{"id":"2de38744-2067-4970-871b-ce9ded7610f9","arxiv_id":"2608.08917","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Including lattice relaxation in a continuum model makes WSe2/WS2 heterobilayers topological: the pseudomagnetic field alone opens a Chern-number ±1 gap between the third and fourth valence bands.","lead":"Moiré TMD heterobilayers like WSe2/WS2 are usually modeled as electrically trivial, but this paper shows that atomic lattice relaxation, which always occurs in these materials, generates a strain-induced pseudomagnetic field that opens topological gaps. A new continuum model predicts Chern bands and fractional Chern insulator candidates where earlier models saw none.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The strain field used to build the pseudomagnetic field is computed from the relative displacement Δu rather than the WSe2 layer displacement u2, so the sole driver of the claimed topological gap may be mis-specified.","rationale":"The paper's central single-particle claim—that the pseudomagnetic field from relaxation alone opens a topological gap between the third and fourth valence bands—is carefully isolated in Fig. 2, and the derivation in SM S1.2–S1.3 is otherwise coherent. The load-bearing weakness lies in the link between the relaxed geometry and the electronic strain Hamiltonian. The derivation requires the strain of the layer that carries the low-energy holes, which is WSe2 (layer 2). The explicit statement in SM S3.2 that the strain entering the electronic model is computed from the relative displacement Δu = u1 − u2 does not follow from the derivation: the moiré potential legitimately depends on Δu, but the intralayer pseudogauge field depends on the gradient of the active layer's own displacement. Since WS2 and WSe2 have comparable elastic constants, the equilibrium u1 and u2 are both nonzero, so ε(Δu) differs from ε(u2) in magnitude and spatial pattern. In the twist-only limit the overestimate can approach a factor of two, which is very likely enough to move the small (meV-scale) topological gap and the gap-closing boundaries that define the Chern phase diagram in Fig. 3. This is a structural defect, not a parameter-accuracy issue: it would remain even for a perfect GSFE. The proposed test is a direct post-processing check of the existing relaxation solver and should settle the matter. I therefore move the verdict from CONDITIONAL to UNVERDICTED pending confirmation that the published strain convention reproduces the reported topological phase diagram.","tokens_in":30447,"tokens_out":7921,"duration_ms":81843,"concrete_test":"Modify the relaxation post-processing to output layer-resolved displacement fields u1(r) and u2(r); compute ε^(2)_ij = (∂_i u2_j + ∂_j u2_i)/2 via spectral differentiation; rebuild the strain Hamiltonian of Eq. (4) and Eq. (5) using this WSe2-layer strain; and recompute the band structure, Δ34, and Chern numbers at θ=1.0°, δ=0.02 (Fig. 2) and across the (θ,δ) phase diagram. Also compare the peak |B| at θ=2°, δ=0 against the reported ~900 T: if the relative-displacement convention overestimates by about a factor of 2, the corrected pseudomagnetic field will be roughly halved and the topological region in Fig. 3 will likely shrink or vanish.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"SM S1.2 derives the strain Hamiltonian from the intralayer deformation of the layer hosting the low-energy holes: Eq. (S10) modifies the hopping by ∇u with u the displacement of that layer. For WSe2/WS2 the active layer is WSe2, so the pseudogauge field in Eq. (5) should be built from ε^(2)_ij = (∂_i u2_j + ∂_j u2_i)/2. However, SM S3.2 states: 'The strain tensor εij(r) used in the continuum electronic model is computed from the relative displacement Δu = u1 − u2 via spectral differentiation.' If u1 ≠ 0 (the elastic constants for WS2 and WSe2 are comparable, so both layers relax), ε(Δu) is not the WSe2 strain. In the twist-only limit u1 ≃ −u2, the pseudomagnetic field would be overestimated by a factor of about 2, and its spatial pattern would be altered whenever the acoustic component is nonzero. Since Fig. 2(b) attributes the C3=+1/C4=−1 gap entirely to this pseudomagnetic field, the central claim is directly affected. This is an internal modeling issue, not merely parameter sensitivity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops a continuum model for moiré TMD heterobilayers that includes lattice relaxation through three channels: a relaxed moiré potential with higher Fourier harmonics, a scalar deformation (pseudoelectric) potential, and a pseudomagnetic vector potential. Applying the model to WSe2/WS2, the authors find that the pseudomagnetic field alone opens a topological gap between the third and fourth valence bands with Chern numbers C3=+1 and C4=-1 over a broad range of twist angle and lattice mismatch, while the other two channels flatten bands and enhance gaps. The paper also computes quantum-geometric diagnostics, predicts favorable conditions for fractional Chern insulators, and reports neural-network variational Monte Carlo charge gaps at filling ν=2 for the rigid and relaxed models.","tokens_in":30677,"tokens_out":3681,"duration_ms":40142,"significance":"If the central claim holds, the paper would resolve a notable discrepancy: prior continuum models of TMD heterobilayers found trivial bands, whereas this work argues that intrinsic lattice relaxation alone generates topology. The derivation from a tight-binding starting point is careful and the decomposition into three relaxation channels is a useful conceptual framework. The authors provide Chern-number phase diagrams over dense grids, DFT-parametrized inputs for both the moiré potential and the generalized stacking fault energy, and public code and data availability statements for the many-body calculations. The end-to-end pipeline from DFT energetics to continuum bands to many-body observables is a genuine strength. However, one load-bearing modeling choice in the strain Hamiltonian appears inconsistent with the derivation, and the many-body test addresses the ν=2 charge gap rather than the topological gap Δ34, so the current manuscript overstates some of its support.","major_comments":[{"comment":"The strain Hamiltonian is derived from the intralayer deformation of the layer hosting the low-energy holes. In SM Eq. (S10), the hopping correction is proportional to u_ij = ∂_i u_j with u the displacement of that layer, so the pseudogauge field in Eq. (5) should be constructed from the WSe2 layer strain ε^(2)_ij = (∂_i u2_j + ∂_j u2_i)/2. However, SM S3.2 states: 'The strain tensor εij(r) used in the continuum electronic model is computed from the relative displacement Δu = u1 − u2 via spectral differentiation.' Since the moiré potential correctly uses Δu but the intralayer strain correction does not, this appears to be an internal inconsistency. In the twist-dominated limit the two layers relax in an approximately opposite manner, so ε(Δu) can overestimate the WSe2 strain by a factor of order 2, and the spatial pattern is altered whenever the acoustic component is nonzero. Because Fig. 2(b) attributes the C3=+1/C4=-1 gap entirely to this pseudomagnetic field, the central topological claim and the phase diagrams in Fig. 3 depend directly on this choice. The authors should recompute the strain Hamiltonian using the active-layer displacement field (or clearly justify why the relative-displacement strain is the correct object for the intralayer hopping correction).","section":"SM S1.2 and SM S3.2; main-text Eq. (5)"},{"comment":"The many-body charge gap in Fig. 1(e) is computed at filling ν=2, which probes the top two valence bands. Those bands are topologically trivial throughout the parameter range studied, so this calculation tests the relaxation enhancement of a trivial gap, not the survival of the topological gap Δ34 or the Chern numbers C3=±1. The abstract's statement that the bandgap enhancements 'survive many-body interactions using neural-network variational Monte Carlo calculations' is therefore supported only for the top two bands. The authors should either soften the claim to specify the charged gap at ν=2 or provide a many-body calculation that directly addresses the topological gap between the third and fourth valence bands.","section":"Fig. 1(e)-(f) and abstract"}],"minor_comments":[{"comment":"The main text says gray regions indicate bands with adjacent bands within a 3 meV tolerance, while the Fig. 3 caption says Δ34 < 0.3 meV and SM S3 states a 0.8 meV threshold; these tolerances should be harmonized and defined in one place.","section":"Main text, Fig. 3 caption"},{"comment":"The plane-wave basis is truncated at 163 moiré reciprocal lattice vectors; a convergence test in the number of basis functions for the Chern numbers and the gap Δ34 would strengthen the phase diagram claims, particularly near the gap-closing boundaries.","section":"SM S2.4"},{"comment":"Reference [54] contains the placeholder identifier 'arXiv:2509.XXXXX' and should be completed before publication; reference [11] should also be checked for completeness.","section":"References"},{"comment":"The label 'Alle,ects' in the fourth panel of Fig. 2 contains a typo; it should read 'All effects'.","section":"Fig. 2 label"}],"recommendation":"major_revision","confidential_remarks":"The strain-field concern raised by the internal reading is genuine and is the main reason I cannot recommend acceptance in the current form: the paper derives the pseudogauge coupling from the active layer's intralayer strain but then computes that strain from the relative displacement field. This is a fixable modeling inconsistency, but it directly affects the central topological phase diagram. If the authors recompute with the WSe2 strain field and the C3=+1/C4=-1 region survives with comparable breadth, the paper would be a strong contribution. The many-body claim in the abstract also needs narrowing or additional calculation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's my take on Luskin et al. The paper has a genuinely useful idea: treat lattice relaxation in TMD heterobilayers through three channels—modified moiré potential, scalar deformation, and pseudomagnetic gauge field—and show that the last of these can open a topological gap between the third and fourth valence bands. The decomposition in Fig. 2 is clear, and the framework connecting DFT relaxation to a continuum model to NNVMC is a real step forward. I hadn't seen this combination before, and the FCI diagnostics for the third band are a practical addition.\n\nThe soft spot is bigger than the reader's report suggests. The supplemental derivation (S1.2) says the hopping correction in the active layer is built from the displacement of that layer, u2. But S3.2 states the strain tensor used in the electronic model is computed from the relative displacement Δu = u1−u2. Those are different objects. WS2 and WSe2 have comparable elastic constants, so both layers relax. In the twist-only limit u1 ≈ −u2, so ε(Δu) ≈ 2 ε(u2), doubling the pseudomagnetic field and flipping its sign. For combined twist+mismatch, the pattern changes too. Since the entire topological phase diagram is driven by that field, the quantitative claims—Chern numbers over a broad range, the ν=6 QAH prediction—are not trustworthy until this is recomputed. The mechanism may survive the fix, but the paper as written does not give the correct strain to the electronic Hamiltonian.\n\nThe reader is right that the many-body claim overreaches: the NNVMC charge gap is for ν=2 (the top two bands), not the topological Δ34 gap. The abstract's phrase 'survive many-body interactions' refers to the top-band gap, not the topological one. That should be reworded.\n\nWhat else is good: the derivation of the strain Hamiltonian from tight-binding is careful, the GSFE and elastic inputs are documented, the code is public, and the phase diagrams are thorough. The trace-condition and quantum metric analysis is a nice addition. No invented entities or circular fitting; the model parameters are all external to the target result.\n\nBottom line: this deserves a serious referee, but it needs major revision. I would send it out, with the specific instruction to the authors to reconcile S1.2 and S3.2 and recompute all results with the correct strain. If the topological gap survives with ε(u2), this becomes an important paper. As is, I wouldn't cite the numbers.\n\nBest.","headline":"Compelling framework, but the strain tensor is likely built from the wrong displacement field—recompute before trusting the topology.","tokens_in":31254,"tokens_out":3908,"would_cite":false,"duration_ms":39814,"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":"Lattice relaxation alone opens topological bands in moiré WSe2/WS2.","keywords":["moiré heterobilayers","lattice relaxation","pseudomagnetic field","Chern numbers","flat bands","transition metal dichalcogenides","quantum anomalous Hall effect","fractional Chern insulators"],"falsifier":"A direct test is to measure the Hall conductance of a 3R-stacked WSe2/WS2 device near $\\theta = 1^\\circ$ and $\\delta = 0.02$ as a function of filling up to $\\nu = 6$; the model predicts a quantized plateau $\\sigma_{xy} = e^2/h$ when the chemical potential lies in the gap between the third and fourth valence bands. A null result across the predicted topological region would indicate the strain field, the stacking-fault landscape, or the band calculation is overestimating the gap.","tokens_in":1848,"feed_emoji":"🧲","tokens_out":4298,"duration_ms":111236,"temperature":0.7,"pith_summary":"Standard continuum models of moiré transition metal dichalcogenide heterobilayers produce topologically trivial bands, and this paper argues the reason is that they leave out an intrinsic effect: atomic relaxation creates strain that acts on electrons like a magnetic field. The paper builds a continuum model that splits relaxation into three channels—a relaxed moiré potential with higher Fourier harmonics, a scalar deformation potential, and a pseudomagnetic vector potential—and shows in WSe2/WS2 that the pseudomagnetic channel alone opens a topological gap between the third and fourth valence bands, with Chern numbers +1 and −1 over a broad range of twist angle and lattice mismatch. The scalar and potential corrections narrow bandwidths and enlarge gaps, and the enhanced charge gap survives neural-network variational Monte Carlo calculations with Coulomb interactions. If these results hold, moiré heterobilayers are a new class of topological materials whose topology comes from relaxation itself, not from external strain or an intrinsic rigid-limit Chern band.","feed_headline":"Relaxation alone turns moiré WSe2/WS2 bands topological","feed_subtitle":"A strain pseudomagnetic field gives the third and fourth valence bands Chern numbers +1 and −1 over wide twist angles.","key_machinery":"The load-bearing object is the relaxed continuum Hamiltonian $$\\hat H = \\hat H_{\\mathrm{mono}} + V_{m,\\mathrm{relax}}(\\mathbf r) + \\hat H_\\epsilon,$$ where the strain contribution is written in pseudo-gauge form as $\\Phi(\\mathbf r) + \\{\\hat k_\\alpha, A_\\alpha(\\mathbf r)\\}$. Relaxation enters through a displacement field obtained by minimizing elastic energy plus a DFT-parameterized generalized stacking fault energy; the strain tensor then produces a scalar deformation potential $\\Phi = C\\epsilon_0$ and a vector potential $\\mathbf A = -\\tfrac12 D(\\epsilon_1, \\epsilon_2)$ whose curl is the pseudomagnetic field $\\mathbf B = \\nabla\\times\\mathbf A$. The $\\mathbf A$ term is what opens the topological gap at the quadratic band touching, while the relaxed moiré potential and scalar potential are topology-neutral but flatten bands and enlarge gaps.","core_discovery":"On the paper's own terms, the central discovery is that lattice relaxation in moiré TMD heterobilayers is not a minor correction but the mechanism that makes the bands topological. The pseudomagnetic field generated by the relaxation-induced strain—not the relaxed moiré potential and not the scalar deformation potential—opens a gap between the third and fourth valence bands with Chern numbers $C_3 = +1$ and $C_4 = -1$ over a connected region of the twist-angle and lattice-mismatch plane, including the natural mismatch of WSe2/WS2 near $\\delta \\approx 0.04$. The top two valence bands remain trivial, which matches the experimentally observed trivial Mott insulator at filling $\\nu = 1$, while the third band becomes a Chern band whose Berry curvature and quantum metric are smoothed by relaxation toward the ideal limit. The topological gap survives many-body interactions in neural-network variational Monte Carlo calculations at filling $\\nu = 2$, and the net pseudomagnetic flux per moiré cell vanishes, so the topology arises from redistributed Berry curvature rather than net Landau-level quantization.","pith_inferences":["If the central claim holds, many existing heterobilayer samples may already contain unexplored topological states at fillings near $\\nu = 6$, since the pseudomagnetic field is strongest in the same small-angle regime where flat bands are flattest.","The decomposition into three relaxation channels suggests a tuning knob: substrate coupling, pressure, or twist angle could move a sample across the topological phase boundary, giving an in-situ switch between trivial and Chern regimes.","A testable extension would be to probe the predicted alternating-sign pseudomagnetic field pattern at domain-wall intersections, looking for position-dependent spectral shifts despite the vanishing net flux per moiré cell."],"forward_implications":["At integer filling $\\nu = 6$, the model predicts a quantum anomalous Hall state with quantized Hall conductance $\\sigma_{xy} = e^2/h$ in 3R-stacked WSe2/WS2 across a broad, connected region of twist angle and lattice mismatch.","Partial filling of the third valence band is a candidate for fractional Chern insulator states, because relaxation reduces the bandwidth and drives the trace-condition violation and Berry-curvature fluctuation toward the ideal Chern limit.","The top two valence bands remain topologically trivial, consistent with the observed trivial Mott insulator at $\\nu = 1$; the relaxation-enhanced charge gap at $\\nu = 2$ is a quantitative many-body prediction accessible to transport and compressibility measurements.","Previous continuum heterobilayer models found trivial bands because they omitted the strain-induced gauge fields; the paper implies that including them is necessary, not optional, for describing topology in these systems."],"supporting_citations":[{"why":"Supplies the rigid moiré-potential parameterization and the single-band Hubbard construction whose trivial bands the paper compares against.","marker":"[18]"},{"why":"The existing continuum model that includes relaxation only through the moiré potential while omitting strain gauge fields; the central contrast case.","marker":"[25]"},{"why":"Provides the strained-monolayer tight-binding parameters, including the effective mass and the strain coefficients $C$ and $D$, that fix the pseudoelectric and pseudomagnetic couplings.","marker":"[56]"},{"why":"Supplies the configuration-space continuum relaxation framework used to compute the equilibrium displacement field and strain.","marker":"[3]"},{"why":"Demonstrates in twisted bilayer graphene that relaxation generates a strain pseudomagnetic field reshaping flat bands, the precedent mechanism the paper transfers to heterobilayers.","marker":"[2]"},{"why":"Reports a quantum anomalous Hall effect in a MoTe2/WSe2 heterobilayer, the experimental context that the predicted $\\nu=6$ state extends.","marker":"[36]"}],"fun_headline_variants":["Relaxation-driven pseudomagnetic field yields Chern bands in moiré","Moiré topology emerges from lattice relaxation alone","Strain-induced Chern numbers in WSe2/WS2 moiré","Intrinsic lattice relaxation sets moiré band topology"],"cache_read_input_tokens":33280,"weakest_assumption_plain":"The prediction rests on the strain field computed from a continuum elasticity model whose generalized stacking fault energy and elastic constants come from DFT; if that stacking-energy landscape, especially the AB/BA asymmetry, is inaccurate, the pseudomagnetic field and the topological gap it opens would change.","fun_headline_variants_meta":{"raw":{"variants":["Relaxation-driven pseudomagnetic field yields Chern bands in moiré","Moiré topology emerges from lattice relaxation alone","Strain-induced Chern numbers in WSe2/WS2 moiré","Intrinsic lattice relaxation sets moiré band topology"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000385,"raw_usage":{"total_tokens":2099,"prompt_tokens":1070,"completion_tokens":1029,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":686,"completion_tokens_details":{"reasoning_tokens":957}},"tokens_in":686,"tokens_out":1029,"duration_ms":11428,"temperature":1.0,"reasoning_tokens":957,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:20:55.703715+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test is to measure the Hall conductance of a 3R-stacked WSe2/WS2 device near $\\theta = 1^\\circ$ and $\\delta = 0.02$ as a function of filling up to $\\nu = 6$; the model predicts a quantized plateau $\\sigma_{xy} = e^2/h$ when the chemical potential lies in the gap between the third and fourth valence bands. A null result across the predicted topological region would indicate the strain field, the stacking-fault landscape, or the band calculation is overestimating the gap.","supporting_citations":[{"cited_title":"Angeli, G","cited_arxiv_id":null,"evidence_quote":"The existing continuum model that includes relaxation only through the moiré potential while omitting strain gauge fields; the central contrast case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the strained-monolayer tight-binding parameters, including the effective mass and the strain coefficients $C$ and $D$, that fix the pseudoelectric and pseudomagnetic couplings."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports a quantum anomalous Hall effect in a MoTe2/WSe2 heterobilayer, the experimental context that the predicted $\\nu=6$ state extends."}],"review_version":1}