{"id":"e999a549-b419-4b2d-8a5f-bdcb57c596a0","arxiv_id":"1908.09014","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"DFT+U calculations identify a small Q2-type orthorhombic distortion of Na-O octahedra as the source of the electric field gradient observed by NMR in Ba2NaOsO6.","lead":"The paper computes, from first principles, the local electric-field distortions around sodium atoms in the magnetic insulator Ba2NaOsO6 and compares them with NMR measurements. It identifies a specific octahedral stretch pattern (the Q2 mode) as the cause of the measured distortion, and shows that the type of magnetic order does not affect this local distortion.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Q2 assignment rests on a hand-imposed 0.52% distortion that is not a DFT equilibrium; Appendix E shows relaxation cuts the EFG roughly in half, so the conclusion 'explicitly show' is not supported.","rationale":"The reader's weakest assumption is exactly the representativeness of the hand-imposed distortion, and Appendix E, which the paper itself includes, confirms the concern: relaxation lowers the energy by 0.3 eV and halves the EFG. That makes the 0.52% amplitude a fit, not a prediction. This is the most load-bearing issue because the headline scientific claim is causal: a specific Q2 distortion is said to be the main source of the NMR-observed EFG. If the distortion is not stable in the same DFT+U theory, the calculation only demonstrates that one possible charge configuration can reproduce the NMR parameters; it does not demonstrate that this configuration is the physical one. I also note two supporting weaknesses: PBE0 without SOC gives η ≈ 0.46 rather than 0.88, and PBE+SOC+U gives νQ ≈ 128 kHz instead of 190 kHz, so the quantitative match is functional-dependent. These do not change the verdict because the mode-assignment part of the claim retains some support: only Q2-type Models A and F2 produce the observed principal-axis orientation and near-unity η, and the EFG is shown to be insensitive to cFM versus FM110 order. The paper is transparent about the static nature of the calculations and about the lack of proof for the Kramers-doublet-lifting mechanism. The appropriate assessment remains conditional: the Q2 identification is plausible and worthy of follow-up, but the causal conclusion as written overstates the evidence. Since the reader already reached CONDITIONAL, no verdict change is needed.","tokens_in":14965,"tokens_out":8159,"duration_ms":89052,"concrete_test":"Take the fully relaxed geometry already mentioned in Appendix E (same GGA+SOC+U, PP6 pseudopotential, and cFM setup), decompose the oxygen displacement pattern around Na into Q2 and Q3 normal modes of the octahedron, and recompute the EFG tensor at this relaxed geometry. If the relaxed structure retains a Q2 amplitude near 0.52% and yields νQ ≈ 190–200 kHz with η ≈ 0.88, the central claim survives; if it instead has a smaller Q2 amplitude or an EFG roughly half the experimental values, then the 0.52% Model A.3 distortion is not a DFT equilibrium and the conclusion must be downgraded from 'explicitly show' to 'consistent with a fitted Q2 model.'","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on Model A.3: a static, hand-imposed orthorhombic distortion of the Na-O octahedra, with Na-O bonds elongated and compressed by about 0.52%, placed on the experimental cubic structure. Appendix E reports that full relaxation of this structure lowers the total energy by 0.3 eV and reduces the computed EFG parameters to roughly half. The fitted distortion is therefore not a stationary point of the same DFT+U functional, and the amplitude is an adjustable parameter rather than a first-principles output. The NMR match (νQ ≈ 190–200 kHz, η ≈ 0.88) is obtained only at this chosen amplitude; at the theory's own relaxed geometry the EFG would not match. The quantitative identification is also functional-dependent: PBE0 without SOC gives η ≈ 0.46 for the same Model A.3 (Table X), and PBE+SOC+U gives νQ ≈ 128 kHz (Appendix D). Symmetry does give some independent support: among Models A–F2, only A and F2 place Vzz along crystalline axes with near-unity η, so the Q2 mode assignment is plausible. But reproducing an EFG with an imposed structural model does not establish that the actual BLPS phase contains that distortion, and the paper's own Appendix E shows the model is not energetically preferred. The conclusion that the Q2 distortion 'is the main source' and 'lifts the j = 3/2 quartet' goes beyond what the calculations demonstrate; the authors also concede in Section IV A that they cannot provide proof of the Jahn-Teller-lifting hypothesis.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports DFT+U and hybrid DFT calculations of the electric field gradient (EFG) tensor at the 23Na sites in the double perovskite Ba2NaOsO6, a 5d1 Mott insulator with strong spin-orbit coupling. The authors construct several model distortions (Models A, B, C, C2, D, E, F, F2) based on earlier point-charge work, compute νQ and η for each, and compare with NMR values of νQ ≈ 190–200 kHz and η ≈ 0.88. They find that a uniform orthorhombic distortion of the Na-O octahedra, with Na-O bonds elongated/compressed by about 0.52% along crystalline axes (Model A.3), best reproduces the NMR EFG parameters, and they identify this as a Q2 distortion mode. They also report that the EFG is insensitive to the type of magnetic ordering and only weakly dependent on U and J over the ranges tested. The conclusion asserts that this Q2 distortion is the main source of the EFG and that it lifts the j = 3/2 quartet into two Kramers doublets before the onset of canted ferromagnetic order.","tokens_in":15244,"tokens_out":4566,"duration_ms":47140,"significance":"If the central claim were fully supported, the paper would provide a valuable first-principles-based identification of the local structural distortion in the broken local point symmetry (BLPS) phase of a strongly spin-orbit-coupled Mott insulator, complementing NMR data and testing a Jahn-Teller-type mechanism in a 5d^1 system. The work has clear strengths: it goes beyond the earlier point-charge approximation by computing EFG from the self-consistent DFT charge density; it systematically tests six families of distortion models, including negative results for Models C, D, and E; it checks PAW pseudopotential convergence, k-point and cutoff convergence, and U/J sensitivity (Table XI); and it openly reports functional dependence and the results of structure relaxation in Appendix E. These sensitivity studies and the qualitative symmetry argument based on η ≈ 1 with principal axes along crystalline axes give nontrivial support to the Q2-mode assignment.","major_comments":[{"comment":"The amplitude of the Model A distortion is a fitted, hand-imposed parameter rather than a first-principles output. The text states that distortions of 0.53–0.55% are chosen because they \"can produce the desired EFG parameters,\" and Appendix E reports that full relaxation of Model A.3 lowers the total energy by 0.3 eV and reduces the EFG parameters to roughly half of their unrelaxed values. Consequently, the structure that matches the NMR data is not a stationary point of the same DFT+U functional, and the quantitative agreement (νQ ≈ 183–203 kHz, η ≈ 0.79–0.99 in Table II) is obtained by construction. The conclusion in Section V that the calculations \"explicitly show\" the Q2 distortion to be the main source of the EFG therefore overstates what is demonstrated. The revision should either provide an independent determination of the distortion amplitude (for example, a relaxed low-temperature structure in which the Q2 mode is a genuine local minimum) or explicitly reframe the result as a consistency/fitting study and identify what additional evidence would make the assignment predictive.","section":"Section IV A, Table II, and Appendix E"},{"comment":"The quantitative EFG prediction is strongly functional-dependent. For the same Model A.3 distortion, PBE0 gives η = 0.46 and νQ = 260 kHz (Table X), PBE+SOC+U gives νQ = 128 kHz with η = 0.803 (Table XII in Appendix D), while the GGA+SOC+U calculation gives νQ in the 183–203 kHz range with η ≈ 0.79–0.99. The match to experiment is thus achieved only with one exchange-correlation functional at a fitted distortion amplitude. The paper should quantify this functional uncertainty and justify why GGA+SOC+U is the appropriate level for EFG prediction in this strongly correlated system, or temper the claim that the Q2 assignment is robustly established.","section":"Section IV C, Table X, and Appendix D"},{"comment":"The claim that the Q2 distortion \"lifts the j = 3/2 quartet to two Kramer doublets\" is not derived from the EFG calculations. The authors explicitly acknowledge on page 8 that \"we cannot provide proof of this hypothesis since a systematic theoretical framework for the description of the spin orbit channels in the strong SOC case is lacking.\" The conclusion section should clearly separate the supported statement (the experimental EFG is consistent with a static Q2-type local distortion) from the speculative electronic-level mechanism, and should not present the level-splitting scenario as a demonstrated result of the calculations.","section":"Section IV A (Q2 discussion) and Section V"}],"minor_comments":[{"comment":"The sign convention for the distortion is contradictory: the text defines elongation as positive and compression as negative, while the Table II caption states \"Positive distortions indicate compression and negative distortions indicate elongation.\" Please make the convention consistent throughout.","section":"Section IV A and Table II caption"},{"comment":"The table is titled \"Model B.3\" while the surrounding text refers to Model B.2; please correct the label so the table matches the text.","section":"Appendix D, Table XIII"},{"comment":"The summary of the relaxation calculations is too brief to be useful. Please report the relaxed lattice parameters, the residual distortion amplitudes, and the resulting EFG tensor components rather than only stating that the values are \"roughly half\" of the unrelaxed values.","section":"Appendix E"},{"comment":"The word \"orthohombic\" should be \"orthorhombic\".","section":"Introduction, page 2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is best viewed as a careful consistency/fitting study that discriminates between candidate local distortions using NMR EFG data. Its main quantitative result depends on a hand-imposed distortion amplitude that is not a DFT equilibrium structure, and the functional dependence of the EFG is substantial; the present wording of the abstract and conclusion goes beyond the evidence. I do not recommend rejection because the systematic model comparison and the negative results for Models C–E are valuable, and the authors have been transparent about the relaxation and functional issues. A major revision that reframes the central claim and addresses the three points above would make the paper publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The useful part of this paper is the systematic first-principles EFG survey of six distortion models for BNOO. It confirms, with real charge densities rather than point charges, that only Models A and F2 put Vzz along the crystalline axes with near-unity eta, and among those, Model A matches the NMR nuQ and eta best. That is a genuine step beyond Ref. 20, and the insensitivity of the EFG to the assumed magnetic order is a new, clean result. The paper is also honest about its own machinery: U/J dependence is tabulated, pseudopotential choices are tested, and the PBE0 and PBE+SOC+U disagreements are shown rather than hidden.\n\nWhere it overreaches is the phrase “explicitly show.” The 0.52% distortion in Model A.3 is not an ab initio prediction; it is an amplitude selected to make the computed splitting land in the experimental window. Appendix E says that fully relaxing that structure costs 0.3 eV in energy and cuts the EFG roughly in half, meaning the static model is not a stationary point of the same functional. So the calculation does not demonstrate that the actual BLPS phase contains that distortion—it demonstrates that among the static models considered, one specific Q2-type amplitude reproduces the NMR observables. The mode assignment is plausible and symmetry supports it, but the quantitative match is by construction. The functional sensitivity (PBE0 gives eta 0.46, PBE+SOC+U gives nuQ 128 kHz) further weakens the claim that the distortion is uniquely pinned. The Jahn-Teller / Kramers-doublet language in the conclusion is also speculative, and the authors themselves concede in Section IV A that they cannot prove it.\n\nAll that said, this is still a worthwhile paper for the osmate-double-perovskite community. The calculations are reproducible, the uncertainty is documented, and the structural conclusion is consistent with the prior point-charge work. The right fix is to retitle the central claim: Model A is the best among the candidates, not the proven source. I would send it to a serious referee. The overclaim is correctable, and the underlying data are solid enough to merit publication in a reasonable journal after revision.","headline":"A careful DFT+U study that tentatively validates the Q2 distortion in Ba2NaOsO6, but the central claim is overstrong because the distortion amplitude is fitted and the relaxed structure would not reproduce the NMR data.","tokens_in":15864,"tokens_out":2625,"would_cite":true,"duration_ms":30665,"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":"The paper claims that the electric field gradient seen in Na NMR of Ba2NaOsO6 is caused by a static orthorhombic Q2 distortion of the Na-O octahedra, not by the magnetic order, and that this distortion lifts the j=3/2 quartet into two…","keywords":["Ba2NaOsO6","electric field gradient","NMR","DFT+U","spin-orbit coupling","Mott insulator","Q2 distortion mode","broken local point symmetry"],"falsifier":"A direct structural measurement of the local Na-O bond lengths in the BLPS phase (for example, EXAFS or X-ray pair distribution function analysis) that does not find the about 0.52% Q2 pattern would falsify the claim; so would a fully relaxed DFT structure, which has lower energy, that still fails to reproduce the observed $\\nu_Q \\approx 190$–200 kHz splitting.","tokens_in":14709,"feed_emoji":"🔬","tokens_out":11465,"duration_ms":98457,"temperature":0.7,"pith_summary":"The paper sets out to identify the structural change that breaks local cubic symmetry in the $5d^1$ Mott insulator Ba2NaOsO6 before it enters its magnetically ordered state. It combines 23Na NMR quadrupole data with DFT+U calculations of the electric field gradient (EFG) tensor at the sodium site, testing six families of octahedral distortion. The conclusion is that the NMR signal is produced by a static orthorhombic Q2 distortion of the Na-O octahedra — about 0.52% elongation along one cubic axis and compression along another — and that this distortion, not the magnetic or spin-orbit state, is what generates the electric field gradient. The result matters because it turns the broken local point symmetry phase into a concrete lattice distortion and links that distortion to the lifting of the $j = 3/2$ quartet into two Kramers doublets before canted ferromagnetic order appears.","feed_headline":"Distortion, not magnetism, creates the NMR signature in Ba2NaOsO6","feed_subtitle":"First-principles EFG calculations trace the NMR splitting to a Q2 octahedral distortion, not to magnetism.","key_machinery":"The central object is the electric field gradient tensor at the 23Na nucleus, computed from the DFT+U charge density; its largest eigenvalue $V_{zz}$ and asymmetry parameter $\\eta$ are the direct theoretical counterparts of the NMR observables $\\nu_Q$ and $\\eta$. The comparison is carried across the six distortion models proposed in an earlier point-charge study, and the argument hinges on which model, at what distortion amplitude, reproduces both the observed splitting and the cubic alignment of the principal axes. The winning structure is Model A, a static Q2 distortion mode of the Na-O octahedra — the Jahn-Teller-active displacement that elongates one octahedral axis, compresses another, and leaves the third unchanged, giving orthorhombic local symmetry. That mode is what converts a charge redistribution into a finite, experimentally visible EFG, and it is also the symmetry lowering that the paper connects to splitting the $j = 3/2$ quartet into two Kramers doublets.","core_discovery":"On the paper's own terms, the central discovery is that a specific local orthorhombic distortion of the Na-O octahedra reproduces the experimentally observed EFG parameters at the 23Na site: a Q2 mode with Na-O bonds elongated along the $a$ axis and compressed along the $c$ axis by roughly 0.52%, giving $\\nu_Q \\approx 190$–200 kHz, $\\eta \\approx 0.8$–1, and the principal EFG axis aligned with a cubic axis. The EFG is nearly identical whether the underlying magnetic order is canted ferromagnetic, FM[110], or absent altogether, so the distortion alone sets the quadrupole splitting; magnetic order and spin-orbit coupling only broaden the NMR lines. Rotational, tilt, and GdFeO3-type distortions are ruled out because their principal axes point along diagonal directions and their splittings disagree with experiment. The paper therefore claims that the broken local point symmetry phase in BNOO is a static Q2 Jahn-Teller-like distortion that lifts the $j = 3/2$ quartet into two Kramers doublets before long-range magnetic order sets in.","pith_inferences":["If the static Q2 distortion is the true order parameter of the BLPS transition, the transition itself is a local structural (Jahn-Teller-type) event that can occur independently of magnetism; the canted ferromagnetic easy axis may then be set by the coupling of this strain to spin-orbit-coupled moments rather than by exchange alone.","Since full DFT relaxation lowers the energy by 0.3 eV and halves the computed EFG, the fitted 0.52% distortion is not a DFT equilibrium structure; the real BLPS phase may involve a dynamic or partially averaged distortion, or the functionals may underestimate the Jahn-Teller coupling, so mapping the potential-energy surface along the Q2 coordinate would be a direct test.","Carrying the same EFG-based structure identification to other $5d^1$ double perovskites could reveal whether a Q2 distortion generally precedes magnetic order in this family, turning NMR quadrupole parameters into a routine structural probe of octahedral deformations."],"forward_implications":["The broken local point symmetry phase in BNOO is a static Q2 orthorhombic distortion with Na-O bond changes of about 0.52%, not a rotation, tilt, or charge-only effect.","The quadrupole splitting seen in 23Na NMR can be read as a direct measure of the local Q2 distortion amplitude, essentially independent of the magnetic order present.","The rotated and tilted distortions favored by the earlier point-charge analysis are ruled out, since their principal axes do not align with the cubic axes.","The same DFT+NMR comparison can be applied to sister compounds such as Ba2LiOsO6 to test whether their different magnetic ground states reflect different local distortions.","The Q2 distortion provides the symmetry lowering that lifts the $j = 3/2$ moment into two Kramers doublets before canted ferromagnetic order appears."],"supporting_citations":[{"why":"Reports the 23Na NMR discovery of the broken local point symmetry phase and the canted ferromagnetic order; supplies the experimental phase sequence and magnetic structure the calculations reproduce.","marker":"[18]"},{"why":"Provides the measured EFG parameters and rotation patterns and defines the six distortion models A–F compared here; it is the experimental baseline and the earlier point-charge study this paper supersedes.","marker":"[20]"},{"why":"Supplies the U = 3.3 eV and J = 0.5 eV values and the magnetic entropy analysis indicating the j = 3/2 quartet splits into two Kramers doublets.","marker":"[15]"},{"why":"Earlier first-principles result finding FM[110] magnetism in BNOO, the magnetic state used to test the EFG's sensitivity to magnetic order.","marker":"[24]"},{"why":"Later first-principles calculation of the FM[110] state that this paper compares against when showing the EFG is insensitive to magnetic order.","marker":"[26]"},{"why":"Defines the Q2 and Q3 octahedral distortion modes used to identify the winning Model A distortion as a Q2 Jahn-Teller-active mode.","marker":"[40]"}],"fun_headline_variants":["Q2 distortion, not magnetism, drives NMR splitting in Ba2NaOsO6","Distortion alone sets NMR quadrupole in Ba2NaOsO6","EFG calc: Q2 mode, not magnetism, keys NMR line shape","Ba2NaOsO6: Distortion, not spin order, sets the EFG"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the assumption that the real BLPS structure is a static Q2 distortion of about 0.52%, even though fully relaxing the structure in DFT lowers the energy by 0.3 eV and reduces the computed EFG to roughly half the measured value.","fun_headline_variants_meta":{"raw":{"variants":["Q2 distortion, not magnetism, drives NMR splitting in Ba2NaOsO6","Distortion alone sets NMR quadrupole in Ba2NaOsO6","EFG calc: Q2 mode, not magnetism, keys NMR line shape","Ba2NaOsO6: Distortion, not spin order, sets the EFG"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000603,"raw_usage":{"total_tokens":2869,"prompt_tokens":1052,"completion_tokens":1817,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":668,"completion_tokens_details":{"reasoning_tokens":1729}},"tokens_in":668,"tokens_out":1817,"duration_ms":11812,"temperature":1.0,"reasoning_tokens":1729,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:24:16.161241+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct structural measurement of the local Na-O bond lengths in the BLPS phase (for example, EXAFS or X-ray pair distribution function analysis) that does not find the about 0.52% Q2 pattern would falsify the claim; so would a fully relaxed DFT structure, which has lower energy, that still fails to reproduce the observed $\\nu_Q \\approx 190$–200 kHz splitting.","supporting_citations":[{"cited_title":"Lu , author M","cited_arxiv_id":null,"evidence_quote":"Reports the 23Na NMR discovery of the broken local point symmetry phase and the canted ferromagnetic order; supplies the experimental phase sequence and magnetic structure the calculations reproduce."},{"cited_title":"Liu , author R","cited_arxiv_id":null,"evidence_quote":"Provides the measured EFG parameters and rotation patterns and defines the six distortion models A–F compared here; it is the experimental baseline and the earlier point-charge study this paper supersedes."},{"cited_title":"Erickson , author S","cited_arxiv_id":null,"evidence_quote":"Supplies the U = 3.3 eV and J = 0.5 eV values and the magnetic entropy analysis indicating the j = 3/2 quartet splits into two Kramers doublets."},{"cited_title":"\\ Lee \\ and\\ author W","cited_arxiv_id":null,"evidence_quote":"Earlier first-principles result finding FM[110] magnetism in BNOO, the magnetic state used to test the EFG's sensitivity to magnetic order."},{"cited_title":"Gangopadhyay \\ and\\ author W","cited_arxiv_id":null,"evidence_quote":"Later first-principles calculation of the FM[110] state that this paper compares against when showing the EFG is insensitive to magnetic order."},{"cited_title":"Khomskii ,\\ @noop title Transition metal compounds \\ ( publisher Cambridge University Press ,\\ year 2014 ) NoStop","cited_arxiv_id":null,"evidence_quote":"Defines the Q2 and Q3 octahedral distortion modes used to identify the winning Model A distortion as a Q2 Jahn-Teller-active mode."}],"review_version":1}