{"id":"dd2bfd99-f0cc-4a02-accb-8b665202ac68","arxiv_id":"2602.23570","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Domain-wall networks in strained graphene bilayers spontaneously form chiral geometries during relaxation, and this chirality shifts low-energy electrons from junctions to asymmetric wall channels.","lead":"Using atomistic simulations, the paper shows that domain-wall networks in strained graphene bilayers can relax into chiral (twisted) geometries — straight, mono-chiral, or dual-chiral — depending on strain and how flexibly the bottom layer is held. A reader might care because this network geometry changes where low-energy electrons sit, turning chirality into a possible design knob for one-dimensional electronic channels.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Phase diagram's 'spontaneous' chiral selection rests on MD potential accuracy and finite candidate sampling; energy gaps between morphologies are unreported and may be below empirical potential error.","rationale":"The reader's weakest assumption correctly identifies that the phase diagram is built from MD with empirical potentials and finite sampling, leaving open the possibility that the chiral ground states are artifacts. My stress-test sharpens this into a concrete falsifiable condition: the energy gaps between morphologies must exceed the potential's expected error and the annealing protocol must at least reach the claimed global minimum among possible chiralities. Without such evidence, the phrase 'spontaneous' is not justified, and the phase boundaries are only tentative. This supports the existing CONDITIONAL verdict rather than changing it. The proposed test is computationally feasible at the small-strain end using a modern machine-learned potential, and would directly probe whether the central claim is robust or an artifact of potential accuracy and sampling.","tokens_in":10827,"tokens_out":9833,"duration_ms":103471,"concrete_test":"At three representative phase-diagram points (ε=0.1% z-rigid, claimed mono-chiral; ε=0.20% z-rigid, claimed dual-chiral; ε=0.24% z-free, claimed straight), rerun the relaxation with an independent interlayer potential (e.g., a machine-learned potential trained on DFT-vdW data or the Ouyang 2018 registry-dependent potential) and 20 independent seeding anneals starting from straight, mono-chiral, and dual-chiral templates. If the lowest-energy morphology at any point changes, or if the energy gap between the claimed ground state and the competitor is below 1 meV/atom (the approximate accuracy of the empirical potentials), the phase diagram and the 'spontaneous' claim are not robust. Also report the energy gaps between the three morphologies at all sampled strains to allow assessment of significance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that TDW networks spontaneously select chiral equilibrium morphologies is supported entirely by the strain–flexibility phase diagram in Fig. 2A, built from MD relaxation with REBO + Kolmogorov–Crespi potentials and a finite annealing/candidate-generation protocol (Methods: 'MD dynamics simulation details' and 'Configuration identification'). For the claim to hold, the MD energy ordering among straight, mono-chiral, and dual-chiral configurations must correctly reflect the true energetic ordering at biaxial strains of 0.1–0.8%. This is insecure for two reasons. First, the KC potential, while registry-dependent, is a fit to limited graphitic data; its accuracy at strains below 1% and at the subtle orientation-dependent line energies that drive chirality is untested. The orientation-dependent formation-energy contrast (Fig. 3A) changes by ~100 meV/Å between edge- and screw-like walls — but the total-energy differences between competing chiral networks are not reported, so the phase boundaries could lie within the potential's error. Second, 'Configuration identification' enumerates only a handful of initial stacking phases and anneals for 100 ns; there is no guarantee that the true global minimum has been found, so the word 'spontaneous' is not established. If the energy separation between the claimed ground state and the next-lowest morphology is less than ~1 meV/atom at the phase boundaries, the diagram is not a reliable equilibrium phase diagram. The electronic-regime claim inherits this uncertainty because the straight/dual-chiral comparison at 0.20% z-rigid is performed on the MD-identified equilibrium structure.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies moiré domain-wall networks in strained graphene bilayers using classical MD with REBO and Kolmogorov–Crespi potentials, followed by tight-binding LDOS calculations. It reports a strain–flexibility phase diagram containing three TDW network morphologies: straight, mono-chiral, and dual-chiral. It further argues that these geometric differences, not stacking topology alone, control the spatial localization of low-energy boundary states: straight networks concentrate spectral weight at AA-stacked topological crossing points, while chiral networks redistribute it along asymmetric wall edges. The paper introduces geometric descriptors (chiral angle α, curvature localization Rc, domain asymmetry Ad) and explains the mechanism via orientation-dependent TDW formation energies from dislocation theory. It concludes that network geometry is an active, programmable degree of freedom in moiré electronic structure.","tokens_in":11137,"tokens_out":5560,"duration_ms":58304,"significance":"If the phase diagram and electronic-structure comparison are quantitatively supported, this is a useful contribution to the moiré-graphene literature. The central idea — that the network geometry of TDWs, not just local stacking registry, controls the spatial distribution of topological boundary states — is well motivated and is likely to be of interest to experimental groups studying strained bilayer graphene. The geometric descriptors are clear, and the TB LDOS maps provide falsifiable signatures distinguishing straight from chiral networks. The manuscript’s principal quantitative support is, however, a direct MD energy minimization with a finite candidate-generation protocol, and the paper does not currently report the energy differences that justify the claimed ground-state morphology at each point in Fig. 2A. That gap makes the phase boundaries hard to assess. The paper is therefore significant if it holds, but the load-bearing numerical evidence needs strengthening.","major_comments":[{"comment":"The phase diagram in Fig. 2A is built by directly comparing total energies of candidate configurations, but no energy differences between the lowest-energy morphology and the next-lowest competitor are reported. At the phase boundaries near ε≈0.23% and ε≈0.83%, the assignment of straight, mono-chiral, or dual-chiral is meaningful only if the energy separation exceeds the empirical-potential uncertainty. Please provide, for each (ε,k) point, the energy difference per atom (or per supercell) between the selected morphology and the runner-up. Also report how many distinct initial stacking phases and annealing repeats were used, and demonstrate that the 100 ns protocol converges to the same state from independent seeds/histories. Without this, the word “spontaneous” for the chiral equilibrium morphology is not fully established.","section":"Methods: Configuration identification"},{"comment":"The central phase diagram rests on REBO + Kolmogorov–Crespi interlayer energetics. The chirality-driving contrast in Fig. 3A is on the order of 100 meV/Å between edge- and screw-like walls, but the KC potential is a fit to limited graphitic data and its registry-dependent accuracy at sub-percent strain is not demonstrated. This is a concrete correctness risk, not a claim of inconsistency. Please validate at least representative straight/mono-chiral/dual-chiral configurations against DFT or a second interlayer potential, or otherwise quantify the sensitivity of the phase boundaries to the interlayer-potential parametrization. At minimum, decompose the total-energy ordering into intralayer and interlayer contributions and show that the ordering is not driven by an artifact of the KC registry term.","section":"Methods: MD dynamics simulation details"},{"comment":"The mechanism in Fig. 3B uses the fitted orientation-dependent formation energy Ef(θ) from Eq. (3), computed with the same MD potential that produced the relaxed chiral networks, and then applies it via Eq. (4) to ‘explain’ the observed curvature. This is partly circular. I recommend making the test predictive: take a relaxed chiral network, constrain or reconstruct the corresponding straight network at identical strain, and show that the total-energy difference is positive and is quantitatively accounted for by the integrated orientation-dependent line-energy gain. That would convert the Ef(θ) curve from a rationalization into an independent energy budget for the chirality transition.","section":"Energetic origins / Eq. (4)"},{"comment":"The text states that at ε=0.1% “the formation energy reduces from –50 eV/Å down to v200 eV/Å.” The symbol “v200” appears to be a typographical corruption, and the sign pattern is unclear from the sentence alone. Because Fig. 3A is the quantitative basis for the orientation-contrast mechanism, the correct values and units must be stated accurately in the text; the current garbled sentence is load-bearing and must be fixed.","section":"Fig. 3A / TDW model and energy analysis"}],"minor_comments":[{"comment":"Typo: “mono-chiral and due-chiral” should be “dual-chiral.”","section":"Results near Fig. 1"},{"comment":"The text refers to “the phase map in Fig. 3C (right)”, but the relevant panel appears to be Fig. 2C (right). Please correct the cross-reference.","section":"Results near Fig. 2C"},{"comment":"The descriptors α' and α'' are introduced in the text and in the caption but are not quantitatively defined. Please give explicit definitions and state how they are computed from the relaxed atomic coordinates.","section":"Fig. 1B–C"},{"comment":"The flexibility is parameterized by 1/k, but Eq. (1) uses a harmonic restraint with spring constant k. For k→0 the inverse stiffness diverges, which is expected, but the range of k values used for the intermediate points in Fig. 2A is not listed. Please specify the k values sampled and how the z-free limit is approached.","section":"Methods: Implementation of interlayer flexibility"},{"comment":"The statement that code is “available from the authors on reasonable request” is weaker than current reproducibility standards. Consider depositing the analysis scripts and structure-generation protocols in a permanent repository.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":"I think the paper’s core idea is interesting and within scope for cond-mat.mes-hall. The main barrier is that the central MD phase diagram is not yet quantitatively supported in a way that lets a reader judge whether the morphological transitions are real equilibrium transitions or artifacts of potential accuracy and finite sampling. The requested energy-difference tables and robustness checks are within the scope of a major revision rather than requiring a new project. I would not recommend rejection at this stage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Plain take: this paper gives the moiré community a genuinely new organizing principle — network geometry (straight vs. mono-chiral vs. dual-chiral) selected by strain and interlayer flexibility — and shows that this geometry reshapes where low-energy states live. The MD phase diagram and the TB comparison are the core evidence, and they are plausible but not yet sharply quantified.\n\nWhat's new: the strain–flexibility phase diagram, the chiral-angle/curvature-localization descriptors, the mono/dual-chiral classification, and the chirality-dependent LDOS redistribution (TCP-centered vs wall-centered). Prior experiments saw curved/swirl domain walls, but this is the first systematic phase diagram I know. The dislocation-theory energy explanation (edge vs screw character driven by strain and flexibility) is a nice explanatory frame and fits Fig. 3.\n\nWhat it does well: the methods are standard (REBO + Kolmogorov–Crespi for MD; Slater–Koster TB with literature parameters). The paper is careful to note that opposite handedness is degenerate without symmetry-breaking fields. It does not overclaim about topological protection; it says topology guarantees edge modes but geometry controls their localization. That's a reasonable and useful distinction.\n\nSoft spots, in proportion: first, the phase-diagram boundaries are the heart of the paper, but the total energy differences between competing morphologies are not reported. The stress-test observation is right: if the separation is below the potential-error threshold at 0.1–0.8% strain, the boundaries are only qualitative. That's a real gap, but it is fixable; the authors should report ΔE per atom or per area at representative points and show convergence with annealing time. Second, candidate generation is finite (a handful of stacking phases, 100 ns anneals), so calling the result 'spontaneous' is stronger than the evidence supports. I'd soften that wording. Third, the electronic comparison is at a single strain and a single flexibility value, so the claimed 'regimes' are illustrative rather than fully demonstrated. Minor: no error bars or sensitivity tests on TB parameters.\n\nNone of these are load-bearing errors. The central phase diagram is a direct MD result, not fitted to anything, and the qualitative trend (rigid layer preserves chirality to higher strain; flexibility straightens walls) is physically sensible. I would take the claim as likely correct in broad strokes, needing quantitative support at the boundaries.\n\nWho is it for: anyone working on relaxation and transport in moiré bilayers, especially the experimental groups that have seen swirl textures. It deserves a serious referee. I would send it out and then push for a revision that reports energy gaps, convergence tests, and makes code/data available.","headline":"A genuinely new organizing principle for moiré domain-wall networks, with a plausible but under-quantified MD phase diagram.","tokens_in":11673,"tokens_out":2156,"would_cite":true,"duration_ms":21413,"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":"Strained bilayer graphene domain-wall networks can spontaneously break mirror symmetry and become chiral, and this geometric chirality—which topology alone does not determine—redistributes low-energy electrons from junctions to curved one-d","keywords":["moiré bilayer graphene","topological domain wall networks","chirality","symmetry breaking","lattice relaxation","dislocation energy","tight-binding local density of states","strain engineering"],"falsifier":"An STM/LDOS line scan along a TDW in a 0.20% biaxially strained bilayer on a rigid substrate: if the chiral-network picture is correct, the TCP-centred −110 meV doublet weakens and the edge-mode LDOS becomes asymmetric and switches sides along a mono-chiral wall; a straight network would keep near-symmetric edge channels for the same strain and stacking.","tokens_in":10702,"feed_emoji":"🌀","tokens_out":4509,"duration_ms":44872,"temperature":0.7,"pith_summary":"The paper tries to establish that, in biaxially strained graphene bilayers, the network of topological domain walls is not fixed by topology: energy minimization can drive the whole network into chiral geometries (mono-chiral or dual-chiral) that break the straight network's mirror and inversion symmetry. These morphologies are selected by a competition between orientation-dependent line energy of each wall (screw-like walls are cheaper than edge-like walls) and the connectivity constraints of the moiré superlattice. Because interlayer flexibility changes that energy contrast, biaxial strain and bottom-layer flexibility form a phase diagram with straight, mono-chiral, and dual-chiral regions. The paper then argues this geometry controls low-energy electronic states: straight networks keep states concentrated at AA-stacked junction points with balanced edge channels, while chiral networks suppress junction peaks and push spectral weight into asymmetric, curvature-selected edge channels. A sympathetic reader would care because it separates topological protection of boundary modes from their spatial distribution, making network geometry a new control knob.","feed_headline":"Graphene moiré networks can twist into chiral domain-wall patterns","feed_subtitle":"Geometry, not topology alone, chooses where low-energy electrons localize in moiré bilayers.","key_machinery":"The load-bearing object is the topological domain wall described as a partial basal-plane dislocation with a Burgers vector, whose formation energy per unit length depends strongly on whether the wall is edge-like, screw-like, or mixed: screw-like orientations cost far less energy. In a connected moiré network the walls cannot all orient their favorite way because they must meet at AA-stacked crossing points separated by the moiré period, so the system minimizes total line energy over the whole network; the chiral angle α and curvature-localization parameter Rc quantify the resulting curvature. This orientation-dependent line energy, plus the flexibility of the bottom layer (which relaxes ou","core_discovery":"Central claim: after stacking-registry symmetry breaking forms AB/BA domains separated by topological domain walls, the domain-wall network itself undergoes a second, independent symmetry breaking. Because a wall's formation energy depends on its angle to its Burgers vector—more screw-like orientations are cheaper—walls want to curve; but walls meet at AA-stacked crossing points, so curvature must be optimized collectively. The global minimum can be a straight network, a network with one handedness around all crossings (mono-chiral), or a network with alternating handedness row by row (dual-chiral), depending on biaxial strain and out-of-plane flexibility of the lower layer. In the chiral st","pith_inferences":["If the predicted phase boundaries transfer to real devices, then choosing a substrate with different out-of-plane compliance (suspended vs supported bilayer) at the same strain should move a sample from dual-chiral to mono-chiral to straight; this is directly testable by low-temperature STM imaging of wall curvature at fixed ~0.3% strain.","Because left- and right-handed mono-chiral states are degenerate, a real sample quenched through the transition could form domains of opposite handedness separated by defect lines; the paper does not discuss what happens at such boundaries.","The curvature-selected edge asymmetry suggests a possible valley-filtering effect: electrons on a curved wall may favour one physical edge depending on propagation direction and local curvature, which would show up as non-reciprocal transport; the paper doesn't claim this, but the geometry points that way.","A quantitative falsifier beyond the paper's potentials: recompute the straight vs chiral energy ordering with a first-principles registry-dependent interlayer functional at 0.1–0.8% strain; if the edge–screw energy contrast is much smaller than the classical potentials predict, the chiral windows would shrink or disappear."],"forward_implications":["If the phase diagram is right, biaxial strain and substrate stiffness can deliberately switch a device between junction-centred and wall-centred electronic behaviour without changing twist angle or stacking sequence.","Chiral networks should show directional low-energy channels along walls whose spatial side (which edge carries weight) is predictable from local curvature and crossing handedness—an in-plane anisotropy knob.","Within the chiral regime, AA-domain size shrinks by roughly a factor of two relative to straight walls, which will affect any property sensitive to the AA-stacked regions, such as interlayer coupling or impurity capture.","The straight-to-chiral transition is a genuine network-level symmetry breaking, so the straight state cannot be recovered simply by relaxing a single domain wall; any modelling that treats walls as isolated line defects will miss the effect.","The same mechanism should appear in other deformable layered systems where connected arrays of topological line defects form, not just graphene bilayers."],"fun_headline_variants":["Moiré walls twist into chirality, redirecting electrons","Strain flips moiré networks into chiral, electron-steering patterns","Chiral moiré networks: a second symmetry break steers electrons","Moiré domain walls go chiral, rerouting low-energy states","Beyond topology: chirality reshapes moiré electronic paths"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole phase diagram rests on the classical interlayer potentials and annealing protocol finding the global minimum: if the registry-dependent energy difference between edge-like and screw-like walls is wrong, or the 100 ns anneals get trapped in metastable chiral states, then the 'spontaneous' chiral ground states are computational artifacts rather than equilibrium phases.","fun_headline_variants_meta":{"raw":{"variants":["Moiré walls twist into chirality, redirecting electrons","Strain flips moiré networks into chiral, electron-steering patterns","Chiral moiré networks: a second symmetry break steers electrons","Moiré domain walls go chiral, rerouting low-energy states","Beyond topology: chirality reshapes moiré electronic paths"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00027,"raw_usage":{"total_tokens":1467,"prompt_tokens":751,"completion_tokens":716,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":495,"completion_tokens_details":{"reasoning_tokens":638}},"tokens_in":495,"tokens_out":716,"duration_ms":6465,"temperature":1.0,"reasoning_tokens":638,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T20:16:29.173365+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An STM/LDOS line scan along a TDW in a 0.20% biaxially strained bilayer on a rigid substrate: if the chiral-network picture is correct, the TCP-centred −110 meV doublet weakens and the edge-mode LDOS becomes asymmetric and switches sides along a mono-chiral wall; a straight network would keep near-symmetric edge channels for the same strain and stacking.","supporting_citations":[],"review_version":1}