{"id":"d8016be5-e878-40c8-b425-648cd0e26647","arxiv_id":"2506.03789","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"High-field ESR on Na2Co2TeO6 finds three spin-wave modes for B||c with split zero-field gaps of 211 and 237 GHz, excluding reported zigzag models and exposing the triple-q model as incomplete.","lead":"Researchers measured the low-energy magnetic excitations in the Kitaev candidate Na2Co2TeO6 using high-frequency electron spin resonance in magnetic fields. The measurements reveal a split zero-field spin-wave gap and show that all published spin models, including the leading triple-q model, fail to capture the full spectrum.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim that the triple-q model is incomplete because c1 is an additional magnon branch presumes a single-domain assignment; HF-ESR is not momentum-resolved and domain multiplicity could produce extra resonances.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing concern: the three resonances c1, c2, and c3 are assumed to be intrinsic magnon modes of a single magnetic domain. I agree with the CONDITIONAL verdict. The central negative result against zigzag models is robust: all published zigzag models predict flat low-field modes and softening above 10 T, in stark contrast to the observed steep slopes and softening near 5 T. The positive comparison with the triple-q model for c2 and c3 is also convincing. However, the inference that c1 is an additional magnon branch (and hence that the triple-q model is incomplete) relies on mode-counting in a non-momentum-resolved experiment. Domain multiplicity is a known confound for this material, which is precisely why single-crystal neutron scattering was needed to argue for triple-q over multi-domain zigzag. The manuscript provides no control for domain-related resonances, and the c1 gap is defined by linear extrapolation. The paper should soften the 'unambiguous proof' wording and explicitly discuss domain multiplicity in the ESR assignment. My recommendation remains CONDITIONAL because the core observations and zigzag exclusion are likely correct and valuable, but the 'extra mode' conclusion is not fully settled.","tokens_in":16149,"tokens_out":1799,"duration_ms":18953,"concrete_test":"Measure HF-ESR on at least two independently grown Na₂Co₂TeO₆ crystals and perform angular-dependent measurements rotating the field slightly away from the c axis. If c1 is a distinct magnon branch, its intensity and frequency evolution should be reproducible across crystals and vary smoothly with angle near c; if c1 is domain-related, its intensity/position should differ between crystals or show branching/discontinuous angular behavior. Additionally, compute the k=0 ESR spectrum of the triple-q state including all three q-domain orientations; if three low-energy resonances with the observed slopes and gaps appear without enlarging the magnetic unit cell, the 'extra mode' conclusion is falsified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The main conclusion in Sec. IV B is that the triple-q model fails because it predicts only two low-energy modes (c2, c3) while three modes (c1, c2, c3) are observed. The inference that c1 is a distinct magnon branch of the bulk single-domain ground state is load-bearing. However, HF-ESR at k=0 has no momentum resolution and cannot by itself separate a distinct magnon branch from resonance features arising from magnetic domain multiplicity, surface excitations, or minority phases. The authors carefully exclude a paramagnetic impurity mode p, but c1 is not subjected to equivalent domain scrutiny. Since the triple-q state is constructed from three zigzag components and the material is known to exhibit multi-domain order in neutron diffraction, a domain-related origin for c1 would invalidate the mode-counting argument for a larger magnetic unit cell. The absence of a control experiment (e.g., multiple crystals, angular dependence, or domain-sensitive comparison) leaves this ambiguity open. A related weakness is that the zero-field gap of c1 is obtained by linear extrapolation from B > 1.5 T; the authors acknowledge that flattening near zero field cannot be excluded, which would weaken the split-gap claim. The zigzag exclusion and the c2/c3 comparisons are well supported, but the 'extra mode' conclusion rests on the single-domain magnon assignment.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports high-frequency/high-field electron spin resonance (HF-ESR) measurements on single-crystal Na2Co2TeO6 for magnetic fields along both the c axis and the in-plane a* axis. For B||c, three distinct spin-wave modes (c1, c2, c3) are observed in the antiferromagnetic phase; linear extrapolations yield zero-field gaps Delta = 211(6) GHz and Delta_2 = 237(5) GHz. Mode c1 softens near 4.7 T, coincident with a kink in the isothermal magnetization. Spin-wave calculations using published extended Heisenberg-Kitaev parameters, with no refitting to the ESR data, show that all published zigzag ground-state models are incompatible with the observed strong field dependence and softening field, while the triple-q model of Krüger et al. reproduces modes c2 and c3 but fails to reproduce mode c1. The authors conclude that the triple-q ground-state model is incomplete and suggest the relevance of interlayer interactions.","tokens_in":16496,"tokens_out":7745,"duration_ms":81861,"significance":"If the mode assignment is correct, the paper provides a decisive experimental test of published spin models for Na2Co2TeO6: all reported zigzag models are excluded by the strong field dependence and by the softening field near 5 T, and the only published triple-q model is shown to describe two of the three observed low-energy modes. The use of literature parameters without refitting makes this a genuine, non-circular model comparison, which is a notable strength. The concurrence of the c1 softening with a magnetization kink at Bc1 = 4.7 T also establishes a well-characterized field-induced transition. These results, if confirmed, constitute an important step toward constraining the microscopic Hamiltonian of a leading Kitaev candidate material.","major_comments":[{"comment":"The inference that c1 is an additional magnon branch of the bulk single-domain ground state is load-bearing for the central claim that the triple-q model is incomplete. HF-ESR is a k=0 probe with no momentum resolution, and the paper does not discuss whether magnetic domain multiplicity, surface excitations, or minority phases could produce an extra resonance. The authors carefully exclude a paramagnetic impurity mode (p), but c1 is not subjected to equivalent scrutiny. Please add an explicit argument that all possible domains of the triple-q state remain equivalent under B||c, or provide an additional control (e.g., a second crystal, angular dependence, or a domain-sensitive comparison). Without this, the mode-counting argument for a larger magnetic supercell is incomplete.","section":"III A and IV B"},{"comment":"The zero-field gaps Delta = 211 GHz and Delta_2 = 237 GHz are obtained by linear extrapolation from data at B > 1.5 T. The paper acknowledges that flattening near zero field cannot be excluded, yet the abstract presents these extrapolated values without qualification. Please quantify the uncertainty in the extrapolation (for example, by fitting alternative functional forms) or explicitly state the range of zero-field gaps consistent with the data. This is not fatal to the extra-mode conclusion, but it is needed to support the quantitative gap values quoted in the abstract and conclusions.","section":"Fig. 3b and Section III A"},{"comment":"The triple-q calculation assumes field-independent effective local fields h and field-independent renormalized bilinear couplings. The paper acknowledges this is an approximation for B != 0. The conclusion that c1 cannot be one of the high-energy modes relies on the qualitative argument that a reordering is 'highly unlikely'. This argument would be strengthened by a controlled test, such as allowing h to vary linearly with field and examining whether a high-energy branch can be brought down to the observed c1 frequency without destroying the agreement for c2 and c3. As written, the exclusion of field-induced mode reordering remains plausible but not demonstrated.","section":"Section IV B and Eq. (3)"}],"minor_comments":[{"comment":"The abstract states that the zero-field excitation gap splits into Delta = 211 GHz and Delta_2 = 237 GHz without noting that these values are extrapolated; please add 'extrapolated' or a similar qualifier, as the main text correctly acknowledges that flattening near zero field cannot be excluded.","section":"Abstract"},{"comment":"The phrase 'spreads over at least 16 magnetic sides for a (2 x 4) cell' appears to contain a typo; it should likely read '16 magnetic sites' or '16 spins'.","section":"Section IV C"},{"comment":"The sentence 'In contrast, upon reducing f below 259 GHz, the resonance mode c1 is found at increasing magnetic fields' is confusingly worded; it would be clearer to state that lower frequencies require higher resonance fields for c1.","section":"Section III A"},{"comment":"The dashed line representing the non-linear extrapolation of c1 is described in the caption, but the functional form (e.g., parabolic or square-root) is not given; specifying the form would help readers assess the extrapolation to Bc1.","section":"Fig. 3b caption"}],"recommendation":"major_revision","confidential_remarks":"The paper makes a valuable and largely convincing experimental case that the triple-q model of Na2Co2TeO6 is incomplete. The main concern is the unaddressed possibility that the extra mode c1 arises from domain or surface effects rather than a distinct bulk magnon branch; this issue is fixable by discussion or additional control measurements. The manuscript is likely appropriate for the journal if the authors address this point and the extrapolation caveat in the abstract."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is a careful HF-ESR study that delivers a solid negative result. The new observation is the split zero-field gap in Na2Co2TeO6 for B||c: three spin-wave modes with gaps at roughly 211 and 237 GHz. The authors then systematically compare these modes against every published extended Heisenberg-Kitaev Hamiltonian, using the published parameters without refitting, and show that all zigzag models fail. The triple-q model of Krüger et al. correctly predicts the two rising modes c2 and c3, including the split gap, but does not produce the softening mode c1. That is a genuine constraint on the Hamiltonian.\n\nThe stress-test worry about domain multiplicity does not hold up on my reading. With B along the c-axis, the in-plane rotations that relate zigzag or triple-q domains leave the spin-wave spectrum invariant, so multiple domains would not add extra resonance lines. The mode count is determined by the magnetic unit cell, not by the domain population. The authors could have made this explicit, but the inference is sound.\n\nThe real caveat is more mundane: the zero-field gaps are obtained by linear extrapolation from data above about 1.5 T. The paper acknowledges flattening near zero field cannot be excluded. That said, the split between Δ and Δ2 is robust because flattening would, if anything, increase the difference between Δ1 and Δ2. I think the phrase 'unambiguously prove' in Sec. III A is a bit strong, but the underlying claim is fine.\n\nThe paper's own suggested resolution—interlayer coupling enlarging the effective unit cell—is speculative, as the authors admit. The central result, that the triple-q model is incomplete, is well supported. A modest revision should add a sentence about domain equivalence and soften the 'unambiguous' phrasing.\n\nWho should read this: anyone working on Kitaev candidates or Co-honeycomb magnets. The split-gap observation will be a benchmark for future model fitting. I would send this to peer review; it deserves referee time and likely publication after minor revision.\n\nBest,","headline":"A careful ESR study that sharpens the case against the triple-q model; the extra-mode argument holds up better than the stress-test suggests.","tokens_in":17002,"tokens_out":6488,"would_cite":true,"duration_ms":62578,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["75.30.Ds","76.30.-v","75.50.Ee"],"model":"deepseek-v4-flash","headline":"The paper establishes that the ordered magnetic ground state of Na2Co2TeO6 carries three low-energy spin-wave branches, a count that rules out every published zigzag model and leaves the triple-q model incomplete.","keywords":["Na2Co2TeO6","Kitaev quantum spin liquid","high-field electron spin resonance","spin waves","triple-q magnetic order","Heisenberg-Kitaev model","zigzag magnetism","magnon softening"],"falsifier":"Measure the low-energy $B \\parallel c$ spin-wave spectrum of a detwinned single crystal with a bulk momentum-sensitive probe such as high-resolution inelastic neutron scattering and show that fewer than three distinct branches exist, or that the third line is an impurity, domain, or two-magnon artifact; conversely, a slightly adjusted or interlayer-extended triple-q model that reproduces all three branches including the softening would confirm the paper's conclusion.","tokens_in":15969,"feed_emoji":"🧲","tokens_out":7983,"duration_ms":86916,"temperature":0.7,"pith_summary":"High-frequency electron spin resonance on Na2Co2TeO6 single crystals resolves the low-energy magnetic excitations and uses them to test proposed spin models. For fields along the c axis, the paper finds three distinct spin-wave modes with a split zero-field gap, $\\Delta = 211$ GHz and $\\Delta_2 = 237$ GHz, where earlier bulk probes had seen only one broad gap. One mode softens at $B_{\\rm c1} = 4.7$ T, matching a kink in the magnetization and indicating a field-induced phase transition. Comparing the measured modes with linear spin-wave calculations, the paper argues that all published zigzag-based extended Heisenberg-Kitaev models fail, while the only reported triple-q model reproduces two modes but misses the softening one. The conclusion is that the triple-q ground-state picture is incomplete and that interlayer interactions are a likely missing ingredient.","feed_headline":"Three magnon modes break the triple-q model of Na2Co2TeO6","feed_subtitle":"High-field ESR finds a third spin-wave branch that zigzag and triple-q Hamiltonians cannot produce.","key_machinery":"The discriminating observable is the number and field dependence of zero-wave-vector magnon modes for $B \\parallel c$, measured by high-frequency electron spin resonance; in an ordered magnet the mode count equals the number of magnetic sublattices in the ground-state unit cell. The calculations are linear spin wave theory on classically relaxed ground states for the extended Heisenberg-Kitaev Hamiltonian, with a Zeeman term built from the measured anisotropic g-factors, and for the triple-q case a non-bilinear ring-exchange term approximated as effective local fields. The split gaps and non-zero effective g-factors are traced to the difference $J_{2A} - J_{2B}$ between next-neighbor couplings on the two Co sublattices.","core_discovery":"The central experimental discovery is that the antiferromagnetic ground state of Na2Co2TeO6 has three low-energy magnon branches for $B \\parallel c$, with zero-field gaps of $\\Delta = 211$ GHz and $\\Delta_2 = 237$ GHz, and that the lowest branch softens at $B_{\\rm c1} = 4.7$ T. Linear spin-wave calculations for every published zigzag Heisenberg-Kitaev parameter set produce at most two low-energy modes, almost field-independent mode frequencies at low field, and softening only above 10 T, in direct contradiction with the data. The triple-q model reproduces the rising modes c2 and c3, including the split-gap structure tied to the sublattice difference $J_{2A} \\neq J_{2B}$, but it does not predict the softening mode c1. Because the number of magnon branches equals the number of magnetic sublattices, the extra mode means the true magnetic unit cell must be larger than the one assumed by the triple-q model; the paper proposes interlayer coupling as the most plausible way to enlarge it.","pith_inferences":["If the missing c1 mode comes from interlayer coupling, the same coupling should create a small but finite out-of-plane dispersion of the spin-wave branches, which a dedicated search along the out-of-plane momentum direction in existing single-crystal neutron data could directly test.","The mode-count argument implies that any correct Hamiltonian must have at least as many magnetic sublattices as observed branches; enlarged in-plane supercells such as $(2\\times4)$ or $(4\\times4)$ would also produce zone-folding signatures in neutron or resonant x-ray scattering, distinguishing them from the interlayer scenario.","Because $J_{2A} - J_{2B}$ controls both the split gaps and the ferrimagnetic moment, a controlled tuning of the two Co sublattices, by strain or chemical substitution, should continuously change the gap difference and the magnetization, a testable prediction the paper does not state.","Applying the same ESR protocol to other honeycomb cobaltates could show whether an extra magnon branch is a generic feature of triple-q Kitaev candidates or specific to Na2Co2TeO6."],"forward_implications":["All published zigzag parameter sets for Na2Co2TeO6 are excluded: they predict at most two low-energy modes, near-zero effective g-factors at low field, and softening only above 10 T for $B \\parallel c$.","The zero-field gap structure of the antiferromagnetic phase is a split pair, $\\Delta = 211$ GHz and $\\Delta_2 = 237$ GHz, which any future model must reproduce.","The softening of mode c1 at $B_{\\rm c1} = 4.7$ T anchors a field-induced phase transition that also appears as a kink in isothermal magnetization and in magnetostriction.","The triple-q model remains partly correct: it captures c2 and c3 and the split-gap structure, but its eight-sublattice unit cell cannot generate the extra c1 branch, so the true magnetic unit cell must be larger, e.g. through interlayer coupling giving 16 modes.","A complete model will need a combined analysis of in-field ESR data with zero-field single-crystal inelastic neutron scattering, including $J_{2A} \\neq J_{2B}$, non-bilinear interactions treated without approximation, and interlayer couplings."],"supporting_citations":[{"why":"Supplies the only reported triple-q model, with ring-exchange interactions implemented as effective local fields, whose predicted spin waves are compared against the measured c1, c2, and c3 modes.","marker":"[15]"},{"why":"Single-crystal neutron diffraction and inelastic neutron scattering work that revised the zero-field ground state to triple-q order; it defines the magnetic unit cell and the INS spectrum the model must simultaneously describe.","marker":"[20]"},{"why":"Provides the tx+ and tx- zigzag Heisenberg-Kitaev parameter sets whose full field-dependent spin-wave spectra are computed in detail as representative zigzag tests.","marker":"[50]"},{"why":"Proposes an extended Heisenberg-Kitaev model with zigzag ground state and gives the anisotropic g-factors used in all Zeeman calculations; it also reported a much smaller polycrystalline ESR gap.","marker":"[14]"},{"why":"Earlier single-crystal HF-ESR study that linearly extrapolated a 4.0 T softening field and a single gap; the present three-mode split-gap observation extends and corrects this baseline.","marker":"[39]"},{"why":"Single-crystal inelastic neutron scattering study that reported one broad low-energy zero-field gap, the comparison point showing that the splitting into two gaps was previously unresolved.","marker":"[18]"}],"fun_headline_variants":["ESR finds third magnon mode that defeats triple-q model","Spin wave spectra break triple-q model for Na2Co2TeO6","Three magnon branches expose limits of triple-q model","New magnon mode rules out triple-q state in Na2Co2TeO6"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The ESR lines c1, c2, and c3 are intrinsic magnon branches of a single magnetic domain in the bulk antiferromagnetic phase, and the triple-q model's fixed parameters do not shift significantly with field; if any line has a non-magnon origin or the model parameters are strongly field-dependent, the mode count that invalidates the triple-q model collapses.","fun_headline_variants_meta":{"raw":{"variants":["ESR finds third magnon mode that defeats triple-q model","Spin wave spectra break triple-q model for Na2Co2TeO6","Three magnon branches expose limits of triple-q model","New magnon mode rules out triple-q state in Na2Co2TeO6"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000266,"raw_usage":{"total_tokens":1688,"prompt_tokens":1099,"completion_tokens":589,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":715,"completion_tokens_details":{"reasoning_tokens":512}},"tokens_in":715,"tokens_out":589,"duration_ms":7551,"temperature":1.0,"reasoning_tokens":512,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:55:25.672697+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the low-energy $B \\parallel c$ spin-wave spectrum of a detwinned single crystal with a bulk momentum-sensitive probe such as high-resolution inelastic neutron scattering and show that fewer than three distinct branches exist, or that the third line is an impurity, domain, or two-magnon artifact; conversely, a slightly adjusted or interlayer-extended triple-q model that reproduces all three branches including the softening would confirm the paper's conclusion.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the only reported triple-q model, with ring-exchange interactions implemented as effective local fields, whose predicted spin waves are compared against the measured c1, c2, and c3 modes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Single-crystal neutron diffraction and inelastic neutron scattering work that revised the zero-field ground state to triple-q order; it defines the magnetic unit cell and the INS spectrum the model must simultaneously describe."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the tx+ and tx- zigzag Heisenberg-Kitaev parameter sets whose full field-dependent spin-wave spectra are computed in detail as representative zigzag tests."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposes an extended Heisenberg-Kitaev model with zigzag ground state and gives the anisotropic g-factors used in all Zeeman calculations; it also reported a much smaller polycrystalline ESR gap."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier single-crystal HF-ESR study that linearly extrapolated a 4.0 T softening field and a single gap; the present three-mode split-gap observation extends and corrects this baseline."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Single-crystal inelastic neutron scattering study that reported one broad low-energy zero-field gap, the comparison point showing that the splitting into two gaps was previously unresolved."}],"review_version":1}