{"id":"a3069af1-c0bd-442f-885f-ba97de9ee01c","arxiv_id":"2608.13435","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A hydrodynamic volume integral equation is combined with characteristic mode analysis to compute excitation-independent resonances of plasmonic nanostructures.","lead":"Scientists have built a new simulation tool that reveals the hidden resonant frequencies of tiny metal particles, including effects from electron motion that simpler models miss. It works like an antenna design method, showing all natural modes a particle supports rather than only the ones a particular light source excites.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The eigensolver inverts the reactance matrix X, which becomes nearly singular at exactly the extinction resonances the method claims to identify, making the reported MS curves numerically unreliable.","rationale":"The reader's weakest_assumption focused on the hard-wall boundary condition, which is a modeling approximation the authors explicitly acknowledge and which does not undermine the internal consistency of the method. The concern I raise is more directly load-bearing for the central numerical claim: the eigenvalue solver inverts a matrix that is singular at the exact resonances the method seeks. The paper provides no evidence that the eigenproblem is solved accurately near these points, and the loose inner-solve tolerance aggravates the risk. The reader's rationale did mention 'no treatment of the ill-conditioned X^{-1} eigenproblem near resonances' as one of several reasons for the conditional verdict, but it was not identified as the primary concern. My analysis agrees with the conditional verdict, but the numerical conditioning should be the primary gate for acceptance, alongside the requested reproducibility artifacts. If the numerical test I propose were to pass (i.e., the generalized eigensolver reproduces the reported MS values), the concern would be resolved and the paper's central claim would be substantially supported. If it fails, the MS curves and mode reconstructions would need to be recomputed with a regularized eigensolver, and the quantitative conclusions could change. Thus the verdict remains CONDITIONAL pending this specific numerical check.","tokens_in":38958,"tokens_out":6434,"duration_ms":66242,"concrete_test":"At the frequency points nearest the MS peaks in Fig. 3 (e.g., ω = 0.65ω_p and ω = 1.13ω_p), assemble the N×N matrix X = Im(Z) from Eq. (11) and compute its smallest singular value and condition number. If σ_min(X) < 1e-6 * ||X|| (or similarly small relative to the TFQMR tolerance), the inversion is numerically singular. Then recompute the characteristic modes by solving the generalized eigenproblem X I = λ R I directly with a dense generalized eigensolver (LAPACK ggev) or a shift-invert iterative solver using a small shift to regularize, and compare the top-10 MS values and mode current patterns with those reported. Also check whether the eigenvalues from Eq. (33) have significant imaginary parts; if they do, the real-mode assumption is violated.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that solving the GEE in Eq. (15), X I = λ R I, yields excitation-independent modes whose MS curves (Eq. 21) reveal intrinsic resonances. The numerical solution, Section 2.5, transforms this into the standard eigenproblem X^{-1} R I = (1/λ) I (Eq. 33) and applies IRAM with TFQMR inner solves. However, the modes of interest have |λ| ≈ 0, meaning X I ≈ 0; at those frequencies X is numerically singular. The paper does not shift, regularize, or precondition this inversion, and the TFQMR tolerance in Eq. (34) is a loose 1e-4. Near a resonance, X is ill-conditioned, so the inner solves become inaccurate or fail silently, and the eigenvectors returned by IRAM may mix genuine physical modes with numerical artifacts. The paper reports MS peaks at such frequencies (e.g., ω = 0.65ω_p and ω = 1.13ω_p in Fig. 3) without demonstrating that the eigen-solution is accurate there. This is a load-bearing weakness because the MS curves and the reconstructed currents (Figs. 4, 7, 8) are the primary evidence for the method's validity. A simple consistency check against the extinction spectra is insufficient, since both the excitation response and the eigenproblem share the same impedance matrix and could exhibit peaks at similar frequencies even if the eigenvectors are corrupted. The absence of any convergence study or condition-number report leaves the central numerical claim unsupported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives an excitation-independent modal analysis for nonlocal plasmonic nanostructures. Starting from the coupled hydrodynamic equation and volume integral equation, the authors reduce the system to a single Hydrodynamic Volume Integral Equation (HDVIE) for the induced hydrodynamic current, discretize it with full SWG basis functions (excluding half-SWG functions to enforce the hard-wall boundary condition), and cast the resulting matrix system into a characteristic-mode generalized eigenvalue problem. The modal significance (MS) curves are then used to identify intrinsic extinction resonances. The method is applied to a nanosphere, a nanorod, and a nanodimer, with the sphere case checked against Mie-series extinction spectra and the rod/dimer cases compared against extinction spectra computed from the same HDVIE solver under plane-wave and dipole excitations.","tokens_in":39244,"tokens_out":6231,"duration_ms":74543,"significance":"If the numerical results are trustworthy, this is a useful contribution: it provides a modal, excitation-independent framework for analyzing nonlocal plasmonic response, including longitudinal resonances absent in local Drude models. The derivation is internally consistent, the sphere validation against Mie theory is a genuine external check, and the modal reconstruction of the RCS and of the hydrodynamic currents demonstrates the intended use of the modes. The authors also state the hard-wall boundary-condition limitation and explicitly flag the subnanometer dimer regime as outside the model's quantitative validity. The main weakness is that the numerical eigensolver used for the central MS curves is not demonstrated to be reliable at the frequencies where the method claims to find resonances, and the non-sphere validations are not independent of the same discretized system being solved.","major_comments":[{"comment":"This issue is load-bearing because the MS curves and the reconstructed currents in Figs. 3, 4, 7, and 8 are the primary evidence for the method's central claim.","section":"Section 2.5, Eqs. (33) and (34)"},{"comment":"The absence of such a study leaves the numerical reliability claim unsupported, particularly for the nonlocal resonances that are the main novelty of the paper.","section":"Section 3, all three examples"},{"comment":"The circularity is not total, since the modal expansion and RCS reconstruction are nontrivial, but the absence of any external reference for the non-sphere cases weakens the validation section.","section":"Sections 3.2 and 3.3, Figs. 9-13"}],"minor_comments":[{"comment":"The MS curves for the 10 tracked modes are plotted without a legend or explicit mode labeling, which makes it difficult to assess the mode-tracking quality and to identify which curve corresponds to the current distributions in Figures 4, 7, 8, 11, and 14. Please add labels or otherwise distinguish the modes.","section":"Figures 3, 10, and 13"},{"comment":"The phrase 'coincide closely' is used repeatedly when comparing MS peaks with extinction peaks, but no quantitative frequency differences are given. Reporting the relative frequency error at each resonance would make the validation more precise.","section":"Sections 3.1-3.3"},{"comment":"The claim that R is positive definite is important for the reality of the eigenvalues and modes, but no justification or numerical check is provided. A brief explanation or a numerical verification of positive definiteness of R for the three examples would remove a potential concern.","section":"Section 2.3, after Eq. (15)"},{"comment":"The statement that excluding half-SWG basis functions enforces the hard-wall boundary condition in Eq. (6) should be made more explicit: the reader must infer that all basis functions have zero normal component on boundary faces because the half functions attached to those faces are removed. This is correct, but it merits a sentence.","section":"Section 2.2, Eq. (9)"}],"recommendation":"major_revision","confidential_remarks":"The core derivation and the central idea are sound, and the requested fixes are feasible within the manuscript's scope: adding eigen-residuals, condition-number or shift-invert results, a mesh-convergence study, and an independent non-sphere check. The conference-paper predecessor [33] is properly cited, and I see no novelty-disclosure concern. The main risk is that the near-singular X inversion at resonance may corrupt the very quantities the paper reports; this can be resolved with additional numerical evidence rather than a rewrite."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a solid extension of the authors' earlier HDVIE-CMA conference paper. The reduction of the coupled HDE/VIE to a single hydrodynamic volume integral equation is clean, and the power-balance interpretation of the characteristic eigenvalues as reactive-to-extinction ratios is physically meaningful. The sphere benchmark against Mie series, for the extinction spectra, is a good sanity check and shows the nonlocal longitudinal resonance above ω_p. The mode reconstruction of RCS from a small number of modes is a useful demonstration of the intended design utility.\n\nThe main soft spot is the numerical eigensolver. The stress-test concern is valid: the modes of interest have |λ| ≈ 0, so the reactance matrix X becomes numerically singular exactly at the resonances the method claims to identify. The paper transforms the GEE into X⁻¹ R I = λ̃ I, applies IRAM with TFQMR inner solves, and stops at a loose relative residual of 1e-4. No shift, regularization, preconditioning, condition-number estimate, or convergence study is reported. That is load-bearing, because the MS curves and reconstructed currents are the primary evidence that the eigenmodes are physical. The fact that the MS peaks line up with extinction peaks is not sufficient, since both share the same impedance matrix; the eigenvectors themselves could be corrupted even if the eigenvalues happen to be right.\n\nOther soft spots are less severe but still material: no mesh-convergence study beyond a passing mention, no code or mesh data, and the nanodimer example with a 0.2 nm gap is explicitly outside the hard-wall model's expected accuracy (the authors concede this, which is honest). Novelty is limited relative to the conference paper [33], which already introduced the framework. The reference list is appropriate, and self-citation is not a problem here.\n\nOverall, the formulation is plausible and the power-balance derivaton is internally consistent. But the numerical claims are not yet reproducible or demonstrably robust. If I were the editor, I would send it to peer review—it deserves referee time—but I would expect major revisions addressing the conditioning of the eigenproblem, convergence against mesh refinement, and ideally a benchmark for the eigenvalues themselves (e.g., against a known analytic case for a cylinder or layered sphere with nonlocality). As it stands, my own verdict would be conditional, not accept.\n\nFor a colleague deciding whether to build on this: use the HDVIE derivation and the mode-expansion framework as a starting point, but reimplement the eigensolver with care and add your own validation before trusting the MS curves.","headline":"A competent but incremental extension of the authors' own conference work; the formulation is sound, but the numerical eigensolver's treatment of the nearly singular reactance matrix undermines the central resonance claims.","tokens_in":39765,"tokens_out":1532,"would_cite":false,"duration_ms":19833,"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 tries to establish that the intrinsic resonances of plasmonic nanostructures can be found without exciting them, by solving a generalized eigenvalue problem built from a hydrodynamic volume integral equation.","keywords":["characteristic mode analysis","hydrodynamic model","volume integral equation","plasmonic nanostructures","modal significance","nonlocal response","nanoantennas","generalized eigenvalue problem"],"falsifier":"Replace the hard-wall boundary condition by a soft or spill-out boundary condition in the same HDVIE discretization and recompute the MS curves for the 0.2 nm-gap dimer and the nanosphere; if the predicted gap resonance or the longitudinal resonance near $\\omega = 1.13\\,\\omega_p$ shifts by more than the linewidth or disappears, the hard-wall CMA modes are not the structure's true intrinsic modes in that subnanometer regime.","tokens_in":38785,"feed_emoji":"🔬","tokens_out":6569,"duration_ms":67864,"temperature":0.7,"pith_summary":"This paper claims that the intrinsic resonances of a plasmonic nanostructure can be extracted without ever choosing an excitation, by carrying out characteristic mode analysis on a hydrodynamic volume integral equation. Existing solvers compute only the response to a particular source, so modes that a given source does not excite stay hidden. The authors reduce the coupled hydrodynamic and volume integral equations, valid for simple metals, to one equation for the induced hydrodynamic current, discretize it, and turn it into a generalized eigenvalue problem. The resulting modal significance curves reveal all extinction resonances of the structure, including the nonlocal longitudinal resonance above the plasma frequency that local Drude models cannot produce. If correct, the method gives nanoantenna designers a source-independent mode basis and explains why some resonances appear in extinction spectra while others stay dark under a chosen illumination.","feed_headline":"Intrinsic resonances of plasmonic nanostructures, no source needed","feed_subtitle":"Hydrodynamic integral equation finds every intrinsic resonance, including nonlocal longitudinal modes local Drude models miss.","key_machinery":"The load-bearing object is the hydrodynamic volume integral equation (HDVIE), Eq. (7): the coupled hydrodynamic equation of the free-electron fluid and the volume integral equation for the scattered field are reduced, for simple metals, to a single equation in the induced hydrodynamic current density $\\mathbf{J}_H$. Discretization uses full Schaubert–Wilton–Glisson (SWG) basis functions on tetrahedra with half-basis functions removed at the boundary; that removal enforces the hard-wall boundary condition $\\hat{\\mathbf{n}} \\cdot \\mathbf{J}_H = 0$, which makes the power-balance derivation and the matrix symmetry work. The characteristic mode machinery is the generalized eigenvalue equation $\\bar{X} \\bar{I}_k = \\lambda_k \\bar{R} \\bar{I}_k$, where $\\bar{R}$ and $\\bar{X}$ are the real and imaginary parts of the discretized HDVIE impedance matrix, so the eigenvalue is the reactive-to-extinction power ratio of each mode and the modal significance $\\sigma_k = 1/|1 + j\\lambda_k|$ peaks at resonance. Mode tracking across frequency connects the eigenpairs into continuous MS curves.","core_discovery":"On the paper's own terms, the central discovery is that characteristic modes—currents and eigenvalues depending only on geometry and material—exist for hydrodynamic (nonlocal) metallic scatterers, and they can be computed from the hydrodynamic volume integral equation (HDVIE). The coupled hydrodynamic equation and volume integral equation are combined into one equation in the induced hydrodynamic current; after Galerkin discretization with full Schaubert–Wilton–Glisson basis functions, with half-basis functions removed to enforce the hard-wall boundary condition, the impedance matrix is split into real and imaginary parts, and the generalized eigenvalue equation $\\bar{X} \\bar{I}_k = \\lambda_k \\bar{R} \\bar{I}_k$ is solved. The eigenvalue equals the ratio of reactive to extinction power of the mode, and modal significance $\\sigma_k = 1/|1 + j\\lambda_k|$ gives a bounded resonance curve. For a 1 nm-radius nanosphere, a nanorod, and a nanodimer with a 0.2 nm gap, the MS curves reproduce every extinction peak seen under plane-wave and dipole excitation and predict additional resonances—one at $\\omega = 1.13\\,\\omega_p$ in the nanosphere above the plasma frequency, identified as a longitudinal resonance supported only by nonlocal response—that no local model can produce. The paper further shows that a resonance can be absent from extinction spectra simply because the source couples weakly to it, while a small number of strongly coupled resonant modes can reconstruct the radar cross section.","pith_inferences":["An implication the authors leave implicit: the same generalized eigenvalue construction with a different weighting matrix could define radiation or absorption characteristic modes for nonlocal scatterers, giving a more complete modal decomposition of lossy plasmonic antennas than extinction modes alone.","A testable extension would be to run the HDVIE-CMA on dimers with gaps between 0.2 nm and 1 nm and compare the predicted gap-mode resonance shift against quantum-corrected models; this would quantify where the hard-wall boundary condition starts to fail.","Because the paper shows that the longitudinal current above the plasma frequency needs far more modes to reconstruct than the transverse current, modal truncation error is mode-type- and frequency-dependent, so an error estimator for mode-count selection would be a natural follow-up.","The method could be used to identify structures that couple efficiently to the usually dark longitudinal resonance, turning a nonlocal response feature into a usable channel for emission or sensing."],"forward_implications":["The induced current on any such nanostructure can be expanded in characteristic modes independent of the source, so quantities like extinction and radar cross section can be reconstructed from a small number of dominant modes.","MS curves identify all intrinsic extinction resonances of the structure, including resonances that a plane wave or dipole does not excite, and explain dark resonances through weak modal excitation coefficients.","For simple metals, the framework includes nonlocal longitudinal modes above the plasma frequency that local Drude models omit, giving a fuller modal picture near the plasma frequency.","Because a resonance appears in the extinction spectrum only when a source couples strongly to it, the choice of excitation is an active design variable for targeting a specific mode in nanoantenna design.","The framework provides a source-free computational tool for analyzing and designing plasmonic nanoantennas whose dimensions require the nonlocal hydrodynamic model."],"supporting_citations":[{"why":"Supplies the coupled volume integral and hydrodynamic equation solver whose formulation is reduced to the single HDVIE.","marker":"[14]"},{"why":"Provides the tetrahedral SWG basis functions used to expand the hydrodynamic current and enforce the boundary condition.","marker":"[32]"},{"why":"Gives the characteristic mode eigenvalue formulation that the HDVIE discretization is cast into.","marker":"[23]"},{"why":"Supplies the modal significance definition and the mode-tracking procedure used to produce continuous MS curves.","marker":"[21]"},{"why":"Supports the choice of weighting matrix for characterizing extinction resonances in lossy structures.","marker":"[29]"},{"why":"Reports the nonlocal longitudinal resonances that motivate identifying the additional resonance above the plasma frequency.","marker":"[35]"},{"why":"Supplies the power computation formulas used to define extinction power and to connect the impedance matrix to power balance.","marker":"[38]"},{"why":"Provides the hydrodynamic model for simple metals, including the pressure term and the power contributions used in the derivation.","marker":"[31]"}],"fun_headline_variants":["Characteristic modes map all plasmonic resonances, no source needed","Nonlocal theory exposes hidden resonances in metal nanoparticles","Eigenmode solver finds every resonance, even unexcited ones","Hydrodynamic model reveals longitudinal modes local theory misses"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The hard-wall boundary condition—that the hydrodynamic current's normal component vanishes at the metal surface, so electrons cannot spill out of the metal—must remain accurate for the structure being studied; the paper itself concedes that for a 0.2 nm gap, spill-out and tunneling may change the modes quantitatively.","fun_headline_variants_meta":{"raw":{"variants":["Characteristic modes map all plasmonic resonances, no source needed","Nonlocal theory exposes hidden resonances in metal nanoparticles","Eigenmode solver finds every resonance, even unexcited ones","Hydrodynamic model reveals longitudinal modes local theory misses"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000429,"raw_usage":{"total_tokens":2262,"prompt_tokens":1080,"completion_tokens":1182,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":696,"completion_tokens_details":{"reasoning_tokens":1114}},"tokens_in":696,"tokens_out":1182,"duration_ms":10337,"temperature":1.0,"reasoning_tokens":1114,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:05:18.936719+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Replace the hard-wall boundary condition by a soft or spill-out boundary condition in the same HDVIE discretization and recompute the MS curves for the 0.2 nm-gap dimer and the nanosphere; if the predicted gap resonance or the longitudinal resonance near $\\omega = 1.13\\,\\omega_p$ shifts by more than the linewidth or disappears, the hard-wall CMA modes are not the structure's true intrinsic modes in that subnanometer regime.","supporting_citations":[{"cited_title":"Uulu, Rui Chen, Lei Chen, Peng Li, and Hakan Bagci","cited_arxiv_id":null,"evidence_quote":"Supplies the coupled volume integral and hydrodynamic equation solver whose formulation is reduced to the single HDVIE."},{"cited_title":"Schaubert, Donald R","cited_arxiv_id":null,"evidence_quote":"Provides the tetrahedral SWG basis functions used to expand the hydrodynamic current and enforce the boundary condition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the characteristic mode eigenvalue formulation that the HDVIE discretization is cast into."},{"cited_title":"John Wiley & Sons, Hoboken, NJ, USA, 2015","cited_arxiv_id":null,"evidence_quote":"Supplies the modal significance definition and the mode-tracking procedure used to produce continuous MS curves."},{"cited_title":"Orthogonality properties of characteristic modes for lossy structures.IEEE Trans","cited_arxiv_id":null,"evidence_quote":"Supports the choice of weighting matrix for characterizing extinction resonances in lossy structures."},{"cited_title":"Asger Mortensen","cited_arxiv_id":null,"evidence_quote":"Reports the nonlocal longitudinal resonances that motivate identifying the additional resonance above the plasma frequency."},{"cited_title":"Polimeridis, Maxwell T","cited_arxiv_id":null,"evidence_quote":"Supplies the power computation formulas used to define extinction power and to connect the impedance matrix to power balance."},{"cited_title":"Gerhardts.Metal Optics Near the Plasma Frequency","cited_arxiv_id":null,"evidence_quote":"Provides the hydrodynamic model for simple metals, including the pressure term and the power contributions used in the derivation."}],"review_version":1}