{"id":"3066969a-e8b4-49cd-9251-305d0f02e87f","arxiv_id":"2607.25359","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Van Hove singularities in twisted double bilayer graphene produce large, tunable, positive Nernst peaks at about 1 K, reproduced by semiclassical Boltzmann theory and a minimal two-band model.","lead":"This paper reports large Nernst thermoelectric signals appearing exactly at the van Hove singularities of twisted double bilayer graphene, with values up to about 40 microvolts per kelvin per tesla at 1 K. The result suggests the Nernst effect can serve as a sensitive probe of Fermi-surface topology in moiré materials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative theory-data agreement relies on an unconstrained DOS-broadening correction; without independent measurement, the 'quantitative' identification of the Nernst peak with the Lifshitz transition is not established.","rationale":"The paper presents a compelling experiment: two devices, independent Hall-based vHS mapping, a BLG control, and a semiclassical calculation that matches the peak positions and the sign of the Nernst signal. The main experimental observation—a large, tunable, positive Nernst peak at the vHS—is well supported. I find no fatal flaw in the data analysis or in the application of the Mott relation; Eq. (1) is a general consequence of the Sommerfeld expansion for the measured conductivity tensor, so the reader's first listed premise (single-Drude mobility) is not the critical assumption. The genuine soft spot is the unquantified DOS-broadening correction used to reconcile the calculated and measured magnitudes. The paper claims quantitative agreement but then introduces an ad hoc broadening without independent justification. This weakens the strength of the central claim but does not overturn it. The reader's verdict of CONDITIONAL is appropriate, and my analysis does not change it.","tokens_in":21897,"tokens_out":21851,"duration_ms":226072,"concrete_test":"Convolve the continuum-model DOS with a Gaussian of width σ, where σ is independently determined (e.g., from the width of the measured Hall-density divergence near the vHS or from the temperature smearing of Sxx around the vHS). Recompute Eq. (1) using the measured σxx(n) and σxy(n) and this broadened ρ(n). If the resulting Syx peak height falls within the experimental uncertainty of the measured Syx, the quantitative claim is supported; if not, the agreement is qualitative only.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanistic claim rests on Eq. (1) and the comparison in Fig. 4d. The calculated Syx is larger than the measured peak, and the paper attributes this to twist-angle inhomogeneity broadening the theoretical DOS, but no independent measurement of that broadening is provided. The Conclusion states the semiclassical analysis 'reproduces this response quantitatively,' yet the only adjustable element in the comparison is the unmeasured DOS broadening, making the magnitude agreement effectively a one-parameter fit. Peak positions and sign are robust, but the quantitative claim is not supported. The reader's additional concern about energy-dependent τ / single-Drude mobility is less damaging: Eq. (1) follows from the standard Mott relation applied to the measured full conductivity tensor, and the Drude mobility extraction is mainly used to justify the large-Hall-angle regime, not as the essential route to Syx. The load-bearing weakness is therefore the unconstrained DOS-broadening correction, which the paper itself flags in the main text ('the magnitude of the measured signal is on the smaller side, which could be due to uncertainty in DOS...').","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports Nernst-effect measurements in twisted double bilayer graphene (tDBLG) devices at twist angles 1.66° and 1.20°, finding sharp positive peaks in the transverse thermoelectric coefficient Syx at carrier densities that coincide with van Hove singularities (vHSs) identified independently via Hall-density divergences and continuum-model DOS. The peak position is tunable by displacement field and the magnitude reaches ~40 µV K⁻¹ T⁻¹ at ~1 K (D2). The authors explain the enhancement using the semiclassical Mott relation [Eq. (1)] fed with the measured conductivity tensor and a theoretical DOS, and they support the mechanism with a minimal two-band triangular-lattice model that reproduces a positive Nernst peak at the vHS. They further claim that the positive sign of Syx around a vHS is universal, making the Nernst effect a probe of Fermi-surface topology in moiré materials.","tokens_in":22156,"tokens_out":9411,"duration_ms":98497,"significance":"If the quantitative claim held, this would be the first equilibrium Nernst study of a twisted moiré system at a vHS and would establish a new experimental probe of Lifshitz transitions. The paper's strengths include: two devices with different twist angles; an independent identification of the vHS trajectory from the Hall-density divergence (Fig. 2c); a BLG control sample (SI S9); comparison of the full thermoelectric tensor α with the Mott prediction (SI S13); and a transparent minimal model that captures the trend. These elements make the central qualitative mechanism (vHS-driven enhancement, positive peak at the Lifshitz transition) plausible. However, the quantitative agreement is not settled: the calculated Syx systematically exceeds the measured value, and the paper's own text attributes the discrepancy to unquantified DOS broadening from twist-angle inhomogeneity. The 'universal sign' claim also goes beyond the evidence presented.","major_comments":[{"comment":"The comparison in Fig. 4d uses Eq. (1) with the measured conductivity tensor and a theoretical DOS. The calculated Syx (orange) is systematically larger than the measured peak, and the main text states this 'could be due to uncertainty in DOS, since the real device with angle inhomogeneity can broaden the DOS.' No independent measurement or even a fitted value of this broadening is provided, and the SI minimal-model section explicitly says that model magnitudes 'should not be compared with the experimental results.' The Conclusion's claim that the semiclassical analysis 'reproduces this response quantitatively' is therefore not supported; the comparison is a consistency check with an unconstrained broadening parameter. Please either remove 'quantitatively' and describe the agreement as qualitative in peak position and sign, or include a quantitative fit over a plausible DOS-broadening ra","section":"Fig. 4d and Conclusion"},{"comment":"The claim that the positive sign of Syx is 'universal around a vHS' goes beyond the evidence. The experiments cover one material (tDBLG at two twist angles), and the minimal model in SI S12 uses a symmetric two-band triangular lattice with equal hoppings (te=th) and a fixed Gaussian broadening. The sign in Eq. (S14) depends on a combination of σxx and σxy derivatives; SI Fig. S22 shows that varying the mass ratio changes peak heights and can introduce negative Nernst regions at high temperature, although the near-vHS peak remains positive in the cases shown. A universality claim requires either a proof from the Boltzmann expression or a systematic scan over representative band parameters and scattering conditions. Otherwise, please soften to 'positive in the systems studied here.'","section":"Abstract, Conclusion, SI S12"},{"comment":"The sign convention for the Nernst coefficient is inconsistent as presented. The introduction defines N = -Syx = -Ey/∇xT (so Syx = Ey/∇xT), and later the paper refers to positive Syx as a positive Nernst signal. In SI S10, Syx is defined as E_y^th/∇xT under the 'vortex convention,' and the sign argument yielding a positive peak is based on that definition. Since the universality claim concerns the sign of Syx, the reader cannot tell whether the reported positive peaks correspond to a positive or negative conventional Nernst coefficient. Please state the convention used for all reported Syx values, define N consistently, and verify that the comparison with Bi and PrFe4P12 uses the same sign convention.","section":"Introduction and SI S10"}],"minor_comments":[{"comment":"The notation ∂tanΘ_H/∂n ρ(n) is redundant/ambiguous; since d/dn = (1/ρ) d/dε, the expression equals ∂tanΘ_H/∂ε but the placement of ρ(n) is confusing. In Fig. 4e, 'without the DOS contribution' is unclear. Please define the derivative variable and state whether ρ(n) is the total DOS per unit area.","section":"Eq. (1) and Fig. 4e"},{"comment":"For D2, the theoretical DOS at D=0 is shown but no comparison like the middle panel of Fig. 2c is given; please add the corresponding nH cut and state the twist-angle uncertainty for both devices.","section":"Fig. 2 and SI S4"},{"comment":"Several typos: 'postive' in SI S7 title; 'the the' in the text around Fig. 1; 'moire band' should be 'moiré band'; missing comma after 'CNP' in Fig. 3a. A careful proofread is needed.","section":"Throughout"},{"comment":"The text refers to 'section 10 of the SI' and 'SI section 11' but the supplementary file does not consistently number sections; ensure cross-references match.","section":"References to SI"},{"comment":"The Hall mobility extraction uses σxx(B)=σ0/(1+µ²B²); please state whether µ extracted this way is consistent with tanΘ_H = σxy/σxx at the same densities, since the large-Hall-angle regime is important for Eq. (1).","section":"Methods / Mobility extraction"}],"recommendation":"major_revision","confidential_remarks":"The experimental results appear solid and the qualitative mechanism is convincing. The main issues are the unquantified DOS broadening used to excuse the magnitude mismatch between theory and experiment, the overclaim of universality, and the sign-convention confusion. These are fixable in revision and do not require new experimental runs unless the authors want to support the quantitative claim more strongly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is the first equilibrium Nernst measurement at a van Hove singularity in a twisted graphene stack, and the central observation holds up. Two tDBLG devices show large, positive, displacement-field-tunable Nernst peaks (about 40 µV/K/T at ~1 K for the 1.2° device) that sit on the same n–D trajectories as the Hall-density divergences used to map the vHS. Seebeck changes sign across the singularity; Nernst peaks at it. The BLG control and the drop in peak height with temperature fit the semiclassical picture. I came away believing the qualitative claim.\n\nWhat is actually new: no equilibrium Nernst study at a vHS in any twisted platform appears in the literature; ref. 38 is photo-Nernst, a different regime. The minimal two-band model is a useful addition—it shows a vHS alone produces a positive Nernst peak, with the sign fixed by the effective-mass reversal at the saddle point.\n\nSoft spots, in order of real softness. First, the 'quantitative' language in the Conclusion overstates what Fig. 4d supports. Eq. (1) is the Mott relation applied to the measured conductivity tensor and a theoretical DOS. It nails peak positions and sign, but the magnitude mismatch is absorbed by an unmeasured DOS-broadening correction. The authors flag this in the main text ('could be due to uncertainty in DOS'), which is honest, but it makes the magnitude agreement a one-parameter fit, not a parameter-free prediction. 'Consistent with' is the accurate claim. Second, the main Syx values lack error bars, and the 'universal positive Nernst at vHS' conclusion rests on two devices. Raw data and code are not deposited. These are fixable.\n\nOne place I'd push back on the reader's report: the worry about energy-dependent τ and single-Drude mobility is less damaging than it looks. Eq. (1) uses the measured full conductivity tensor; the Drude extraction only justifies the µB > 1 regime. The fragile premise is the DOS broadening, not the mobility model.\n\nThis paper is for anyone using Nernst to probe Fermi-surface topology in moiré systems. It deserves a serious referee: send it out, and ask for error bars, data/code, and calmer wording on the quantitative claim.","headline":"First equilibrium Nernst data at a van Hove singularity in a twisted platform — real result, but the 'quantitative' theory agreement is a Mott-relation consistency check with an unmeasured DOS-broadening knob.","tokens_in":22715,"tokens_out":4225,"would_cite":true,"duration_ms":38372,"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":"Van Hove singularities in twisted double bilayer graphene generate a giant, universally positive Nernst signal.","keywords":["van Hove singularity","Nernst effect","twisted double bilayer graphene","Lifshitz transition","moiré materials","semiclassical Boltzmann transport","Mott relation","thermoelectric transport"],"falsifier":"A decisive check would be to measure the energy-dependent mobility near the vHS directly—for instance, via quantum oscillations that resolve mobility per Landau level—and recompute Eq. (1) with that mobility. If the predicted peak is absent, or if its sign flips at a vHS where the band structure is independently known, the semiclassical attribution to Lifshitz transitions would fail.","tokens_in":21777,"feed_emoji":"🧲","tokens_out":4171,"duration_ms":41077,"temperature":0.7,"pith_summary":"This paper tries to establish that the Nernst effect—the transverse voltage produced by a longitudinal temperature gradient in a magnetic field—is dramatically enhanced at van Hove singularities in twisted double bilayer graphene. It reports positive Nernst peaks at the valence and conduction vHS, tunable by displacement field, reaching about 40 microvolts per kelvin per tesla at about 1 K, comparable to the best-known Nernst materials. The enhancement is attributed to Lifshitz transitions: at the saddle point the effective mass changes sign, so hot and cold carriers deflect to the same side instead of cancelling. Semiclassical Boltzmann transport with the Mott relation quantitatively reproduces the data, and a minimal two-band tight-binding model shows a positive Nernst peak as a universal signature of a vHS. If correct, the paper establishes the Nernst effect as a sensitive probe of Fermi-surface topology in moiré materials.","feed_headline":"Van Hove points create a giant Nernst signal in twisted graphene","feed_subtitle":"Peaks at both conduction- and valence-band saddle points reach 40 µV/K/T at 1 K, tracing Lifshitz transitions.","key_machinery":"The central identity is the semiclassical generalization of the Mott relation, Eq. (1) in the paper: Syx = (π²/3)(k_B²T/e) cos²Θ_H ∂tanΘ_H/∂n ρ(n)|εF, where Θ_H is the Hall angle, ρ(n) is the density of states, and εF is the Fermi energy. This formula converts the measured Hall angle and theoretical DOS into a Nernst signal. Near a vHS, ∂tanΘ_H/∂n is large and the DOS diverges logarithmically, while the large Hall angle (µB > 1) prevents the small-angle cancellation that suppresses the Nernst response in ordinary metals. A supporting two-band tight-binding model demonstrates the same positive Nernst peak from the vHS.","core_discovery":"The paper's central claim is that van Hove singularities in twisted double bilayer graphene produce an unusually large, universally positive Nernst signal that tracks Lifshitz transitions. The sign and magnitude are explained by the effective-mass sign reversal at the saddle point: carriers on either side of the chemical potential deflect in the same transverse direction, so the Nernst response does not cancel. The measured Nernst signal can be computed from the conductivity tensor and theoretical density of states via a semiclassical Mott relation, giving positive peaks at both conduction- and valence-band vHS. This mechanism is distinct from conventional metals, where the Nernst sign depen","pith_inferences":["If the mechanism is generic, other moiré or flat-band materials with saddle-point singularities should show similar Nernst peaks, making the effect a rapid screening tool for detecting vHS without requiring high-resolution spectroscopy.","The paper's sign argument suggests that a negative Nernst peak around a vHS would signal a different origin (e.g., interactions or fluctuations), providing a discriminator in correlated regimes.","The semiclassical framework predicts a specific temperature dependence of the peak height; deviations from the Mott prediction at higher temperatures would indicate the onset of inelastic or multi-band transport.","One testable extension would be to measure the Nernst signal in a strained or tilted moiré system where the saddle-point topology is altered, checking whether the positive sign is indeed universal."],"forward_implications":["In tDBLG, the Nernst signal can be tuned by displacement field, reaching about 40 µV/K/T at ~1 K for a 1.20° twist, placing it among the strongest known Nernst materials.","The positive sign of the Nernst signal around a vHS is argued to be universal, independent of carrier type, unlike the sign behavior in metals and semimetals.","Semiclassical Boltzmann transport quantitatively reproduces the measured thermoelectric coefficients from the measured conductivity and theoretical DOS, including the temperature dependence and B-field linearity.","A minimal two-band tight-binding model with a vHS and Lifshitz transition reproduces the large positive Nernst peak, supporting the mechanism as a generic band-structure effect."],"fun_headline_variants":["Van Hove singularities yield giant Nernst in twisted graphene","Giant Nernst signal from van Hove singularities in twisted double bilayer graphene","Lifshitz transitions at van Hove points drive large Nernst in tDBLG","40 µV/K/T Nernst at 1 K from van Hove singularities"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The central claim rests on treating the mobility as a single, energy-independent, Drude value across the entire n–D plane; if mobility actually varies strongly across the vHS in a way not captured by the σxx(B) fit, the quantitative identification of the Nernst peak with the Lifshitz transition collapses.","fun_headline_variants_meta":{"raw":{"variants":["Van Hove singularities yield giant Nernst in twisted graphene","Giant Nernst signal from van Hove singularities in twisted double bilayer graphene","Lifshitz transitions at van Hove points drive large Nernst in tDBLG","40 µV/K/T Nernst at 1 K from van Hove singularities"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001066,"raw_usage":{"total_tokens":4292,"prompt_tokens":719,"completion_tokens":3573,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":463,"completion_tokens_details":{"reasoning_tokens":3483}},"tokens_in":463,"tokens_out":3573,"duration_ms":23626,"temperature":1.0,"reasoning_tokens":3483,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T02:39:03.513213+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be to measure the energy-dependent mobility near the vHS directly—for instance, via quantum oscillations that resolve mobility per Landau level—and recompute Eq. (1) with that mobility. If the predicted peak is absent, or if its sign flips at a vHS where the band structure is independently known, the semiclassical attribution to Lifshitz transitions would fail.","supporting_citations":[],"review_version":1}