{"id":"c3d476fb-5c0c-4034-b1d0-dcccb049433f","arxiv_id":"2502.01007","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Polarization reversal in sliding ferroelectrics is driven by superlubric, wave-like domain wall motion enabled by off-diagonal Born effective charges at the walls.","lead":"The paper proposes that electric fields switch sliding ferroelectrics by pushing wide, wave-like domain walls across the material, not by sliding whole atomic layers. If correct, this explains the ultrafast, fatigue-free switching seen in bilayer boron nitride and suggests cryogenic-friendly memory devices.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The MD driving force uses a BEC map fitted to rigidly slid unit cells; its validity inside a 10-nm domain wall is unverified, so the 4000 m/s superlubric velocities may be artifacts.","rationale":"The paper's symmetry argument against homogeneous switching is sound, and the proposal that DW motion, driven by off-diagonal Born effective charges at symmetry-broken interfaces, is the switching mechanism is physically plausible. The MLFF simulation and the qualitative match to curved triangular domains provide real but not quantitative support. However, the headline predictions — near-barrierless propagation, v ≈ 3000–4000 m/s, cooling-enhanced speed, and inertial motion — are generated by MD in which the only coupling to the electric field is the Z3j(u) BEC map. That map is built from rigidly slid unit cells and is never checked against DFT forces for actual wall configurations. If it overestimates or underestimates the in-plane forces on wall atoms, the entire superlubric dynamics could be an artifact. The reader's CONDITIONAL verdict with moderate confidence is therefore appropriate, and the concrete DFT-force test would either strengthen the paper to ACCEPT or require substantial revision. No change to the verdict is needed at this stage.","tokens_in":10104,"tokens_out":9886,"duration_ms":112119,"concrete_test":"Extract representative MD snapshots of the Σ0 wall at 100 K and 293 K under E3 = 0.1 and 0.3 V/nm, and compute per-atom in-plane forces from first principles (finite-field DFT or DFPT on the same relaxed geometries). Compare these DFT forces with the forces assigned by the Z3j(u) model used in the MD. If the mean vector error for atoms inside the wall exceeds roughly 0.1 meV/Å, or if rerunning MD with corrected forces changes the steady-state DW velocity by more than 20%, the superlubric-motion claim is not supported. A complementary test: refit the BEC map with additional local descriptors (uy gradient, buckling height, interlayer distance) and check whether one trajectory's velocities shift materially.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim depends on the Z3j(ux, uy) functions, fitted to unit-cell-averaged Born effective charges from rigidly displaced bilayers, being accurate when applied locally inside a moving domain wall under an out-of-plane field. The paper does not show that the field-induced in-plane force on an atom is determined by the local in-plane displacement u alone. Inside a Σ0 wall the local environment also contains gradients of u, strain, possible small buckling, and dynamic electronic reorganization; BECs are sensitive to such degrees of freedom. If the map is wrong for wall atoms, the computed driving force, the near-zero motion barrier, and the superlubric velocities near 4000 m/s could be artifacts. The qualitative reproduction of curved triangular domains is not a quantitative test of force magnitudes or velocities, and Eq. (1) is fitted to the same MD output, so it cannot validate the force model. Without code, data, or a direct DFT-force comparison, this is a load-bearing unverified approximation rather than an internal inconsistency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript argues that polarization reversal in sliding ferroelectrics, modeled by Bernal-stacked bilayer h-BN, cannot proceed by homogeneous interlayer sliding because such a process violates Neumann's principle and carries a prohibitive energy barrier. Instead, switching is governed by the propagation of wide, wave-like Σ0-type 180° domain walls, driven by out-of-plane-field-induced in-plane forces that arise from off-diagonal Born effective charges in C3-broken regions. Using DFT-computed BECs, a fitted interpolation Z3j(u), and machine-learned force-field molecular dynamics, the authors report near-barrierless, superlubric DW motion with velocities up to about 4000 m/s, anomalous cooling-enhanced velocities, and inertial response after field removal. They also report a creep-to-superlubric transition under compressive stress for buckled Σπ/6 walls.","tokens_in":10362,"tokens_out":8607,"duration_ms":88737,"significance":"The symmetry argument is clear and the DFT BEC calculations are carefully presented; the distinction between homogeneous sliding and DW-mediated switching addresses a real inconsistency with experiments on h-BN and related sliding ferroelectrics. If the proposed mechanism is correct, it explains ultrafast, fatigue-free switching, provides the falsifiable prediction of cooling-enhanced DW mobility, and suggests design rules for cryogenic applications. The qualitative reproduction of curved triangular moiré domains is also a notable success. However, the central quantitative claims—superlubric velocities of about 4000 m/s, near-zero motion barrier, cooling enhancement, and the 1/h^2 scaling—currently rest on a fitted BEC map and a heuristic velocity law that has not been independently validated. The stress-test concern about the local validity of the Z3j(u) map inside a moving domain wall is real and load-bearing; it needs to be addressed with direct DFT-force checks or equivalent evidence.","major_comments":[{"comment":"The Z3j(ux, uy) functions are fitted to unit-cell-averaged Born effective charges from rigidly slid bilayers and then applied locally to atoms inside a moving ~10-nm domain wall. This assumes the field-induced in-plane force on an atom depends only on the local in-plane displacement u, independent of strain gradients, residual buckling, and dynamic electronic reorganization inside the wall. The manuscript offers no direct validation (e.g., finite-field DFT forces on the relaxed wall supercell or BEC values for wall atoms). Because the DW switching mechanism and the superlubric velocities are computed with this map, the central quantitative claim is not yet established.","section":"§3 (MD simulations of domain walls in sliding ferroelectrics), Fig. 2 and Supplementary III"},{"comment":"Equation (1) is fitted to the same MD velocities it is claimed to describe; with θ fixed to 1 and c_h and c_T as free parameters, the agreement in Fig. 3e-f is a curve fit, not an independent validation. No statistical uncertainties on v are reported, so the deviations in Fig. 3d cannot be assessed. A direct calculation of the wall-energy barrier versus wall position would provide a much stronger test of the near-zero motion-barrier and superlubricity claims.","section":"§4 (Superlubric motion), Eq. (1) and Fig. 3d-f"},{"comment":"The v ∝ 1/h^2 scaling is extracted from MD data for the buckled Σπ/6 wall using the same u-only BEC interpolation, whose validity is least secure precisely in strongly buckled walls. In addition, E3 = 5 V/nm is far into a regime where the linear BEC ansatz F = Z* E may fail due to nonlinear electronic response. The creep-to-superlubric transition needs support from finite-field DFT forces or from BEC calculations at buckled configurations before it can be accepted.","section":"§5 (Creep-to-superlubric transition), Fig. 4"},{"comment":"Key simulation details (system size, boundary conditions, thermostat or NVE integration, trajectory statistics, and how DW velocities are extracted) are not provided in the manuscript, and the data are only available 'upon reasonable request.' The paper repeatedly defers to Supplementary Sections I-IV, which were not part of the text provided for review. Given that the central claims are quantitative and dynamical, the authors should deposit code and input files and report statistical errors.","section":"Computational methods and Data Availability"}],"minor_comments":[{"comment":"The abstract reports approximately 4000 m/s at room temperature, while the text reports about 3000 m/s at 293 K and E3 = 0.3 V/nm; please reconcile these numbers.","section":"Abstract and §4"},{"comment":"The statement that 'only atoms at the domain walls ... possess non-zero off-diagonal BEC elements' is imprecise; individual atoms in the domains have non-zero off-diagonal BEC tensors, but their C3-symmetric contributions cancel in the unit-cell sum.","section":"Abstract"},{"comment":"Several typographical errors need correction: 'with with' in the Fig. 2 caption, 'sperlubric-like' in the final section, 'We suggests' in the superlubricity discussion, and 'bucking of 38 Å' (presumably 0.38 Å).","section":"Throughout"},{"comment":"The statement that homogeneous switching 'violates Neumann's principle' is conceptually loose; the rigorous statement is that the C3-symmetric initial state has no net first-order in-plane force from E3.","section":"Introduction and §3"},{"comment":"The axis conventions are confusing: the text says the wall extends along y, while Fig. 3e plots wall position along x; please define the orientation conventions explicitly and state how DW velocity is extracted from the MD trajectories.","section":"Fig. 2 and Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a significant open question and the symmetry argument is strong. My main concern is that the quantitative claims (superlubric velocities, zero barrier, cooling enhancement, and the 1/h^2 scaling) all depend on the fitted BEC map and are not validated against direct DFT forces on the domain wall. I recommend requesting those calculations and the deposition of code and data before acceptance. The paper's citation pattern and scope are appropriate for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper is worth taking seriously. The symmetry argument against homogeneous sliding is clean and probably right, and the DW-mediated switching picture is a genuine contribution. The soft spot is the quantitative side: the BEC map used in MD is fitted to rigidly slid unit cells, and its validity inside a moving 10-nm wall is assumed, not demonstrated. Read it for the mechanism, not for the exact 4000 m/s.\n\nWhat's new: the Neumann's principle argument that a C3-symmetric single domain cannot generate an in-plane force under E3, and the off-diagonal BEC mechanism that only acts at symmetry-breaking DWs. That is a nice, falsifiable idea. The DFT BEC tables and the local Z3j(ux,uy) fits look internally consistent, and the MD reproduction of curved triangular domains in twisted h-BN is a good qualitative check. The wave-like coherent propagation picture, with local displacements small and a wide wall, is clearly different from Miller-Weinreich, and the superlubric analogy is worth discussing.\n\nSoft spots: the main one is the one you flagged. The Z3j(u) functions come from unit-cell-averaged BECs of rigidly slid bilayers. Inside a moving DW, the environment has strain, buckling (for non-Sigma0 walls), and dynamic charge reorganization. No direct DFT force comparison inside the wall is shown. So the field-induced forces that drive the 4000 m/s velocities could be off. Eq. (1) is fitted to the same MD data, so it doesn't validate the force model. Velocities have no error bars. No code or data shipped. That's a lot of 'trust us' on the quantitative side. But these are approximations, not internal contradictions. The central switching mechanism doesn't depend on the exact velocity law; it depends on the off-diagonal BECs and the symmetry argument, which are on firmer ground.\n\nThe citation pattern is fine; ref [31] is fairly cited as the uniform-force simulation. The paper is a bit breathless in places (Lorentz-invariant sine-Gordon speculation), but the core is serious.\n\nWho is it for: anyone working on sliding ferroelectrics or 2D Moiré devices. It deserves a real referee. I'd send it out. My own verdict would be: accept the mechanism, require the authors to validate the BEC map inside the wall, or at least show sensitivity to its form, before the quantitative velocities are taken at face value.","headline":"The symmetry argument against homogeneous sliding is the real result; the 4000 m/s DW velocities are a model prediction resting on an unverified BEC map.","tokens_in":10861,"tokens_out":1946,"would_cite":true,"duration_ms":19205,"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":"Sliding ferroelectric switching is driven by wide, wave-like domain-wall motion, not by coherent layer sliding.","keywords":["sliding ferroelectrics","bilayer h-BN","domain walls","Born effective charges","superlubricity","ferroelectric switching","molecular dynamics","twisted h-BN"],"falsifier":"Measure the switching time of a bilayer h-BN ferroelectric device as a function of temperature from 300 K down to a few kelvin at fixed field: the paper predicts faster switching on cooling. Observing slower switching, or temperature-independent switching, at low temperature would contradict the superlubric-domain-wall mechanism.","tokens_in":9887,"feed_emoji":"⚡","tokens_out":13503,"duration_ms":102796,"temperature":0.7,"pith_summary":"This paper argues that the long-accepted picture of switching in sliding ferroelectrics, in which two monolayers rigidly slide past each other under an out-of-plane electric field, cannot be right: the three-fold rotational symmetry of the AB and BA stackings cancels any in-plane force, and a synchronous uniform flip would carry an enormous energy barrier. Using Bernal-stacked bilayer boron nitride as a model system, the authors show instead that polarization reversal is carried by wide, roughly ten-nanometer domain walls separating oppositely polarized domains. Only atoms inside the wall break the symmetry, and their off-diagonal Born effective charges let an out-of-plane field push them sideways; molecular dynamics with these dynamic charges shows the wall gliding as a coherent wave packet at speeds around 3000-4000 m/s with an almost zero motion barrier. The same mechanism reproduces the curved triangular domains seen in twisted bilayers and explains why experiments see ultrafast, fatigue-free switching that even speeds up as the sample is cooled. If right, this reframes sliding ferroelectrics as systems where domain walls, not bulk layers, are the active switching element.","feed_headline":"Sliding ferroelectrics switch by wave-like walls, not sliding","feed_subtitle":"Friction-free domain walls hit 4000 m/s; cooling makes switching faster.","key_machinery":"The central object is the $\\Sigma_0$-type 180-degree domain wall in Bernal-stacked bilayer h-BN: a wall that runs parallel to the in-plane displacement vector $u$ and separates $P^+$ ($u=(0,u_0)$) and $P^-$ ($u=(0,0)$) domains, with a width of roughly ten nanometers. The load-bearing machinery is the tensorial Born effective charge, the quantum response property that describes how an electric field in one direction produces a force on an atom in another direction; the off-diagonal components $Z^*_{31}$ and $Z^*_{32}$ are exactly zero in the $C_3$-symmetric single domains but nonzero in the symmetry-broken wall region, converting an out-of-plane field $E_3$ into in-plane forces on wall atoms. The authors fit DFT-computed unit-cell-averaged Born effective charges to analytic functions $Z_{3j}(u_x,u_y)$ of the local sliding displacement and insert these into a machine-learned force field for finite-field molecular dynamics. The mechanism is the resulting coherent, wave-like propagation: the wall translates because unit cells with decreasing local energy balance those with increasing local energy, so the net motion barrier is nearly zero, and kinetic friction follows a superlubric form in which velocity grows with field and shrinks with thermal corrugation, producing faster walls at lower temperature.","core_discovery":"The central discovery is that polarization reversal in sliding ferroelectrics is governed entirely by the motion of symmetry-breaking domain walls, with no global interlayer translation. In the Bernal-stacked h-BN bilayer, the AB and BA stackings have $C_3$ symmetry, so an out-of-plane electric field produces a strictly zero net in-plane force in single-domain regions; only at the wide $\\Sigma_0$-type 180-degree wall, where $C_3$ is broken, do nonzero off-diagonal Born effective charges convert $E_3$ into in-plane forces. These forces push atoms at the wall collectively, and because the wall is about ten nanometers wide, each atom moves only a small fraction of the lattice constant while the wall travels macroscopic distances, a wave-like coherent propagation rather than the nucleation-and-growth, layer-by-layer switching of perovskite ferroelectrics. Finite-field molecular dynamics with a local-environment-dependent Born effective charge model yields wall velocities of roughly 3000 m/s at 293 K and 0.3 V/nm, up to about 4000 m/s, with a near-zero motion barrier, continued inertial motion after the field is switched off, and velocities that increase as temperature decreases; the authors attribute this to structural superlubricity, with the field balancing a kinetic friction that grows with velocity and thermal corrugation.","pith_inferences":["Because only wall atoms respond to the field, one could pattern or pin domain walls with strain or gates and use them as deterministic multi-state memory elements without stochastic nucleation, an engineering direction the paper sketches but does not develop.","The superlubric friction law suggests a quantitative prediction beyond the paper's data: at fixed temperature, switching time should scale with field as roughly $1/E$ with correction from thermal corrugation, which a field-dependent switching-time measurement could test directly.","The paper's sine-Gordon analogy implies a maximum wall velocity at ultralow temperature; if the motion is truly Lorentz-invariant, a domain wall driven near that speed would show relativistic-like saturation and effective mass growth, potentially observable in time-resolved switching experiments."],"forward_implications":["Electric-field switching in sliding ferroelectrics such as bilayer h-BN is predicted to proceed by domain-wall motion rather than by coherent interlayer sliding, so device switching speed is set by wall velocity, not by the static sliding barrier.","Because only wall atoms carry the off-diagonal Born effective charges that couple $E_3$ to in-plane forces, a single-domain sample in a uniform field should be unswitchable: no new domains can nucleate, which can be exploited for deterministic, stochastic-free polarization control.","The near-zero motion barrier and superlubric friction imply switching that is fatigue-free and extremely fast, consistent with the reported endurance exceeding 10^11 cycles and nanosecond-scale switching.","Domain-wall speed is predicted to increase as temperature decreases, so cryogenic operation should be faster, not slower, in these materials.","The same dynamic-Born-effective-charge mechanism reproduces the curved triangular domain patterns in twisted h-BN, indicating that the theory applies beyond the Bernal bilayer."],"supporting_citations":[{"why":"Reports interfacial ferroelectricity in bilayer h-BN by van der Waals sliding, establishing the model system used here.","marker":"[6]"},{"why":"Demonstrates stacking-engineered ferroelectricity in bilayer boron nitride, the experimental platform whose switching the paper explains.","marker":"[7]"},{"why":"Measures ultrafast high-endurance switching in sliding-ferroelectric memory, the experimental speed and fatigue behavior the paper's mechanism is built to explain.","marker":"[17]"},{"why":"Shows interfacial ferroelectricity and triangular domain patterns in twisted 2D semiconductors, which the MD simulations reproduce.","marker":"[10]"},{"why":"Provides the first-principles picture of narrow 180-degree domain walls in perovskites, the contrasting layer-by-layer switching mechanism.","marker":"[30]"},{"why":"Introduces the classic nucleation-and-growth mechanism for sidewise 180-degree domain-wall motion that the wave-like propagation is compared against.","marker":"[32]"},{"why":"Supplies the concept and formalism of structural superlubricity in 2D interfaces used to interpret the near-zero wall barrier.","marker":"[33]"},{"why":"Gives the kinetic-friction law that the paper adapts into its heuristic velocity-field-temperature relation.","marker":"[34]"},{"why":"Shows facilitated sliding at lower temperatures in superlubric interfaces, supporting the anomalous cooling-promoted wall speed.","marker":"[35]"}],"fun_headline_variants":["Wave-like domain walls replace sliding in ferroelectrics","Superlubric walls drive switching at 4000 m/s","Cooling accelerates ferroelectric switching, not sliding","No sliding: wave-like walls switch ferroelectrics","Superlubricity yields 4000 m/s wall motion"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the sideways force an out-of-plane electric field exerts on each atom is fully determined by that atom's local in-plane displacement, as encoded in Born effective charges fitted to zero-field DFT; if buckling, strain, or dynamic charge transfer inside the moving wall changes these charges, the near-zero barrier and superlubric speeds could be artifacts.","fun_headline_variants_meta":{"raw":{"variants":["Wave-like domain walls replace sliding in ferroelectrics","Superlubric walls drive switching at 4000 m/s","Cooling accelerates ferroelectric switching, not sliding","No sliding: wave-like walls switch ferroelectrics","Superlubricity yields 4000 m/s wall motion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001356,"raw_usage":{"total_tokens":5558,"prompt_tokens":1057,"completion_tokens":4501,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":673,"completion_tokens_details":{"reasoning_tokens":4419}},"tokens_in":673,"tokens_out":4501,"duration_ms":29128,"temperature":1.0,"reasoning_tokens":4419,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T16:54:01.176237+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the switching time of a bilayer h-BN ferroelectric device as a function of temperature from 300 K down to a few kelvin at fixed field: the paper predicts faster switching on cooling. Observing slower switching, or temperature-independent switching, at low temperature would contradict the superlubric-domain-wall mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports interfacial ferroelectricity in bilayer h-BN by van der Waals sliding, establishing the model system used here."},{"cited_title":"Yasuda, E","cited_arxiv_id":null,"evidence_quote":"Measures ultrafast high-endurance switching in sliding-ferroelectric memory, the experimental speed and fatigue behavior the paper's mechanism is built to explain."},{"cited_title":"Weston, E","cited_arxiv_id":null,"evidence_quote":"Shows interfacial ferroelectricity and triangular domain patterns in twisted 2D semiconductors, which the MD simulations reproduce."},{"cited_title":"coherent propagation","cited_arxiv_id":null,"evidence_quote":"Introduces the classic nucleation-and-growth mechanism for sidewise 180-degree domain-wall motion that the wave-like propagation is compared against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the concept and formalism of structural superlubricity in 2D interfaces used to interpret the near-zero wall barrier."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the kinetic-friction law that the paper adapts into its heuristic velocity-field-temperature relation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows facilitated sliding at lower temperatures in superlubric interfaces, supporting the anomalous cooling-promoted wall speed."}],"review_version":1}