{"id":"2406fe17-038a-46cf-9a1f-e27b5146bff3","arxiv_id":"1908.04455","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Ab initio molecular dynamics shows that thermal lattice fluctuations continuously halve the DFT+U band gap of monolayer 1T-TaS2 below the CDW transition, while the static CDW amplitude stays nearly constant.","lead":"In a simulated layer of the material 1T-TaS2, raising the temperature shrinks the electronic gap by half even before the lattice order breaks down. This suggests lattice vibrations alone can strongly soften the Mott insulating state.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim hinges on identifying <E_BO_g> with the observable gap; the thermally averaged spectral function can behave very differently, especially near the level crossings shown in Fig. 4(e-h).","rationale":"The paper is a serious computational study: the MD simulation is carefully described, the Wannier analysis provides a plausible microscopic mechanism (Δcs coupling), and the authors honestly list limitations including the neglect of electronic entropy. The low-temperature DFT+U gap matching STM is real supporting evidence. However, the strongest claim is framed around a specific number: the gap shrinks by half, from ~0.4 eV to ~0.2 eV, before the transition. That number is the time average of instantaneous Γ-point gaps. The reader's weakest assumption identified exactly this estimator problem, and I agree it is the most load-bearing concern. The reason is not that DFT+U is approximate—that is a stated limitation—but that the conversion from a set of instantaneous BO gaps to a finite-temperature spectral gap is not a simple average. In the extreme case of level crossings, configurations with zero gap must contribute metallic weight to the averaged spectral function, so the true gap can close while the averaged gap is still positive. Whether this effect changes the quantitative conclusion at intermediate temperatures is an empirical question that the current paper does not answer. A spectral-function test on the existing trajectories would settle it. If the test shows that <E_BO_g> tracks the spectral gap, the central claim stands; if not, the paper needs a revised estimator or a softened quantitative claim. Since the reader already issued CONDITIONAL, the verdict should remain unchanged: the paper is promising but needs this validation and possibly revised wording about the 'Mott gap'.","tokens_in":8443,"tokens_out":7743,"duration_ms":93377,"concrete_test":"From the stored MD trajectories, compute the thermally averaged Kohn–Sham density of states A(ω) = ⟨Σ_n δ(ω−ε_n(R))⟩ over snapshots at 5, 150, 250, and 275 K, defining the gap as the largest zero-weight interval around the chemical potential (or as the distance between the lower and upper Hubbard peaks if broadened). Compare this gap with <E_BO_g> from Fig. 3(c). If the averaged-spectral gap follows <E_BO_g> and shrinks by half by TC, the estimator is adequate; if A(ω) already shows subgap weight or a closed gap at 250 K, or if the gap is controlled by min_R g(R) rather than the mean, the central claim must be re-quantified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III defines E_BO_g as the instantaneous Γ-point HOMO–LUMO gap from DFT+U and plots its time average in Fig. 3(c). The central claim that dynamical CDW fluctuations shrink the Mott gap by half treats <E_BO_g> as the experimentally observable gap. But the thermally averaged single-particle spectral function—the quantity probed by STM—does not have a gap equal to the time average of instantaneous gap values. For each ionic configuration R, the Kohn–Sham spectrum has a gap g(R); the averaged spectral function A(ω)=∫P(R)A_R(ω)dR has zero weight only in the intersection of the individual gaps. Its gap is bounded by min_R g(R), not by ∫P(R)g(R)dR. When g(R) fluctuates strongly, <E_BO_g> can be much larger than the spectral gap, or can miss a closing gap. Fig. 4(e-h) explicitly shows level crossings at 275 K, so some snapshots have g≈0; the average spectral function should then develop subgap weight or close, while <E_BO_g> remains finite until averaged down. The comparison with STM in Sec. IV therefore validates a proxy, not the measured gap. This is load-bearing because the paper's quantitative prediction (0.4→0.2 eV) and the claim that the reduction is ten times k_B ΔT are both properties of the estimator, not of a directly computed spectral gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents ab initio molecular dynamics simulations of a 1T-TaS2 layer, using DFT+U to compute instantaneous electronic structures along the trajectory. The time-averaged Γ-point HOMO–LUMO gap is found to decrease from about 0.4 eV at 5 K to about 0.2 eV at the CDW transition temperature, while the gap of the time-averaged structure remains nearly constant. The authors interpret this as dynamical CDW fluctuations renormalizing the Mott gap and support this with Wannier analysis identifying the on-site energy Δcs as the key electron–phonon coupled parameter. They also compare with STM data and speculate on possible pseudogap phases.","tokens_in":8738,"tokens_out":5624,"duration_ms":51953,"significance":"If the central claim is correct, the work offers a general computational methodology for quantifying lattice entropy effects on Mott gaps and provides a plausible interpretation of 1T-TaS2's temperature-dependent spectral features. The paper has notable strengths: the Hubbard U is taken from an independent linear-response calculation rather than fitted, the MD equilibration is carefully checked via heating–cooling cycles, and the Wannier analysis provides a transparent physical mechanism (Δcs modulation). The qualitative conclusion that lattice dynamics, not the static CDW amplitude, drives the gap renormalization is compelling. However, the quantitative prediction rests on the identification of the time-averaged instantaneous gap with the observable spectral gap, which is not justified.","major_comments":[{"comment":"The time-averaged instantaneous Γ-point gap ⟨E_BO_g⟩ is not the gap of the thermally averaged single-particle spectral function that STM probes. For each ionic configuration R, the Kohn–Sham spectrum has a gap g(R); the disorder-averaged spectral function A(ω)=∫P(R)A_R(ω)dR has zero weight only in the intersection of the individual gaps, so its gap is at most min_R g(R), not ⟨g(R)⟩. Since Fig. 4(h) shows level crossings at 275 K, some snapshots have near-zero gaps, and the averaged spectral function should develop subgap weight or close, making ⟨E_BO_g⟩ an overestimate of the spectral gap. The quantitative claim of a half reduction (0.4→0.2 eV) and the comparison with STM in Sec. IV therefore validate a proxy, not the measured gap. The authors should either compute the actual spectral function from the MD trajectory or clearly re-frame the claim as a property of the instantaneous gap distribution.","section":"Sec. III, Fig. 3(c) and Sec. IV"},{"comment":"The gap is evaluated only at the Γ point. The authors mention in Sec. IV that the gap melting can be momentum dependent, but the MD trajectory data and the central Fig. 3(c) use only the Γ-point HOMO–LUMO gap. If the minimum gap lies elsewhere in the Brillouin zone, or if the STM dI/dV measurement is sensitive to k≠0 states, the quantitative reduction reported may be different. The authors should justify that the Γ-point gap is representative or at least discuss the uncertainty this introduces.","section":"Sec. II.A and Sec. III"},{"comment":"The authors correctly state that DFT+U is a mean-field approximation and that electronic entropy is missing, yet in Sec. IV they plot ⟨E_BO_g⟩ against experimental STM data as if it were the actual Mott gap. Given the issues above, the comparison is premature. The discussion would be strengthened by explicitly distinguishing the calculated quantity (a BO, mean-field, spin-polarized Γ-point gap) from the experimental Mott gap, and by treating the STM comparison as suggestive rather than confirmatory.","section":"Sec. II.D and Sec. IV"}],"minor_comments":[{"comment":"It is unclear why the initial spin polarization of the four SDs is set to be the same; some discussion of the sensitivity of the results to the spin configuration would be helpful.","section":"Sec. II.A"},{"comment":"The definition of φSD uses the SD positions from the CCDW phase even in the high-T phase; this is stated in the text but should also be noted in the figure caption for clarity.","section":"Sec. III, Fig. 3(a)"},{"comment":"No uncertainties are given for the Wannier-derived parameters. A brief statement about the robustness of the Wannierization would increase confidence in Fig. 3(d).","section":"Table I"},{"comment":"The phrase 'one order of magnitude larger than the lattice temperature variation' is ambiguous; the quantitative statement is Δ⟨E_BO_g⟩/k_B ΔT ≈ −10, which is a dimensionless ratio, not a comparison of energies.","section":"Sec. IV"},{"comment":"There are several typographical errors, including 'Accrodingly' (should be 'Accordingly'), 'investiations' (should be 'investigations'), and 'transtion' (should be 'transition').","section":"Sec. IV"},{"comment":"The reference for Ritschel et al. is given as Phys. Rev. B 11, 328 (2015), which appears incorrect; the proper citation is likely Phys. Rev. B 92, 115142 (2015).","section":"Reference 28"},{"comment":"The term 'site-selective Mott transition' is used; a brief definition or citation would help readers unfamiliar with the rare-earth nickelate literature.","section":"Sec. II.C"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely to be of interest to the condensed matter community, but the central quantitative claim is undermined by the proxy gap issue. The authors need to either compute a spectral function or substantially temper the quantitative statements. Given the emphasis on the number '0.4→0.2 eV', this is more than a presentation issue. I would advise the editor that the paper is not suitable for publication in its present form, but the underlying idea and MD approach are valuable and the authors may be able to address the concern."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper's real content is an AIMD-based protocol for tracking instantaneous DFT+U gaps and Wannier parameters in 1T-TaS2. The claim is that the time-averaged gap drops from ~0.4 eV at 5 K to ~0.2 eV near the transition while the static gap from the time-averaged structure barely moves, which they attribute to dynamical CDW fluctuations rather than a change in the mean CDW amplitude. The Wannier analysis identifies the on-site energy difference Delta_cs, not the hoppings, as the parameter that correlates with the instantaneous gap. That identification is the most interesting and portable result.\n\nThe computational setup is honest and reasonably careful. The supercell accommodates the sqrt(13) x sqrt(13) wavevector, they run heating-cooling cycles and check equilibration, and U is taken from a prior linear-response calculation rather than fitted. The limitations—no electronic entropy, mean-field DFT+U, small cell—are stated explicitly in Sec. II.D. The citation pattern is clean; the only self-citation is the prior Wannier model, used here for interpretation. I have no circularity concern.\n\nThe soft spot is load-bearing. The central observable <EBO_g> is the time average of instantaneous Gamma-point HOMO-LUMO gaps. The thermally averaged spectral function probed by STM has a gap that is bounded by the minimum instantaneous gap, not by the average of gaps. The figures show level crossings at 275 K, so some snapshots have near-zero gap; the true spectral function should develop subgap weight or close, while the average of gaps remains finite until the crossings dominate. Comparing <EBO_g> to STM Hubbard-peak splittings therefore validates a proxy, not the measured gap. That does not make the qualitative trend wrong—large fluctuations will still reduce the spectral gap—but it means the quantitative claim of a halving is an artifact of the estimator. The Gamma-point-only sampling and the use of a Kohn-Sham gap as a stand-in for the Mott gap are additional, smaller caveats.\n\nWho is this for? People working on electron-phonon coupling in CDW-Mott insulators will find the Delta_cs result and the general methodology worth knowing, even if they can't take the quantitative curve at face value. It deserves a serious referee, but the revision needs to either justify <EBO_g> in terms of the occupied spectral function, compute the spectral gap more directly, or soften the Mott-gap wording throughout. As it stands I would not cite the 0.4-to-0.2 eV numbers, but I would cite the method if it survives revision.\n\nMy recommendation: send it to peer review, but flag the estimator issue prominently for the authors.","headline":"A careful AIMD study of 1T-TaS2 whose qualitative mechanism is plausible, but the headline 'gap halved by lattice entropy' rests on a questionable identification of the time-averaged Kohn-Sham gap with the measured spectral gap.","tokens_in":9274,"tokens_out":4332,"would_cite":false,"duration_ms":38712,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Lattice vibrations halve the Mott gap in 1T-TaS2","keywords":["1T-TaS2","Mott gap","lattice entropy","charge density wave","ab initio molecular dynamics","DFT+U","electron-phonon coupling","metal-insulator transition"],"falsifier":"A decisive test would be a scanning-tunneling-spectroscopy measurement that tracks the Mott-gap width continuously from 5 K to above 250 K while a diffraction measurement monitors the CDW amplitude: the paper predicts the gap shrinks by roughly half while the CDW amplitude stays constant, so observing a nearly constant gap until the first-order transition would falsify the claim.","tokens_in":8242,"feed_emoji":"⚛️","tokens_out":9687,"duration_ms":85176,"temperature":0.7,"pith_summary":"This paper argues that lattice vibrations alone can shrink the Mott gap of the layered compound 1T-TaS2 by half before the charge-density-wave order is destroyed. The authors perform first-principles molecular dynamics and record the instantaneous electronic gap at every time step; the time-averaged gap falls from about 0.4 eV at 5 K to about 0.2 eV as the temperature approaches the first-order transition near 250–300 K, while the gap computed from the time-averaged structure stays almost constant. Because the CDW amplitude barely changes in this window, the reduction is attributed to dynamical lattice fluctuations, i.e., lattice entropy, rather than to a weakened static distortion. If correct, the result establishes a general computational route for quantifying lattice-entropy effects in metal-insulator transitions and identifies the orbital-energy difference within the Star-of-David cluster as the electronic parameter that carries the coupling.","feed_headline":"Lattice vibrations halve the Mott gap in 1T-TaS2","feed_subtitle":"Simulations show the gap shrinking from 0.4 to 0.2 eV while the charge-density-wave amplitude stays nearly constant.","key_machinery":"The load-bearing object is the time-averaged Born-Oppenheimer gap $\\langle E_g^{\\text{BO}}\\rangle$, defined as the mean over the molecular-dynamics trajectory of the instantaneous $\\Gamma$-point gap between the highest occupied and lowest unoccupied DFT+U levels. Comparing $\\langle E_g^{\\text{BO}}\\rangle$ with the static gap $E_g^{\\text{static}}$, recomputed for the time-averaged structure at each temperature, isolates the effect of lattice dynamics from that of static distortion. The CDW order parameter $\\varphi_{\\text{SD}} = \\bar{d}_{\\text{inter}} - \\bar{d}_{\\text{intra}}$ distinguishes amplitude changes from fluctuation effects, and maximally-localized Wannier functions map each instantaneous structure onto a tight-binding model whose onsite-energy difference $\\Delta_{cs}$ is the parameter that tracks the gap.","core_discovery":"The central discovery is a large, purely dynamical renormalization of a Mott gap. In a DFT+U molecular-dynamics simulation of a single 1T-TaS2 layer, the authors compare the static gap $E_g^{\\text{static}}$ from the time-averaged lattice structure with the time-averaged Born-Oppenheimer gap $\\langle E_g^{\\text{BO}}\\rangle$, which is the mean over the trajectory of the instantaneous $\\Gamma$-point highest-occupied to lowest-unoccupied gap. $E_g^{\\text{static}}$ remains nearly constant below the transition, whereas $\\langle E_g^{\\text{BO}}\\rangle$ drops from roughly 0.4 eV at 5 K to roughly 0.2 eV at $T_C$, i.e., by about half. Since the CDW order parameter $\\varphi_{\\text{SD}}$ changes little over this range, the gap shrinking is driven by thermal CDW fluctuations rather than by amplitude reduction. A Wannier-function projection of the instantaneous structures shows that the gap correlates with $\\Delta_{cs}$, the onsite energy difference between the central and edge orbitals of the Star-of-David cluster, while the hopping parameters are nearly inert; this identifies $\\Delta_{cs}$ as the key electron-phonon coupling channel, echoing the site-selective Mott mechanism proposed for rare-earth nickelates.","pith_inferences":["Because the simulation uses only the $\\Gamma$ point and a small supercell, the true spectral gap renormalization could be momentum-dependent; extending the same time-averaging to k-point samples would test whether the halving is uniform across the Brillouin zone.","If lattice entropy is the dominant mechanism, then altering the vibrational spectrum—for example by isotope substitution or by strain that does not change the static structure—should shift the gap at a fixed temperature, a prediction the paper does not make.","The paper's use of the mean instantaneous gap as the experimental gap estimator is an approximation; near $T_C$ level crossings occur, so the true spectral gap might close even faster than $\\langle E_g^{\\text{BO}}\\rangle$, possibly producing a pseudogap regime below the first-order transition."],"forward_implications":["In 1T-TaS2, the Mott gap measured below the transition should be interpreted as a thermally renormalized quantity, so STM and transport data must be compared with $\\langle E_g^{\\text{BO}}\\rangle$, not with the static DFT+U gap.","A nearly temperature-independent CDW amplitude does not imply a temperature-independent electronic gap; dynamical fluctuations of that amplitude can dominate the gap's temperature dependence.","The same methodology—time-averaging instantaneous gaps over ab initio molecular dynamics—can be applied to other transition-metal dichalcogenides and oxides to estimate the magnitude of lattice-entropy effects in their metal-insulator transitions.","The identification of $\\Delta_{cs}$ as the relevant coupling parameter suggests that the site-selective Mott scenario (an onsite potential difference from a lattice distortion) is the microscopic channel through which lattice vibrations renormalize the gap."],"supporting_citations":[{"why":"Establishes the Wannier-orbital picture of the Star-of-David cluster and the DFT+U gap at 5 K that matches STM dI/dV spectrum.","marker":"18"},{"why":"Provides the experimental phase diagram and transition temperatures (T_CCDW, T_NC) used to anchor the simulation temperature scale.","marker":"17"},{"why":"Reports STM gap at 78 K that the paper compares with its $\\langle E_g^{\\text{BO}}(T)\\rangle$ curve.","marker":"19"},{"why":"Reports STM spectra at 130 K showing a V-shaped gap that the paper uses to motivate the pseudogap discussion.","marker":"20"},{"why":"Supplies the effective U = 2.27 eV from linear response used in the DFT+U calculations.","marker":"38"},{"why":"Proposes the site-selective Mott transition scenario in nickelates that the paper invokes to interpret the role of $\\Delta_{cs}$.","marker":"44"},{"why":"Provides the prior estimate that electronic entropy alone can reduce the Mott gap by an amount comparable to the lattice entropy effect found here.","marker":"46"},{"why":"Introduces the simplified rotationally invariant DFT+U approach used throughout.","marker":"37"}],"fun_headline_variants":["Thermal fluctuations halve Mott gap in 1T-TaS2","Lattice entropy halves Mott gap in 1T-TaS2","CDW fluctuations shrink Mott gap by half in 1T-TaS2","Instantaneous lattice motion halves 1T-TaS2 Mott gap","Thermal CDW fluctuations reduce 1T-TaS2 gap by half"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that averaging the instantaneous electronic gap over a lattice-vibration trajectory, with spins frozen and electrons at zero temperature, faithfully reproduces the gap measured in experiments.","fun_headline_variants_meta":{"raw":{"variants":["Thermal fluctuations halve Mott gap in 1T-TaS2","Lattice entropy halves Mott gap in 1T-TaS2","CDW fluctuations shrink Mott gap by half in 1T-TaS2","Instantaneous lattice motion halves 1T-TaS2 Mott gap","Thermal CDW fluctuations reduce 1T-TaS2 gap by half"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000591,"raw_usage":{"total_tokens":2794,"prompt_tokens":991,"completion_tokens":1803,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":607,"completion_tokens_details":{"reasoning_tokens":1707}},"tokens_in":607,"tokens_out":1803,"duration_ms":13629,"temperature":1.0,"reasoning_tokens":1707,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:41:48.997755+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be a scanning-tunneling-spectroscopy measurement that tracks the Mott-gap width continuously from 5 K to above 250 K while a diffraction measurement monitors the CDW amplitude: the paper predicts the gap shrinks by roughly half while the CDW amplitude stays constant, so observing a nearly constant gap until the first-order transition would falsify the claim.","supporting_citations":[{"cited_title":"Qiao , author X","cited_arxiv_id":null,"evidence_quote":"Establishes the Wannier-orbital picture of the Star-of-David cluster and the DFT+U gap at 5 K that matches STM dI/dV spectrum."},{"cited_title":"Cho , author Y.-H","cited_arxiv_id":null,"evidence_quote":"Reports STM gap at 78 K that the paper compares with its $\\langle E_g^{\\text{BO}}(T)\\rangle$ curve."},{"cited_title":"Lutsyk , author M","cited_arxiv_id":null,"evidence_quote":"Reports STM spectra at 130 K showing a V-shaped gap that the paper uses to motivate the pseudogap discussion."},{"cited_title":"Darancet , author A","cited_arxiv_id":null,"evidence_quote":"Supplies the effective U = 2.27 eV from linear response used in the DFT+U calculations."},{"cited_title":"Park , author A","cited_arxiv_id":null,"evidence_quote":"Proposes the site-selective Mott transition scenario in nickelates that the paper invokes to interpret the role of $\\Delta_{cs}$."},{"cited_title":"\\ Han , author C","cited_arxiv_id":null,"evidence_quote":"Provides the prior estimate that electronic entropy alone can reduce the Mott gap by an amount comparable to the lattice entropy effect found here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the simplified rotationally invariant DFT+U approach used throughout."}],"review_version":1}