{"id":"0b9bd6bd-7931-4549-8d07-6de2c8ce0941","arxiv_id":"2502.02137","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Domain walls in sliding ferroelectrics are predicted to move like undamped solitons, accelerating uniformly under an electric field and coasting indefinitely after field removal.","lead":"Using machine learning simulations, the paper predicts that domain walls in sliding ferroelectric bilayer 3R-MoS2 move without damping, continuing at constant speed after the electric field is removed. This challenges the usual picture of ferroelectric switching and could enable low-power racetrack memories.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The undamped claim rests on setting Γ=0 in Eq. (2) without a quantitative bound; the short field-off MD window cannot distinguish zero damping from a small but nonzero damping.","rationale":"I read the paper in good faith: the authors combine DFT, a DP potential, large-scale MD, and sine-Gordon field theory to predict that sliding-ferroelectric DWs move without damping. If correct, this is an important result. The reader's CONDITIONAL verdict is appropriate. My stress test focused on the step where the damping coefficient Γ is introduced and then set to zero. The concern is not that the paper is internally inconsistent; it is that the central prediction is more fragile than the headline suggests. The MD simulation is genuinely useful evidence, and the consistency between the quadratic displacement fits and Eq. (2) with Γ=0 is a nontrivial check; the relativistic-like saturation near v_c is also encouraging. However, the field-off portion is short, the DP model is acknowledged to miss long-range interactions, and no independent calculation of Γ or phonon-limited decay is provided. A longer simulation and a phonon-linewidth estimate would settle whether the observed constant velocity is true undamped motion or simply an underdamped transient. This does not require changing the reader's verdict; it reinforces the need for conditional acceptance with these checks.","tokens_in":8818,"tokens_out":6600,"duration_ms":78061,"concrete_test":"Run the published DP-MD setup with the field switched off at 20 ps and continue to at least 1 ns at 300 K in a cell at least twice as large, recording v_DW(t). Fit the field-off branch to v_DW(t)=v0 exp(-t/τ) and report τ with a 95% confidence interval. Also compute the zone-center interlayer shear phonon linewidth from the same DP potential as an independent estimate of Γ. If τ is resolved and is comparable to or shorter than the device transit time (L/v ~ 1 ns), the undamped claim must be downgraded to underdamped; if τ cannot be resolved because it exceeds the accessible window, the paper should state an upper bound on Γ instead of claiming exactly zero.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the passage containing Eq. (2): a phenomenological damping coefficient Γ is written into the soliton equation of motion and then set to zero (\"When the damping coefficient Γ approaches zero, the formula reduces to Eq. S17\"), turning the dynamics into undamped Newtonian motion. The paper never derives Γ=0 from the DP Hamiltonian or from independent phonon-lifetime data. The only direct evidence is the short field-off segment in Fig. 3: the field is removed at 20 ps and the DW is followed for only tens of picoseconds at 1 K and 300 K. With a small nonzero Γ, the field-off velocity decays as v(t)=v0 exp(-t/τ); any τ ≳ 100 ps would look like a flat line over this window. Thus the MD data can at most place an upper bound on Γ, not prove it is exactly zero. The qualitative phonon-scattering argument (weak interlayer shear mode) is not quantitative, and the authors' own note that the DP cutoff radius discards long-range interactions (around Fig. 1) weakens the MD evidence for an exactly dissipation-free limit. A finite periodic MD cell can also reabsorb emitted phonons, artificially removing the phonon radiation channel. The central claim therefore rests on an unverified limit rather than a measured smallness.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript combines DFT, Deep Potential MD, and sine-Gordon field theory to study domain-wall (DW) motion in sliding ferroelectric bilayer 3R-MoS2. It claims that the DW undergoes uniformly accelerated motion under a constant electric field, that its velocity approaches a relativistic-like limit set by the in-plane transverse acoustic phonon speed, and that after field removal the wall continues to move at constant velocity, i.e., undamped soliton-like motion, in contrast to PbTiO3. The authors fit the MD displacement curves, extract a linear acceleration-field relation, and propose a two-terminal experimental geometry to detect the post-pulse inertial motion.","tokens_in":9057,"tokens_out":5257,"duration_ms":47663,"significance":"If the undamped-motion claim is correct, the result is important: it would establish an inertia-dominated, dissipation-free DW regime in a 2D sliding ferroelectric, with direct implications for low-power racetrack-type memories. The paper's strengths are its direct MD evidence of uniform acceleration and continued motion after field removal, the explicit comparison with conventional ferroelectric PbTiO3, and a concrete experimental protocol. However, the central 'undamped' conclusion is not yet established: the analytic model sets the damping coefficient to zero rather than deriving or bounding it, and the MD field-off window is short. These issues are load-bearing and need to be addressed before the claim is supported.","major_comments":[{"comment":"The undamped Newtonian equation, Eq. S17, is obtained by writing a phenomenological damping term proportional to Gamma in Eq. (2) and then taking Gamma -> 0. This is an assumption, not a derivation: the paper does not quantify Gamma from the DP Hamiltonian or from independent phonon-lifetime calculations. Since the central claim is that the DW is undamped, this limit must be supported by a quantitative estimate or by an upper bound extracted from the simulations, rather than imposed by hand.","section":"Eq. (2) and Supplemental Material S2"},{"comment":"The field-off MD window is too short to distinguish zero damping from a small nonzero damping. The field is removed at 20 ps and the wall is followed for only tens of picoseconds; for v(t) = v0 exp(-t/tau), any tau of order 100 ps or larger is visually indistinguishable from a plateau over this window. Please provide longer trajectories at 1 K and 300 K, fit the post-field velocity decay, and report an upper bound on Gamma (or a lower bound on tau).","section":"Fig. 3(a,b,d)"},{"comment":"The manuscript states on page 6 that vc is estimated to be on the order of 3000 m/s, but on page 7 it reports that the DW velocity ceases to increase at around 4000 m/s, 'close to the value estimated from our field theory analysis.' These statements are inconsistent: a plateau at 4000 m/s would exceed the model's limiting speed vc. Please reconcile this discrepancy and report vc, the fitted saturation velocity, and the associated uncertainties.","section":"Fig. 2(e) and text near Eq. (1)"},{"comment":"The periodic simulation cell and the finite cutoff of the Deep Potential model can suppress phonon-radiation damping: a periodic cell can reabsorb phonons emitted by the moving wall, and the authors themselves note that the DP cutoff discards long-range interactions when discussing the static DW width and energy. Please discuss these effects quantitatively, for example by testing the field-off velocity decay as a function of cell size and cutoff.","section":"Simulation setup around Fig. 1(c)"}],"minor_comments":[{"comment":"The text cites 'Fig. 3(b) and (c)' for the PbTiO3 comparison, but Fig. 3(b) shows MoS2 at 300 K; the PbTiO3 data appear to be in Fig. 3(c). Please correct the cross-reference.","section":"Fig. 3 caption / main text"},{"comment":"There are several typos and duplicated phrases, including 'Tthe', 'the strength of the strength of electric field', 'originate s', and 'the undamped DW motion the undamped motion of DW after' near the end of the experimental-proposal paragraph. A careful copyedit is needed.","section":"Throughout"},{"comment":"Please report statistical uncertainties on the fitted accelerations, especially for the 300 K data, so that the linear acceleration-field relation and the field-off plateau can be assessed quantitatively.","section":"Fig. 2(d) and Table S1-2"},{"comment":"The notation E_v = E / gamma is introduced without a derivation or a clear definition of E; please define E explicitly and ensure the typesetting of the gamma factor is correct.","section":"Eq. (1) and surrounding notation"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope and the claimed phenomenon would be significant, but the undamped conclusion is currently supported mainly by an imposed Gamma -> 0 limit and by short field-off MD trajectories. I would like the authors to either provide a quantitative bound on Gamma from longer simulations or independent phonon-lifetime data, or scale back the 'undamped' claim to a statement about very low damping. The inconsistency between vc ~ 3000 m/s and the reported saturation at ~4000 m/s should also be fixed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the core prediction is new and worth taking seriously. The paper shows, via DP-based MD of bilayer 3R-MoS2, that a ferroelectric domain wall accelerates uniformly under an applied field and keeps moving at constant speed after the field is switched off. That behavior is distinct from the damped, terminal-velocity motion seen in conventional ferroelectrics and in magnetic DWs. The static DW structure matches the sine-Gordon field theory, and the relativistic-like saturation near the transverse acoustic speed is an interesting consistency check. For that, the paper gets real credit.\n\nThe soft spot is the zero-damping assumption. The field theory in Eq. (2) includes a damping coefficient Γ, then sets it to zero to obtain the undamped Newtonian equation. That is an input, not a result. The qualitative argument that the weak interlayer shear mode suppresses phonon scattering is plausible but not quantitative. The MD evidence is only tens of picoseconds after field removal; with a small damping, any relaxation time above ~100 ps would produce an almost flat line. The DP cutoff radius also excludes long-range interactions, and a finite periodic cell can reabsorb emitted phonons, both of which could mask dissipation. These concerns are addressable: derive a bound on Γ from the Hamiltonian or phonon lifetimes, run longer and larger simulations, and check convergence.\n\nMinor: the measured saturation velocity (~4000 m/s) and the estimated vc (~3000 m/s) differ more than the text acknowledges. The paper also does not ship the DP model parameters or scripts, so independent reproduction is harder.\n\nBottom line: the paper is a serious candidate for a real effect, but the central claim is not yet nailed down. It deserves a careful referee, not a desk reject. I would ask for a quantitative damping analysis and longer-time simulations before publication. For a reading group, it is a good discussion piece: the physics is clear, the questions are precise, and the claim is falsifiable.","headline":"New and plausible prediction of undamped DW motion in sliding ferroelectrics, but the zero-damping limit is assumed rather than derived; deserves refereeing.","tokens_in":9624,"tokens_out":2704,"would_cite":false,"duration_ms":27008,"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":"This paper predicts that domain walls in bilayer 3R-MoS2 accelerate uniformly under an electric field and keep moving at constant speed after the field is removed, unlike conventional ferroelectrics.","keywords":["sliding ferroelectricity","domain wall motion","soliton","sine-Gordon equation","molecular dynamics","machine learning potential","3R-MoS2","undamped dynamics"],"falsifier":"In a bilayer 3R-MoS2 device, apply a picosecond electric-field pulse long enough to launch a domain wall but short enough that the wall does not cross the sample, then track the wall position after the pulse. If its velocity decreases after pulse removal, or if the delayed-switching interval grows with the distance the wall must travel, the undamped claim is falsified.","tokens_in":8566,"feed_emoji":"⚡","tokens_out":7460,"duration_ms":68299,"temperature":0.7,"pith_summary":"The paper predicts that domain walls in the sliding ferroelectric bilayer 3R-MoS2 behave as undamped solitons. Under a constant electric field they accelerate uniformly, following Newton's second law; after the field is switched off they keep moving at constant speed instead of stopping. This contrasts with conventional ferroelectrics, where damping makes walls reach a terminal velocity and halt when the drive ends. The authors argue this happens because sliding ferroelectric switching is carried by a weak interlayer shear mode that couples weakly to phonons, so the damping coefficient $\\Gamma$ is nearly zero. If correct, the result means domain-wall racetrack memory in sliding ferroelectrics could run on short pulses rather than continuous currents.","feed_headline":"Domain walls in sliding ferroelectrics glide on after the field is off","feed_subtitle":"Simulations show undamped, inertia-driven walls that keep moving once the drive stops.","key_machinery":"The central object is the (1+1)-dimensional sine-Gordon soliton for the interlayer sliding field $u_s(x,t)$, whose static waveform and energy distribution match the machine-learned molecular dynamics result. The perturbed equation of motion for the soliton's center includes a damping term proportional to $\\Gamma$; setting $\\Gamma=0$ reduces it to an undamped Newton's second law, giving uniform acceleration under a constant field and inertial motion after the field is off. The critical velocity $v_c=\\sqrt{\\lambda/\\rho}$ is the speed of the in-plane transverse acoustic phonon and plays the role of the speed of light, while the prefactor multiplying the electric field acts as the domain wall's mass-to-charge ratio.","core_discovery":"The central claim is an undamped, 'relativistic-like' soliton motion: the domain wall in bilayer 3R-MoS2 obeys an undamped Newtonian equation of motion (obtained from the perturbed sine-Gordon equation when $\\Gamma\\to 0$), accelerates linearly with applied field, and retains its velocity indefinitely after the field is removed. Molecular dynamics at 1 K and 300 K show the displacement is quadratic in time under constant field, the acceleration scales linearly with field strength, and the wall continues at constant velocity after field switch-off, while a comparison PbTiO3 wall stops. At large fields the velocity saturates near the in-plane transverse acoustic phonon speed (about 3000 to 4000 m/s), the analogue of the speed of light for this soliton. The authors therefore propose that sliding ferroelectrics offer a dissipation-free mechanism for domain-wall-based memory devices.","pith_inferences":["Beyond the paper, the same undamped mechanism should apply to other sliding ferroelectrics such as bilayer h-BN, since the authors note the universal character of their domain walls but do not run the driven-dynamics simulation for those systems.","A testable extension is to vary the pulse width in the proposed delayed-voltage experiment: undamped motion predicts the delay equals film length divided by constant wall speed, independent of pulse width, whereas any damping would make the delay pulse-width dependent.","The claim is an idealization at exactly $\\Gamma=0$; even a very small damping would produce slow deceleration over long distances, so experiments bounding the velocity decay rate would quantify how close real materials are to the idealized limit.","The relativistic-like analogy implies the wall's effective mass grows as it approaches $v_c$; measuring the precise velocity saturation curve could test the Lorentz-factor energy scaling that the paper invokes but does not directly measure."],"forward_implications":["If correct, racetrack memory built from sliding ferroelectrics could be operated with field pulses only, eliminating the continuous current and associated heating that limits conventional domain-wall devices.","The predicted wall speed of roughly a kilometre per second is an order of magnitude faster than typical ferromagnetic domain walls, so read/write operations would be much quicker.","Because the wall accelerates uniformly, its arrival time at a target electrode is a deterministic quadratic function of pulse duration, which could simplify timing in device operation.","At very high fields the acceleration fades as the wall approaches the transverse acoustic phonon speed, setting an intrinsic upper bound on wall velocity that device designs must respect."],"supporting_citations":[{"why":"This review defines sliding ferroelectricity and the stacking-dependent polarization mechanism that the domain-wall model is built on.","marker":"[2]"},{"why":"This work supplies the machine-learned interlayer potential used in the large-scale molecular dynamics simulations and the fatigue-resistant switching context.","marker":"[5]"},{"why":"This reference provides the perturbed sine-Gordon soliton dynamics used to derive the domain-wall equation of motion and the damping term.","marker":"[20]"},{"why":"This molecular dynamics study of conventional ferroelectric domain walls provides the damped terminal-velocity behavior that the paper contrasts with undamped motion.","marker":"[21]"},{"why":"This study of stacking-engineered ferroelectrics shows the universal static domain-wall properties in sliding systems that motivate the general claim.","marker":"[30]"},{"why":"This paper establishes the machine-learned molecular dynamics method that gives near-ab-initio forces for the large-scale simulations.","marker":"[31]"},{"why":"This reference supplies the transverse acoustic phonon properties of MoS2 used to estimate the critical soliton speed.","marker":"[40]"}],"fun_headline_variants":["Sliding ferroelectric domain walls coast on after field removal","Undamped soliton walls in sliding ferroelectrics ignore the brake","Relativistic-like speed limit for domain walls in sliding ferroelectrics","Domain walls in sliding ferroelectrics accelerate then glide forever","No damping: sliding ferroelectric walls move on after drive stops"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the damping coefficient $\\Gamma$ in the equation of motion is effectively zero, supported only by the qualitative argument that the interlayer shear mode couples weakly to phonons; if any real damping exists, the wall would decelerate after the field is removed and the central claim would fail.","fun_headline_variants_meta":{"raw":{"variants":["Sliding ferroelectric domain walls coast on after field removal","Undamped soliton walls in sliding ferroelectrics ignore the brake","Relativistic-like speed limit for domain walls in sliding ferroelectrics","Domain walls in sliding ferroelectrics accelerate then glide forever","No damping: sliding ferroelectric walls move on after drive stops"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1661,"prompt_tokens":885,"completion_tokens":776,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":501,"completion_tokens_details":{"reasoning_tokens":687}},"tokens_in":501,"tokens_out":776,"duration_ms":7623,"temperature":1.0,"reasoning_tokens":687,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T13:11:02.988269+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In a bilayer 3R-MoS2 device, apply a picosecond electric-field pulse long enough to launch a domain wall but short enough that the wall does not cross the sample, then track the wall position after the pulse. If its velocity decreases after pulse removal, or if the delayed-switching interval grows with the distance the wall must travel, the undamped claim is falsified.","supporting_citations":[{"cited_title":"Wu and J","cited_arxiv_id":null,"evidence_quote":"This review defines sliding ferroelectricity and the stacking-dependent polarization mechanism that the domain-wall model is built on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This work supplies the machine-learned interlayer potential used in the large-scale molecular dynamics simulations and the fatigue-resistant switching context."},{"cited_title":"Fogel, S","cited_arxiv_id":null,"evidence_quote":"This reference provides the perturbed sine-Gordon soliton dynamics used to derive the domain-wall equation of motion and the damping term."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This molecular dynamics study of conventional ferroelectric domain walls provides the damped terminal-velocity behavior that the paper contrasts with undamped motion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This study of stacking-engineered ferroelectrics shows the universal static domain-wall properties in sliding systems that motivate the general claim."},{"cited_title":"Zhang, J","cited_arxiv_id":null,"evidence_quote":"This paper establishes the machine-learned molecular dynamics method that gives near-ab-initio forces for the large-scale simulations."},{"cited_title":"Kaasbjerg, K","cited_arxiv_id":null,"evidence_quote":"This reference supplies the transverse acoustic phonon properties of MoS2 used to estimate the critical soliton speed."}],"review_version":1}