{"id":"ef19d5dc-26db-43a7-8ba0-d6b958f508ba","arxiv_id":"2607.09466","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"Atomic-resolution VEELS maps of volume plasmons, plexcitons and interband transitions in CaNb2O6 reveal contrast reversal versus HAADF and direct [NbO6] contributions under on- and off-axis collection.","lead":"Researchers achieved atomic-resolution imaging of low-energy electronic excitations (plasmons, plexcitons, interband transitions) in CaNb2O6 crystals using advanced STEM-EELS. This shows how to map bonding and electronic structure at individual atomic columns despite long-standing delocalization limits.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper's strongest claim is an experimental demonstration that Cs-STEM-EELS with hybrid-pixel detection can produce energy- and atom-resolved maps of low-loss features (plexcitons ~7.3 eV, VPs ~6/15 eV, and the 3.2–3.5 eV interband) in a complex oxide, with contrast that systematically tracks collection geometry (on-axis HABF-like reversal vs weak-beam Z-contrast). That claim is supported by multi-modal data (aloof spectra + KKA dielectric function, position-dependent SI, BF/HAADF simulations, DFT-PDOS, and the off-axis Kikuchi-band control). The reader's identified weakest assumption correctly flags the absence of full inelastic multislice of the SI cubes, which would more rigorously separate elastic preservation from residual inelastic delocalization or ZLP tails. However, the paper already performs the key experimental isolations (collection-angle series, IBF correction, weak-beam geometry) that make the elastic-preservation interpretation the most economical reading of the data. Because those controls are present and consistent, the missing simulations do not undermine the central experimental result. Data unavailability and specialized instrumentation remain practical limitations for independent reproduction, but they do not constitute an internal logical or evidential failure. Verdict therefore stays ACCEPT; no adjustment is warranted.","tokens_in":18401,"tokens_out":683,"duration_ms":8343,"concrete_test":"Re-process one on-axis SI cube (e.g., the dataset underlying Fig. 3 or 4b) by dividing the 15 ± 1 eV VP map by the simultaneously acquired IBF (1–60 eV) image pixel-by-pixel and compare the residual contrast amplitude at Nb vs interstitial sites to the raw VP map; if residual atomic contrast remains >20 % of the raw amplitude after this elastic correction, the elastic-preservation interpretation needs revision, otherwise the paper's claim is reinforced.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (elastic-contrast preservation as the dominant origin of atomic contrast in on-axis ZLP/VEELS, without full inelastic multislice of the SI datasets) is a real methodological gap, but it is not load-bearing against the central claim. The paper already supplies independent experimental controls that isolate the same conclusion: (i) systematic variation of collection angle β relative to α (Figs. 4a–c) that progressively recovers HABF-like contrast, (ii) IBF elastic correction that leaves residual delocalized VP intensity (Fig. 5), and (iii) weak-beam/off-axis geometry that reverses contrast back to Z-contrast while still yielding usable VEELS spectra (Fig. 6). These multi-geometry results, plus the aloof-to-bulk spectral series and DFT-PDOS assignment of the 3.2–3.5 eV feature to [NbO6], make the atomic-resolution VEELS demonstration robust even without quantitative inelastic multislice. The missing simulations would strengthen the mechanistic interpretation but are not required for the experimental claim to hold.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports atomic-resolution valence EELS (VEELS) imaging of single-crystal CaNb2O6 using Cs-corrected STEM with a hybrid-pixel direct electron detector. Above the ~3.8 eV bandgap, volume plasmons (~6 eV, ~15 eV) and a ~7.3 eV feature assigned as a plexciton (mixed plasmon–exciton) are identified via aloof-beam spectra, Kramers–Krönig dielectric functions, and DFT PDOS. On-axis ZLP and VEELS maps reverse contrast relative to HAADF (HABF-like), while weak-beam (off-axis) collection recovers Z-contrast; systematic variation of collection angle β relative to convergence α and IBF elastic correction are used to address elastic-contrast preservation. Interband maps at 3.2–3.5 eV are linked to [NbO6] octahedra, with a possible Čerenkov contribution at interstitial sites. The central claim is that Cs-STEM-EELS can deliver energy- and atom-resolved low-loss contrast for structure, bonding, and electronic properties of complex oxides.","tokens_in":18687,"tokens_out":1121,"duration_ms":14861,"significance":"Atomic-resolution VEELS has remained difficult because of inelastic delocalization and ZLP tails; prior STEM results on graphene were largely delocalized, while EFTEM and limited STEM work on Si and oxide heterostructures showed mixed outcomes. This work supplies a multi-geometry experimental demonstration (on-axis vs weak-beam, β vs α series, IBF correction) on a technologically relevant mixed-cation oxide, with independent spectral assignment from dielectric analysis and DFT. The hybrid-pixel detector and CEFID spectrometer enable usable S/N at atomic columns. If the contrast mechanisms hold, the approach is a practical route to site-specific low-loss mapping of bonding and defects in complex crystals, complementary to core-loss and phonon mapping. Strengths include thorough experimental controls, multislice HAADF/BF/CBED simulations, and transparent reporting of intensity trends and acquisition parameters.","major_comments":[{"comment":"The ~7.3 eV feature is assigned as a plexciton (and surface exciton polariton) on the basis of ε2 > ε1 > 0, surface-like energy shifts in aloof spectra, and literature on plasmon–exciton hybrids. For bulk CaNb2O6 this assignment is nonstandard; the manuscript should either strengthen it with a quantitative comparison (e.g., oscillator strength, dispersion, or comparison to optical data) or soften the language to “mixed interband/plasmon character” so that the central imaging claim does not rest on a contested label.","section":null},{"comment":"Elastic-contrast preservation is invoked as the dominant origin of atomic contrast in on-axis ZLP/VEELS (Figs. 3–5 and discussion in §III). The experimental controls (β series, IBF correction, weak-beam reversal) are strong, but the paper lacks quantitative inelastic multislice simulations of the full SI datasets that would separate residual delocalization, ZLP tails, and elastic interference. Adding such simulations (or a clear statement of why they are beyond scope) would make the mechanistic interpretation load-bearing rather than qualitative.","section":null}],"minor_comments":[{"comment":"Energy-integration windows for SI maps (e.g., 7 ± 1 eV, 15 ± 1 eV, 3.25 ± 0.25 eV) should be stated once in Methods and justified relative to peak widths and energy resolution (0.7 eV).","section":null},{"comment":"Fig. 4: experimental vs simulated BF images differ in O/interstitial contrast at β = 20 mrad; a short note on residual aberrations or TDS treatment would help readers assess the comparison.","section":null},{"comment":"The optical bandgap linear-fit procedure is cited to prior work; a one-sentence description of the intercept method and uncertainty (already given as ±0.05 eV) would make the manuscript self-contained.","section":null},{"comment":"Typographical consistency: “plexcitons” vs “plexciton excitations”; “weak-beam (off-axis)” is used repeatedly—define once and use a single term thereafter.","section":null},{"comment":"Table 1 (refinement) is referenced but the numerical content is not fully reproduced in the provided text; ensure lattice parameters, Wyckoff sites, and DW factors are complete and match the HAADF simulations.","section":null},{"comment":"Data-availability statement is restrictive; if raw SI cubes or reduced spectra can be shared on request, a repository DOI would strengthen reproducibility.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The experimental design is solid and the multi-geometry controls largely address the skeptic’s concern about missing inelastic multislice. The plexciton label is the main novelty risk; if the journal prefers conservative spectral language, a minor revision to tone that assignment down would be sufficient. Fit for a materials/condensed-matter microscopy audience is good. No circularity or citation-pattern issues of concern."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a careful experimental paper that actually delivers atomic-column VEELS maps on CaNb2O6 under both on-axis and weak-beam geometries. That is the new result. Prior atomic VEELS/plasmon work exists for Si, graphene (mostly EFTEM), and a few oxide heterostructures; what they add is a systematic collection-angle and on/off-axis study on a technologically relevant complex oxide, plus site-specific intensity at the 3.2–3.5 eV interband feature tied to [NbO6].\n\nThey do the work properly. Aloof spectra, Kramers–Kronig dielectric functions, DFT PDOS, multislice HAADF/BF/CBED simulations, β varied relative to α, IBF elastic correction, and weak-beam Kikuchi-band collection are all there. Spectral assignments (VPs near 6 and 15 eV, the ~7.3 eV feature as SEP/plexciton, CR at interstitial sites) rest on standard criteria and are cross-checked. The contrast systematics are the strongest part: on-axis ZLP/VEELS reverse relative to HAADF (HABF-like), larger β cleans the interference, IBF leaves residual delocalized VP intensity, and off-axis recovers Z-contrast while still giving usable VEELS. That multi-geometry package makes the experimental claim robust.\n\nThe soft spot the reader flagged is real but not load-bearing. They lean on elastic-contrast preservation as the origin of the atomic contrast without quantitative inelastic multislice of the full SI datasets. That would tighten the mechanism. It is not required for the demonstration itself, because the experimental controls already isolate the same conclusion. Data are not public (cost/feasibility note), and the instrument is specialized, so independent reproduction will be limited. The “plexciton” label for the bulk 7.3 eV feature is a bit of an extension of the surface-exciton language; minor. Free parameters (bandgap intercept, integration windows) are ordinary for the field.\n\nThis is for electron microscopists and oxide electronic-structure people who care about low-loss mapping beyond core-loss elemental maps. It deserves a serious referee. I would accept it for peer review and would cite the geometry systematics and the [NbO6] interband maps if I were working in the area.","headline":"Solid multi-geometry demonstration that atomic-resolution VEELS maps are achievable on a complex oxide, with clear on-axis contrast reversal and off-axis Z-contrast recovery; the elastic-contrast interpretation is supported by controls even without full inelastic multislice.","tokens_in":19296,"tokens_out":592,"would_cite":true,"duration_ms":6921,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Atomic-resolution valence EELS maps plasmons and plexcitons in CaNb2O6, with contrast that flips or follows Z-contrast depending on collection geometry.","keywords":["atomic-resolution VEELS","plexciton","volume plasmon","CaNb2O6","Cs-STEM-EELS","elastic contrast preservation","weak-beam EELS","hybrid-pixel detector"],"falsifier":"A full multislice inelastic simulation of the same spectrum-imaging data set that predicts the observed on-axis contrast reversal and its disappearance under weak-beam conditions; if the simulated inelastic maps lack atomic contrast while experiment shows it, the elastic-preservation account fails.","tokens_in":19348,"feed_emoji":"🔬","tokens_out":667,"duration_ms":7341,"temperature":0.7,"pith_summary":"Atomic-resolution imaging of valence electronic excitations has long been blocked by the delocalized nature of inelastic scattering at low energy loss. This paper shows that modern Cs-corrected STEM paired with a hybrid-pixel direct detector can still produce energy- and atom-resolved maps of volume plasmons, plexcitons, and near-gap interband transitions in single-crystal CaNb2O6. On-axis collection yields zero-loss and valence images whose contrast is reversed relative to HAADF (atomic columns dark), which the authors attribute to preserved elastic contrast; increasing the collection angle or switching to a weak-beam (off-axis) geometry recovers ordinary Z-contrast. The same data set also shows that the 3.2–3.5 eV interband intensity is localized on the [NbO6] octahedra. The practical claim is that low-loss STEM-EELS can now be used to visualize bonding and electronic structure at individual atomic columns, interstitial sites, and defects in complex oxides.","feed_headline":"Atomic maps of plasmons flip contrast with collection geometry","feed_subtitle":"Cs-STEM valence EELS images plexcitons and [NbO6] interband states in calcium niobate","key_machinery":"On-axis versus weak-beam (off-axis) collection geometry, combined with systematic variation of the EELS collection semi-angle relative to the probe convergence angle, which controls whether elastic contrast is preserved or suppressed in the low-loss images.","core_discovery":"In CaNb2O6, Cs-STEM-EELS with hybrid-pixel detection produces atom-resolved maps of volume plasmons (~6 eV and ~15 eV) and plexcitons (~7.3 eV). On-axis images reverse HAADF contrast; weak-beam images restore Z-contrast. The 3.2–3.5 eV interband maps are dominated by the [NbO6] octahedra.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Atom-resolved plasmon maps reverse HAADF contrast in CaNb2O6","On-axis VEELS flips Z-contrast for volume plasmons and plexcitons","Cs-STEM-EELS maps [NbO6] interband states atom by atom","Weak-beam setup restores Z-contrast in atomic VP images","Hybrid-pixel EELS resolves plexcitons at ~7.3 eV atomically"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The atomic contrast seen in on-axis zero-loss and valence maps is assumed to come mainly from preserved elastic scattering rather than residual inelastic delocalization or zero-loss-peak tails.","fun_headline_variants_meta":{"raw":{"variants":["Atom-resolved plasmon maps reverse HAADF contrast in CaNb2O6","On-axis VEELS flips Z-contrast for volume plasmons and plexcitons","Cs-STEM-EELS maps [NbO6] interband states atom by atom","Weak-beam setup restores Z-contrast in atomic VP images","Hybrid-pixel EELS resolves plexcitons at ~7.3 eV atomically"]},"model":"grok-4.5","effort":"low","cost_usd":0.007848,"raw_usage":{"total_tokens":1998,"prompt_tokens":939,"num_sources_used":0,"completion_tokens":107,"cost_in_usd_ticks":78480000,"prompt_tokens_details":{"text_tokens":939,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":952,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":939,"tokens_out":107,"duration_ms":9837,"temperature":1.0,"reasoning_tokens":952,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T02:48:45.497425+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A full multislice inelastic simulation of the same spectrum-imaging data set that predicts the observed on-axis contrast reversal and its disappearance under weak-beam conditions; if the simulated inelastic maps lack atomic contrast while experiment shows it, the elastic-preservation account fails.","supporting_citations":[],"review_version":1}