{"id":"77b63330-5a6f-40fc-8a68-28df3fdf4cf9","arxiv_id":"2607.27862","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"In 6H-SiC co-doped with nitrogen and beryllium, W-band EPR/ENDOR resolves site-specific hyperfine and quadrupole parameters (A(Be)=5.5–5.7 MHz, A(N)=33.2–33.6 MHz) and coherence times of 8.5–13.5 µs at 30 K.","lead":"This paper measures the electronic and nuclear spin properties of nitrogen and beryllium impurities together in 6H-silicon carbide using high-field EPR and ENDOR, reporting hyperfine couplings, quadrupole splittings, and spin coherence times. A smart generalist might read it to see whether co-doped SiC can host both quantum-grade spin defects and conventional doping functions in one crystal.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Be EPR and ENDOR hyperfine values are mutually inconsistent, undercutting the h- vs k-site assignment.","rationale":"The reader's weakest assumption (assignment of Be quartets to h-site vs k1/k2 from g-factor and hyperfine comparison with [10], plus a Jahn–Teller averaging argument) is legitimate. My read identifies an even more basic issue in the same data: the reported EPR and ENDOR hyperfine numbers are not mutually convertible. If the EPR quartet spacing were 2.45 mT, the hyperfine coupling would be ~68.6 MHz; if the ENDOR A=5.5–5.7 MHz is correct, the EPR spacing would be ~0.2 mT. This is not a subtle discrepancy; it is a factor of ~12 in A and a factor of ~100 in the mT/MHz conversion. Because the W-band EPR paragraph uses those splittings to identify the 9Be quartets and to separate h-site vs quasi-cubic k sites, the site assignment is undercut by the inconsistency. The issue is addressable—it may be a units/typo error—but as written, the central claim of fingerprinting two Be configurations cannot be verified from the reported numbers. I therefore keep the reader's CONDITIONAL verdict and do not move it; the condition should explicitly include resolving this hyperfine conversion and providing a joint EPR/ENDOR fit. My agreement is partial because the reader's weakest_assumption and my concern are both about Be site assignment but focus on different vulnerabilities: assignment provenance versus internal numerical self-consistency.","tokens_in":17402,"tokens_out":10157,"duration_ms":93915,"concrete_test":"Digitize or refit the W-band EPR spectra in Fig. 4a: measure the peak-to-peak spacing of the four lines in each Be quartet, convert to MHz via A = 28.025 MHz/mT × ΔB, and jointly fit the EPR and ENDOR spectra with a single spin Hamiltonian (S=1/2, I=3/2, A and Q as free parameters). If the EPR-constrained A does not match 5.5–5.7 MHz (or if no single A fits both datasets), the central Be assignment and A(Be) values are unsupported; the authors should then state explicitly which numbers are typographical errors.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing flaw is an internal numerical inconsistency in the Be hyperfine parameters, located in the same §3 paragraph as the site assignment. The W-band EPR text states that the low-field 9Be quartet has a 'hyperfine splitting constant of 2.45 mT (686 kHz)' and the high-field quartet '2.2 mT (616 kHz)'. For S=1/2, g≈2, a field splitting ΔB is related to A by A = g_e μ_B ΔB/h ≈ 28.0 MHz/mT × ΔB. Thus 2.45 mT corresponds to ≈68.6 MHz, not 686 kHz; 2.2 mT corresponds to ≈61.6 MHz, not 616 kHz. The ENDOR section later reports A(Be)=5.5–5.7 MHz. Even if '686 kHz' were the intended A, the mT value should be ≈0.0245 mT, not 2.45 mT; if A=5.5–5.7 MHz is intended, the EPR quartet spacing should be ≈0.196–0.204 mT. As written, the two measurements cannot refer to the same 9Be centers: an ENDOR-detected A≈5.5 MHz would not produce 2.45 mT EPR spacing at W-band, and an EPR A≈68 MHz would place ENDOR lines far outside the 17–23 MHz range in Fig. 7a. Since the attribution of each quartet to h-site vs k1/k2 sites is made by comparing these (inconsistent) magnitudes with literature [10], the most load-bearing quantitative anchor for that assignment is missing. This may be a typographical or unit error, but until resolved, the central claim is not self-consistent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a multi-frequency (X- and W-band) EPR/ENDOR/TRIPLE study of a 6H-SiC crystal co-doped with nitrogen and beryllium. It claims to identify two nitrogen donor configurations (Nk1, Nk2) and two beryllium acceptor configurations (hexagonal h and quasi-cubic k1/k2), to extract site-specific hyperfine and quadrupole parameters (A(Nk1)=33.23(3) MHz, A(Nk2)=33.55(3) MHz, A(Be)=5.5–5.7 MHz, Q(Be)=220 kHz), and to determine coherence times (T2(N)=13.5 µs, T2(Be)=8.5 µs, T1(N)=20 ms, T1(Be)=143 µs at 30 K). It further interprets TRIPLE data as evidence of coupled nuclear spin subspaces and argues that the N/Be pair provides complementary functionality for defect engineering and quantum sensing in SiC.","tokens_in":17789,"tokens_out":11458,"duration_ms":104663,"significance":"If the site assignments and interaction parameters hold, this work provides a useful quantitative reference for N and Be spin defects in 6H-SiC and demonstrates the resolving power of W-band ENDOR/TRIPLE on a dual-doped sample. The main strengths are the direct measurement of multiple interaction parameters in one crystal and the explicit N-donor/Be-acceptor comparison of hyperfine and quadrupole interactions. However, the load-bearing inconsistency between the reported Be EPR hyperfine splittings and the ENDOR A(Be) values must be resolved before the site-specific conclusions can be accepted. The paper does not provide raw spectra or simulation files, but the reported precision of the nitrogen parameters is a strong point.","major_comments":[{"comment":"The text states that the low-field 9Be quartet has a hyperfine splitting constant of 2.45 mT (686 kHz) and the high-field quartet 2.2 mT (616 kHz). For S=1/2, g≈2, ΔB(mT) and A(MHz) are related by A ≈ 28.0 MHz/mT × ΔB, so 2.45 mT corresponds to ≈68.6 MHz, not 686 kHz, and 2.2 mT to ≈61.6 MHz. Conversely, the ENDOR section reports A(Be)=5.5–5.7 MHz, which would give an EPR quartet spacing of ≈0.20 mT. If 686 kHz were the intended A, the spacing would be ≈0.0245 mT, not 2.45 mT. As written, the EPR and ENDOR data cannot refer to the same 9Be centers, and the h-site vs k1/k2 assignment anchored to these magnitudes and to comparison with ref. [10] is not internally consistent. Please supply corrected values and the conversion used, or revise the site-specific interpretation.","section":"§3, W-band EPR assignment paragraph (Fig. 4a)"},{"comment":"The main plot of Fig. 4b reports T2(Be)=8.5 ms and T2(N)=13.5 ms, and the inset reports T1(Be)=143 ms and T1(N)=20 ms, while the text states T2(Be)=8.5 µs, T2(N)=13.5 µs, T1(Be)=143 µs, T1(N)=20 ms. This factor-of-1000 unit discrepancy must be resolved, since the coherence times are a central quantitative claim of the paper. Please also report uncertainties for the fitted relaxation times.","section":"§3, Fig. 4b relaxation data"},{"comment":"The separation of the two Be quartets into a 'shallow axial' h-site and 'quasi-cubic' k1/k2 acceptors is inferred from the magnitude of g-factors and hyperfine constants relative to ref. [10] and from a dynamic Jahn–Teller averaging argument, not from angular-dependent ENDOR or a direct structural determination. Given the inconsistency identified above, the word 'unambiguous' used for this assignment is too strong. Either provide supporting angular data or present the site labels as tentative. The abstract and conclusions also state that TRIPLE spectra 'verify coupled nuclear spin subspaces,' while the body text says the population-transfer mechanisms 'require further study'; please make the wording consistent.","section":"§3, Be site assignment; Abstract and Conclusions"}],"minor_comments":[{"comment":"The acceptor Be concentration is given as '10 cm−3'; this lacks an exponent (presumably 10^17 or 10^18 cm−3) and conflicts with the abstract's '1018 cm−3'. Please correct.","section":"§2.1, Sample"},{"comment":"The spin Hamiltonian in Eq. (1) omits the nuclear Zeeman and nuclear quadrupole terms that are used later in the ENDOR analysis. Either include them explicitly or state that they are added for the ENDOR simulations.","section":"Eq. (1)"},{"comment":"The text says the separation of NMR lines corresponds to 2Q = 3.795 MHz and 'translates to' Cq = 2.53 MHz, while Table 1 gives Cq = 2.426(3) MHz for the NV center. For I=1, the formula Q = 3/(4I(2I−1))Cq gives Cq ≈ 5.06 MHz from Q = 1.8975 MHz. The formula, the line separation, or the table entry must be reconciled.","section":"§3, quadrupole comparison text after Table 1"},{"comment":"The text mentions 'ten distinct spectral features' but labels them A(1) through A(9). The numbering should match the stated number of features.","section":"§3, 29Si ENDOR description"},{"comment":"Reference [37] combines two separate citations (Holiatkina et al. and Goldfarb/Stoll) into a single numbered entry. These should be split.","section":"References"},{"comment":"Fig. 3b caption says 'dashed lines' while the text says 'dashed-dotted lines'; Fig. 4a lacks units on the y-axis. The data availability statement is too generic; please identify where raw EPR/ENDOR spectra or simulation input files can be obtained.","section":"Figure captions and data availability"}],"recommendation":"major_revision","confidential_remarks":"The paper contains promising W-band ENDOR data and a useful donor-acceptor comparison, but the internal numerical inconsistencies and overstatements discussed above must be resolved. I recommend major revision rather than rejection; the issues appear fixable if the original data are rechecked and the text is tightened."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nQuick take: this paper has genuinely new W-band ENDOR/TRIPLE data on co-doped 6H-SiC, and the N-donor results (A(Nk1)=33.23 MHz, A(Nk2)=33.55 MHz, with no resolvable quadrupole) look solid and consistent with prior work. The same sample gives relaxation times, T2(N)=13.5 µs, T2(Be)=8.5 µs at 30 K, which are plausible numbers for a defect-engineering reference. That part is worth having.\n\nThe problem is the Be hyperfine assignment. In the W-band EPR section, the two quartets are reported with hyperfine splittings of 2.45 mT and 2.2 mT, converted in the text to 686 kHz and 616 kHz. For a S=1/2, g≈2 center, 2.45 mT corresponds to ~69 MHz, not 686 kHz, and 686 kHz would be a splitting of ~0.024 mT. Then the ENDOR section reports A(Be)=5.5–5.7 MHz for the same centers. No two of these numbers match. The ENDOR value is the most direct measurement, so my guess is the mT and/or kHz values are typos, but as written the site assignment is anchored to numbers that are mutually inconsistent. The paper attributes each quartet to the h-site vs quasi-cubic k1/k2 sites by comparing these hyperfine magnitudes to Ref [10]; with that anchor missing, the assignment is not self-consistent.\n\nOther, smaller problems: the abstract claims TRIPLE spectra \"verify coupled nuclear spin subspaces\" while the text says the mechanism \"requires further study\"; the T2 values are labeled in ms in Figure 4b but µs in the text; the Be concentration \"10 cm−3\" is missing an exponent; and the Cq values for the NV center differ between table and text (2.426 vs 2.53 MHz). None of these are fatal on their own, but they add up.\n\nSo the paper reads like a useful experimental dataset with an over-ambitious interpretation and a load-bearing numerical error. I'd send it to a referee, but I'd ask the authors to fix the Be hyperfine inconsistency, reconcile the figure/table values, deposit raw spectra if possible, and tone down the TRIPLE claim to match what the data actually show. As it stands, I would not cite the Be parameters.\n\nRecommendation: send for review, but the revision should be major.","headline":"Useful W-band ENDOR data on N and Be in 6H-SiC, but the Be hyperfine numbers are internally inconsistent and the site assignment rests on them.","tokens_in":18436,"tokens_out":4420,"would_cite":false,"duration_ms":37156,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["76.30.-v","76.70.-r"],"model":"deepseek-v4-flash","headline":"High-field EPR/ENDOR separates and assigns nitrogen donors and beryllium acceptors in one 6H-SiC crystal, yielding site-specific spin parameters for quantum-memory and strain-sensing applications.","keywords":["6H-SiC","EPR","ENDOR","TRIPLE resonance","nitrogen donor","beryllium acceptor","hyperfine interaction","nuclear quadrupole interaction"],"falsifier":"Rotate the 6H-SiC crystal at W-band and record the angular dependence of the two Be quartets: an h-site axial acceptor should show one smoothly varying pattern, while quasi-cubic k1/k2 centers should split into distinct orientation-dependent patterns. If the low-field quartet shows non-axial behavior or the high-field quartet lacks the predicted Jahn-Teller averaging, the site assignment collapses. Alternatively, resolve the 29Si ENDOR shell structure for each quartet and compare with calculated hyperfine patterns for BeSi at h versus k1/k2.","tokens_in":17234,"feed_emoji":"🧲","tokens_out":5590,"duration_ms":55019,"temperature":0.7,"pith_summary":"Working in a single 6H-SiC crystal co-doped with nitrogen and beryllium, the paper sets out to show that high-field (94 GHz) pulsed EPR and ENDOR can resolve the two impurities by lattice site and extract quantitative spin-Hamiltonian parameters for each. It assigns two nitrogen donor configurations (k1 and k2) with hyperfine constants near 33 MHz, two beryllium acceptor configurations (hexagonal h-site and quasi-cubic k1/k2) with hyperfine 5.5–5.7 MHz and quadrupole 220 kHz, and reports coherence times of 13.5 µs for N and 8.5 µs for Be at 30 K. If correct, this establishes dual-doped 6H-SiC as a material where a stable donor spin memory and a Jahn-Teller-sensitive acceptor strain probe can coexist, and it supplies the quantitative inputs needed to engineer such centers.","feed_headline":"High-field spin resonance separates N and Be defects in SiC","feed_subtitle":"Site-specific hyperfine and quadrupole data show which defect suits quantum memory and which suits strain sensing.","key_machinery":"The central tool is W-band (94 GHz) pulsed EPR with Mims ENDOR and TRIPLE resonance, interpreted through an axial spin Hamiltonian containing g, hyperfine A, and nuclear quadrupole Cq terms. The high magnetic field separates sites whose g-factors differ only slightly, while ENDOR's narrow nuclear transitions resolve hyperfine differences of a few hundred kilohertz. TRIPLE resonance then distinguishes true multiple sites from quadrupole splittings by showing controlled population transfer between nuclear spin sublevels of different coordination shells.","core_discovery":"The central claim is that high-frequency ENDOR, aided by the Zeeman resolution of the W-band at 3.4 T, can fingerprint two complementary functional impurities in one 6H-SiC sample and tie each spectral feature to a specific crystallographic site. For beryllium, the low-field quartet is assigned to the shallow axial acceptor at the hexagonal h-site, while the high-field quartet is assigned to the quasi-cubic k1/k2 sites, where a dynamic Jahn-Teller hop partially averages the spin density and lowers the observed hyperfine splitting. ENDOR yields A(Be)=5.5–5.7 MHz with a quadrupole splitting of about 220 kHz, indicating that the hole density sits mostly on neighboring carbons rather than on the","pith_inferences":["A direct test of the Be site model would be angular-dependent W-band ENDOR on a rotating crystal; the paper reports spectra at fixed orientation, so the full hyperfine and quadrupole tensors for the two Be configurations remain to be mapped.","The measured T2 values come from ensemble measurements and do not by themselves prove single-defect addressability; a natural next step is optically detected magnetic resonance on individual N or Be centers in co-doped 6H-SiC.","Because the 14N quadrupole coupling is two orders of magnitude larger for NV than for isolated N donors, ENDOR-resolved quadrupole data could serve as a local fingerprint of nearby silicon vacancies in co-doped material.","The 9Be nuclear spin (I=3/2) coupled to a short-T1 electron could act as a long-lived nuclear memory; one testable extension is a nuclear spin echo or coherence measurement at 30 K extending beyond the electronic T2."],"forward_implications":["Co-doped 6H-SiC can host a stable N-donor spin state and a Be-acceptor state in the same crystal, with distinct microscopic identities instead of one overlapped spectrum.","The nearly equal Nk1/Nk2 hyperfine constants and the 10 kHz quadrupole upper bound place a quantitative limit on the host electric field gradient at carbon substitution sites.","The small Be Fermi contact (5.5–5.7 MHz) and large 9Be quadrupole (220 kHz) support the off-center BeSi acceptor model and explain why the Be center is sensitive to local strain.","The measured 30 K coherence times, 13.5 µs for nitrogen and 8.5 µs for beryllium, give concrete design targets for quantum memory and sensing sequences in this material.","The two-orders-of-magnitude difference in quadrupole coupling between the N donor and the NV center traces the EFG at nitrogen to the adjacent silicon vacancy, providing a quantitative handle on vacancy-induced charge redistribution."],"fun_headline_variants":["W-band EPR/ENDOR separates N and Be in SiC","Hyperfine data distinguish N and Be roles in SiC","Site-specific spin data for N and Be in SiC","ENDOR resolves N and Be sites in 6H-SiC","High-field ENDOR tags N and Be defects in SiC"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The assignment of the two beryllium EPR quartets to the hexagonal h-site versus the quasi-cubic k1/k2 sites rests on comparing g-factors and hyperfine magnitudes with earlier data and on a dynamic Jahn-Teller averaging argument, not on a direct structural measurement or angular-dependent ENDOR.","fun_headline_variants_meta":{"raw":{"variants":["W-band EPR/ENDOR separates N and Be in SiC","Hyperfine data distinguish N and Be roles in SiC","Site-specific spin data for N and Be in SiC","ENDOR resolves N and Be sites in 6H-SiC","High-field ENDOR tags N and Be defects in SiC"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000673,"raw_usage":{"total_tokens":2919,"prompt_tokens":783,"completion_tokens":2136,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":527,"completion_tokens_details":{"reasoning_tokens":2049}},"tokens_in":527,"tokens_out":2136,"duration_ms":12674,"temperature":1.0,"reasoning_tokens":2049,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T00:07:09.855629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Rotate the 6H-SiC crystal at W-band and record the angular dependence of the two Be quartets: an h-site axial acceptor should show one smoothly varying pattern, while quasi-cubic k1/k2 centers should split into distinct orientation-dependent patterns. If the low-field quartet shows non-axial behavior or the high-field quartet lacks the predicted Jahn-Teller averaging, the site assignment collapses. Alternatively, resolve the 29Si ENDOR shell structure for each quartet and compare with calculated hyperfine patterns for BeSi at h versus k1/k2.","supporting_citations":[],"review_version":1}