{"id":"714d01b9-39b6-4203-9f3b-1a5c95591e98","arxiv_id":"2412.09985","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"A rhombohedral hexalayer graphene/hBN moiré device hosts a doping-switchable Chern insulator, three competing isospin insulators at v=2, and charge-density-wave states at fractional fillings.","lead":"Experiments on a tiny stack of six graphene layers on a boron nitride crystal reveal several switchable, insulating quantum states, including one whose topological character flips when the electron count is nudged. These results map a rich competition among magnetic, valley, and charge-order phases controlled by electric and magnetic fields.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero-field anomalous Hall loops show the same sign on both doping sides of the v=1 gap, so the claimed doping-controlled Chern-number reversal is only established at finite B and may reflect a field-driven transition rather than an electrical switch.","rationale":"The reader's weakest assumption concerned the v=2 isospin classification, which is a secondary result. My concern targets the paper's primary novelty: the switchable Chern insulator at v=1. The text itself states that the electron-doped side has the same sign anomalous Hall near zero field as the hole-doped side, so this is not a hypothetical risk but a direct inconsistency with the stated doping-controlled reversal at zero field. A single zero-field measurement with proper background subtraction would settle whether the Chern number actually reverses with doping or whether the electron side undergoes a field-driven transition. Because the data may still support a field-induced Chern transition, the paper remains conditionally acceptable but with a required revision of the central claim and its interpretation. The verdict therefore stays CONDITIONAL, unchanged from the reader's assessment, though for a different reason than the reader identified.","tokens_in":9796,"tokens_out":12434,"duration_ms":143391,"concrete_test":"Re-analyze the raw Rxy(B) sweeps at v=0.95 and v=1.05 (Figs. 2e-f): subtract a contact offset and ordinary Hall contribution, then extract the zero-field remanent Hall after saturating at ±8T. Also perform field-cooling from ±8T to zero field on both doping sides and measure the zero-field Hall. If the zero-field remanent Hall has opposite signs on the hole and electron sides, the doping-controlled reversal is supported; if the signs are identical, the claim must be revised to a field-induced transition on the electron side and the abstract/title adjusted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of an electrically switchable Chern insulator (C=-1 under hole doping, C=+1 under electron doping at v=1) is based on Hall measurements at B=1T (Fig. 2a-b) and Landau fans. However, the zero-field Hall loops in Figs. 2e-f are inconsistent with a pure doping-controlled sign reversal: the electron-doped side (v=1.05) shows the same sign anomalous Hall near zero field as the hole-doped side (v=0.95), and only reverses to +h/e² by 2.5T. Thus at zero magnetic field both doping sides have the same sign, and the apparent 'switch' in Fig. 2b at 1T occurs because the electron side has already undergone a field-driven transition to C=+1. This also contradicts the proposed mechanism (M changes sign across the gap, which would give opposite signs at zero field). Additionally, the hole-side Rxy=15kΩ is not quantized (h/e²≈25.8kΩ), so the C=-1 assignment relies on sign alone. Because the switchable Chern insulator is the paper's headline result, this is a load-bearing concern.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports low-temperature transport measurements on two rhombohedral hexalayer graphene/hBN moiré devices. In device I (θ=0.51°), the authors identify at filling v=1 a Chern insulator whose Chern number appears to reverse with doping (C=-1 on the hole side, C=+1 on the electron side) at BꞱ=1 T, and they attribute the reversal to a sign change of the total orbital magnetization across the gap (Refs 31, 32). At v=2 they observe three insulating states, labeled spin-antiferromagnetic, spin-polarized, and valley-polarized based on the linear-in-B evolution of their transport gaps under in-plane and out-of-plane magnetic fields. In device II (θ=1.29°), they report charge-density-wave insulators at v=1/3 and 2/3 at zero field, a field-induced stripe phase at v=1/2, and they estimate CDW unit-cell areas from the field dependence of the phase boundaries.","tokens_in":10106,"tokens_out":13300,"duration_ms":148865,"significance":"If correct, the doping-controlled reversal of the Chern number at v=1 would be a notable demonstration of the orbital-magnetization switching scenario in a rhombohedral graphene moiré system, and the v=2 isospin phase map would provide a useful benchmark for competing orders in multilayer rhombohedral graphene. The paper has real strengths: two complementary devices, a direct comparison with a moiré-free r-6LG sample, Streda-slope analysis in the Landau fan that supports the Chern assignments at finite field, and an internally consistent slope-ratio argument for the CDW phases. The manuscript is, however, a purely experimental report whose central claims are interpretive, and the switching claim in particular rests on evidence that is partly inconsistent at zero field (see major comments). The v=2 classification and the CDW observations are independent of that claim and retain value regardless of how the v=1 issue is resolved.","major_comments":[{"comment":"The claim that the Chern number at v=1 is reversed by doping is not supported at zero magnetic field. The text itself states that on the electron-doped side \"an anomalous Hall signal with same sign is detected near zero magnetic field (Fig. 2f)\", with the reversal to +h/e² developing only by BꞱ = 2.5 T. The zero-field loops therefore show the same Chern sign on both doping sides, which is what one expects if C=-1 persists on both sides at zero field. This contradicts a purely doping-controlled switch and also contradicts the proposed mechanism, since a sign change of M across the gap would produce opposite Hall signs at zero field. The opposite Streda slopes in Fig. 2c-d do establish opposite Chern numbers at finite field, but the headline \"switchable Chern insulator\" and the mechanism paragraph must be reconciled with the zero-field data. The authors should either demonstrate opposite remanent Hall signs after field poling on the two doping sides, or explicitly reframe the result as a doping-dependent field-stabilized transition and revise the title/abstract accordingly.","section":"Fig. 2e-f and the 'Switchable Chern insulator' paragraph"},{"comment":"The hole-side state is assigned C=-1 from a zero-field Hall resistance of about 15 kΩ, well below the quantized value h/e² ≈ 25.8 kΩ, and the assignment therefore relies on the sign alone. Because Figs. 2e-f are raw, unprocessed data (Methods S4 states that only Fig. 2c-d and Extended Data Fig. 5 were symmetrized/anti-symmetrized), contact misalignment mixing Rxx and Rxy could contribute sizeably to the zero-field signal. Please provide the anti-symmetrized Hall loops for both doping sides and state what mechanism accounts for the ~40% shortfall relative to h/e² (partial gap occupancy, series resistance, or mixing).","section":"Fig. 2e and Methods S4"},{"comment":"The identification of insulators I, II, and III as antiferromagnetic, spin-polarized, and valley-polarized assumes that the entire magnetic-field dependence of the transport gap is the Zeeman energy g μ_B B of a single isospin flavor. Orbital coupling of BꞱ to the moiré bands, Landau-level formation, and field-induced changes of the Hartree-Fock order parameter can all shift a measured activation gap linearly or nonlinearly in B, in which case the extracted g-factors (≈ -2, +2, +9) would be effective values rather than isospin Zeeman constants. In particular, insulator III is characterized only against BꞱ in Fig. 3d; no in-plane-field gap evolution is reported for this phase, so the large g=9 does not by itself isolate valley order. Please report the B|| response of insulator III or provide an explicit argument separating Zeeman and orbital contributions.","section":"Fig. 3d and the 'Isospin competitions' paragraph"},{"comment":"The flux-ratio argument requires the CDW unit-cell areas, but the assignments are not derived and are stated in a confusing way: the text says that for v=1/3 (2/3) the area A is twice the moiré unit cell and for v=1/2 it is three times, whereas commensurability at v=1/3 (one electron per three moiré cells) would naturally suggest a three-moiré-cell CDW unit cell. In addition, the comparison ratio of the fitted slopes, 0.02:0.014 = 1.43, is asserted to agree with 3:2 = 1.5 without a tolerance; error bars on the slopes are needed to substantiate this claimed agreement.","section":"Fig. 4e and the CDW field-shift discussion"}],"minor_comments":[{"comment":"The abstract uses \"r-6G/hBN\" while the main text uses \"r-6LG/hBN\"; please unify the notation.","section":"Abstract"},{"comment":"The Methods sections are numbered S1, S2, S4; section S3 is missing from the submitted manuscript.","section":"Materials and Methods"},{"comment":"The text cites 'Extended Data Fig. 5' for the measurements at D = 0.62 V/nm and D = 0.70 V/nm, but this figure is not included in the submitted material.","section":"Extended Data"},{"comment":"Reference 27 is incomplete: it gives only a title and no journal, volume, or preprint identifier; please complete all references before resubmission.","section":"References"},{"comment":"The g-factors are described as 'estimated', but the fit ranges, the number of data points, and the uncertainty on each g-factor are not given; please specify them so that the linearity claim can be assessed.","section":"Fig. 3d"},{"comment":"The symmetrization section refers to contacts 1-4 on one side of the channel and contact 5 on the opposite side, but the contact geometry is not defined until Fig. S2; please clarify the contact layout in the text itself.","section":"Methods S4"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the zero-field inconsistency in Fig. 2e-f is the crux of the paper's headline claim. I treat it as a major-revision issue rather than grounds for rejection because the finite-field Streda-slope evidence, the v=2 isospin classification, and the CDW observations are independent of the zero-field behavior and retain value. However, if the authors cannot produce opposite zero-field remanent Hall signs after poling on the two doping sides, the title, abstract, and mechanism paragraph should be softened to describe field-assisted or field-stabilized Chern states rather than an electrical switch. Please also verify that Extended Data Fig. 5 and Methods S3 were not lost in submission, and note that the paper draws its magnetization mechanism from Refs 31-32 but invokes it beyond what the zero-field data currently support."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi [Colleague],\n\nQuick read on 2412.09985. The paper is a clean experimental study of correlated phases in rhombohedral hexalayer graphene/hBN moiré. What's genuinely new: the same device family shows a v=1 Chern insulator, three D-tunable isospin insulators at v=2, and fractional-filling CDW/stripe states in the second device. Each ingredient has precedent in other graphene moirés, but the combination in r-6LG/hBN with these particular tunings is new, and the transport data are of good quality.\n\nThe main thing to know before you trust the abstract: the ‘electrically switchable Chern insulator’ claim is only established at finite B. In Fig. 2e and 2f, the zero-field anomalous Hall loops on both the hole side (v=0.95) and the electron side (v=1.05) have the same sign. The sign flip appears only after the electron-doped side is pushed to about 2.5 T. So this looks like a field-driven transition on the electron side, not a doping-controlled reversal of the orbital Chern number at zero field. That contradicts the stated mechanism (M changes sign across the gap, which would give opposite signs at zero field). Also, the C=-1 assignment rests on a 15 kΩ Rxy—not quantized—and the sign. So the headline claim needs a rewrite or much stronger zero-field evidence.\n\nThe v=2 isospin classification is reasonable but not airtight. The gap-vs-field slopes are fit to a single Zeeman g-factor for each state, so the names (AFI, SPI, VPI) are only as good as that assumption. Orbital or Landau-level contributions could shift the gaps and change the assignment. That's a moderate worry, not a fatal one.\n\nThe fractional states in device II are classified as CDW from gaps and a plausible flux-slope ratio argument, but there are no Hall measurements to check whether they carry Chern numbers. Minor.\n\nCredit where due: two devices, careful moiré characterization, Streda fits, symmetrized Hall data, and gaps extracted from Arrhenius plots. The references include the relevant theory (Zhu, Su, MacDonald) and cite prior Chern insulators in multilayer graphene.\n\nWho this is for: anyone working on correlated and topological phases in rhombohedral graphene moirés. It deserves a serious referee, but the editor should expect major revision to fix the switchable-Chern claim and qualify the isospin assignments. I'd bring it to reading group to debate the zero-field inconsistency.","headline":"Solid second-generation moiré graphene transport study, but the headline ‘switchable Chern insulator’ only holds at finite B, so the claim needs a rewrite.","tokens_in":10660,"tokens_out":2962,"would_cite":true,"duration_ms":34415,"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":"Transport measurements show that a correlated insulator at one electron per moiré cell in rhombohedral hexalayer graphene on hBN is a Chern insulator whose Chern number switches sign with doping.","keywords":["rhombohedral hexalayer graphene","hBN moiré superlattice","switchable Chern insulator","orbital magnetization","isospin order","transport gap","charge density wave","anomalous Hall effect"],"falsifier":"Compute the bulk and edge contributions to the orbital magnetization $M$ of the rhombohedral hexalayer graphene/hBN moiré band at $\\nu=1$: if $M$ does not change sign between the hole and electron sides of the gap, the proposed mechanism for a doping-switchable Chern number is ruled out. On the experimental side, a direct test is to check the $\\nu=2$ gap evolution in a tilted field with the perpendicular component held fixed: a spin-polarized state should respond only to the in-plane component, while a valley-polarized state should not respond to the in-plane component at all.","tokens_in":1976,"feed_emoji":"🧲","tokens_out":2052,"duration_ms":112027,"temperature":0.7,"pith_summary":"This paper reports transport experiments on rhombohedrally stacked hexalayer graphene aligned with hexagonal boron nitride, forming a moiré superlattice, and claims that at filling $\\nu=1$ the correlated insulator is a Chern insulator whose Chern number can be switched by doping: $C=-1$ on the weak-hole side and $C=+1$ on the weak-electron side of the gap. The paper attributes the sign reversal to the total orbital magnetization $M=M_{\\rm bulk}+M_{\\rm edge}$ changing sign as the Fermi level crosses the gap, so that both Chern signs are stable in a magnetic field. At $\\nu=2$, three insulating states are distinguished by their transport-gap response to in-plane and perpendicular magnetic fields, and are identified as spin-antiferromagnetic, spin-polarized, and valley-polarized insulators. In a larger-twist device, insulating charge-density-wave states appear at $\\nu=1/3$ and $2/3$ at zero field and a stripe insulator at $\\nu=1/2$ in a perpendicular field. If correct, these observations make a single rhombohedral graphene moiré device a tunable platform for orbital topology, spin-valley order, and charge order.","feed_headline":"Doping direction flips the Chern number of a graphene moiré insulator","feed_subtitle":"In one rhombohedral graphene device, weak doping reverses the Chern number while other fillings reveal competing spin and valley orders.","key_machinery":"The central object is the orbital Chern insulator's total magnetization, $M=M_{\\rm bulk}+M_{\\rm edge}$, whose sign can reverse as the Fermi level crosses the gap; the paper uses this reversal to explain why both $C=-1$ and $C=+1$ can be realized on opposite doping sides of the $\\nu=1$ insulator. For the $\\nu=2$ classification, the operative mechanism is the distinct Zeeman response of spin versus valley order: spin couples to both in-plane and perpendicular magnetic fields with $g\\approx2$, while valley order responds only to the perpendicular component, so the measured gap slopes in $B_\\parallel$ and $B_\\perp$ identify each insulator's isospin flavor. For the fractional states, the relevant quantity is the ratio of magnetic flux through the charge-density-wave unit cell to flux through the moiré unit cell, which gives a predicted slope ratio of $3:2$ between the $\\nu=1/3$ (or $2/3$) and $\\nu=1/2$ states, matching the measured phase-boundary slopes.","core_discovery":"On the displacement-field side where electrons are pushed away from the moiré superlattice, the paper finds that the $\\nu=1$ correlated insulator in its $0.51^\\circ$ device hosts an anomalous Hall effect with Chern number $C=-1$ when the filling is slightly hole-doped and $C=+1$ when slightly electron-doped, with the two states following the Streda formula in a Landau fan. The sign switch is explained as a doping-controlled reversal of the total magnetization $M$: an orbital Chern insulator has both bulk and edge contributions to $M$, and when the edge contribution dominates, $M$ changes sign across the gap, making both Chern numbers robust. At $\\nu=2$ the same device shows two zero-field insulators whose gaps evolve with magnetic field with effective $g$-factors of roughly $-2$ and $+2$, identifying them as spin-antiferromagnetic and spin-polarized insulators, while a third state that emerges at higher displacement field has a $g$-factor near $9$ and is identified as a valley-polarized insulator because valley order responds only to the perpendicular field. In the second, larger-twist device, insulating states at $\\nu=1/3$ and $2/3$ appear at zero field; their positions shift linearly with perpendicular field at a rate consistent with the ratio of magnetic flux through a charge-density-wave unit cell to flux through the moiré cell, and a $\\nu=1/2$ stripe phase appears above about $5$ T.","pith_inferences":["By extension, the same bulk-versus-edge magnetization competition should be present in other rhombohedral multilayer graphene/hBN moiré systems, making a doping-switchable Chern number a plausible generic feature of orbital Chern insulators with flat bands, though the paper demonstrates it only in hexalayer devices.","A testable consequence not pursued in the paper is that the valley-polarized insulator with $g\\approx9$ should show a correspondingly large valley susceptibility in complementary probes such as circular dichroism or capacitance spectroscopy.","The linear field shift of the charge-density-wave states implies their real-space superlattice unit cells should be imageable by scanning probes, and the measured slopes predict the supercell area ratio of two versus three moiré cells, providing a direct microscopic check."],"forward_implications":["A single device can be toggled between Chern numbers $C=-1$ and $C=+1$ purely by shifting the gate voltage, providing an electrical switch for orbital topology.","At $\\nu=2$, the isospin ground state can be cycled through antiferromagnetic, spin-polarized, and valley-polarized orders using the displacement field and magnetic field as independent knobs.","Zero-field correlated insulators at $\\nu=1/3$ and $2/3$ with distinct spin textures, plus a field-induced $\\nu=1/2$ stripe phase, show that fractional charge order coexists with isospin order in the same moiré flat bands.","Any theory of rhombohedral multilayer graphene moiré bands must reproduce a doping-reversible Chern number at $\\nu=1$ and the $g$-factor hierarchy of roughly $-2$, $+2$, and $9$ at $\\nu=2$."],"supporting_citations":[{"why":"Supplies the theory that an orbital Chern insulator's total magnetization has bulk and edge parts and can change sign across the gap, which the paper invokes to explain the reversible Chern number.","marker":"[31]"},{"why":"Reports electrical switching of magnetic order in an orbital Chern insulator, the experimental precedent for doping-controlled Chern-number reversal.","marker":"[32]"},{"why":"Predicts fractional quantum anomalous Hall physics in rhombohedral multilayer graphene, providing theoretical context for the observed Chern insulator.","marker":"[33]"},{"why":"Describes interaction-driven Chern bands and fractional quantum Hall states in rhombohedral multilayer graphene, supporting the identification of the $\\nu=1$ state as a Chern insulator.","marker":"[34]"},{"why":"Develops the theory of fractional quantum anomalous Hall phases in pentalayer rhombohedral graphene moiré structures, another basis for expecting a topological $\\nu=1$ insulator.","marker":"[35]"},{"why":"Shows topological flat bands in rhombohedral multilayer graphene/hBN moiré superlattices, the band-narrowing effect that enables the correlated and topological phases.","marker":"[25]"},{"why":"Provides the large graphene/hBN moiré band gap used to explain why the layer-antiferromagnetic insulator is absent in the first device.","marker":"[36]"},{"why":"Reports half- and quarter-metals in rhombohedral trilayer graphene, the symmetry-broken isospin phases used to benchmark the $\\nu=2$ insulator classification.","marker":"[16]"}],"fun_headline_variants":["Doping direction flips Chern number in graphene insulator","Reversible Chern numbers in rhombohedral graphene moiré","Switchable Chern insulator and isospin phases in graphene","Doping toggles Chern number and reveals isospin orderings","Graphene moiré hosts switchable Chern and charge density waves"],"cache_read_input_tokens":12800,"weakest_assumption_plain":"The classification of the three insulators at two electrons per moiré cell assumes that magnetic fields shift their transport gaps only through the Zeeman energy of a single spin or valley flavor, with spin responding to both field directions and valley responding only to the perpendicular field, so if orbital or band-structure effects from the field also shift the gaps, the inferred spin and valley orders would change.","fun_headline_variants_meta":{"raw":{"variants":["Doping direction flips Chern number in graphene insulator","Reversible Chern numbers in rhombohedral graphene moiré","Switchable Chern insulator and isospin phases in graphene","Doping toggles Chern number and reveals isospin orderings","Graphene moiré hosts switchable Chern and charge density waves"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00139,"raw_usage":{"total_tokens":5714,"prompt_tokens":1120,"completion_tokens":4594,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":736,"completion_tokens_details":{"reasoning_tokens":4507}},"tokens_in":736,"tokens_out":4594,"duration_ms":37768,"temperature":1.0,"reasoning_tokens":4507,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:28:50.583153+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the bulk and edge contributions to the orbital magnetization $M$ of the rhombohedral hexalayer graphene/hBN moiré band at $\\nu=1$: if $M$ does not change sign between the hole and electron sides of the gap, the proposed mechanism for a doping-switchable Chern number is ruled out. On the experimental side, a direct test is to check the $\\nu=2$ gap evolution in a tilted field with the perpendicular component held fixed: a spin-polarized state should respond only to the in-plane component, while a valley-polarized state should not respond to the in-plane component at all.","supporting_citations":[{"cited_title":"& MacDonald, A","cited_arxiv_id":null,"evidence_quote":"Supplies the theory that an orbital Chern insulator's total magnetization has bulk and edge parts and can change sign across the gap, which the paper invokes to explain the reversible Chern number."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports electrical switching of magnetic order in an orbital Chern insulator, the experimental precedent for doping-controlled Chern-number reversal."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows topological flat bands in rhombohedral multilayer graphene/hBN moiré superlattices, the band-narrowing effect that enables the correlated and topological phases."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the large graphene/hBN moiré band gap used to explain why the layer-antiferromagnetic insulator is absent in the first device."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports half- and quarter-metals in rhombohedral trilayer graphene, the symmetry-broken isospin phases used to benchmark the $\\nu=2$ insulator classification."}],"review_version":1}