{"id":"b204bdce-582c-42a1-8771-b7cb6d11d5dd","arxiv_id":"2505.17742","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Twistronic stacking induces flexoelectric polarization vortices in metallic SrRuO3 membranes, a material class where such ordering was previously thought impossible.","lead":"Twisted stacking of metallic SrRuO3 membranes creates periodic vortex-like patterns of atomic displacements, even though free electrons would normally wipe out such polarization. The effect may turn a metallic ferromagnet into a 'multiferroic metal' where magnetic and polar orders compete.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim rests on 1–20 pm Ru displacement maps whose artifact floor is not quantified; the unannealed and depth-sectioning controls are suggestive but not exculpatory, so an independent bias test is needed before the vortices can be taken as real.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the measured Ru displacement maps are the existential evidence for the paper's central claim. I agree with that assessment. Alternative concerns, such as the meaning of 'polarization' in a metal or the qualitative nature of the flexoelectric reconstruction, are secondary because they presuppose that the displacement pattern is real. If the displacement map is an artifact, the headline claim collapses; if the map survives an independent bias test, the remaining caveats are about interpretation and extrapolation, not existence. The paper does include meaningful controls: single-layer SRO is non-polar, unannealed twisted bilayers are non-polar despite a visible moiré, and DFT with 0-degree stacking is non-polar. These controls raise confidence and justify a conditional rather than negative verdict. However, the quantitative claim is anchored on displacements near the precision limit of the technique, and the controls do not isolate all systematic imaging biases. The proposed multislice simulation directly measures the artifact floor of the exact analysis pipeline under moiré conditions and would settle whether the 1–20 pm signal is real. For these reasons, I would keep the reader's CONDITIONAL verdict rather than upgrading to accept or downgrading to reject.","tokens_in":14487,"tokens_out":4605,"duration_ms":54207,"concrete_test":"Deposit the raw STEM-HAADF series for the 3.0-degree sample and run a multislice STEM simulation of the annealed twisted bilayer with zero Ru displacements, using the same thickness, defocus, scan pixelation, and Wiener-filter/Atomap pipeline as the experiment. If the zero-displacement simulation already produces a periodic delta-Ru map of comparable amplitude (e.g., >3 pm mean or >5 pm maximum) or the same vortex/antivortex topology, then the experimental vortex map is an imaging artifact. Additionally, re-fit the same experimental images with a center-of-mass peak finder without Wiener filtering and with a reverse scan direction; a true lattice distortion must reproduce the same vortex pattern and orientation, whereas a scan or filtering artifact will shift or rotate with the analysis settings.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is the inference from 1–20 pm Ru displacements in top-layer STEM-HAADF images (Fig. 2f–i) to 'polarization vortices' in a metal. Every downstream claim, including twist-angle dependence, flexoelectric correlation, and multiferroic competition, inherits this inference. The unannealed-bilayer control (Supplementary Fig. S9) and the depth-sectioning FFT control (Supplementary Fig. S7) are good and should be credited, but they do not fully bound the error budget. The top-layer images come from a two-layer stack whose lower interface is only about one unit cell away; depth-sectioning shows that bottom-layer Bragg spots vanish at top focus, yet sub-Bragg moiré contrast from the buried interface can still bias Gaussian centroid fits at the few-picometre level without producing distinct Bragg spots. The annealed-versus-unannealed comparison changes not only interlayer coupling but also interface chemistry, crystallinity, and possibly local thickness, so it is not a perfect counterfactual for every imaging artifact. The reported displacements are close to the expected precision floor of column fitting under realistic membrane conditions, and no error bars, scan-direction reversals, or alternative peak detectors are reported. A moiré-periodic systematic error in centroid positions, whether from scan distortion, Wiener filtering, or residual interface contrast, would produce exactly the observed periodic vortex/antivortex maps. The DFT vortices at 18.9–28.1 degrees in idealized monolayer models support plausibility but do not independently validate the 3.0-degree experimental amplitude or the exact vortex layout.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the observation of periodic arrays of polarization vortices in twisted bilayer SrRuO3 membranes, based on picometer-scale Ru displacements extracted from planar STEM-HAADF images of the top layer. The vortex spacing matches the moiré periodicity, the displacement magnitude decreases with increasing twist angle, and unannealed bilayers and single-layer membranes show no such displacements. The authors also report a correlation between the vortex pattern and shear strain gradients, reproduce vortices in DFT for twisted but not 0-degree bilayers, and show twist-angle-dependent changes in ferromagnetism and transport that they interpret as multiferroic competition. The central claim is that twisted stacking induces dipolar vortices in a metallic ferromagnet despite free-carrier screening.","tokens_in":14763,"tokens_out":3152,"duration_ms":35672,"significance":"If the picometer-scale displacements are real structural distortions, the result extends polarization topology into metallic systems and provides a potentially important platform for flexo-Rashba and multiferroic-metal physics. The paper has notable strengths: the single-layer and unannealed-bilayer controls (Supplementary Figs. S9, S3), the depth-sectioning FFT control (Supplementary Fig. S7), the four twist angles measured, and the ab initio DFT calculations that produce vortices in twisted but not 0-degree bilayers. These controls make the observation credible, but the absolute magnitude of the displacements sits close to the expected precision floor of column-fitting analysis, and the manuscript does not quantify that floor. The flexoelectric correlation is also derived from the same STEM images as the displacement maps, so it is partly a consistency check. The central claim is therefore defensible but needs an explicit artifact-floor analysis before it can be regarded as established.","major_comments":[{"comment":"The central evidence is the 1–20 pm Ru displacement maps obtained by Gaussian fitting of Wiener-filtered STEM-HAADF images, but no error bars, replicate statistics, scan-direction reversals, or alternative peak detectors are reported. The unannealed-bilayer control (Supplementary Fig. S9) and the depth-sectioning FFT control (Supplementary Fig. S7) are useful and should be credited, but they do not bound the few-picometre systematic error that a moiré-periodic centroid bias (from residual interface contrast, scan distortion, or filtering) would produce. Please add an explicit artifact-floor estimate, for example from images of an undistorted single layer processed through the same analysis pipeline, from simulated images with known sub-picometre displacements, or from repeated acquisitions with opposite scan directions, and show that the vortex amplitude exceeds this floor.","section":"Fig. 2f-i; Methods, 'Polarization vortices and strain analysis'"},{"comment":"The flexoelectric field maps are reconstructed from shear strain gradients measured by GPA on the same STEM-HAADF images that provide the displacement maps, so the reported correlation between E_flexo and the vortex pattern is partially a consistency check rather than an independent test. Moreover, Eq. (S1) uses a unit flexoelectric coefficient, so the reconstruction is only qualitative. An independent strain measurement (for example from a different detector geometry or from simulations) would strengthen the claim. At minimum, the authors should state explicitly that this correlation does not by itself prove causation, and should discuss how the effective flexoelectric coefficient for a metal would enter the comparison.","section":"Supplementary Text I; Fig. 2j-l; Eq. (S1)"},{"comment":"The DFT calculations are performed at twist angles of 18.92°, 22.62°, and 28.07°, whereas the experiments cover 3.0°–10.4°. The calculated monotonic decrease of Ru displacement with twist angle is therefore not directly comparable to the experimental trend, and the agreement claimed in Fig. 3d and in the text is an extrapolation. The authors should either perform a calculation at a commensurate angle in or near the experimental range, or explicitly discuss the expected evolution of the mechanism from small to large angles, in particular because the moiré periodicity at small angles may introduce relaxation effects not captured by the high-angle supercells.","section":"DFT section; Fig. 3d; Supplementary Fig. S13"},{"comment":"The multiferroic competition claim is based on comparing TC, Ms, and δRu across different samples, but no error bars or replicate measurements are reported for the magnetic data, and the experimental TC values in Fig. 4f are obtained by linear interpolation. This is acceptable for a trend, but the statement that polarization and ferromagnetism 'compete' would be more convincing with repeated samples or with measurements performed on the same sample before and after a perturbation that changes the polarization. Please provide at least a statement of the sample-to-sample variability or additional data points.","section":"Fig. 4c,f; Methods, 'Electrical transport and magnetization measurements'"}],"minor_comments":[{"comment":"The caption refers to 'Figs. 3j-l' for the flexoelectric field maps, but these maps are shown in Fig. 2j-l; please correct the cross-reference.","section":"Fig. 2 caption"},{"comment":"There is a typo: 'sourcementer' should be 'sourcemeter' in the sentence describing the Keithley 2614B connection.","section":"Methods, 'Electrical transport and magnetization measurements'"},{"comment":"Reference 34 duplicates reference 21 (Junquera et al., Rev. Mod. Phys. 95, 025001 (2023)); please remove the duplicate and renumber.","section":"References"},{"comment":"The statement that 'the polarization systematically switches towards the opposite direction upon the reversal of the strain gradients' is supported only by superimposed maps; please rephrase to reflect that this is an observed correlation, not an established causal relation.","section":"Supplementary Text I"},{"comment":"The notation for the effective flexoelectric coefficient is unclear: the superscript and subscript on f are inconsistent with the text. Please define all indices in full.","section":"Eq. (S1)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is likely to be of interest to the journal's readership, but the central STEM artifact concern is load-bearing and needs to be resolved with additional experiments or explicit error budgeting. The authors should be encouraged to provide the artifact-floor analysis and error bars, and to address the DFT angle mismatch before publication. The note about a competing group is not a reason to change the technical bar."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The new thing here is real: polarization vortices in a metallic ferromagnet, SrRuO3, induced by twisted stacking. That extends polar topology into metals, where free-carrier screening should suppress ordinary electrostatic polarization, so if it holds up it is a significant step beyond the BaTiO3 and SrTiO3 twisted-bilayer work. The paper also earns credit for taking the artifact question seriously. The single-layer and unannealed-bilayer controls show no displacement pattern, the depth-sectioning FFT shows the bottom layer's spots vanish at top-layer focus, and the effect reproduces across four twist angles. DFT adds plausibility: twisted bilayers develop vortices, 0-degree bilayers do not, and the calculated twist dependence matches the observed trend. That is a solid experimental and theoretical package for a first report.\n\nThe soft spots are in proportion. The displacements are 1–20 pm, which is close to the expected precision floor of Gaussian column fitting in STEM-HAADF under realistic conditions. There are no error bars, replicate statistics, scan-direction reversals, or alternative peak detectors. The depth-sectioning control rules out Bragg-scattering interference from the bottom layer, but sub-Bragg moiré contrast from the buried interface, or scan distortion, could still bias centroid positions at the few-picometre level and produce a periodic vortex-like map. The unannealed control is suggestive, but annealing changes interface chemistry and contact quality, not only interlayer coupling, so it is not a perfect counterfactual. An independent structural probe, like X-ray or electron diffraction refinement, would settle this. The second issue is conceptual: what exactly does “polarization” mean in a metal? The paper asserts rather than defines it. The flexoelectric correlation is also partly internal, since the shear-strain gradients are extracted from the same STEM images as the displacements, making the agreement a consistency check rather than an independent test. The multiferroic competition is plausible but rests on DFT with manually off-centered Ru and idealized high-angle models; the experimental TC and Ms shifts are consistent but not conclusive.\n\nBottom line: this is a paper for a serious referee, not a desk reject. The claim is important, and the authors have done more than most to control for artifacts. What is missing is statistical rigor and independent validation of the absolute displacement scale. I would send it to review, with the request that the authors deposit raw images and fitting code, add error analysis, and address what polarization means in a screened metal.","headline":"Twisted SrRuO3 membranes showing polar vortices in a metal is a genuinely new result with decent controls, but the 1–20 pm displacement signal sits close to the imaging artifact floor and needs independent confirmation before the claim is fully credible.","tokens_in":15428,"tokens_out":1595,"would_cite":true,"duration_ms":20170,"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":"Twisted stacking creates polarization vortices in a metallic ferromagnet.","keywords":["polarization vortices","twistronics","SrRuO3","flexoelectricity","multiferroic metal","moiré","ferromagnetic metal"],"falsifier":"An independent diffraction experiment—for example synchrotron X-ray diffuse scattering or precession electron diffraction on the same 3.0° twisted bilayer—that looks for the moiré-periodic off-centering of Ru atoms would settle the claim: if no such 1–20 pm periodic displacement is found, the vortex interpretation collapses.","tokens_in":14263,"feed_emoji":"🌀","tokens_out":8094,"duration_ms":61812,"temperature":0.7,"pith_summary":"This paper reports that stacking two freestanding membranes of the metallic oxide SrRuO3 with a relative twist produces a periodic array of polarization vortices and antivortices in the top layer, with clockwise vortices at AA-stacked sites and anticlockwise vortices at AB-stacked sites. The authors argue these vortices are flexoelectric in origin: the moiré pattern creates shear strain gradients that act as pseudo-electric fields, polarizing the lattice even though free carriers should screen ordinary electrostatic forces. The Ru displacement grows as twist angle shrinks, and ferromagnetism weakens as polarization strengthens, indicating a multiferroic metal with competing orders. If correct, this extends polarization topology from insulating ferroelectrics into metals and correlated electron systems.","feed_headline":"Twisted metal membranes grow polarization vortices","feed_subtitle":"Even with free electrons that screen charges, twisted SrRuO3 develops chiral polarization that competes with its ferromagnetism.","key_machinery":"The central object is the off-center displacement of Ru relative to the four surrounding Sr columns, written $\\delta_{\\mathrm{Ru}}$, extracted from planar STEM-HAADF images by Gaussian fitting. Maps of $\\delta_{\\mathrm{Ru}}$ are superimposed on the toroidal moment $Q=(1/N)\\sum \\mathbf{r}_i \\times \\delta_i$ to expose vortex and antivortex arrays, and compared with flexoelectric field maps reconstructed from shear strain gradients ($\\varepsilon_{xy,x}$, $\\varepsilon_{xy,y}$) obtained by geometric phase analysis. Flexoelectricity—polarization induced by a strain gradient—is the mechanism invoked. The unannealed twisted bilayer and the single-layer membrane are the null controls, while the DFT supercells with twist angles 18.92°, 22.62°, and 28.07° provide the theoretical demonstration that vortex formation is intrinsic to the twisted stacking.","core_discovery":"The paper's central claim is that twisted stacking of two freestanding SrRuO3 membranes produces a moiré-periodic pattern of Ru off-centering that constitutes polarization vortices and antivortices, in a metal where free-carrier screening would normally suppress dipolar order. The Ru displacement in the top layer, measured relative to the surrounding Sr square, reaches about 20 pm at 3.0° twist and forms clockwise vortices at AA-stacked sites and anticlockwise vortices at AB-stacked sites. The vortices disappear in unannealed twisted bilayers and in single layers, and their maps correlate with flexoelectric fields reconstructed from shear strain gradients, which the authors take as evidence of a mechanical, flexoelectric origin. They further report that the bilayers remain metallic and ferromagnetic, with the ferromagnetic transition temperature and saturation magnetization decreasing as the polar displacement increases, interpreted as a polarization–magnetism competition yielding a multiferroic metal.","pith_inferences":["Editorial inference: because the displacement values come from Gaussian fitting of STEM images of a single layer, a non-imaging structural probe on the same samples would test whether the roughly 1–20 pm moiré-periodic shifts are real; the paper's controls weaken artifact explanations but do not pin down absolute magnitudes.","Editorial inference: the authors note that quantitative comparison of resistance curves across twist angles is limited by conductive-area uncertainty from microcracks, so the transport-kink evidence should be read as qualitative until area-calibrated devices are measured.","Editorial inference: if flexoelectric strain gradients are the mechanism, the same twisted-stacking recipe should produce polar vortices in other metallic perovskite membranes, and varying twist angle in situ would amount to a non-chemical knob for tuning magnetism; neither follow-up is demonstrated here."],"forward_implications":["Polarization vortices form in a metal, not just in insulators, and their periodicity tracks the moiré lattice for twist angles from 3.0° to at least 10.4°.","The amplitude of the Ru displacement decreases with increasing twist angle, so twist angle is a tuning knob for the polar order.","Below the ferromagnetic transition, polarization and magnetism coexist in the same metallic membrane, and their opposite twist-angle trends indicate competition rather than independence.","The DFT analysis attributes the magnetic weakening to Ru off-centering reducing the Ru–O–Ru bond angle and exchange interaction, consistent with the measured drop in Curie temperature and saturation magnetization.","Because the mechanism is flexoelectric and therefore symmetry-universal, similar vortex states should appear in other twisted metallic bilayers."],"supporting_citations":[{"why":"It supplied the first observation of polarization vortices in twisted BaTiO3 membranes and the flexoelectric hypothesis this paper adapts.","marker":"[8]"},{"why":"It showed twist-induced polar vortices in paraelectric SrTiO3, extending the comparison to a cubic, non-polar material.","marker":"[12]"},{"why":"It reported flexoelectric polarization in ferromagnetic metal SrRuO3, providing the metal-specific precedent that motivates expecting polar effects despite screening.","marker":"[17]"},{"why":"It established the water-soluble sacrificial layer method used to make the freestanding SRO membranes.","marker":"[5]"},{"why":"It describes the Gaussian-fitting software used to extract atomic column positions and compute Ru displacements.","marker":"[38]"},{"why":"It provides the geometric phase analysis method used to obtain the shear strain and strain-gradient maps from which flexoelectric fields are reconstructed.","marker":"[39]"},{"why":"It supplies the plane-wave DFT total-energy method used for the twisted-bilayer relaxations and electronic-structure calculations.","marker":"[40]"},{"why":"It gives the twisted bilayer supercell construction method adopted in the DFT simulations.","marker":"[45]"},{"why":"It provides the self-consistent Gaussian approximation used to convert DFT exchange parameters into estimates of the Curie temperature.","marker":"[47]"}],"fun_headline_variants":["Twisted metal films defy screening to host polarization vortices","Moiré twist creates polarization vortices in ferromagnetic metal","Polarization vortices appear in twisted metallic membranes","Twistronics induces chiral polarization in a ferromagnetic metal","Chiral polarization vortices emerge in twisted metal"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's entire vortex claim rests on the 1–20 pm Ru displacements measured from STEM-HAADF images of the top layer being real atomic shifts rather than artifacts of moiré interference, focus-depth mixing, or Gaussian peak fitting.","fun_headline_variants_meta":{"raw":{"variants":["Twisted metal films defy screening to host polarization vortices","Moiré twist creates polarization vortices in ferromagnetic metal","Polarization vortices appear in twisted metallic membranes","Twistronics induces chiral polarization in a ferromagnetic metal","Chiral polarization vortices emerge in twisted metal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000971,"raw_usage":{"total_tokens":4091,"prompt_tokens":868,"completion_tokens":3223,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":484,"completion_tokens_details":{"reasoning_tokens":3144}},"tokens_in":484,"tokens_out":3223,"duration_ms":16946,"temperature":1.0,"reasoning_tokens":3144,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:40:30.115111+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An independent diffraction experiment—for example synchrotron X-ray diffuse scattering or precession electron diffraction on the same 3.0° twisted bilayer—that looks for the moiré-periodic off-centering of Ru atoms would settle the claim: if no such 1–20 pm periodic displacement is found, the vortex interpretation collapses.","supporting_citations":[],"review_version":1}