{"id":"ab20595b-855e-461b-845e-2339129d53e0","arxiv_id":"2608.01553","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Layer-hybridized generalized Wigner crystals survive and strengthen in H-stacked MoSe2/WS2 as an out-of-plane field tunes the bands through resonance.","lead":"Electric-field tuning drives electrons in a MoSe2/WS2 moiré bilayer into a layer-hybridized regime where generalized Wigner crystal states persist and even strengthen at fillings above one electron per moiré cell. The result opens a new electrical knob for engineering correlated and potentially topological states in 2D materials.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Absolute filling calibration is the load-bearing link: a single ΔV=7.7 V assignment sets every n label, with no independent density check and a possible dual-gate lever-arm ambiguity.","rationale":"The reader's weakest-assumption analysis identifies exactly the step I find most load-bearing: the conversion of a 7.7 V feature spacing into an absolute moiré density. Every quantitative filling label flows from this one point, and the paper's headline claims are stated in terms of those labels. The paper does offer some internal consistency: at zero field the MoSe2 trion M1− intensity peaks between the same two features and vanishes at the feature labeled n=2, and the WSe2 2s sensor shows correlated-state dips at rational fillings. This makes the assignment plausible but not independently established. Because the sensor features themselves are the calibration reference, there is a circularity risk: any extrinsic feature spacing (e.g., from a gate-dependent sensor response) would be interpreted as a filling interval. I therefore agree with the reader's conditional verdict rather than escalating to reject: the data are rich and internally coherent, but the absolute filling axis needs an independent anchor. A direct recalc from the raw gate-voltage pairs is the cheapest decisive check; a compressibility or Landau-fan measurement would be the strongest one. Since the reader already assigned CONDITIONAL for this reason, my stress-test pass does not move the recommended verdict.","tokens_in":10235,"tokens_out":6237,"duration_ms":65581,"concrete_test":"Recompute the moiré density from the raw top and bottom gate voltages at the two sensing features identified as n=1 and n=2 using the charge-basis combination n ∝ (V_TG/d1 + V_BG/d2) with the stated d1=d2=55 nm and εBN=3.5; if the resulting density per feature interval differs from 2.71e12 cm^-2 by more than ~20%, all filling labels in the paper must be re-derived before testing the central n>1 layer-separation claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central narrative—n=1 hybridized Mott insulator, n<1 GWC weakening, and n>1 layer-separated GWC strengthening—rests on assigning the gate-voltage interval ΔV=7.7 V between two WSe2 2s sensing features to exactly one electron per moiré cell (Methods, 'Determination of twist angle'). This single assignment fixes n_moiré = 2.71e12 cm^-2 and therefore every filling label in Figs. 1–4. The paper provides no independent density calibration: no Landau fan, no capacitance or compressibility measurement, no STM imaging. The conversion formula n = ε0 εBN ΔV/(e d_BN) uses a single-gate lever arm, but in the dual-gated geometry used here, the charge-basis combination is (V_TG + V_BG); how the two gates are swept while holding the displacement field fixed determines whether the correct lever arm is larger by up to a factor of 2. If the factor is 2, the inferred density, twist angle, and all occupancy labels shift, and the claim that states at n>1 are stabilized by layer-separated double occupation is no longer supported by the data as labeled. The M1− intensity behavior and the two-level hybridization fit provide useful internal consistency checks, but they are interpreted using the same filling labels, so they do not independently validate the calibration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental study of an H-stacked MoSe2/WS2 moiré heterobilayer in a dual-gated device, using doping- and electric-field-dependent reflectance contrast spectroscopy and magnetic circular dichroism. The authors claim that tuning an out-of-plane electric field drives a type-I to type-II band-alignment transition, with an intermediate regime in which the conduction-band minima of the two layers hybridize. They interpret a doping-dependent evolution of MoSe2 trion (M1−) intensity and WS2 exciton-polaron features as evidence of interlayer hybridization with an extracted energy t ≈ 1.95 meV, much smaller than the inferred Hubbard repulsions UMo ≈ 55 meV and UW ≈ 22 meV. They further use a WSe2 2s exciton sensing layer to claim that at fillings below one electron per moiré cell the generalized Wigner crystal states are weakened by hybridization, while for fillings above one electron per cell the correlated states are strengthened because electrons on doubly occupied moiré sites preferentially occupy opposite layers. The n = 1 state is interpreted as a layer-hybridized Mott insulator that retains finite magnetism, as evidenced by a reduced but nonzero MCD susceptibility.","tokens_in":10554,"tokens_out":4779,"duration_ms":45673,"significance":"If the central claims hold, this work would demonstrate a new electrically tunable platform in which interlayer hybridization coexists with charge-ordered states, and it would provide the first evidence that fractional correlated insulators can survive hybridization while n > 1 states are actually stabilized by layer-separated occupation. The experiment is technically demanding, the data are systematic across electric field and doping, and the source data for the main figures are promised, which are strengths. The interpretation is also falsifiable in principle: the filling labels, the hybridization energy, and the asymmetry between n < 1 and n > 1 are all testable with independent measurements. However, as detailed in the major comments, the load-bearing filling calibration rests on a single gate-voltage interval, and the quantitative parameters t, UMo, UW, and Δoffset are all extracted from the same spectroscopic features used to define the phases, so the current evidence does not yet uniquely establish the narrative.","major_comments":[{"comment":"The entire filling-factor axis (n in Figs. 1–4) is fixed by the single assignment of ΔV = 7.7 V between the n = 1 and n = 2 features in the WSe2 2s sensing spectra to exactly one electron per moiré cell. The paper provides no independent density calibration (no Landau fan, capacitance or compressibility measurement, or STM imaging), and the formula n = ε0 εBN ΔV/(e dBN) assumes a single-gate lever arm. In the dual-gated geometry used here, the relevant charge-basis combination involves both top and bottom gates, and the lever arm may be larger by up to a factor of two depending on how the gates are swept while holding the displacement field fixed. If the factor is two, the inferred moiré density of 2.71 × 10^12 cm^-2 and every filling label shift, and the central claims about n > 1 layer-separated Wigner crystals and the n = 1 hybridized Mott insulator are no longer supported by the data as labeled. This calibration must be justified with an independent measurement or at least a careful derivation of the dual-gate lever arm.","section":"Methods, 'Determination of twist angle'"},{"comment":"The hybridization energy t = 1.95 ± 0.42 meV is extracted from a two-level model that explicitly assumes the M1− peak intensity is proportional to the probability of the electron residing in MoSe2. This proportionality is cited from ref 32 but is not independently verified in this device, so the fit does not uniquely determine t. Likewise, the Hubbard parameters UMo = 55 meV, UW = 22 meV, and the offset Δoffset = 113 meV are obtained from the electric-field evolution of the same M1− and WS2 exciton features (Supplementary Sections 3 and 4) that are used to identify the band alignments. The inference that t << U is therefore based on parameters that all derive from the same set of spectra; an independent measurement of the density of states or compressibility would be needed to break this circularity.","section":"Extended Data Fig. 1 and 'Electrically tunable and layer-hybridized conduction bands'"},{"comment":"The conclusion that correlated states in the 0 < n < 1 range are weakened while those in 1 < n < 2 are strengthened is based on the visual amplitude of the WSe2 2s sensing features and the presence or absence of dips in the line traces (Fig. 3e). These features are not quantitatively linked to the charge gap or compressibility, so 'less pronounced' and 'more pronounced' do not yet constitute a quantitative measure of correlation strength. This is particularly important because the reported disappearance of the n = 2/3 state and the survival of n = 1/2 in the hybridization regime is a striking claim; a capacitance measurement or a density-dependent linewidth analysis would provide a concrete test of the proposed asymmetry.","section":"Fig. 3 and 'Interlayer-hybridized generalized Wigner crystal states'"}],"minor_comments":[{"comment":"Please define all symbols in the density formula (εBN, dBN) and clarify whether the same h-BN thickness and dielectric constant used for the nominal electric field are used in the density conversion; a reader should be able to reproduce the 2.71 × 10^12 cm^-2 value without cross-referencing other sections.","section":"Methods, 'Determination of twist angle'"},{"comment":"The caption of Fig. 2 refers to red and grey dashed lines in panels (d), (h), and (l), but these are not visible in the text version; please ensure the figure panels are legible and the marks are clearly labeled.","section":"Fig. 2"},{"comment":"The MCD susceptibility values (χMCD = 0.08%, 0.03%, and 0.01% T^-1) are quoted without uncertainties or the fitting range; please provide error bars and state the linear region used for the slope.","section":"Magnetic circular dichroism measurements"},{"comment":"The term 'charge-transfer insulator' is used in the Fig. 3f caption but is not explicitly defined in the main text; please define it in the context of the n = 2 state.","section":"Throughout"},{"comment":"Reference 29 (Polovnikov et al.) is cited for Hubbard parameters but is an arXiv preprint; please check whether an updated, peer-reviewed version is available and cite that instead.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The filling calibration is the linchpin of the paper: a single ΔV = 7.7 V assignment, with no independent density check, fixes all n labels. If the dual-gate lever arm is off by a factor of two, the narrative about n > 1 layer-separated Wigner crystals collapses. I recommend the editor require an independent calibration (e.g., Landau fan, capacitance, or a careful two-gate lever-arm analysis) before publication. Also note that the M1− intensity probe is cited to ref 32, which is an in-press Nano Letters paper; the authors should confirm that the proportionality assumption is established there and not simply an in-house assumption."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Rough take: this is a solid experimental paper, possibly a nice result, but the filling calibration is the whole ballgame and it is not nailed down. The data show a systematic set of reflectance and MCD measurements across the type-I-to-type-II transition in H-stacked MoSe2/WS2. The central observation—that GWC features persist in the hybridization regime and that a state appears near n=3/2 which is absent outside it—is genuinely new. The claim that n>1 states are stabilized by Coulomb-driven layer separation is physically sensible and consistent with the M1- intensity behavior. The MCD inheritance at n=1 is a nice addition.\n\nThe soft spot is the density calibration. The moiré density, and hence every filling label, comes from assigning the voltage interval ΔV=7.7 V between two features in the WSe2 2s sensor to exactly one electron per moiré cell. The formula used is the single-gate lever arm, but the device is dual-gated. If the density sweep is performed by moving both gates while holding the displacement field fixed, the correct lever arm is the sum of the two gate contributions, which can be twice as large as the single-gate value. The paper does not specify the sweep protocol, and there is no independent check (Landau fan, capacitance, STM). If the lever arm is off by 2, the n=1 Mott state becomes n=2, the n=1/2 GWC becomes n=1, and the whole n<1 vs n>1 narrative collapses. This is not a pedantic point; it is the load-bearing axis of the paper.\n\nThe two-level model fit for t and the Hubbard U values are reasonable but fitted to the same spectra that define the states, so they don't provide independent validation. That's a minor issue.\n\nBottom line: the paper deserves a serious referee. The experimental work looks careful and the physics is interesting. But the calibration needs to be addressed head-on, and the authors should provide the exact gating protocol used for the density axis, plus at least one independent density calibration. As written, I would not cite the filling-dependent narrative without first getting that straight.","headline":"Potentially important observation of hybridized Wigner crystals, but the density calibration rests on a single unconvincing lever arm and needs to be nailed down before the filling narrative can be trusted.","tokens_in":11107,"tokens_out":6355,"would_cite":false,"duration_ms":62204,"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":"An electric field can make generalized Wigner crystals in a MoSe2/WS2 moiré bilayer survive—and even strengthen—interlayer hybridization above one electron per moiré cell.","keywords":["moiré superlattice","MoSe2/WS2 heterobilayer","generalized Wigner crystal","Mott insulator","interlayer hybridization","band alignment transition","magnetic circular dichroism","correlated insulator"],"falsifier":"Measure the moiré density of the same heterostructure by an independent method, for example the period of the Landau fan in a dual-gated transport device or the moiré period by scanning tunneling microscopy, and compare it with the density obtained from the $\\Delta V = 7.7$ V interval; if the two disagree by more than the experimental uncertainty, every filling factor, including the $n = 3/2$ correlated insulator, would need to be renumbered.","tokens_in":10055,"feed_emoji":"⚡","tokens_out":14408,"duration_ms":109206,"temperature":0.7,"pith_summary":"This paper reports that an electric field can tune a MoSe2/WS2 moiré bilayer through a band-alignment transition and into a regime where the two flat conduction bands hybridize and electrons are shared between layers. In that regime the authors observe a layer-hybridized Mott insulator at $n = 1$ and generalized Wigner crystals whose stability depends strongly on filling. Below one electron per cell, hybridization delocalizes electrons and weakens the charge-ordered states; above one electron per cell, Coulomb repulsion pushes the two electrons on a doubly occupied site into opposite layers, which strengthens the Wigner crystals and even produces a correlated insulator at $n = 3/2$ that is absent without hybridization. The result matters because it shows that interlayer hybridization, usually expected to destroy fragile charge order, can instead stabilize it when repulsion dominates, and it opens an electrically tunable route toward engineered lattice symmetries and topological correlated phases.","feed_headline":"Electrons split across layers, strengthening Wigner crystals","feed_subtitle":"Why shared-layer electrons make stronger Wigner crystals, not weaker ones.","key_machinery":"The central mechanism is the competition between interlayer tunneling and Coulomb repulsion inside the hybridized flat bands of H-stacked MoSe2/WS2. The two conduction-band minima, offset by about $113$ meV at zero field, are driven through resonance by an out-of-plane electric field; the authors fit the field dependence of the MoSe2 trion intensity with a two-level model to obtain an interlayer hybridization energy of about $1.95$ meV. This hybridization is the control knob. Because it is far smaller than the on-site Hubbard repulsions, a moiré cell holding two electrons can lower its energy by placing them in opposite layers, converting what would be a conductive double-occupied state into a layer-separated correlated insulator. The experimental readouts carrying the argument are the doping- and field-dependent reflectance contrast of intralayer moiré excitons in both layers, the 2s exciton of a nearby WSe2 monolayer used as a remote sensor of charge ordering, and magnetic circular dichroism at the MoSe2 resonance.","core_discovery":"The central discovery is that interlayer hybridization in an H-stacked MoSe2/WS2 moiré superlattice reorganizes correlation-driven charge order instead of simply destroying it, and the outcome is controlled by filling. By applying an out-of-plane electric field, the authors bring the conduction-band minima of the two layers into resonance and extract an interlayer tunneling amplitude of roughly $1.95$ meV, much smaller than the on-site Hubbard repulsions (about $55$ meV in MoSe2 and $22$ meV in WS2). At $n = 1$ the state is a hybridized Mott insulator with the electron shared between layers; magnetic circular dichroism shows its magnetic susceptibility is reduced but finite compared with the pure MoSe2 state. For $0 < n < 1$ the hybridization increases kinetic energy, and fractional states such as $n = 2/3$ lose their insulating character while $n = 1/2$ survives. For $1 < n < 2$ the paper argues that a doubly occupied moiré cell lowers its Coulomb energy by placing the two electrons in opposite layers, stabilizing the generalized Wigner crystals and producing a new correlated insulator at $n = 3/2$; the $n = 2$ state is best described as a layer-separated charge-transfer insulator rather than a hybridized Mott insulator.","pith_inferences":["The density calibration rests on a single voltage interval, so an independent measurement of the moiré density (for example by a Landau fan or STM) would either strengthen or shift every filling label; the $n = 3/2$ insulator is the most distinctive state to check, since it is predicted to exist only in the hybridized regime.","The same hybridized-band physics should appear in other TMDC heterobilayers with nearby conduction-band minima, such as MoSe2/WSe2 or MoS2/WS2, where the ratio of interlayer tunneling to on-site repulsion would determine whether $n>1$ charge order survives.","If the layer-separation picture is correct, the strength of the $n = 3/2$ insulator should depend non-monotonically on electric field: it should vanish far from resonance, peak near the hybridization regime, and weaken again as one layer moves far off resonance.","The contrasting fates of the $n = 1/2$ and $n = 2/3$ states under hybridization suggest that the effect of tunneling on a Wigner crystal depends on the wavefunction overlap of the specific charge configuration; capacitance or noise measurements could map this state-by-state sensitivity."],"forward_implications":["At fillings above one electron per moiré cell, correlated insulating states appear in the hybridization regime that are absent in either non-hybridized layer, including a generalized Wigner crystal at $n = 3/2$.","The $n=1$ hybridized Mott insulator retains a tunable magnetic response, so the applied electric field can shift the magnetic character of the state toward or away from the magnetic $n=1$ MoSe2 configuration.","Because doubly occupied sites separate across the two layers, the effective charge-ordered lattice in the hybridization regime has a different geometry from the single-layer Wigner crystals, potentially enabling electrically tunable emergent honeycomb lattices.","The measured Hubbard parameters ($U_{\\mathrm{Mo}} \\approx 55$ meV, $U_{\\mathrm{W}} \\approx 22$ meV) and the band offset of $113$ meV give a quantitative benchmark for models of correlated states in this heterobilayer."],"supporting_citations":[{"why":"Supplies the reference Mott and generalized Wigner crystal states in a WSe2/WS2 moiré bilayer that this paper extends to the hybridized regime.","marker":"3"},{"why":"Gives the Hubbard model description of TMDC moiré bands used to interpret Hubbard-band splitting and the Mott insulator.","marker":"4"},{"why":"Introduces the WSe2 2s exciton sensing technique used throughout to detect charge ordering and to set the filling scale.","marker":"13"},{"why":"Demonstrates a hybridized interlayer Mott insulator in a trilayer moiré system, the phenomenon extended here to a bilayer and to fractional fillings.","marker":"15"},{"why":"Provides the trilayer moiré system and the two-exciton layer-separation concept that the paper analogizes to layer-separated electrons at n > 1.","marker":"16"},{"why":"Shows the electric-field-driven type-I to type-II band-alignment transition in MoSe2/WS2 that the paper tunes through.","marker":"26"},{"why":"Supports the band-alignment picture with electrical switching between exciton dissociation and funneling in MoSe2/WS2.","marker":"27"},{"why":"Reports the magnetism of the n = 1 MoSe2/WS2 moiré state that the MCD measurements are interpreted against.","marker":"28"},{"why":"Provides field-induced hybridization of moiré excitons in MoSe2/WS2, supporting the hybridization interpretation of the trion intensity.","marker":"31"},{"why":"Supplies the WSe2 2s sensing and moiré-density determination method used to extract the twist angle and all filling factors.","marker":"36"}],"fun_headline_variants":["Hybridization strengthens Wigner crystals above one electron per cell","Layer hybridization strengthens Wigner crystals at high fillings","New Wigner crystal at 3/2 filling from layer hybridization","Moiré hybridization tunes Wigner crystal stability by filling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that the $\\Delta V = 7.7$ V voltage interval between features in the WSe2 2s sensing spectra corresponds to exactly one electron per moiré cell, and from that single calibration it derives the moiré density and every filling factor $n$ used in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Hybridization strengthens Wigner crystals above one electron per cell","Layer hybridization strengthens Wigner crystals at high fillings","New Wigner crystal at 3/2 filling from layer hybridization","Moiré hybridization tunes Wigner crystal stability by filling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001365,"raw_usage":{"total_tokens":5578,"prompt_tokens":1028,"completion_tokens":4550,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":644,"completion_tokens_details":{"reasoning_tokens":4481}},"tokens_in":644,"tokens_out":4550,"duration_ms":33541,"temperature":1.0,"reasoning_tokens":4481,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:08:55.163403+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the moiré density of the same heterostructure by an independent method, for example the period of the Landau fan in a dual-gated transport device or the moiré period by scanning tunneling microscopy, and compare it with the density obtained from the $\\Delta V = 7.7$ V interval; if the two disagree by more than the experimental uncertainty, every filling factor, including the $n = 3/2$ correlated insulator, would need to be renumbered.","supporting_citations":[],"review_version":1}