{"id":"2a53eb71-a417-457f-b6ec-1a9c3d924919","arxiv_id":"1908.02420","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Monolayer graphene at ν=0 exhibits a disorder-induced metal-insulator transition and transport signatures consistent with helical edge states in the canted antiferromagnetic phase.","lead":"Experiments on graphene at filling factor ν=0 show a metal-insulator transition near 8.9 T, with conductance plateaus and nonlocal signals interpreted as helical edge transport induced by disorder. The result suggests that one does not need special substrates or large magnetic fields to get quantum spin Hall-like edge channels in graphene.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing step is equating the zero-field intercept of the Δ(B) fit (Γ≈67 K) with CAF edge-gap broadening; the comparison to u⊥ from Ref. [21] does not robustly establish Γ>Δ_edge below Bc.","rationale":"The reader's weakest-assumption analysis identifies exactly the same load-bearing concern: the Γ≈67 K intercept and its comparison to u⊥. I agree with that assessment. The paper presents coherent transport signatures—conductance saturation near e2/h and 2e2/h in the two contact configurations, zero-mode splitting, and a nonlocal resistance pattern—that are suggestive of helical edge transport. However, the paper's own mechanism for obtaining helical transport in the CAF phase is quantitative: disorder broadening Γ must exceed the edge gap Δ_edge. That inequality is built from an extrapolated fit parameter and an order-of-magnitude theoretical range, with no independent measurement of Γ and no error analysis. If the intercept is a fitting artifact or if the true u⊥ is at the high end of its range, the central claim loses its quantitative support. This does not invalidate the data, but it does prevent a clear ACCEPT. The reader's CONDITIONAL verdict is appropriate; no verdict change is needed, provided the condition is understood to be an independent check of Γ and Δ_edge.","tokens_in":8355,"tokens_out":8451,"duration_ms":94516,"concrete_test":"Refit the raw Rxx(T) data at each magnetic field with a full Arrhenius-plus-background model, e.g. Rxx(T)=R0+R1 exp(Δ_B/2T), using bootstrap or weighted least squares with reported uncertainties, then fit Δ_B versus B only over the gapped range and compute the zero-field intercept with its 95% confidence interval. Independently estimate Γ in the same device from the temperature width of the ν=0↔ν=±1 quantum Hall transition or from the width of the Rxx minimum near 5 T. If the bootstrap lower bound of Γ does not exceed the upper bound of u⊥(B) from Ref. [21] at B≈5–8 T, or if the independent Γ differs from 67 K by more than a factor of two, then Γ>Δ_edge is not established and the central disorder-closing mechanism is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires Γ≈67 K, extracted as the zero-field intercept of the linear fit in Fig. 1(b), to exceed the CAF edge gap Δ_edge≈u⊥. That inequality is the only quantitative basis for the statement that the edge gap is 'effectively closed.' It is not secure. First, Γ is a fit parameter: the activation gap Δ(B) is measured only for B above the MIT at Bc≈8.9 T, and the line is extrapolated to B=0, across the phase boundary into a regime where the CAF gap does not exist. The intercept is therefore not a directly measured level broadening. Second, Δ(B) is extracted from bulk Rxx, while the claim concerns the edge gap; no model is given connecting the bulk activation intercept to the edge quantity Δ_edge. Third, the comparison u⊥≈1-10B[T] K from Ref. [21] spans an order of magnitude (at B≈8 T, u⊥≈8–80 K), so whether Γ=67 K exceeds Δ_edge depends on which branch of the theoretical range is used; for the upper branch, Γ<Δ_edge even near Bc. Finally, no error bars or confidence intervals are given for Γ, Bc, or the activation energies, so the inequality cannot be statistically assessed. Because the helical-transport interpretation rests on this inequality, a fitting artifact or an alternative u⊥ value would remove the quantitative foundation of the disorder-closing mechanism.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports magnetotransport experiments on hexagonal boron nitride encapsulated monolayer graphene in an H-bar geometry, focusing on filling factor ν=0 as a function of perpendicular magnetic field. The authors observe a metal-insulator transition at Bc≈8.9 T. Above Bc, the behavior is consistent with the canted antiferromagnetic (CAF) insulating phase, with a thermal activation gap that increases linearly with field. Below Bc, the system is metallic: the zeroth Landau level splits, two-terminal conductance saturates near e2/h or 2e2/h depending on contact geometry, and nonlocal resistance shows a single-peak to double-peak evolution. The central claim is that the CAF edge gap is effectively closed by disorder-induced Landau level broadening (Γ≈67 K, extracted from the zero-field intercept of the activation-gap fit), so that helical edge transport appears without the need for large parallel fields or proximity effects. The parallel-field experiments show a reduction of a measured activation energy with tilt and enhanced backscattering, which the authors interpret as spin-related scattering between helical edge states.","tokens_in":8651,"tokens_out":6679,"duration_ms":74803,"significance":"If the central claim is correct, the paper is significant: it proposes that a certain amount of disorder can turn the CAF phase at ν=0 into a platform for helical edge transport in pristine graphene, offering a simpler alternative to the FM phase induced by large parallel fields or by proximity to magnetic materials. The experimental dataset is rich, including local transport with two different contact configurations, parallel-field dependent measurements, and nonlocal resistance maps. The contact-geometry-dependent conductance saturation (close to 2e2/h and e2/h) is a concrete, nontrivial signature that is consistent with helical edge states, and the nonlocal data are checked against classical Ohmic and flavor-Hall alternatives. However, the quantitative foundation of the disorder-closing mechanism is not secure: the comparison between Γ and the edge gap relies on an indirect, extrapolated fit with no error bars and on a theoretical estimate for u⊥ that spans an order of magnitude. The significance of the paper is therefore conditional on resolving this quantitative gap.","major_comments":[{"comment":"The identification of Γ≈67 K as the disorder-induced broadening of the CAF edge gap is an uncontrolled extrapolation. The activation data used for the linear fit exist only for fields above Bc≈8.9 T, where the insulating CAF phase is present, and the fit is extended through the metal-insulator transition down to B=0, into a regime where the gap is claimed to be closed. No error bars or confidence intervals are provided for Γ, Bc, or the activation energies, so the central inequality Γ>Δ_edge cannot be statistically assessed. The authors should provide a fit with uncertainties, justify the linear extrapolation, and preferably obtain Γ from a directly measured broadening, such as Landau-level linewidths or an independent transport probe.","section":"Section 2, paragraph after Fig. 1(b)"},{"comment":"The comparison with the valley isospin anisotropy energy u⊥≈1−10B[T] K from Ref. [21] does not, by itself, establish Γ>Δ_edge below Bc. At B≈Bc the theoretical estimate spans roughly 9–89 K, so Γ=67 K lies inside that range; the claim requires selecting the lower branch. The authors should specify which u⊥ value they use and why, or demonstrate that the conclusion is robust over the entire theoretical range. As written, the inequality is not decided by the cited theory.","section":"Section 2, same paragraph"},{"comment":"The activation gap Δ is extracted from bulk Rxx measurements (Fig. 1), whereas the disorder-closing claim concerns the edge gap Δ_edge. No model or experimental procedure connects the bulk activation intercept to the edge quantity; the schematic in Fig. 1(b) explicitly distinguishes bulk and edge, but this distinction is not carried through the analysis. This missing link is load-bearing because the central mechanism is specifically an edge-gap closure.","section":"Sections 2–3"},{"comment":"The activation energies measured at B⊥=5 T in the regime below Bc (59 K at 41.6° and 97 K at 90°) are in tension with the claim that disorder (Γ≈67 K) closes the edge gap: at 90° the measured activation energy exceeds Γ. If this activation energy corresponds to the edge gap, then Γ<Δ_edge and the mechanism fails; if it corresponds to a different gap (e.g., a bulk gap or a backscattering gap), the paper should state that explicitly and explain the physical distinction. As written, the two quantitative statements are not reconciled.","section":"Section 3, Fig. 3(c)"}],"minor_comments":[{"comment":"There is a typo in the sample fabrication paragraph: 'perdiodic modulation' should be 'periodic modulation'.","section":"Sample fabrication"},{"comment":"The word 'configuraten' in the text near Fig. 2(c) is a typo and should read 'configuration'.","section":"Section 2, Fig. 2 caption and text"},{"comment":"The abstract contains an extra space in 'm etal-insulator transition'; please correct it.","section":"Abstract"},{"comment":"The axis label for the vertical axis of Fig. 1(b) appears corrupted in the manuscript rendering ('40/g39/g3 (K)'); please ensure the final figure has a legible label such as 'Δ (K)'.","section":"Fig. 1(b)"},{"comment":"The contact resistance correction (Rc≈0.3 kΩ) is matched at ν=−2 only; the sensitivity of the conductance plateaus in Fig. 2(c) and (d) to this correction should be stated so that the reader can judge the closeness of the saturation to e2/h and 2e2/h.","section":"Section 2, contact resistance"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting set of transport data with a plausible but not yet quantitatively secured interpretation. The main obstacle to acceptance is the unsupported inequality Γ>Δ_edge, which is the linchpin of the disorder-closing mechanism. The authors should be asked to provide error analysis for the activation-gap fit, clarify the relation between the bulk activation gap and the edge gap, and reconcile the activation energies at B⊥=5 T with the claim that the edge gap is closed. If these points can be addressed, the paper would merit publication in a mesoscopic transport journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper reports a coherent set of experiments and a plausible new claim: in a single hBN-encapsulated monolayer graphene device, the ν=0 state undergoes a metal–insulator transition at Bc≈8.9 T, and below Bc the transport signatures—conductance saturation, zeroth Landau level splitting, nonlocal resistance—look like helical edge transport in the canted antiferromagnetic phase, with the edge gap closed by disorder rather than by proximity effects or large parallel fields. As far as I know, that specific observation is new, and it is a meaningful step toward simpler routes to helical edge channels in graphene.\n\nThe paper does several things well. The conductance saturation values in two different contact configurations (1.3 e²/h and e²/h) line up with the floating-contact formula for helical edges, which is a nice internal consistency check. The parallel-field experiments are a good control: Bc is invariant in perpendicular field and the activation gap shrinks with added parallel field, which argues against a Zeeman mechanism. The nonlocal data are at least qualitatively consistent with helical edge transport, and the authors explicitly rule out the classical Ohmic contribution and the flavor Hall effect. The writing is careful, and the claims are appropriately hedged.\n\nWhere it gets soft is the quantitative foundation. The disorder-closing mechanism rests on Γ≈67 K, extracted as the zero-field intercept of a linear fit to activation gaps measured only above Bc. That intercept is an extrapolation across a phase boundary, and no model connects the bulk activation gap to the edge gap Δ_edge. The comparison to u⊥≈1–10B[T] K from Kharitonov is a wide target: at B=8.9 T that is roughly 8–90 K, so whether 67 K actually exceeds the edge gap depends on which branch of theory you take. There are no error bars on Γ or Bc. The stress-test note is right that the inequality Γ>Δ_edge is not secure.\n\nThat said, the qualitative picture does not collapse without the inequality. The MIT itself is empirical, and the conductance saturation is the key observation; the Γ comparison is an interpretation of why the MIT occurs, not the evidence that it occurs. A revision that treats the quantitative analysis as suggestive rather than demonstrative, includes error propagation, and ideally shows a second device would make the paper much stronger.\n\nI would send this to a serious referee. It is the kind of result that will be discussed and cited regardless, and a careful referee will help the authors sharpen the argument. My own verdict would be conditional on revision, not rejection.","headline":"Solid experimental paper with a genuinely new result—helical edge transport in the CAF phase of monolayer graphene reached via disorder—though the quantitative case for why disorder closes the edge gap rests on a shaky extrapolation.","tokens_in":9186,"tokens_out":2654,"would_cite":true,"duration_ms":30778,"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":"Disorder turns graphene's ν=0 state into a helical conductor","keywords":["monolayer graphene","filling factor ν=0","canted antiferromagnetic phase","helical edge transport","quantum spin Hall effect","Landau level broadening","zeroth Landau level","nonlocal transport"],"falsifier":"Measure $B_c$ and the activation gap in a set of graphene devices with controlled disorder (for example, varying irradiation dose); if $B_c$ does not shift to lower fields as $\\Gamma$ increases, or if a device with $\\Gamma$ clearly below $u_\\perp$ still shows conductance saturation near $2e^2/h$ below $B_c$, the disorder-closing mechanism is falsified.","tokens_in":8148,"feed_emoji":"🧲","tokens_out":9211,"duration_ms":81228,"temperature":0.7,"pith_summary":"This paper reports that at the charge-neutrality point (filling factor $\\nu=0$) of monolayer graphene, the insulating canted antiferromagnetic phase develops a metallic, helical edge transport regime below a critical perpendicular field $B_c\\approx 8.9$ T. The authors attribute the metal–insulator transition to disorder-induced Landau level broadening, which effectively closes the edge gap in the canted antiferromagnetic phase. They support this with conductance saturation close to $2e^2/h$ or $e^2/h$ depending on contact configuration, a splitting of the zeroth Landau level, and nonlocal resistance patterns. If correct, the result implies that helical (quantum spin Hall-like) edge channels can be realized in pristine graphene without large parallel fields or proximity to magnetic materials.","feed_headline":"Disorder turns graphene's ν=0 state into a helical conductor","feed_subtitle":"Conductance saturates near 2e²/h or e²/h, a fingerprint of quantum spin Hall edge modes in pristine graphene.","key_machinery":"The mechanism is the competition between the valley isospin anisotropy energy $u_\\perp\\approx 1$–$10B[T]$ K, which opens the CAF edge gap $\\Delta_{\\mathrm{edge}}$, and disorder-induced Landau level broadening $\\Gamma\\approx 67$ K, extracted from the zero-field intercept of the linear activation gap. When $\\Gamma>\\Delta_{\\mathrm{edge}}$ (below $B_c$), the edge gap closes and helical edge states with oppositely canted antiferromagnetic spin textures form; the conductance for these states follows the contact-equilibration formula $G=e^2/h\\left((N_1+1)^{-1}+(N_2+1)^{-1}\\right)$.","core_discovery":"The central claim is that in the canted antiferromagnetic (CAF) phase of monolayer graphene at $\\nu=0$, disorder-induced Landau level broadening $\\Gamma\\approx 67$ K exceeds the valley-isospin anisotropy energy that sets the edge gap for fields below $B_c\\approx 8.9$ T. As a result the edge gap is effectively closed, leaving counter-propagating helical edge channels that carry quantized conductance. Evidence includes a metal–insulator transition in $R_{xx}$ and two-terminal conductance at the same $B_c$, a saturation of $G$ near $2e^2/h$ (one floating contact on each edge) and near $e^2/h$ (two floating contacts), the appearance of a zeroth-Landau-level splitting, and nonlocal resistance that is inconsistent with Ohmic or flavor-Hall origin but matches helical edge transport.","pith_inferences":["If the mechanism is generic, there should be an optimal disorder strength for helical transport: too clean a sample stays insulating, while too much disorder would localize the edge modes; this could be tested with graded-disorder devices.","The same gap-closing logic may apply to other quantum Hall ferromagnets where an isospin-anisotropy gap competes with disorder broadening, possibly explaining earlier reports of metallic $\\nu=0$ behavior in lower-mobility samples.","The parallel-field-induced backscattering suggests the helical edge modes have a spin texture that can be manipulated magnetically; spin-resolved or noise measurements could directly probe this texture.","A quantitative re-analysis of the activation gap assuming a field-dependent $u_\\perp$ might refine $B_c$ and produce a sharper prediction for where the conductance plateau should appear."],"forward_implications":["Helical edge transport at $\\nu=0$ can be obtained in pristine monolayer graphene by tuning disorder, without parallel magnetic fields or proximity to magnetic insulators.","The metal–insulator transition at $B_c\\approx 8.9$ T reflects the crossover where the CAF edge gap is washed out by Landau-level broadening.","Conductance saturation values of $2e^2/h$ and $e^2/h$ in different contact geometries provide a quantitative fingerprint of helical edge modes with equilibration at floating contacts.","Nonlocal resistance measurements can distinguish helical edge transport from classical Ohmic and flavor Hall contributions, since the single- and double-peak evolution tracks the gap opening.","The reduction of the thermal activation gap under in-plane magnetic fields points to enhanced backscattering between the counter-propagating spin-textured edge states."],"supporting_citations":[{"why":"Supplies the valley isospin anisotropy energy $u_\\perp\\approx 1$–$10B[T]$ K that sets the CAF edge gap compared with the disorder broadening.","marker":"[21]"},{"why":"Established the CAF phase at $\\nu=0$ and its transition to the FM phase under parallel fields, the basis for expecting helical edge states in the CAF phase.","marker":"[24]"},{"why":"Reported similar conductance saturation and contact equilibration formula in graphene electron–hole bilayers that the present two-terminal conductance is compared with.","marker":"[38]"},{"why":"Provided the earlier activation-gap measurement of the CAF phase that the linear $\\Delta$ vs $B$ fit here extends.","marker":"[22]"},{"why":"Theoretical study of finite-temperature conductance and backscattering of helical edge states that the observed deviation from perfect quantization is compared with.","marker":"[29]"},{"why":"Demonstrated nonlocal resistance as a signature of helical edge (quantum spin Hall) transport, used to interpret the nonlocal data.","marker":"[42]"},{"why":"Showed helical-edge-like nonlocal response in polycrystalline graphene with grain-boundary scattering, an alternative interpretation considered.","marker":"[41]"},{"why":"Describes the flavor Hall effect as a competing explanation for nonlocal resistance, which the paper rules out by the observed peak structure.","marker":"[40]"}],"fun_headline_variants":["Disorder closes graphene's gap, reveals helical edges","Helical edge transport at ν=0: disorder's role","Graphene ν=0 shows quantum spin Hall from disorder","Metal-insulator transition at ν=0 signals helical edges","Disorder-driven helical edge modes in graphene"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative foundation is the assumption that the zero-field intercept $\\Gamma\\approx 67$ K of the thermal-activation-gap fit truly measures the disorder broadening of the CAF edge gap, and that comparing this with $u_\\perp$ correctly predicts the gap closure below $B_c$.","fun_headline_variants_meta":{"raw":{"variants":["Disorder closes graphene's gap, reveals helical edges","Helical edge transport at ν=0: disorder's role","Graphene ν=0 shows quantum spin Hall from disorder","Metal-insulator transition at ν=0 signals helical edges","Disorder-driven helical edge modes in graphene"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000722,"raw_usage":{"total_tokens":3173,"prompt_tokens":810,"completion_tokens":2363,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":426,"completion_tokens_details":{"reasoning_tokens":2284}},"tokens_in":426,"tokens_out":2363,"duration_ms":20057,"temperature":1.0,"reasoning_tokens":2284,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:44:05.208417+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure $B_c$ and the activation gap in a set of graphene devices with controlled disorder (for example, varying irradiation dose); if $B_c$ does not shift to lower fields as $\\Gamma$ increases, or if a device with $\\Gamma$ clearly below $u_\\perp$ still shows conductance saturation near $2e^2/h$ below $B_c$, the disorder-closing mechanism is falsified.","supporting_citations":[{"cited_title":"Kharitonov, Phys","cited_arxiv_id":null,"evidence_quote":"Supplies the valley isospin anisotropy energy $u_\\perp\\approx 1$–$10B[T]$ K that sets the CAF edge gap compared with the disorder broadening."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established the CAF phase at $\\nu=0$ and its transition to the FM phase under parallel fields, the basis for expecting helical edge states in the CAF phase."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reported similar conductance saturation and contact equilibration formula in graphene electron–hole bilayers that the present two-terminal conductance is compared with."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provided the earlier activation-gap measurement of the CAF phase that the linear $\\Delta$ vs $B$ fit here extends."},{"cited_title":"Tikhonov, E","cited_arxiv_id":null,"evidence_quote":"Theoretical study of finite-temperature conductance and backscattering of helical edge states that the observed deviation from perfect quantization is compared with."},{"cited_title":"Ribeiro, S","cited_arxiv_id":null,"evidence_quote":"Showed helical-edge-like nonlocal response in polycrystalline graphene with grain-boundary scattering, an alternative interpretation considered."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the flavor Hall effect as a competing explanation for nonlocal resistance, which the paper rules out by the observed peak structure."}],"review_version":1}