{"id":"b3778d4f-7d0e-44e9-a648-ffd8adb75115","arxiv_id":"1908.01045","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"MBE-grown ScN on GaN, AlN, and 6H-SiC forms cubic rocksalt (111) twinned films with degenerate n-type carrier concentrations around 1e20 cm^-3.","lead":"Scandium nitride films grown by molecular beam epitaxy on hexagonal silicon carbide, gallium nitride, and aluminum nitride form a cubic crystal structure with a (111) orientation and twinned domains. The results provide a growth template for integrating ScN and ScAlN alloys with III-nitride devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the cubic rocksalt (111) twinned ScN claim is supported by mutually consistent RHEED, EBSD, and XRD data, with the GaN XRD overlap honestly disclosed.","rationale":"The reader's weakest-assumption analysis correctly identifies the ScN-on-GaN XRD overlap as the main evidential gap, and the paper itself acknowledges that TEM would be needed for definitive interface and twin-nucleation questions. On review, however, this gap does not rise to a load-bearing flaw. The RHEED pattern is a distinct, previously reported signature of cubic (111) twinning, the EBSD data are indexed to cubic ScN with pole figures showing the expected six-fold symmetry from twin domains, and the XRD phi scans on SiC provide an independent confirmation of the twin model. The absence of a resolved ScN (111) peak on GaN is explained quantitatively by the near-coincidence with GaN (002), and the authors state this explicitly. The electrical half of the claim is also sound: constant carrier concentration down to 20 K on both GaN/SiC and AlN/Al2O3 supports degeneracy, and the AlN measurement is on an insulating substrate, so parallel conduction through a GaN template cannot be the sole source of the signal. Overall, the evidence is coherent, the limitations are transparently stated, and no internal inconsistency or unsupported leap is apparent. The reader's ACCEPT verdict is therefore unchanged.","tokens_in":7630,"tokens_out":5962,"duration_ms":67005,"concrete_test":"Perform cross-sectional TEM with selected-area diffraction (SAD) and STEM-EDS on the ScN/GaN/SiC stack. If the layer is verified to be single-phase rocksalt ScN with (111) out-of-plane orientation and only twin-related domains, with no interfacial phase or minority orientation, the remaining XRD-overlap ambiguity is closed; if a minority orientation appears, the 'exclusive (111)' claim would need qualification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No significant objection identified. The paper's central claim—that MBE-grown ScN on GaN, AlN, and SiC forms cubic rocksalt films with exclusive (111) out-of-plane orientation and twin-related in-plane domains—is supported by mutually consistent in-situ RHEED, EBSD, and XRD evidence. The one genuinely soft spot is that on GaN/SiC the ScN (111) XRD reflection overlaps the GaN (002) peak, so for that substrate the orientation and phase assignment rests on RHEED and EBSD alone. This is not fatal: the RHEED pattern shows the characteristic cubic twin signature with paired (111) and (002) spots, EBSD indexing to Fm-3m ScN yields continuous (111) orientation with 60° in-plane misorientation, and the paper explicitly discloses both the overlap and the need for future TEM. The electrical claim is independently supported by temperature-independent Hall data on AlN/Al2O3, which is an insulating substrate, so the degenerate n-type carrier concentration does not stand or fall with the GaN measurement. The paper's self-identified limitation and future TEM request are statements of scope, not internal inconsistencies. The central argument therefore holds under scrutiny.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports plasma-assisted MBE growth of ~30 nm ScN films on c-plane GaN/SiC, AlN/Al2O3, and 6H-SiC substrates, and characterizes their structure, surface morphology, and transport. The central claim is that ScN adopts the rocksalt (cubic) structure with exclusive (111) out-of-plane orientation and two twin-related in-plane domains on all three hexagonal substrates. Supporting evidence includes in-situ RHEED patterns with paired (111)/(002) spots, EBSD indexing to Fm-3m ScN with 60° in-plane misorientation, XRD 2theta-omega and phi scans on SiC and AlN, and AFM showing sub-nm roughness. Hall measurements on GaN/SiC and AlN/Al2O3 show temperature-independent n-type carrier concentrations near 1×10^20 cm^-3 with mobilities of 11–23 cm^2/Vs, interpreted as degenerate doping. The paper also offers qualitative bonding arguments for the stability of rocksalt ScN and speculates about polar/non-polar interface charges as a possible source of mobile carriers.","tokens_in":7839,"tokens_out":6986,"duration_ms":74444,"significance":"The structural result is significant for the III-nitride community because (111) ScN is nearly lattice-matched to GaN and is the limiting binary for emerging ScAlN/GaScN ferroelectric and piezoelectric heterostructures. The manuscript's main strengths are the complementary, mutually consistent characterization (RHEED, EBSD, XRD phi scans), the explicit and honest disclosure of the unresolved ScN (111) reflection on GaN templates, and the electrical measurement on an insulating AlN/Al2O3 substrate, which independently supports the degenerate carrier concentration. The paper makes no over-claim about the origin of the carriers; the polar-interface explanation is clearly labeled speculative, and the need for TEM is acknowledged. These factors make the central claim credible even though composition, stoichiometry, and minority-phase exclusion are not exhaustively addressed.","major_comments":[],"minor_comments":[{"comment":"The sentence 'Grain misorientation statistics (not shown) indicated the ScN epitaxial film was entirely (111) oriented' is central to the exclusive-orientation claim, especially for the GaN template where XRD is inconclusive; please either include these statistics as a figure/table or provide a summary in the text.","section":"§2 (EBSD)"},{"comment":"The unresolved ScN (111) reflection on GaN/SiC is disclosed in the XRD section, but the conclusions repeat the claim of 'solely cubic twinned (111) orientation' without repeating this caveat; please add one sentence in the conclusions noting that the GaN-based assignment relies on RHEED and EBSD, not XRD.","section":"§3 (XRD, Fig. 3C)"},{"comment":"The text states that 'XPS results (not shown)' verify Sc–N bonding and that 'optical absorption measurements (not shown)' show a band edge near 2.1 eV; these data should either be presented in a supplementary figure or the statements removed, because claims based on unavailable data cannot be evaluated.","section":"§2 (XPS/absorption)"},{"comment":"Report error bars or measurement uncertainties for the Hall carrier concentration, Hall mobility, and rms roughness values; single-point values are not sufficient for quantitative comparison with literature.","section":"§3 (Hall and AFM)"},{"comment":"Please state whether the GaN template could contribute a parallel conduction path in the ScN/GaN Hall measurement; the similarity of the carrier concentration to the AlN/Al2O3 sample suggests this is not a problem, but the assumption should be made explicit.","section":"§3 (Hall on GaN)"},{"comment":"In the caption, 'domains with 180˚ orientation device' appears to be a typo for '180° orientation difference'; please correct it.","section":"Fig. 2 caption"},{"comment":"The grant number is listed as 'AFSOR FA9550-17-1-0048'; the standard abbreviation is AFOSR, so please correct the agency name.","section":"Acknowledgments"}],"recommendation":"minor_revision","confidential_remarks":"This is a well-scoped growth-and-characterization report that fits the journal well. The main risk is the strength of the exclusive-(111) claim on GaN templates, but the RHEED and EBSD evidence is mutually consistent and the XRD overlap is honestly disclosed. I would not require new experiments; the requested clarifications and presentation fixes are sufficient."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis one is worth a careful read. It reports the first MBE growth of ScN on AlN and the first in-situ RHEED observation of cubic twinned ScN(111), with EBSD and XRD corroborating the twin-domain picture. That is a genuine step toward integrating ScN and ScAlN with III-nitride devices.\n\nThe central structural claim holds up. RHEED shows the paired (111)/(002) spots expected for cubic twinning on a hexagonal substrate; EBSD on SiC and GaN/SiC templates indexes to FCC ScN and gives sixfold (111) pole figures with 60-degree in-plane misorientation; XRD on SiC and AlN resolves the ScN(111) peak, and the (224) phi scan shows six peaks. On GaN, the ScN peak is buried under the GaN(002) peak, and the authors say so openly. That is the real soft spot: for the GaN case the phase and orientation assignment rests on RHEED and EBSD alone. The evidence is mutually consistent and reasonable, so the claim is credible, but TEM or reciprocal-space mapping would remove the residual doubt. The paper itself asks for future TEM.\n\nThe electrical data is thinner. The Hall measurement on AlN/Al2O3 is not contaminated by a conducting substrate, and the temperature-independent high n-type concentration is credible. But there are no error bars, no direct composition measurement, no stoichiometry check, and the XPS and optical data are mentioned but not shown. The ionized-impurity scattering estimate relies on an assumed mass, and the polar-interface explanation is explicitly speculative. These are minor-to-moderate issues, not load-bearing.\n\nThe citation pattern looks fine. Self-citations are motivational or a textbook reference for the polar-interface aside; the structural claim is not built on them.\n\nWho is this for? Anyone working on ScN, ScAlN, or III-nitride heteroepitaxy. It is a solid experimental paper that deserves a serious referee and acceptance after minor revision. I would ask for error bars, a supplement with the XPS and optical data, and softer wording about 'solely' on GaN where the XRD does not resolve the film peak. The core result is not in doubt.","headline":"Solid MBE growth study with first ScN-on-AlN and twinned-cubic (111) ScN evidence; publish after minor revision.","tokens_in":8400,"tokens_out":2921,"would_cite":true,"duration_ms":30434,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Plasma-assisted MBE of scandium nitride on hexagonal GaN, AlN, and 6H-SiC produces smooth, twinned cubic rocksalt films with (111) orientation and degenerate n-type conduction.","keywords":["scandium nitride","molecular beam epitaxy","rocksalt structure","cubic twinning","III-nitride semiconductors","RHEED","EBSD","degenerate n-type doping"],"falsifier":"Cross-sectional transmission electron microscopy and reciprocal-space mapping of the ScN/GaN and ScN/AlN interfaces, combined with a high-resolution scan for non-(111) reflections such as ScN (200) or (220), would reveal whether minority orientations or secondary phases exist; a companion measurement of oxygen concentration versus carrier density would test the oxygen-donor explanation.","tokens_in":1815,"feed_emoji":"🔬","tokens_out":1890,"duration_ms":80928,"temperature":0.7,"pith_summary":"The paper reports that scandium nitride, a transition-metal nitride usually stable in the cubic rocksalt phase, can be grown by plasma-assisted molecular beam epitaxy as extremely smooth epitaxial thin films on the hexagonal surfaces of GaN, AlN, and 6H-SiC. The central claim is that the films adopt a face-centered cubic structure with the (111) axis pointing along the substrate c-axis, and that this (111) orientation is twinned, meaning in-plane domains are rotated by 60 degrees with respect to one another. On all three substrates the films are unintentionally doped n-type with a nearly temperature-independent electron concentration around $10^{20}$ cm$^{-3}$, i.e. degenerate metallic behavior. Because ScN is the endpoint binary for Sc-containing III-nitride alloys such as ScAlN and GaScN, which are of current interest for piezoelectric and ferroelectric applications, the result provides a baseline for integrating cubic ScN layers into the III-nitride device family.","feed_headline":"Cubic ScN grows on hexagonal GaN, AlN, and SiC","feed_subtitle":"MBE films align (111) to the c-axis, form twin domains, and stay degenerately n-type.","key_machinery":"The load-bearing observation is the in-situ RHEED signature: during growth on the (110) azimuth, the hexagonal substrate's first-order streaks evolve into pairs of symmetric spots indexed as (111) and (002) reflections of a face-centered cubic lattice, giving four spots on each side of the zone center instead of two. This pattern, previously reported for twinned Cu(111) films on sapphire, is the signature of a (111)-oriented cubic crystal with twin domains. The twinned microstructure is corroborated by EBSD pole figures showing sixfold (111) symmetry and by ScN (224) $\\varphi$ scans with six peaks separated by 60 degrees. The paper also uses a molecular-orbital bonding diagram in which the three outer electrons of Sc fill bonding states formed by N 2p and the crystal-field-split $t_{2g}$ orbitals, leaving the antibonding states empty, to explain why the cubic phase is stable even when grown on hexagonal substrates.","core_discovery":"The paper argues that scandium nitride grown by plasma-assisted MBE on c-plane GaN, AlN, and 6H-SiC does not adopt a hexagonal structure but instead forms a face-centered cubic rocksalt phase oriented with the (111) direction parallel to the substrate 0001 c-axis. The films are entirely (111) oriented with twinned in-plane domains, a conclusion supported by the appearance of paired (111) and (002) spots in RHEED, by two-color striped EBSD topography with sixfold {111} pole figures, and by sixfold symmetry in ScN (224) X-ray $\\varphi$ scans. The films show sub-nanometer RMS roughness at the 750°C growth temperature, and Hall measurements give a temperature-independent n-type carrier concentration near $1\\times10^{20}$ cm$^{-3}$ with mobilities around 20 cm$^2$/Vs, indicating degenerate doping. The paper interprets the stability of the rocksalt phase through a molecular-orbital picture in which N 2p and Sc $t_{2g}$ orbitals hybridize to fill bonding states, and it attributes the high carrier density primarily to oxygen impurities while noting that the polar/nonpolar interface may also contribute mobile charge.","pith_inferences":["If oxygen is the dominant donor, exposing the growth surface to a controlled oxygen flux during MBE should increase the electron concentration in a predictable way, and switching to higher-purity Sc sources should lower it.","A cross-sectional TEM study comparing ScN on GaN, AlN, and SiC could determine whether twin-domain size is set by substrate step-terrace spacing or by threading dislocations, a question the paper leaves open.","The polar/nonpolar interface at ScN/GaN and ScN/AlN may itself supply mobile charge; growing ScN on nonpolar or isostructural cubic substrates would separate interface contributions from bulk impurity doping."],"forward_implications":["Because ScN(111) is nearly lattice-matched to GaN(0001), the cubic (111) orientation is the expected epitaxial relationship even where the ScN(111) XRD peak is masked by the GaN(002) reflection.","The temperature-independent carrier concentration near $10^{20}$ cm$^{-3}$ means the as-grown ScN layers behave as degenerate metallic conductors, so contact and buffer applications do not require intentional doping.","The first MBE growth of ScN on AlN provides a limiting binary datum for the ScAlN alloy system, whose piezoelectric and ferroelectric properties depend on Sc content.","The twinned (111) microstructure, with 60-degree rotated domains, is the natural product of growing a threefold-symmetric cubic plane on a sixfold-symmetric hexagonal surface; the domain size depends on the substrate and may be tunable."],"supporting_citations":[{"why":"Establishes the near lattice match between ScN(111) and wurtzite GaN, which explains the expected orientation and the overlapping XRD peaks.","marker":"[19]"},{"why":"Supplies the literature position of the ScN(111) XRD peak near 34.5 degrees used to assign the ScN reflections on SiC and AlN.","marker":"[25]"},{"why":"Reports higher oxygen incorporation in twinned (111) ScN grown on c-plane sapphire than in untwinned (111) ScN on MgO, supporting the impurity discussion for (111) films.","marker":"[26]"},{"why":"Provides the HVPE-grown ScN comparison with higher mobility and lower carrier concentration, used to argue that impurities limit transport.","marker":"[31]"},{"why":"Motivates the nitrogen-rich growth conditions used here by showing that they suppress nitrogen-vacancy formation.","marker":"[32]"},{"why":"Documents the same paired RHEED spot pattern in twinned Cu(111) films, the key analog used to interpret the ScN RHEED signature.","marker":"[33]"},{"why":"Gives DFT results that oxygen substitutional defects have lower formation energy than other donor candidates, supporting the oxygen-donor explanation.","marker":"[36]"},{"why":"Underlies the suggestion that the polar/nonpolar ScN/GaN or ScN/AlN interface may contribute mobile carriers.","marker":"[40]"}],"fun_headline_variants":["ScN MBE films go cubic rocksalt on hexagonal substrates","Cubic (111) ScN twins on GaN, AlN, and SiC templates","Rocksalt ScN (111) on c-plane nitrides: twins & degenerate n-type","ScN MBE films: cubic phase, twin domains, degenerate carriers"],"cache_read_input_tokens":10624,"weakest_assumption_plain":"The conclusion that every ScN grain is single-phase cubic rocksalt with only (111) out-of-plane orientation and twin-related in-plane domains rests on surface and diffraction evidence rather than direct cross-sectional imaging; a hidden minority orientation, secondary phase, or interfacial layer would weaken the central claim.","fun_headline_variants_meta":{"raw":{"variants":["ScN MBE films go cubic rocksalt on hexagonal substrates","Cubic (111) ScN twins on GaN, AlN, and SiC templates","Rocksalt ScN (111) on c-plane nitrides: twins & degenerate n-type","ScN MBE films: cubic phase, twin domains, degenerate carriers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000804,"raw_usage":{"total_tokens":3542,"prompt_tokens":968,"completion_tokens":2574,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":584,"completion_tokens_details":{"reasoning_tokens":2487}},"tokens_in":584,"tokens_out":2574,"duration_ms":20233,"temperature":1.0,"reasoning_tokens":2487,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:24:17.279829+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cross-sectional transmission electron microscopy and reciprocal-space mapping of the ScN/GaN and ScN/AlN interfaces, combined with a high-resolution scan for non-(111) reflections such as ScN (200) or (220), would reveal whether minority orientations or secondary phases exist; a companion measurement of oxygen concentration versus carrier density would test the oxygen-donor explanation.","supporting_citations":[{"cited_title":"Nörenberg, C.T Foxon, and C.J Humphreys, J","cited_arxiv_id":null,"evidence_quote":"Establishes the near lattice match between ScN(111) and wurtzite GaN, which explains the expected orientation and the overlapping XRD peaks."},{"cited_title":"Bohnen, and P.R Hageman, J","cited_arxiv_id":null,"evidence_quote":"Supplies the literature position of the ScN(111) XRD peak near 34.5 degrees used to assign the ScN reflections on SiC and AlN."},{"cited_title":"le Febvrier, N","cited_arxiv_id":null,"evidence_quote":"Reports higher oxygen incorporation in twinned (111) ScN grown on c-plane sapphire than in untwinned (111) ScN on MgO, supporting the impurity discussion for (111) films."},{"cited_title":"Oshima, E.G Villora, and K","cited_arxiv_id":null,"evidence_quote":"Provides the HVPE-grown ScN comparison with higher mobility and lower carrier concentration, used to argue that impurities limit transport."},{"cited_title":"Hamad, H","cited_arxiv_id":null,"evidence_quote":"Motivates the nitrogen-rich growth conditions used here by showing that they suppress nitrogen-vacancy formation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the same paired RHEED spot pattern in twinned Cu(111) films, the key analog used to interpret the ScN RHEED signature."},{"cited_title":"Cetnar, A","cited_arxiv_id":null,"evidence_quote":"Gives DFT results that oxygen substitutional defects have lower formation energy than other donor candidates, supporting the oxygen-donor explanation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underlies the suggestion that the polar/nonpolar ScN/GaN or ScN/AlN interface may contribute mobile carriers."}],"review_version":1}