{"id":"5b7d8585-0f61-4281-ac91-c40c35b6a47b","arxiv_id":"2505.03060","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Atom-resolved electron ptychography maps thermal vibration amplitudes in twisted bilayer WSe2, showing spatially localized anisotropic motion attributed to moiré phasons.","lead":"Using electron ptychography with sub-15 picometer resolution, researchers mapped the thermal jitter of individual atoms in twisted bilayer WSe2, revealing ultra-soft moiré phason vibrations that localize at solitons and AA regions. The work gives experimental access to a previously hidden branch of moiré phonon physics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. S4's single 15.5 pm global baseline cannot support the reported AA–AB thermal contrast; a 1 pm calibration or region-dependent static-broadening error exceeds the 0.8 pm claim.","rationale":"The reader's weakest assumption—that a single phonon-free simulation provides a valid global baseline for Gaussian deconvolution—is indeed the load-bearing point. My analysis quantifies why: the raw width contrast underlying the reported AA-AB thermal difference is only ~0.3 pm, far below the stated 1-2 pm systematic uncertainty and below plausible region-dependent static broadening. The ptychographic control using MD vs isotropic frozen phonons is a thoughtful internal check that the reconstruction can transfer a nonuniform thermal signal, but it does not remove the baseline problem, and the MD itself predicts a much smaller thermal contrast than reported. I therefore do not see a fatal flaw; the qualitative observation of nonuniform, anisotropic atomic profiles near solitons and AA regions is plausible and worth conditional acceptance pending a region-resolved calibration of the phonon-free baseline and uncertainty propagation. The requested test directly checks whether the global-baseline assumption is satisfied, so the reader's CONDITIONAL verdict should stand.","tokens_in":19150,"tokens_out":8620,"duration_ms":94504,"concrete_test":"Using the existing phonon-free ptychography simulation behind Fig. S7, bin the fitted single-W-atom Gaussian widths and ellipticities by stacking region (AA, soliton, AB) with the same watershed/2D-Gaussian pipeline; if the regional spread in sigma_phonon-free or ellipticity is comparable to the experimental raw-width contrast (~0.3 pm) or to the observed soliton ellipticity, then Eq. S4's global-baseline subtraction is invalid and the thermal amplitude maps in Figs. 3-4 cannot establish AA/soliton enhancement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that phasons dominate thermal vibrations rests on extracting per-atom thermal amplitudes from fitted Gaussian widths via Eq. S4: sigma_thermal = sqrt(sigma_fit^2 - sigma_phonon-free^2), with sigma_phonon-free = 15.5 pm taken from one phonon-free ptychography simulation. The signal is extremely small: the reported regional contrast sigma_AA - sigma_AB = 0.8 pm in thermal amplitude corresponds to only ~0.3 pm difference in the raw fitted widths (since d sigma_fit/d sigma_thermal ~ sigma_thermal/sigma_fit ~ 0.35), against a baseline of 15.5 pm and a 9.79 pm pixel size. The authors' own systematic uncertainty in the deconvolution is 1-2 pm, so the AA-AB difference is within the calibration error. Moreover, Eq. S4 subtracts a global scalar; it does not remove region-dependent static broadening from the reconstructed moiré structure (different W-W column offsets in AA/soliton/AB, local strain, EDF artifacts). A phonon-free atom in a soliton or AA region may already be wider or more elliptical than one in AB, and this static contribution is exactly what the single baseline would misattribute to thermal motion. The MD-vs-isotropic frozen-phonon ptychography control tests whether the reconstruction pipeline can transfer a prescribed nonuniform thermal map, but it does not calibrate per-region static baselines; and the MD-predicted thermal variation (sigma_AB=5.5, sigma_AA=5.6 pm) is an order of magnitude smaller than the experimental contrast (5.8 vs 6.8 pm), so the quantitative 'dominate' claim is not supported by the simulations as presented.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports atom-by-atom imaging of moiré phasons in twisted bilayer WSe2 using multislice electron ptychography (MEP). The authors acquire 4D-STEM data on 1.7°, 2.45°, and 6.0° twisted bilayers, reconstruct high-resolution phase images with extended depth of field, and fit 2D anisotropic Gaussians to individual W (and Se) columns to obtain per-atom widths, ellipticities, and orientations. They interpret these as time-averaged thermal vibration ellipsoids and report increased, anisotropic vibrations localized at solitons and near AA-stacked regions, with quantitative amplitude maps obtained by deconvolving a simulation-derived phonon-free width of σ_phonon-free = 15.5 pm (Eq. S4). Lattice-dynamics calculations and molecular dynamics simulations are used to identify low-energy phason modes and to compare predicted versus measured vibration amplitudes as a function of twist angle. The central claims are that phasons dominate thermal vibrations in low-angle twisted bilayers and that ptychography provides a new route to spatially resolved atomic vibration measurements.","tokens_in":19429,"tokens_out":5804,"duration_ms":60348,"significance":"If the quantitative extraction is reliable, the paper would be significant: it would demonstrate direct real-space observation of moiré phasons, a mode class that has so far been inferred only indirectly, and would extend electron ptychography from static structure determination to atomically resolved thermal motion. The raw observations in Fig. 2—anisotropic, soliton-localized atom shapes with preferred orientation along the soliton—are visually compelling and largely independent of the deconvolution model. The paper also contains useful controls, including ptychography simulations with MD phonons versus isotropic frozen phonons (Fig. S8), and a comparison of experimental maps with MD-derived trajectories. However, the quantitative claims about thermal amplitudes and their regional contrast rest on a single global deconvolution baseline and on error bars that are comparable to or larger than the reported AA–AB contrast. The present level of evidence supports the qualitative phenomenology but not the stated quantitative conclusion that phasons dominate the measured thermal vibrations.","major_comments":[{"comment":"The quantitative thermal-amplitude maps rely on Eq. S4 with a single global σ_phonon-free = 15.5 pm taken from one phonon-free ptychography simulation. The authors report systematic errors of ±1–2 pm in σ, while the measured AA–AB thermal-amplitude contrast is 0.8 pm (6.8 vs 6.0 pm; this corresponds to roughly 0.3 pm in raw fitted widths because dσ_fit/dσ_thermal ≈ σ_thermal/σ_fit ≈ 0.35). A global scalar baseline cannot remove region-dependent static broadening that differs between AA, soliton, and AB sites—different interlayer column overlaps, local strain, and EDF artifacts would already produce different apparent widths in a phonon-free reconstructed structure. No control establishes that a phonon-free atom is equally wide in all regions, so the reported AA–AB contrast is within the stated calibration uncertainty. The claim that AA regions vibrate more than AB regions is not established at the stated precision.","section":"Methods (Deconvolution), Eq. S4; Fig. 3D"},{"comment":"The validation against MD is not quantitatively consistent with the experimental contrast. The MD values are σ_AB = 5.5 pm, σ_soliton = 5.5 pm, and σ_AA = 5.6 pm—an AA–AB difference of about 0.1 pm—whereas the experiment yields 6.0, 5.8, and 6.8 pm, an order of magnitude larger. Describing this as “remarkable agreement” is not supported by these numbers. Either the deconvolution is amplifying the regional contrast or the MD model is missing the dominant mechanism; in either case, the comparison cannot by itself validate the claim that the experimental amplitudes are dominated by phasons. A mode-resolved decomposition of the MD trajectories (phason modes versus ordinary phonons) is needed to support the attribution to phasons.","section":"Fig. 3D/E and the section “Thermal vibrations across the moiré superlattice”"},{"comment":"The AA-region statistics are obtained from regions that, as the text acknowledges, “approach AA stacking but do not reach the AA centers.” Because the central signature is enhanced vibration at AA sites, the histograms and mean values labeled “AA” should either be restricted to well-characterized AA centers with a clearly defined selection criterion, or the claim should be reworded as a near-AA effect. As presented, the assignment of atoms to the AA category is ambiguous and weakens the central contrast claim.","section":"Fig. 3A, Fig. 3D, and the section “Thermal vibrations across the moiré superlattice”"},{"comment":"For the twist-angle comparison in Fig. 4, σ_phonon-free is scaled linearly by the ratio of the furthest identifiable FFT peak positions (Eq. S5). This assumes that the non-thermal baseline is exactly proportional to information transfer and that no other acquisition- or reconstruction-dependent blur contributes. Neither assumption is tested. Because the twist-angle trend in Fig. 4 is a central part of the phason-evolution narrative, the scaling must be validated—for example, by applying the same procedure to simulated data with controlled resolution degradation—or the twist-angle conclusions should be restricted to the qualitative trend visible in the raw maps.","section":"Methods (Deconvolution), Eq. S5; Fig. 4"}],"minor_comments":[{"comment":"The equation is printed as “Eqn. SEqn. S4”; it should read “Eq. S4.”","section":"Materials and Methods, Deconvolution"},{"comment":"The main text refers to “experiment (Fig. 4D–E) and MD (Fig. 4F–G)”, while the caption defines panels D–F as experimental and G–I as MD. The panel references should be corrected for consistency.","section":"Fig. 4 caption and main text"},{"comment":"The quantity σ_ave used in Fig. 4 is not defined. Specify whether it is the geometric mean of the long- and short-axis Gaussian widths before deconvolution or of the deconvolved thermal amplitudes, and give the exact formula.","section":"Fig. 4 and Materials and Methods"},{"comment":"Fig. 2 presents raw long-axis Gaussian widths without deconvolution, while Fig. 3 and Fig. 4 use deconvolved thermal amplitudes. State clearly in each caption which quantity is plotted, so that the qualitative anisotropy evidence and the quantitative amplitude analysis are not conflated.","section":"Fig. 2 and Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"I see no integrity concern: the raw imaging results appear to be honestly reported, and the deconvolution issue is a calibration problem rather than a fabrication issue. The paper is potentially important, but the quantitative central claim—that phasons dominate thermal vibrations and that AA regions vibrate more than AB regions—needs either much stronger calibration controls or a substantial re-scoping to qualitative observations. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is the first time anyone has claimed to image moiré phasons atom by atom, and the raw data are genuinely pretty. The ptychographic images show atoms near solitons visibly elongated along the soliton direction, wider atoms near AA regions, and similar behavior in conventional MEP projections without the extended-depth-of-field processing. The control experiment using MD-phonon ptychography versus isotropic frozen-phonon ptychography is a good idea: it shows that the reconstruction pipeline does not invent the spatial pattern by itself. The lattice dynamics eigenvectors and the twist-angle trend are also well integrated and appropriate. The citation pattern is reasonable, and the paper does not oversell its novelty in the intro.\n\nThe soft spots are real but concentrated in the quantitative layer. The headline regional contrast, sigma_AA - sigma_AB = 0.8 pm after deconvolution, is smaller than the authors' own stated systematic uncertainty of 1-2 pm. In the raw fitted widths the contrast is roughly 0.3 pm on a 15.5 pm baseline with a 9.79 pm pixel size. Equation S4 subtracts one global phonon-free scalar, so any region-dependent static broadening—soliton strain, local column offsets, EDF artifacts—gets silently absorbed into the thermal amplitude. That is exactly the kind of error that could produce an AA vs AB difference of this size.\n\nMore telling, the MD simulations predict an AA-AB short-axis contrast of about 0.1 pm, almost an order of magnitude smaller than the measured 0.8 pm. The mismatch is in the direction that would implicate static disorder or an overly soft interlayer potential, and the paper does not address it. The abstract's claim that phasons dominate the thermal vibrations is therefore stronger than the evidence supports. The qualitative statement—that the observations are consistent with phason contributions—would be fair, but the quantitative 'dominate' needs qualification.\n\nI would not block the paper on these grounds. The central observation, localized anisotropic soliton vibrations with enhanced AA amplitudes, is probably correct and is new. But the quantitative amplitude maps should be presented with the deconvolution uncertainty propagated through to the figures, the static-disorder alternative should be addressed explicitly, and the MD-vs-experiment amplitude mismatch should be discussed rather than glossed over. Code and data should also be released; right now all data is only promised upon publication.\n\nSend this to a serious referee. It needs revision rather than desk rejection, and it could become a very useful reference for ptychographic measurement of localized vibrations.","headline":"First direct imaging claim for moiré phasons is plausible and deserves referee time, but the quantitative thermal-amplitude maps lean on a single simulation baseline and a 0.8 pm contrast that sits inside the stated systematic error.","tokens_in":20078,"tokens_out":1560,"would_cite":true,"duration_ms":19995,"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":"By resolving every atom in twisted bilayer WSe2 with electron ptychography, this paper claims direct images of moiré phasons, seen through their localized, anisotropic thermal vibrations.","keywords":["electron ptychography","moiré phasons","twisted bilayer WSe2","thermal vibrations","atomic displacement parameters","soliton network","molecular dynamics","lattice dynamics"],"falsifier":"Repeat the ptychography measurement on the same 1.7° and 2.45° twisted bilayers at low temperature, where the predicted phason motion nearly freezes: if the atom widths do not approach the phonon-free value of about 15.5 pm, or if the AA and soliton excess widths persist, static disorder or reconstruction artifacts are producing the contrast rather than phason vibrations.","tokens_in":18912,"feed_emoji":"🔬","tokens_out":11358,"duration_ms":106947,"temperature":0.7,"pith_summary":"Electron ptychography is a scanning imaging method that reconstructs the sample's electrostatic potential from overlapping diffraction patterns. The paper tries to establish that moiré phasons—the predicted ultrasoft shear modes of the soliton network in twisted bilayers—can be seen directly as atom-by-atom variations in thermal vibration. Using ptychography at sub-15-pm resolution on twisted bilayer WSe2 at 1.7°, 2.45°, and 6.0° twist, the authors fit every W atom with a 2D Gaussian and deconvolve the non-thermal blur with a phonon-free simulation, converting atom widths into per-atom vibrational amplitudes. The maps show enhanced, anisotropic vibration at soliton domain walls and AA-stacked regions, matching molecular dynamics and lattice dynamics, and the paper concludes that phasons dominate room-temperature thermal motion in low-angle moirés. This matters because phasons have been implicated in superlubricity, charge transport, and low-temperature thermal properties, but had not been directly visualized before.","feed_headline":"Atom-by-atom images expose moiré phasons in twisted WSe2","feed_subtitle":"Electron ptychography maps each atom's thermal vibration, showing ultrasoft moiré modes dominate at low twist angles.","key_machinery":"The load-bearing object is the per-atom 2D anisotropic Gaussian fit to the ptychographic phase image: each atomic column is characterized by long-axis width $\\sigma_x$, short-axis width $\\sigma_y$, and orientation $\\theta$, and the thermal amplitude is isolated by $\\sigma_{\\mathrm{thermal}} = \\sqrt{\\sigma_{\\mathrm{fit}}^2 - \\sigma_{\\mathrm{phonon-free}}^2}$ with a phonon-free simulation width $\\sigma_{\\mathrm{phonon-free}} = 15.5$ pm, rescaled by the Fourier-transform peak position for the twist-angle series. Around this sit molecular-dynamics trajectories converted to per-atom probability density functions, harmonic lattice-dynamics eigenvectors that identify the two near-degenerate phason modes, and multislice ptychography simulations with phonon-free, isotropic-frozen-phonon, and MD-phonon inputs; only the MD-phonon simulation reproduces the experimental spatial pattern. The extended-depth-of-field processing sharpens the projection enough to resolve atom pairs separated by 14.7–19.6 pm.","core_discovery":"The central claim is that the size and shape of each atom in a ptychographic image is a time-averaged thermal ellipsoid, so atom-by-atom Gaussian fitting yields a real-space map of vibrational amplitude. In 1.7°-twisted WSe2, the W atoms inside soliton domain walls are elongated parallel to the wall, while AB-region atoms are rounder and randomly oriented. In the 2.45° sample, short-axis amplitudes are largest near AA sites, with experimental means of $\\sigma_{\\mathrm{AA}} = 6.8 \\pm 0.2$ pm, $\\sigma_{\\mathrm{AB}} = 6.0 \\pm 0.2$ pm, and $\\sigma_{\\mathrm{soliton}} = 5.8 \\pm 0.3$ pm, close to molecular-dynamics values of about 5.5–5.6 pm. The lowest-energy harmonic eigenmodes of the relaxed moiré are two nearly degenerate phason modes localized at solitons and AA sites, and ptychography simulations using MD phonon configurations reproduce the measured contrast while isotropic frozen-phonon simulations do not. The paper concludes that phasons dominate thermal vibrations in reconstructed low-angle moirés and that this ptychographic route can image spatially non-uniform vibrations in general.","pith_inferences":["A natural testable extension is to repeat the measurement at cryogenic temperature: the phason contribution should nearly freeze out, and if the AA/soliton excess width persists, static disorder is at least partly responsible.","The same atom-by-atom vibration mapping could be applied to moiré heterobilayers and to grain boundaries or interfaces, where soft, anisotropic modes are predicted but have not been spatially resolved.","Because ptychography reads vibration from atom shape rather than frequency, it could in principle detect phasons in samples too small or disordered for phonon spectroscopy, including locally heterogeneous moiré lattices.","The finding implies that thermal motion in low-angle moirés is spatially inhomogeneous, so effective Debye-Waller factors and local thermal conductivities may need to become position-dependent quantities in moiré device models."],"forward_implications":["Moiré phasons move from prediction to observable: their characteristic spatial signatures—localized, anisotropic, soliton-aligned motion and enhanced amplitudes at AA sites—are visible in real space at room temperature.","Electron ptychography becomes a general atomically resolved probe of thermal vibrations, able to measure per-atom anisotropic displacement parameters in any material where sub-15-pm resolution can be reached.","In low-angle twisted WSe2, thermal motion is dominated by ultrasoft phason modes rather than ordinary acoustic phonons, so models of transport, thermal conductivity, and specific heat in these systems must include them.","The measured short-axis amplitudes set quantitative benchmarks that future simulations of moiré phonons should reproduce: roughly 6 pm in AB and soliton regions and 6.8 pm near AA sites for the 2.45° sample.","The weakening of regional vibration contrast at 6.0° twist is consistent with the predicted crossover from a reconstructed, phason-dominated moiré to near-rigid bilayer behavior."],"supporting_citations":[{"why":"It predicts the existence and character of moiré phonons, the family of modes this paper aims to observe.","marker":"[3]"},{"why":"It provides the phason picture as soft, localized shear modes and electron-phason coupling, giving the predicted signature tested here.","marker":"[4]"},{"why":"It predicts ultrasoft phasons in twisted transition-metal dichalcogenide bilayers and the twist-angle regimes used in the experiment.","marker":"[5]"},{"why":"It establishes that electron ptychography reaches the resolution limit set by lattice vibrations, and supplies the deconvolution logic the paper extends to per-atom maps.","marker":"[44]"},{"why":"It supplies the calculated room-temperature in-plane vibrational amplitudes of monolayer WSe2, used as simulation input and numerical benchmark.","marker":"[57]"},{"why":"It supplies the interlayer potential that controls the soliton network and drives the molecular-dynamics and lattice-dynamics behavior being compared to experiment.","marker":"[69]"},{"why":"It is used for finite-displacement harmonic phonon calculations, yielding the phason eigenvectors and dispersion.","marker":"[70]"},{"why":"It is used for the molecular-dynamics trajectories that generate the per-atom thermal probability densities and RMSD maps.","marker":"[71]"},{"why":"It provides the multislice ptychography simulation that produces the phonon-free reference width and the control reconstructions.","marker":"[74]"}],"fun_headline_variants":["Electron ptychography maps phason vibrations atom by atom","Atom-resolved phasons: imaging thermal motion in twisted WSe2","Ptychography reveals ultrasoft phason modes in twisted bilayers","Seeing phasons: atom-by-atom vibrational maps in moiré WSe2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole measurement rests on the assumption that the width of each atom in the image is the microscope's static blur and the atoms' thermal motion added together as independent Gaussian spreads, so any extra width from static disorder, reconstruction defects, or imperfect deconvolution is misread as larger vibration.","fun_headline_variants_meta":{"raw":{"variants":["Electron ptychography maps phason vibrations atom by atom","Atom-resolved phasons: imaging thermal motion in twisted WSe2","Ptychography reveals ultrasoft phason modes in twisted bilayers","Seeing phasons: atom-by-atom vibrational maps in moiré WSe2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000156,"raw_usage":{"total_tokens":1238,"prompt_tokens":985,"completion_tokens":253,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":174}},"tokens_in":601,"tokens_out":253,"duration_ms":2896,"temperature":1.0,"reasoning_tokens":174,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:00:08.393120+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the ptychography measurement on the same 1.7° and 2.45° twisted bilayers at low temperature, where the predicted phason motion nearly freezes: if the atom widths do not approach the phonon-free value of about 15.5 pm, or if the AA and soliton excess widths persist, static disorder or reconstruction artifacts are producing the contrast rather than phason vibrations.","supporting_citations":[{"cited_title":"Koshino, Y.-W","cited_arxiv_id":null,"evidence_quote":"It predicts the existence and character of moiré phonons, the family of modes this paper aims to observe."},{"cited_title":"Ochoa, Moiré-pattern fluctuations and electron-phason coupling in twisted bilayer graphene","cited_arxiv_id":null,"evidence_quote":"It provides the phason picture as soft, localized shear modes and electron-phason coupling, giving the predicted signature tested here."},{"cited_title":"Maity, M","cited_arxiv_id":null,"evidence_quote":"It predicts ultrasoft phasons in twisted transition-metal dichalcogenide bilayers and the twist-angle regimes used in the experiment."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It establishes that electron ptychography reaches the resolution limit set by lattice vibrations, and supplies the deconvolution logic the paper extends to per-atom maps."},{"cited_title":"Hovden, H","cited_arxiv_id":null,"evidence_quote":"It supplies the calculated room-temperature in-plane vibrational amplitudes of monolayer WSe2, used as simulation input and numerical benchmark."},{"cited_title":"Thibault, A","cited_arxiv_id":null,"evidence_quote":"It supplies the interlayer potential that controls the soliton network and drives the molecular-dynamics and lattice-dynamics behavior being compared to experiment."},{"cited_title":"Odstrčil, P","cited_arxiv_id":null,"evidence_quote":"It is used for finite-displacement harmonic phonon calculations, yielding the phason eigenvectors and dispersion."},{"cited_title":"Aguet, D","cited_arxiv_id":null,"evidence_quote":"It is used for the molecular-dynamics trajectories that generate the per-atom thermal probability densities and RMSD maps."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the multislice ptychography simulation that produces the phonon-free reference width and the control reconstructions."}],"review_version":1}