{"id":"8f93abf6-32f8-4f68-8f85-247178e51e4b","arxiv_id":"2507.20857","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A non-iterative metasurface design method uses blob detection on random light-intensity maps to place nano-antennas that independently control scattering angle and spectrum.","lead":"This paper introduces a way to design patterned surfaces for light control using a common image-processing trick, blob detection, to place tiny antennas at positions that scatter light into chosen angles. The approach is non-iterative and lets designers tune the color and direction of scattered light independently.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Independence of angular and spectral control is not actually demonstrated: the FDTD analysis integrates over the target k-range and compares to an angle-integrated cross-section, so it cannot rule out form-factor-induced angular-spectral coupling.","rationale":"The reader's weakest assumption identified the same load-bearing premise: the decorated point pattern behaves as independent, non-interacting resonators, so angular and spectral responses can be designed separately. My stress-test sharpens this into a specific, falsifiable gap. The paper's FDTD comparisons do not test factorization at the level of angle-resolved scattering; they integrate over the target angular band and compare with an angle-integrated single-resonator cross-section. That procedure cannot detect a form factor that changes the angular distribution within the band, even if the band-integrated power tracks the isolated-resonator spectrum. The manuscript contains internal evidence of the coupling: S.3 notes that the decoration unit affects the SDF at high spatial frequencies, Fig. 3b shows the ring pattern's SDF deviating from the others, and Sec. 3.3 explicitly suspects lattice interactions. Because the central claim of simultaneous and independent control rests on this factorization, and because the validation discards the angular information needed to test it, the concern is real. However, it is a validation and precision gap rather than a demonstrated contradiction: the proposed test could confirm the factorization, in which case the claim would stand. Therefore the reader's CONDITIONAL verdict remains appropriate, and I recommend no change. The lack of public code and data and the absence of error bars further support keeping the verdict conditional rather than raising it to acceptance.","tokens_in":13192,"tokens_out":7723,"duration_ms":100268,"concrete_test":"Using the existing FDTD workflow, compute the full angle-resolved far-field I(k,ω) for one fixed point pattern decorated separately with the disk, ring, and star units of Fig. 3, without k-band integration. Also compute the isolated single-unit angle-resolved scattering form factor |F_unit(k,ω)|^2 with the same FDTD setup. Evaluate the residual R(k,ω)=I_ensemble(k,ω)/[SDF_pattern(k)|F_unit(k,ω)|^2] over k∈[10,20] µm^-1 and the resonant energy range. If R varies by more than a predefined tolerance (say 10%) with k at fixed ω, or with ω at fixed k, the angular and spectral channels are coupled and the independence claim fails. For the superposition claim, additionally simulate the two subpatterns of Fig. 5 separately, sum their angle-resolved spectra, and compare with the combined-pattern simulation to test additivity quantitatively.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the far-field scattering of the decorated pattern factorizes into a pattern-controlled structure factor and a resonator-controlled form factor, with negligible cross-talk. That condition is not established. Section 2.1 states that the factorization holds only in the first Born approximation for low index-contrast structures, yet the demonstrations use n=3.5 dielectric and Au on Si. More importantly, the quantitative support in Sec. 3.3 (Fig. 3d,e) and Sec. 3.5 (Fig. 5e,f) compares band-integrated ensemble power with the angle-integrated forward scattering cross-section of the isolated resonator. An angle-integrated comparison is blind to whether the angular profile within the designed k-band changes with the decoration unit or with wavelength. Since the single-resonator form factor is generally k-dependent—and S.3 explicitly states that the decoration unit modifies the SDF at high spatial frequencies, with the ring pattern visibly deviating in Fig. 3b—the two control channels are not shown to be independent. The paper's own remark that 'slight spectral differences are likely due to lattice interactions' points to unmodeled cross-talk. Thus the load-bearing assumption is asserted but not quantitatively verified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a non-iterative reverse-design method for correlated-disordered metasurfaces: starting from a target spectral density function (SDF), a Gaussian random field is generated in Fourier space, and Laplacian-of-Gaussians blob detection extracts point coordinates that are subsequently decorated with nano-resonators. The point pattern is intended to control angular scattering through its structure factor, while the resonator geometry controls the spectral response through its form factor. The authors fabricate Au nanopillar arrays, characterize angular scattering with Fourier microscopy at 640 nm, and find ring-shaped scattering patterns matching the input SDFs. They further use FDTD simulations for dielectric (n=3.5) resonator patterns with different shapes to show that spectral peaks track the isolated-resonator forward scattering cross-section. Finally, they demonstrate real-space superposition of two subpatterns with different decoration units as a route to simultaneous angular and spectral control, supported by Fourier microscopy for the angular part and FDTD for the spectral part.","tokens_in":13437,"tokens_out":5077,"duration_ms":58401,"significance":"If fully established, the method would be a valuable fast, non-iterative alternative to optimization-based inverse design for correlated-disordered metasurfaces with tailored angular and spectral scattering. The paper's concrete strengths are the experimental validation of angular ring control with SEM and Fourier microscopy for Au nanopillars, the demonstration of SDF matching for several input profiles in the Supplementary Information, and the real-space superposition concept that allows different subpatterns to be decorated independently. However, the central claim of independent angular and spectral control is not quantitatively supported: the spectral evidence is angle-integrated, the factorization assumption is invoked outside its stated low-contrast regime, and the experimental spectral demonstration is absent. These gaps are addressable with additional analysis, so the work is promising but needs revision.","major_comments":[{"comment":"The evidence for spectral control is based on comparing the ensemble power integrated over the entire k∈[10,20] µm⁻¹ band (Fig. 3d) with the angle-integrated forward scattering cross-section Qscat of the isolated resonator (Fig. 3e). This metric is blind to whether the angular distribution within the designed band changes with photon energy or resonator geometry, so it cannot establish that the angular and spectral channels are independent. Please add an angle-resolved comparison, for example normalized scattering intensity versus k at several photon energies for each resonator shape, or a direct comparison of the FDTD far-field pattern with the product of the designed SDF and the single-resonator angular form factor.","section":"Sec. 3.3, Figs. 3d-e"},{"comment":"The factorization of the far-field scattering into a lattice structure factor and a single-resonator form factor is introduced under the first Born approximation for low-index-contrast structures, yet the FDTD study uses n=3.5 resonators and the experiments use Au nanopillars on Si. The claim of independent control therefore requires a quantitative check of the factorization in this regime. The paper already notes \"slight spectral differences are likely due to lattice interactions\" (§3.3) and S.3 shows that the decoration unit modifies the SDF at high spatial frequencies; please quantify this cross-talk, e.g., by comparing FDTD angle- and wavelength-resolved scattering with the structure-factor × form-factor prediction, and by varying areal density or separation to estimate the strength of near-field and lattice couplings.","section":"Secs. 2.1 and 3.3"},{"comment":"The abstract claims \"individual and independent control over light scattering in angular and spectral terms,\" but the spectral control is demonstrated only in FDTD simulations of dielectric resonators; the fabricated Au nanopillar samples are characterized only at a single wavelength (640 nm) by Fourier microscopy. As written, the claim overstates the experimental support. Either add angle-resolved spectral measurements of at least one decorated sample or clearly restrict the independence claim to the simulation-level demonstration and describe the experimental validation as angular-only.","section":"Abstract and Conclusion"},{"comment":"The superposition of patterns in real space assumes \"linear, non-interacting scattering contributions,\" but the manuscript does not verify that the overlaid point patterns avoid overlapping or closely spaced decoration units. The S.3 discussion already notes that point density increases with target wavevector and limits the maximum decorator size; without a report of the minimum center-to-center spacing in the combined patterns (e.g., Fig. 5a), it remains possible that the two subpatterns interact near coincidence points. Please provide this statistical check or an analysis of its impact on the summed response.","section":"Secs. 3.4 and 3.5"}],"minor_comments":[{"comment":"The red curve is described as the ring subpattern integrated over k∈[15,20] µm⁻¹, but the ring subpattern is designed for k∈[5,10] µm⁻¹; please correct the range or color assignment.","section":"Fig. 5e caption"},{"comment":"The text states \"rinner = 75 nm, rinner = 125 nm\" for the ring; the second quantity should be router. In addition, the ring color is called red in the Fig. 3 caption and orange in the text; please make the color naming consistent.","section":"Sec. 3.3"},{"comment":"The phrase \"spana 500 nm\" should be \"span 500 nm\"; please proofread for similar grammatical typos throughout.","section":"Sec. 2.5"},{"comment":"The SDF comparisons are normalized and vertically shifted, so the figures support agreement in peak position and relative shape but not in absolute scattering intensity; please state this explicitly in the figure discussions, since the text uses phrases like \"agree very well\" without noting the normalization.","section":"Figs. 2e and 4j-l"}],"recommendation":"major_revision","confidential_remarks":"The central claim of independent angular and spectral control is not yet quantitatively supported, and the current angle-integrated analysis leaves a real gap. I nevertheless recommend major revision rather than rejection because the design method itself is simple, non-iterative, and experimentally validated for angular control; the missing angle-resolved comparison and interaction check appear feasible within the scope of the manuscript. The authors should also be asked to avoid overclaiming experimental support for spectral independence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core idea here is solid and worth knowing: instead of thresholding a Gaussian random field to get channel-type structures, the authors run Laplacian-of-Gaussians blob detection to get isolated points, then decorate each point with a resonator. That turns the SDF design philosophy into a toolbox with two separate knobs — point pattern for angular, resonator shape for spectral. The angular control is demonstrated directly: fabricated gold nanopillars show Fourier microscopy rings that match the input SDF, and the SEM Fourier transforms agree. The real-space superposition of two designs also works and produces the expected double-ring response. That is a real, experimentally grounded advance over iterative optimization for this class of problems.\n\nWhere the paper is softer is the \"independent\" control claim. The factorization into structure factor times form factor is exact only in the first Born, low-contrast limit, while the demonstrations use n=3.5 dielectrics and gold pillars. The spectral evidence in Figs. 3 and 5 compares angle-integrated scattered power with the isolated resonator's forward cross-section — that cannot detect whether the angular profile inside the designed k-band shifts with wavelength or with resonator shape. The authors themselves note \"slight spectral differences are likely due to lattice interactions,\" and the supplementary material admits the decoration unit modifies the SDF at high spatial frequencies. So independence is not proven; it is an unverified assumption. I would not call that fatal, because the design method works for angular scattering regardless, and spectral tuning clearly does follow the resonator response to first order. But the abstract's \"individual and independent control\" overstates what the data support.\n\nMinor issues: no public code or data (\"available upon request\" is weak), and the radially averaged profiles have no error bars or ensemble statistics. The writing is clear and the citations to prior SDF and hyperuniform work look appropriate; the novelty claim about blob detection is not undermined by the references.\n\nThis paper deserves a serious referee. The method is useful, the experimental validation is real, and the remaining question — how far the form-factor/structure-factor factorization holds for dense high-index or plasmonic patterns — is exactly the kind of thing referees should push on. I would recommend conditional acceptance after the authors either provide angle-resolved spectral data across the designed k-range or soften the independence claim to \"largely separable.\"","headline":"A genuinely useful non-iterative design route for correlated-disordered metasurfaces, with convincing experimental angular scattering data; the independence of spectral and angular control is plausible but not quantitatively established.","tokens_in":13986,"tokens_out":1716,"would_cite":true,"duration_ms":21426,"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":"This paper introduces a non-iterative, blob-detection-based reverse design method for correlated-disordered metasurfaces that separates angular control (point pattern) from spectral control (resonator decoration) and validates it with…","keywords":["metasurface","correlated disorder","spectral density function","blob detection","Gaussian random field","reverse design","angular scattering","spectral tuning"],"falsifier":"Measure the angular scattering of one point pattern decorated with two different resonator shapes at their respective resonant wavelengths, and also vary the pattern density while keeping the decoration fixed; if the scattering ring position or width shifts with resonator shape, or the spectral peak shifts with density, the structure-factor/form-factor product is violated.","tokens_in":13004,"feed_emoji":"💡","tokens_out":5452,"duration_ms":59847,"temperature":0.7,"pith_summary":"The paper introduces a reverse-design recipe for correlated-disordered metasurfaces that avoids iterative optimization. The idea is to start from a desired angular scattering profile, encoded as a spectral density function; generate a correlated Gaussian random field from it; extract point positions with blob detection; and then place individually chosen nano-resonators at those points. The point pattern is claimed to control the angular scattering, the resonator shape and material the spectral response, and real-space superposition of two such patterns to produce a scattering response that is the sum of the parts. If correct, this gives a fast, non-iterative inverse-design route with independent angular and spectral control, demonstrated experimentally on Au nanopillars and computationally on dielectric resonators.","feed_headline":"One-pass design steers light angle and color independently","feed_subtitle":"A Fourier-space method turns a target scattering pattern into a point layout and sets color with resonator shape alone.","key_machinery":"The spectral density function (SDF), the radially averaged power spectral density of a real-space structure, is the design input and metric: it encodes the target angular scattering. The generation pipeline is the mechanism: construct a complex Fourier space from the SDF with random phases, inverse-transform to a correlated Gaussian random field, and apply Laplacian-of-Gaussians blob detection to read out correlated point coordinates. The physical factorization that makes the design work is the first-Born separation of the far-field scattering into a structure factor (from the point lattice) times a form factor (from each decoration unit); real-space superposition then exploits linearity of non-interacting scatterers to add scattering responses.","core_discovery":"The central claim is that angular and spectral light scattering from a correlated-disordered metasurface can be designed separately and simultaneously in a single non-iterative procedure. An input spectral density function fixes the in-plane momentum content; a Gaussian random field with random phases is generated from it, and a Laplacian-of-Gaussians blob detector locates high-correlation points whose own radially averaged power spectrum closely matches the input. Those points are then decorated with nano-resonators whose geometry sets the spectral response, under the first-Born assumption that the far field is the product of the point-pattern structure factor and the single-resonator form factor. The paper validates the angular part with Fourier microscopy on fabricated Au nanopillar arrays and the spectral part with FDTD simulations of disk, ring, and star dielectric resonators, and shows that overlaying two independent decorated patterns in real space yields double-ring scattering with independently tunable ring intensities.","pith_inferences":["Editorial: a direct quantitative test of the paper's independence premise would compare the measured far-field of a decorated pattern with the product of the pattern's measured structure factor and the single-resonator forward-scattering cross-section; any systematic deviation with wavelength or density would localize where the factorization fails.","Editorial: because the method separates lattice geometry from unit-cell geometry, it should transfer to other linear wave systems—acoustic, elastic, or microwave—where a desired angular response can be prescribed by a two-point correlation function.","Editorial: the paper's own supplementary observation that the decoration unit modifies the SDF at high spatial frequencies implies the angular and spectral channels are not perfectly independent; the practical limit is set by how strongly the form factor overlaps the target spatial-frequency band."],"forward_implications":["Metasurface designs for light trapping, directional emission, or filtering can be generated in a single pass from a target scattering profile, without optimization or iteration.","The spectral response can be tuned by swapping the decoration unit while leaving the angular response fixed, enabling multi-resonant devices.","Real-space superposition of several decorated patterns produces a single-layer metasurface whose scattering is the sum of the component responses, with independent control of relative ring intensities via resonator size.","The same re-decoration step could be extended to phase or polarization control per point, since only the unit cell is changed.","Since point density increases with target spatial frequency, designs at high angles must use smaller resonators; the usable angular range is bounded by this trade-off."],"supporting_citations":[{"why":"Establishes the spectral-density-function and Gaussian-random-field framework that the blob-detection method builds on and that thresholded patterns are compared against.","marker":"[11, 12]"},{"why":"Supplies the Gaussian random field level-set construction used to generate the correlated real-space field.","marker":"[29]"},{"why":"Provides the Laplacian-of-Gaussians blob detection implementation used to read out point coordinates.","marker":"[36]"},{"why":"Shows the precedent of decorating correlated-disordered or hyperuniform point patterns with engineered scatterers, which the re-decoration step extends.","marker":"[9, 25]"},{"why":"Basis for treating each decorator as an individual resonator whose resonances set the spectral response.","marker":"[42, 43]"},{"why":"Commercial FDTD solver used for the far-field scattering and forward cross-section calculations.","marker":"[41]"},{"why":"Cited by the paper when it suspects slight spectral differences arise from lattice interactions, marking the limitation of the independent-resonator assumption.","marker":"[43, 44]"}],"fun_headline_variants":["Blob detection designs metasurfaces for angle and color control","Fourier-space recipe sets scattering angle and color independently","Single-step metasurface design: independent angular and spectral scattering","Blob finding carves metasurfaces that aim light and color independently","Fourier-space blob design gives independent angle and color control"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that each decorated point scatters as an independent resonator, so the ensemble far field factorizes into a structure factor from the point pattern and a form factor from the unit cell; if near-field coupling or collective lattice resonances become significant, independent angular and spectral control degrades.","fun_headline_variants_meta":{"raw":{"variants":["Blob detection designs metasurfaces for angle and color control","Fourier-space recipe sets scattering angle and color independently","Single-step metasurface design: independent angular and spectral scattering","Blob finding carves metasurfaces that aim light and color independently","Fourier-space blob design gives independent angle and color control"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000556,"raw_usage":{"total_tokens":2639,"prompt_tokens":930,"completion_tokens":1709,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":1623}},"tokens_in":546,"tokens_out":1709,"duration_ms":15145,"temperature":1.0,"reasoning_tokens":1623,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T13:11:14.818081+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the angular scattering of one point pattern decorated with two different resonator shapes at their respective resonant wavelengths, and also vary the pattern density while keeping the decoration fixed; if the scattering ring position or width shifts with resonator shape, or the spectral peak shifts with density, the structure-factor/form-factor product is violated.","supporting_citations":[{"cited_title":"Teubner, ``Level Surfaces of Gaussian Random Fields and Microemulsions ,'' Europhysics Letters (EPL) 14 , 403--408 (1991)","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian random field level-set construction used to generate the correlated real-space field."},{"cited_title":"Van Der Walt, J","cited_arxiv_id":null,"evidence_quote":"Provides the Laplacian-of-Gaussians blob detection implementation used to read out point coordinates."},{"cited_title":", FDTD Solutions - 3D Electromagnetic Simulator","cited_arxiv_id":null,"evidence_quote":"Commercial FDTD solver used for the far-field scattering and forward cross-section calculations."}],"review_version":1}