{"id":"4a34ad1a-e9f9-4fc8-a038-031b052ee553","arxiv_id":"2601.14014","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In twisted monolayer–rhombohedral N-layer graphene, the zero-field quantum anomalous Hall Chern number equals N (3,4,5), and electric fields can reverse its sign or switch a (2+4)L device between C=3 and C=4.","lead":"This paper reports zero-field quantum anomalous Hall states in twisted monolayer–rhombohedral multilayer graphene in which the Chern number equals the number of rhombohedral layers (3, 4, or 5), and shows the Chern number can be switched electrically in a single device. The work matters because it turns a topological invariant into a designable, gate-tunable property, a step toward topological transistors and high-Chern fractional states.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The C=N rule rests on one device per N at different twist angles and displacement fields; unless C=N is shown robust to θ/D variations, layer-number control is not distinguished from device-specific parameters.","rationale":"The reader's weakest assumption correctly identifies the single-device-per-stack and stack-identity issue. My concern is more specific: even if the stacks are exactly as labeled, the C=N inference is confounded because twist angle and displacement field are not held constant across the three devices. Extended Data Fig. 10 makes this concrete by showing Chern number depends on both θ and Δ with gapless regions, so the theoretical support is parameter-dependent rather than a parameter-free layer-engineering theorem. This does not invalidate the paper; the authors do show overlapping D ranges where D1 and D2 give different C, which argues against D being the sole control, and the (2+4)L device provides a within-device D-driven transition. However, with only one device per N and no fixed-θ comparison, the central design claim is not fully established. The proposed computational check is directly feasible and would settle whether the C=N regions are robust. Since the reader already assigned CONDITIONAL and my concern reinforces that rather than overturning it, the verdict should remain UNCHANGED. I agree partially with the reader because they focus on stack identity/defects, whereas I emphasize the θ/D confound within the stated device parameters.","tokens_in":15810,"tokens_out":4835,"duration_ms":59693,"concrete_test":"Use the authors' continuum/Hartree–Fock model to recompute the Chern number of the first conduction band on a fine grid in (θ, Δ) covering the experimental ranges θ = 1.0°–1.5° and Δ = 0–20 meV for N=3,4,5, and overlay the exact device parameters (D1, D2, D3, D4) from Extended Data Fig. 10. Quantify the area of the C=N phase and the distance of each device from a C≠N or gapless boundary. If any device sits within the experimental uncertainties in θ (±0.05°) or D (±0.02 V/nm) of a boundary where C changes, then the one-device-per-N data cannot distinguish layer-number control from θ/D control. If the C=N regions are broad and contain all device points, the concern is substantially weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the layer number N programs the Chern number C=N at ν=1—is supported by exactly one device for each N (D1, D2, D3), with N=5 reused from the authors' prior preprint (ref 39). Crucially, these devices do not share the same twist angle or displacement field: θ = 1.29°, 1.16°, 1.39° and optimized D = −0.470, −0.508, −0.620 V/nm for N=3,4,5. The accompanying Hartree–Fock calculations (Fig. 1b) are run at these specific (θ, Δ) values, and Extended Data Fig. 10 explicitly shows that the valley Chern number depends on both θ and Δ, with gapless regions marked by crosses. Therefore the observation of C=3,4,5 could, in principle, be controlled by the different twist angles or by the different D values rather than by N itself. This is not an internal inconsistency, but it is the weakest load-bearing point: if the C=N law is not robust over a range of θ and D at fixed N, the paper's design principle ('layer-engineered Chern number') collapses into a statement about three particular devices. The authors do not provide a second device for any N, nor do they show that C=N persists when θ is varied at fixed N, so the layer-engineering rule is underdetermined by the presented data.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports quantum anomalous Hall (QAH) states in twisted monolayer-rhombohedral N-layer graphene, denoted (1+N)L, with N = 3, 4, 5. At moiré filling ν = 1, the authors observe quantized Hall resistances h/3e², h/4e², and h/5e², vanishing longitudinal resistance, magnetic hysteresis to zero field, and Landau-fan slopes consistent with Chern numbers C = 3, 4, 5, respectively. They further report a sign-switchable |C| = 3 state at ν ≈ 3 in the (1+3)L device, controlled by doping or displacement field, and a displacement-field-driven transition between C = 3 and C = 4 at ν = 1 in twisted Bernal bilayer-rhombohedral tetralayer graphene, (2+4)L. Self-consistent Hartree-Fock continuum-model calculations are presented as supporting the layer-dependent Chern numbers and the D-driven transition.","tokens_in":16103,"tokens_out":4779,"duration_ms":56177,"significance":"If the central claim holds, this is a significant advance: it would establish layer number as a programmable knob for high Chern numbers and demonstrate in-situ electrical switching between distinct QAH states in a single device. The transport evidence is internally consistent, with quantized plateaus, hysteresis, and Streda slopes all pointing to the same Chern indices. The calculations provide a plausible theoretical framework. However, the paper's headline 'C = N layer-engineering rule' is currently supported by exactly one device per N, with different twist angles and displacement fields across devices, and by calculations that explicitly show topology depends on θ and Δ. The D-driven transition in (2+4)L is a strong experimental result, but the quantitative connection between measured D and calculated Δ is not given. These gaps undermine the 'programmable layer-engineering' conclusion as presently stated.","major_comments":[{"comment":"The central claim that C = N is set by the layer number N rests on one device per N (D1, D2, D3), at different twist angles (θ = 1.29°, 1.16°, 1.39°) and different optimized displacement fields (D = -0.470, -0.508, -0.620 V/nm). Device D3 is reused from ref. 39. Extended Data Fig. 10 explicitly shows that the calculated valley Chern number depends on both θ and Δ, with gapless regions (crosses). Since no second device at fixed N and no in-situ θ variation at fixed N is provided, the observed C = 3, 4, 5 could in principle be controlled by the different θ or D values rather than by N. This is not an internal inconsistency, but it is load-bearing for the paper's 'layer-engineered Chern number' design principle. The authors should either supply additional devices at the same N with different θ/D, or present a clear argument—with data—that the observed C is robust across the accessible θ-D p","section":"Fig. 1c; Extended Data Fig. 10"},{"comment":"The claimed agreement between the D-driven C = 3 ↔ 4 transition in device D4 and the Hartree-Fock calculation is not quantitatively testable from the text. Fig. 4a shows band structures at Δ = 11.5 meV and Δ = 10.0 meV, while the experimental transition occurs between D = -0.669 and D = -0.602 V/nm. No mapping from the experimental displacement field D to the model interlayer potential Δ is provided, nor is the built-in offset D0 and its uncertainty discussed. The statement 'agrees well with band structure calculations' therefore lacks a directly checkable basis. This does not invalidate the experimental observation of two quantized states at different D, but it weakens the theoretical-interpretation component of the central advance.","section":"Methods (Band structure calculation) and Fig. 4a"},{"comment":"The claims of an unconventional Chern insulator at ν = 3/2 and an incipient fractional Chern insulator near ν = 7/3 are presented as supporting the platform's promise, but the data shown (anomalous Hall with hysteresis, dispersive fan diagrams) do not exhibit quantized ρxy at zero field. The authors attribute this to disorder or contacts, which is reasonable. However, these observations are not load-bearing for the main claims, and the language should be consistently labeled as 'signatures' rather than established states.","section":"Extended Data Fig. 8/9"}],"minor_comments":[{"comment":"The main text (Section 'Displacement-field-induced Chern number switching') refers to a 'red vertical line' in Fig. 4b, while the Fig. 4f caption says 'yellow vertical line'. Please make these consistent.","section":"Fig. 4f caption / main text"},{"comment":"The statement that device D3 was previously studied in ref. 39 should be highlighted earlier, ideally in the main text or in the Fig. 1 caption, because the N = 5 data point is not a new independent measurement. This is relevant to the one-device-per-N concern.","section":"Methods / Device D3"},{"comment":"A table of devices with all relevant parameters (θ, W/L, contact quality, measurement temperature, D range) would help the reader assess device-to-device variability. Currently only twist angle and some D values are scattered in captions.","section":"Extended Data Tab. 1"},{"comment":"Several typographical and formatting issues should be corrected: 'Cuire temperature' should be 'Curie temperature'; 'moire' should be 'moiré' consistently; and the Streda formula is named after Středa, so 'Streda' is acceptable but should be consistent.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The experimental core—quantized Hall plateaus, hysteresis, and Streda slopes—is convincing for the existence of high-Chern QAH states in specific devices. The main issue is that the paper's headline design principle ('C = N' as a layer-engineering rule) is under-supported by one device per N, and the theoretical agreement lacks a D-to-Δ mapping. This is fixable by adding data or reframing the claims as device-specific demonstrations rather than a general rule. If the authors cannot provide a second device at fixed N, the paper should be revised to tone down the 'programmable by layer number' language. The overlap with refs 54 and 55 also needs to be addressed clearly in the final version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real news here is experimental: quantized Hall plateaus at h/3e², h/4e², and h/5e² with vanishing longitudinal resistance and zero-field hysteresis, plus a displacement-field-driven switch between C=3 and C=4 in a single (2+4)L device. The transport characterization looks careful—Streda slopes match the Chern numbers, Landau fans are shown, and the symmetrized/anti-symmetrized resistances are handled properly. I have no serious internal-consistency concern. The Hartree-Fock calculations use per-device Δ and screening parameters chosen to mimic experiment, but the Chern numbers are read off transport, not from the calculation, so circularity is low. The paper also honestly flags concurrent independent studies in a note added.\n\nThe soft spot is exactly where the stress-test note lands: the C=N rule for (1+N)L devices is supported by one device per N, at different twist angles (1.29°, 1.16°, 1.39°) and different displacement fields. Extended Data Fig. 10 shows the valley Chern number depends on both θ and Δ, so the observed C=3,4,5 could in principle be controlled by those parameters rather than by N itself. The authors do not show a second device at fixed N, nor that C=N persists across a range of θ at fixed N. That is a genuine gap between the data and the phrase \"layer-engineered.\" It is not fatal—the trend across three devices plus the (2+4)L transition strongly suggests N matters—but it underdetermines the design rule. Also structurally: the N=5 data come from the authors' own prior preprint, and raw data/code are not deposited, so independent verification is limited.\n\nWho should read this: anyone working on rhombohedral graphene or high-Chern moiré systems. The paper deserves a serious referee. The single-device-per-configuration issue should be pushed in review—ask for more devices or a clear statement that the rule is provisional. Even with that caveat, the (2+4)L C=3↔4 transition is a strong standalone result and I'd cite this work if I were writing on tunable Chern insulators.","headline":"Solid transport evidence for high-Chern QAH states in twisted rhombohedral graphene, but the C=N rule rests on one device per layer number and needs more devices before it becomes a design law.","tokens_in":16722,"tokens_out":1534,"would_cite":true,"duration_ms":17907,"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":"Layer-engineered quantum anomalous Hall states realize Chern number C = N in twisted rhombohedral graphene, with electrical switching between C = 3 and C = 4 in one device.","keywords":["twisted rhombohedral graphene","quantum anomalous Hall effect","Chern number","layer engineering","moiré superlattice","displacement field","topological phase transition","orbital Chern insulator"],"falsifier":"Fabricate a (1+3)L stack with an independent rhombohedral-order check and a different twist angle, then measure the Chern number at ν=1; if it is not 3, or if a stacking fault in any device can be shown to control the observed C, the C=N design rule is falsified.","tokens_in":1300,"feed_emoji":"🧲","tokens_out":2691,"duration_ms":91873,"temperature":0.7,"pith_summary":"This paper tries to show that the Chern number — the integer counting a quantum anomalous Hall insulator's dissipationless edge channels — can be programmed by construction and retuned by voltage. In twisted monolayer-on-rhombohedral-N-layer graphene, denoted (1+N)L, the authors report quantized Hall resistance h/(Ne²) at one electron per moiré cell, meaning C=N for N=3, 4, and 5. In twisted Bernal-bilayer-on-rhombohedral-tetralayer graphene, (2+4)L, a displacement field drives the same device between two distinct QAH states with C=3 and C=4. They also show that near moiré filling ν=3 in the (1+3)L device, the sign of C can be flipped by doping or displacement field. If the layer-engineering rule holds, Chern number becomes a design parameter rather than an accidental material property.","feed_headline":"Chern number equals graphene layer count in twisted stacks","feed_subtitle":"Stacking a monolayer on rhombohedral N layers gives C=3,4,5 at zero field; a voltage flips C between 3 and 4.","key_machinery":"The mechanism is a small-angle (~1.2°) moiré superlattice formed between a monolayer or Bernal bilayer and an N-layer rhombohedral (ABC-stacked) graphene stack. In the continuum-model band structure, the first conduction band becomes an isolated flat band whose valley Chern number is inherited from the layer-dependent Berry curvature of rhombohedral graphene and equals N. Strong Coulomb interactions polarize spin and valley, producing an orbital Chern insulator at odd integer moiré fillings. Dual graphite gates independently tune carrier density n and displacement field D, and Chern numbers are read out from quantized Hall resistance and from Landau-fan slopes through the Streda formula.","core_discovery":"The central claim is that the first moiré conduction band in twisted monolayer-rhombohedral N-layer graphene carries a valley Chern number equal to N, realized as a zero-field quantum anomalous Hall state at moiré filling ν=1: quantized Hall resistivity h/(Ne²) and vanishing longitudinal resistivity for N=3,4,5, with the Streda formula independently giving C=N. In a (2+4)L stack, one device hosts C=3 at one displacement field and C=4 at another, so the absolute Chern number is electrically switchable in situ. Near ν≈3 in (1+3)L, the sign of C reverses with doping, displacement field, or magnetic-field sweep.","pith_inferences":["A direct test of the edge-channel picture is two-terminal conductance: a (1+4)L device in the quantized state should show a plateau near 4e²/h, and switching D in (2+4)L should change that conductance in integer steps.","The C=N rule is a plateau, not a universal constant: the paper's own Hartree-Fock maps show the valley Chern number depends on twist angle and interlayer potential, so thicker stacks or different θ will require optimized D to stay on the plateau — mapping that plateau for each N is a natural next step.","The layer-engineering logic may transfer beyond carbon: any material family with rhombohedral-stack-like layer-dependent Berry curvature could in principle use the same twist recipe to pre-set a Chern number."],"forward_implications":["Chern number becomes a design parameter: choosing N fixes C, so (1+3)L, (1+4)L, and (1+5)L give C=3, 4, and 5 at ν=1.","A single device can be reconfigured: in (2+4)L, displacement field switches between C=3 and C=4 QAH states, and in (1+3)L near ν=3 the sign of C can be reversed.","Higher Chern numbers mean more dissipationless chiral edge channels, enabling a gate-controlled topological transistor whose edge-channel count changes by voltage.","High-C Chern bands are candidate hosts for fractional Chern insulators and non-Abelian excitations; the paper reports hints of unconventional Chern insulators at ν=3/2 and ν=7/3.","The recipe extends to thicker rhombohedral stacks, so even larger programmable Chern numbers are in reach."],"fun_headline_variants":["Chern number equals graphene layer count","Layer count sets quantum Hall number in twisted graphene","Voltage flips Chern number between 3 and 4","Designer QAH: C = N, tunable by field","Twisted graphene maps layers to Hall conductivity"],"cache_read_input_tokens":17920,"weakest_assumption_plain":"Each measured device is exactly the stack it is labeled as — a clean, single rhombohedral domain at the intended twist with no accidental h-BN alignment — so the Chern number inferred from transport is set by N rather than by disorder, stacking faults, or device-specific displacement field; this premise is checked mainly by Raman mapping, with one new device per N and the N=5 sample carried over from earlier work.","fun_headline_variants_meta":{"raw":{"variants":["Chern number equals graphene layer count","Layer count sets quantum Hall number in twisted graphene","Voltage flips Chern number between 3 and 4","Designer QAH: C = N, tunable by field","Twisted graphene maps layers to Hall conductivity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000149,"raw_usage":{"total_tokens":1040,"prompt_tokens":767,"completion_tokens":273,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":198}},"tokens_in":511,"tokens_out":273,"duration_ms":3895,"temperature":1.0,"reasoning_tokens":198,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T09:20:46.043458+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate a (1+3)L stack with an independent rhombohedral-order check and a different twist angle, then measure the Chern number at ν=1; if it is not 3, or if a stacking fault in any device can be shown to control the observed C, the C=N design rule is falsified.","supporting_citations":[],"review_version":1}