T. Tran, M. Kianinia, J. Scott, C. Bradac, I.Aharonovich, M. Toth, ACS Nano 2019, 13, 3132.[44] N. V. Proscia, Z. Shotan, H. Jayakumar, P. Reddy, C. Cohen,M. Dollar,A. Alkauskas, M. Doherty, C. A. Meriles, V. M. Menon, Optica 2018,5, 1128.[45] A. Dietrich, M. Bürk, E. S. Steiger, L. Antoniuk, T. T. Tran, M. Nguyen,I. Aharonovich, F. Jelezko, A. Kubanek, Phys. Rev. B 2018, 98,081414.[46] D. Zhong, S. Gao,M. Saccone, J. R. Greer,M. Bernardi, S. Nadj-Perge,A. Faraon, Nano Lett. 2024, 24, 1106.[47] T. T. Tran, C. Elbadawi, D. Totonjian, C. J. Lobo, G. Grosso, H. Moon,D. R. Englund, M. J. Ford, I. Aharonovich, M. Toth, ACS Nano 2016,10, 7331.[48] Y. Chen, M. T. Westerhausen, C. Li, S. White, C. Bradac, A. Bendavid,M. Toth, I. Aharonovich, T. T. Tran, ACS Appl. Nano Mater. 2021, 4,10449.[49] N. R. Jungwirth, B. Calderon, Y. Ji, M. G. Spencer, M. E. Flatté, G. D.Fuchs, Nano Lett. 2016, 16, 6052.[50] H. Akbari, W.-H. Lin, B. Vest, P. K. Jha, H. A. Atwater, Phys. Rev. Appl.2021, 15, 014036.[51] M. Fischer, J. M. Caridad, A. Sajid, S. Ghaderzadeh, M. Ghorbani-Asl,L. Gammelgaard, P. Bøggild, K. S. Thygesen, A. V. Krasheninnikov, S.Xiao, M. Wubs, N. Stenger, Sci. Adv. 2021, 7, abe7138.[52] A. Chatterjee, A. Biswas, A. S. Fuhr, T. Terlier, B. G. Sumpter, P. M.Ajayan, I. Aharonovich, S. Huang, Sci. Adv. 2025, 11, adv2899.[53] K. Li, T. J. Smart, Y. Ping, Phys. Rev. Mater. 2022, 6, L042201.[54] L. Weston, D. Wickramaratne, M. Mackoit, A. Alkauskas, C. G. Vande Walle, Phys. Rev. B 2018, 97, 214104.[55] P. Vinchon, X. Glad, G. R. Bigras, R. Martel, L. Stafford, Nat. Mater.2021, 20, 49.[56] X.-M. Bai, A. F. Voter, R. G. Hoagland, M. Nastasi, B. P. Uberuaga,Science 2010, 327, 1631.[57] Y. Liu, X. Zou, B. I. Yakobson, ACS Nano 2012, 6, 7053.[58] A. Zobelli, C. P. Ewels, A. Gloter, G. Seifert, Phys. Rev. B 2007, 75,094104.[59] H. Nozaki, S. Itoh, Phys. Chem. Solids 1996, 57, 41.[60] J. H. Thorpe, J. L. Kilburn, D. Feller, P. B. Changala, D. H. Bross, B.Ruscic, J. F. Stanton, J. Chem. Phys. 2021, 155, 184109.[61] J.-Y. Roncin, F. Launay, M. Larzilliere, Phys. Rev. Lett. 1984, 53, 159.[62] R. Peter, A. Bozanic, M. Petravic, Y. Chen, L.-J. Fan, Y.-W. Yang, J. Appl.Phys. 2009, 106, 083523.[63] X. Yang, S. Nitta, M. Pristovsek, Y. Liu, K. Nagamatsu, M. Kushimoto,Y. Honda, H. Amano, Appl. Phys. Express 2018, 11, 051002.[64] H. Solmon, D. Robinson, R. Dieckmann, J. Am. Ceram. Soc. 1994, 77,2841.[65] M. Snure, Q. Paduano, A. Kiefer, Cryst. Growth 2016, 436, 16.[66] S. Mohn, N. Stolyarchuk, T. Markurt, R. Kirste, M. P. Hoffmann, R.Collazo, A. Courville, R. Di Felice, Z. Sitar, P. Vennéguès, M. Albrecht,Phys. Rev. Appl. 2016, 5, 054004.[67] J. Li, Y. Yuan, M. Lanza, I. Abate, B. Tian, X. Zhang, Adv. Mater. 2024,36, 2310921.[68] F. Zeng, R. Wang, W. Wei, Z. Feng, Q. Guo, Y. Ren, G. Cui, D. Zou, Z.Zhang, S. Liu, K. Liu, Y. Fu, J. Kou, L. Wang, X. Zhou, Z. Tang, F. Ding,D. Yu, K. Liu, X. Xu, Nat. Commun. 2023, 14, 6421.[69] X. Yang, S. Nitta, K. Nagamatsu, S.-Y. Bae, H.-J. Lee, Y. Liu, M.Pristovsek, Y. Honda, H. Amano, J. Cryst. Growth 2018, 482, 1.[70] K. Hamanaka, T. Tachiki, T. Uchida, Jpn. J. Appl. Phys. 2009, 48,125502.[71] A. Ohtake, X. Yang, Cryst. Growth Des. 2023, 23, 5001.Adv. Sci. 2025, 12, e09354 e09354 (11 of 11) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2025, 46, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202509354 by National Institute For, Wiley Online Library on [11/12/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.com Highly Oriented Epitaxial Hexagonal Boron Nitride Multilayers on High-Temperature-Resistant Single-Crystal Aluminum Nitride (0001) 1. Introduction 2. Results and Discussion 2.1. MOVPE of hBN Multilayers on AlN and Post-Growth Annealing Control 2.2. Mechanism of Defect Healing in hBN via High Temperature Annealing 2.2.1. Vacancy Migration-Mediated Defect Healing 2.2.2. Control of Oxygen-Diffusion-Induced Damage from Underlying Sapphire 3. Conclusion 4. Experimental Section Supporting Information Acknowledgements Conflict of Interest Data Availability Statement Keywords