{"id":"942dd4de-9b25-4af9-b955-1c4fba48f317","arxiv_id":"2607.06547","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":3,"one_line_summary":"A half-filling fractional quantum Hall state at ν_tot = -1/2 in decoupled twisted double bilayer graphene transitions from a two-component Halperin (331) state to a one-component non-Abelian state under displacement field.","lead":"This paper reports observing a fractional quantum Hall state at half-filling in twisted double bilayer graphene, and a displacement-field-driven transition from a two-component Abelian state to a one-component non-Abelian state. If confirmed, this provides a tunable platform for studying and controlling non-Abelian topological order.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"The non-Abelian identification at finite D is a pure inference from layer polarization, with no topological evidence distinguishing Pfaffian from Abelian one-component states.","rationale":"The reader correctly identified the load-bearing concern: the non-Abelian identification is inferential, not measured. I agree with the reader's verdict of CONDITIONAL. The experimental observation of a D-field-driven transition from a D-sensitive (likely interlayer coherent) state to a D-insensitive (likely layer-polarized) state at ν_tot = -1/2 is genuinely interesting and worth reporting. The zero-D (331) assignment is reasonably motivated. But the headline claim of a transition to a non-Abelian state overstates what the data show. The data establish one-component character, not non-Abelian topological order. This is a meaningful distinction because the paper's abstract and conclusion emphasize the non-Abelian nature as the key result. The paper would be more accurately described as observing a transition from a two-component to a one-component state at ν_tot = -1/2, with the one-component state being a candidate for non-Abelian order based on analogy with single-BLG experiments. The small activation gaps and limited temperature range further constrain the strength of the topological-order claim. No adjustment to the verdict is needed — CONDITIONAL with moderate confidence is appropriate, though I would note the confidence could be slightly lower given that even the one-component characterization at finite D relies on extrapolation of integer filling-factor indices to the fractional regime without direct layer-charge measurement.","tokens_in":10122,"tokens_out":2027,"duration_ms":131644,"concrete_test":"Search for Pfaffian-characteristic daughter states at ν_tot = -1/2 ± 1/4 (and ±1/8 if accessible) in the finite-D regime, using high-resolution density sweeps at base temperature. If these daughter states appear, the Pfaffian identification is substantially strengthened. If they do not appear despite sufficient sensitivity, or if only Abelian-hierarchy daughter states (±1/6, ±1/4 from Jain-type sequences) are found, the non-Abelian claim is weakened. Additionally, extract the activation gap of the -1/2 state at finite D from Arrhenius analysis over a wider temperature range; if the gap is comparable to or smaller than the -5/2 gap (44 mK), the state is too fragile to support a confident topological-order assignment.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is a transition from a two-component Abelian (331) state to a one-component non-Abelian (Pfaffian) state at ν_tot = -1/2. The zero-D identification as (331) is reasonably supported: the state exists at balanced charge distribution, is sensitive to D field (consistent with interlayer coherence), and matches theoretical expectations for the strong-coupling bilayer regime. However, the finite-D identification as non-Abelian rests on a single sentence (Section IV): 'Since ν = -1/2 state in bilayer graphene hosts the non-Abelian state, based on previous theoretical and experimental results, we claim that a transition from two-component Abelian state to one-component non-Abelian state is observed.' This is an analogy, not evidence. The experimental observations at finite D — layer polarization, weak D-dependence of R_xx, and survival to higher temperature — establish only that the state is one-component and incompressible. They do not distinguish between non-Abelian (Pfaffian/anti-Pfaffian) and Abelian one-component candidates. A one-component ν = 1/2 state could in principle be Abelian (e.g., a single-component composite fermion Fermi liquid that becomes incompressible due to Landau level mixing, or other Abelian paired states). The prior experiments in BLG cited [24–26] established non-Abelian order through specific signatures (e.g., daughter state hierarchies, gap scaling, or shot-noise measurements), not merely by observing a one-component incompressible state. Furthermore, the system remains a double-layer structure even at finite D; the empty second layer modifies the Coulomb interaction relative to a standalone BLG, so direct equivalence to single-BLG physics is not guaranteed. The activation gaps reported are extremely small (44 mK for the -5/2 state; the -1/2 gap is not explicitly stated), and the temperature range is limited (230 mK–1 K), further constraining what can be concluded about topological order. The paper would need at least one of: (","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This manuscript reports magnetotransport measurements in large-angle (8°) twisted double bilayer graphene (TDBG), operated as a decoupled double quantum Hall system. The authors observe fractional quantum Hall states at both odd and even denominators on the hole-doped side, including Jain-sequence states near ν_tot = -2 and -3 and a half-filling state at ν_tot = -1/2. The central claim is that the ν_tot = -1/2 state undergoes a displacement-field-driven transition from a two-component Abelian Halperin (331) state at D = 0 to a one-component non-Abelian (Pfaffian) state at finite D. The (331) identification at D = 0 is argued on the basis of the state's sensitivity to displacement field (consistent with interlayer coherence), while the non-Abelian identification at finite D is argued by analogy to prior bilayer graphene results.","tokens_in":11107,"tokens_out":1261,"duration_ms":165761,"significance":"The observation of a displacement-field-tunable half-filling state in a decoupled TDBG platform is a genuine experimental contribution, and the device quality is demonstrated by the observation of multiple fractional states including 2-flux and 4-flux Jain sequences. The identification of the D = 0 state as a two-component (331) state is supported by the displacement-field sensitivity data shown in Fig. 5, which is a reasonable diagnostic. The paper provides a clear experimental pathway for tuning between two-component and one-component regimes in a system with sub-nanometer interlayer separation, which is a physically interesting regime. However, the central claim of a transition specifically to a non-Abelian state is not supported by the data presented in this manuscript, as discussed below.","major_comments":[{"comment":"Section IV, final paragraph: The claim that the finite-D state is non-Abelian rests entirely on the sentence: 'Since ν = -1/2 state in bilayer graphene hosts the non-Abelian state, based on previous theoretical and experimental results, we claim that a transition from two-component Abelian state to one-component non-Abelian state is observed.' This is an inference from analogy, not a direct measurement. The experimental observations at finite D — layer polarization, weak D-dependence of R_xx, and survival to higher temperature — establish that the state is one-component and incompressible, but they do not distinguish between non-Abelian (Pfaffian/anti-Pfaffian) and Abelian one-component candidates. The cited prior experiments in BLG (refs 24–26) established non-Abelian order through specific signatures such as daughter-state hierarchies or gap scaling, not merely by observing a one-comer","section":null},{"comment":"Section IV, Fig. 5: The temperature dependence data are presented as evidence that the half-filling states at zero and nonzero D have different origins. However, the argument is qualitative: the D = 0 state undergoes a transition to an insulating state at 1 K, while the finite-D state survives to higher temperature. No activation gap is extracted for the ν_tot = -1/2 state at either zero or finite D, in contrast to the gaps reported for the -8/3, -7/3, and -5/2 states in Section III. Without a quantitative gap comparison, the claim that the two regimes host fundamentally different topological orders is suggestive but not definitive.","section":null},{"comment":"Section III, Fig. 2(b): The activation gaps for the -8/3, -7/3, and -5/2 states are reported as 80 mK, 64 mK, and 44 mK respectively. These values are extremely small (sub-Kelvin) for a 10 T magnetic field, and no error bars or discussion of uncertainty are provided. The claim of particle-hole symmetry around -5/2 based on 'identical gaps' of -8/3 and -7/3 is not convincing given that the gaps differ by 25% and neither has a stated uncertainty. This weakens the quantitative claims in the paper.","section":null}],"minor_comments":[{"comment":"Abstract and throughout: 'Pffafian' should be 'Pfaffian'.","section":null},{"comment":"Abstract and Section IV: 'tunning' should be 'tuning'.","section":null},{"comment":"Section I: 'seperation' should be 'separation'; 'tuability' should be 'tunability'.","section":null},{"comment":"Section III: 'unzero' should be 'nonzero'.","section":null},{"comment":"Section II: The global gate voltage of -77 V is mentioned without explanation of why it is necessary or what it compensates for. This should be clarified.","section":null},{"comment":"Section IV: The phrase 'the half state at non-zero D field can survive at higher temperature, consistent with previously calculated large energy gaps [41]' is ambiguous — ref 41 refers to BLG, not TDBG. The connection should be stated more precisely.","section":null},{"comment":"Figures: It would help to show line cuts of R_xx vs D at ν_tot = -1/2 at a fixed temperature, rather than only the 2D color map, to make the contrast between the D-sensitive and D-insensitive regimes clearer.","section":null},{"comment":"Reference [26] is cited as supporting the non-Abelian identification in BLG, but the authors should clarify which specific signatures from that work are being invoked, since the present experiment does not reproduce those signatures.","section":null}],"recommendation":"major_revision","confidential_remarks":"The central experimental result — a displacement-field-tunable half-filling state in TDBG — is solid and publishable. The problem is the framing: the non-Abelian claim is an overreach given the evidence presented. If the authors reframe the finite-D state as 'one-component' rather than 'non-Abelian,' or if they can provide additional evidence (e.g., daughter states, gap scaling, or shot-noise measurements), the paper would be substantially strengthened. As written, the non-Abelian identification is a label imported from prior work without independent verification."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful and constructive report. The referee correctly identifies that the experimental observation of a displacement-field-tunable half-filling state in decoupled TDBG is a genuine contribution, and we agree that the identification of the finite-D state as specifically non-Abelian is not directly established by our data. We address each comment below and describe revisions we will make.","responses":[{"response":"The referee is correct. Our data at finite D establish that the ν_tot = -1/2 state becomes layer-polarized (one-component) and incompressible, with weak D-dependence and survival to higher temperature. These observations are consistent with a one-component state but do not by themselves distinguish between a non-Abelian Pfaffian/anti-Pfaffian state and an Abelian one-component candidate (e.g., a particle-hole-symmetric Pfaffian or a 113-type state). The identification as non-Abelian was based on analogy to prior bilayer graphene experiments (refs 24–26), where the non-Abelian order was established through additional signatures such as daughter-state hierarchies and gap scaling — signatures we do not present here. We will revise the manuscript to soften this claim. Specifically, we will reframe the finite-D state as a one-component state that is consistent with a non-Abelian Pfaffian origin by analogy to BLG, while explicitly stating that our data do not rule out Abelian one-component alternatives. The abstract and conclusion will be revised accordingly to replace 'non-Abelian state' with 'one-component state' in the experimental claims, with the non-Abelian interpretation discussed as a motivated but not definitive assignment.","revision_made":"yes","referee_comment":"The claim that the finite-D state is non-Abelian rests entirely on inference from analogy to bilayer graphene, not direct measurement. The experimental observations establish one-component incompressible behavior but do not distinguish non-Abelian from Abelian one-component candidates."},{"response":"This is a fair point. The temperature dependence data in Fig. 5 are presented qualitatively: the D = 0 state transitions to an insulating behavior by 1 K, while the finite-D state survives to higher temperature. We did not extract activation gaps for the ν_tot = -1/2 state because the resistance minima are shallow and the accessible temperature window (230 mK to ~1 K) is limited, making a reliable Arrhenius fit difficult — particularly at D = 0, where the state appears to compete with an insulating background. We acknowledge that without quantitative gaps, the argument for fundamentally different topological orders at zero and finite D remains suggestive. In the revised manuscript, we will add the Arrhenius analysis where feasible and explicitly state the limitations where a reliable gap cannot be extracted. We will also temper the language from 'different origins' to 'consistent with different origins' to accurately reflect what the data support.","revision_made":"partial","referee_comment":"No activation gap is extracted for the ν_tot = -1/2 state at either zero or finite D, in contrast to gaps reported for -8/3, -7/3, and -5/2. Without quantitative gap comparison, the claim of different topological orders is suggestive but not definitive."},{"response":"The referee raises a valid concern. The gaps are indeed small, which reflects the relatively limited temperature range available in our dilution refrigerator setup and the fact that disorder broadening is non-negligible in this device. We agree that the claim of 'identical gaps' for -8/3 and -7/3 is not justified given the 25% difference and the absence of stated uncertainties. In the revised manuscript, we will: (1) add error bars from the Arrhenius fits, estimated from the scatter of the data points and the fit residuals; (2) remove the claim of 'identical gaps' and instead state that the gaps are 'comparable within experimental uncertainty,' if the error bars support this, or simply report the values without the particle-hole symmetry claim if they do not; and (3) add a brief discussion of possible reasons for the small gap values, including disorder and LL mixing effects. We will not claim particle-hole symmetry unless the data with uncertainties support it.","revision_made":"yes","referee_comment":"The activation gaps for -8/3, -7/3, and -5/2 (80, 64, 44 mK) are extremely small for 10 T, no error bars or uncertainty discussion, and the claim of 'identical gaps' for -8/3 and -7/3 is unconvincing given 25% difference."}],"tokens_in":10036,"tokens_out":1201,"duration_ms":111182,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"The headline result is a displacement-field-driven transition at ν_tot = -1/2 in decoupled twisted double bilayer graphene (8° twist), from a state at zero D that looks like a two-component (331) Halperin state to a layer-polarized one-component state at finite D. The experimental observation itself — that you can tune between these regimes with a displacement field in this platform — is new and worth reporting. The device quality is solid: multiple Jain states at both 2-flux and 4-flux sequences, even-denominator states at -5/2, and a reasonable extraction of filling-factor combinations across the two BLG layers. The zero-D identification as (331) is reasonably motivated: the state exists at balanced charge distribution, is strongly sensitive to D field (consistent with interlayer coherence), and matches expectations for the strong-coupling bilayer regime. That part of the argument holds up. The soft spot is exactly where the stress-test says it is. The finite-D state is identified as non-Abelian in a single sentence: 'Since ν = -1/2 state in bilayer graphene hosts the non-Abelian state, based on previous theoretical and experimental results, we claim that a transition from two-component Abelian state to one-component non-Abelian state is observed.' This is an analogy. The data at finite D — layer polarization, weak D-dependence of R_xx, survival to higher temperature — establish that the state is one-component and incompressible. They do not distinguish Pfaffian from Abelian one-component candidates. The prior BLG experiments they cite (refs 24–26) established non-Abelian order through specific signatures like daughter-state hierarchies or gap scaling, not merely by observing a one-component incompressible state. Furthermore, the system remains a double-layer structure at finite D; the empty second layer modifies the Coulomb interaction, so direct equivalence to standalone BLG is not guaranteed. The activation gaps are also very small (44 mK for -5/2; the -1/2 gap is not explicitly stated), and the temperature range is limited (230 mK–1 K), which constrains what can be concluded about topological order. This is a paper with genuine experimental content and an interesting platform, but the central claim overshoots the evidence. The non-Abelian identification needs either direct topological signatures (daughter states, quasiparticle interferometry, shot noise) or a serious theoretical argument tied to the specific device parameters, or it needs to be softened to 'one-component state consistent with Pfaffian.' The 2C-to-1C transition observation stands on its own as a result without the non-Abelian label. This deserves a serious referee — the experimental work is real and the platform is promising — but the referee should push hard on the non-Abelian claim.","headline":"New platform for 2C-to-1C transition at half-filling in TDBG, but the non-Abelian identification is an inference, not evidence","tokens_in":11014,"tokens_out":1261,"would_cite":false,"duration_ms":92044,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["73.43.-f","73.22.Pr"],"model":"glm-5.2","headline":"Displacement field drives Abelian-to-non-Abelian transition at half filling","keywords":[],"falsifier":"A direct experimental test of non-Abelian statistics — such as a Fabry-Pérot interferometry measurement showing anyonic braiding phase shifts inconsistent with Abelian quasiparticles, or a shot-noise measurement revealing charge e/4 quasiparticles characteristic of the Moore-Read state — would falsify or confirm the non-Abelian identification. Short of that, measuring the topological degeneracy of the ground state on a torus (genus-1) geometry would distinguish Abelian from non-Abelian order. Additionally, if numerical exact diagonalization or density-matrix renormalization group calculations","tokens_in":10416,"feed_emoji":"🔌","tokens_out":1222,"duration_ms":208212,"temperature":0.7,"pith_summary":"The paper reports magnetotransport measurements in a large-angle (8°) twisted double bilayer graphene device, where two Bernal-stacked bilayer graphene sheets are rotated relative to each other so that interlayer tunneling is negligible while the physical separation remains sub-nanometer. This creates a strongly coupled double quantum Hall system with independent gate control of each layer. At total filling factor ν_tot = -1/2 and zero displacement field, the authors observe a half-filling state whose symmetry properties and strong sensitivity to charge imbalance between layers are consistent with the two-component Abelian Halperin (331) state, an interlayer-coherent state where each layer hosts charge carriers at fractional filling. When a finite displacement field is applied, charges polarize toward one layer, the interlayer coherence breaks down, and the state transitions to a layer-polarized, one-component configuration. The authors identify this finite-field state as a non-Abelian Pfaffian (Moore-Read) state by analogy to prior theoretical and experimental results showing that ν = -1/2 in single bilayer graphene hosts non-Abelian order. The paper also catalogs multiple odd- and even-denominator fractional quantum Hall states at fillings near -2 and -3, including Jain-sequence states following both 2-flux and 4-flux composite-fermion series, and a -5/2 state that emerges only above a displacement-field threshold. Temperature-dependent measurements yield activation gaps for several of these states and reveal that the zero-field and finite-field half-filling states have different thermal behavior, supporting the claim that they have distinct topological origins.","feed_headline":"Displacement field drives Abelian-to-non-Abelian transition at half filling","feed_subtitle":"A twisted double bilayer graphene device shows a gate-tuned transition at ν = -1/2 from a two-component Halperin (331) state to a one-state.","key_machinery":"The central object is the 8° twisted double bilayer graphene heterostructure: two Bernal-stacked bilayer graphene sheets twisted by a large angle so that interlayer tunneling is suppressed while the physical layer separation remains sub-nanometer, enabling strong interlayer Coulomb coupling. The control knob is the displacement field D, which tunes the charge distribution between the top and bottom bilayer graphene layers. At D = 0, balanced charge distribution favors interlayer-coherent two-component states (Halperin 331). At finite D, charge polarization toward one layer breaks interlayer coherence and drives the system toward one-component states (Pfaffian). The filling factor ν_tot = -1/","core_discovery":"The central finding is a displacement-field-driven transition at ν_tot = -1/2 in decoupled twisted double bilayer graphene: at zero displacement field the half-filling state behaves as a two-component Abelian Halperin (331) state with strong sensitivity to layer-charge imbalance, while at finite displacement field it becomes a robust one-component state that the authors identify as a non-Abelian Pfaffian state. The paper establishes that this transition is driven purely by electrostatic control of the interlayer charge distribution, without requiring changes to magnetic field, sample geometry, or tunneling strength. The device platform — large-angle twisted double bilayer graphene — is the载体","pith_inferences":[],"forward_implications":["If the transition is genuinely from Abelian to non-Abelian topological order, this platform could serve as a tunable route to creating and manipulating non-Abelian quasiparticles using only electrostatic gates, which would be relevant for topological quantum computation architectures.","The displacement-field knob provides a continuous parameter to study the quantum phase transition between competing topological orders at half filling, potentially allowing mapping of the critical behavior and universality class of the Abelian-to-non-Abelian transition.","The observation of both 2-flux and 4-flux Jain sequences, plus even-denominator states at -1/2 and -5/2, in a single device demonstrates that twisted double bilayer graphene can realize a broad family of fractional quantum Hall states, making it a versatile platform for systematic studies of competing topological orders.","The thermal activation gaps reported (44–80 mK at 10 T) are modest; if device quality improves, larger gaps could stabilize the non-Abelian state at higher temperatures, making it more accessible for interferometric or tunneling experiments that directly probe exchange statistics."],"fun_headline_variants":["Gate tuning drives Halperin-to-Pfaffian transition at half filling in TDBG","Displacement field switches half-filled state from Abelian to non-Abelian","Twisted double bilayer graphene tunes half-filling from 331 to Pfaffian","Electrostatic control drives 2C-to-1C transition at ν = -1/2 in TDBG","Displacement field tunes half-filling from two-component to one-component"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The identification of the finite-displacement-field state as non-Abelian rests on an inference: because ν = -1/2 in single bilayer graphene is known to host non-Abelian states, the layer-polarized state observed here at finite displacement field is assumed to be non-Abelian as well. The paper does not directly measure quasiparticle exchange statistics or perform topological-order-sensitive probes such as interferometry; the non-Abelian identification is based on analogy and间接","fun_headline_variants_meta":{"raw":{"variants":["Gate tuning drives Halperin-to-Pfaffian transition at half filling in TDBG","Displacement field switches half-filled state from Abelian to non-Abelian","Twisted double bilayer graphene tunes half-filling from 331 to Pfaffian","Electrostatic control drives 2C-to-1C transition at ν = -1/2 in TDBG","Displacement field tunes half-filling from two-component to one-component"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":1377,"prompt_tokens":594,"completion_tokens":783,"prompt_tokens_details":null},"tokens_in":594,"tokens_out":783,"duration_ms":16669,"temperature":1.0,"reasoning_tokens":617,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T02:05:29.522659+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"A direct experimental test of non-Abelian statistics — such as a Fabry-Pérot interferometry measurement showing anyonic braiding phase shifts inconsistent with Abelian quasiparticles, or a shot-noise measurement revealing charge e/4 quasiparticles characteristic of the Moore-Read state — would falsify or confirm the non-Abelian identification. Short of that, measuring the topological degeneracy of the ground state on a torus (genus-1) geometry would distinguish Abelian from non-Abelian order. Additionally, if numerical exact diagonalization or density-matrix renormalization group calculations","supporting_citations":[],"review_version":1}