{"id":"7402a45c-9a41-4b86-9578-994d7995ec76","arxiv_id":"2607.23950","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Near-field infrared microscopy is quantitatively described by discrete probe-cavity eigenmodes ('nano-gap polaritons') whose poles explain measured spectra and enable fast extraction of optical constants.","lead":"This paper models the sharp metal tip of a near-field infrared microscope together with the sample under it as a single coupled optical cavity made of discrete 'probe-cavity eigenmodes'. If the model is right, these microscopes can be turned from imaging tools into fast, quantitative spectrometers that extract a material's optical constants from one scan.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Probe idealization—axisymmetric PEC hyperboloid with only apex radius calibrated—is the load-bearing assumption; no independent full-wave check against the real pyramidal PtSi probe is provided, so systematic geometry error would bias extracted optical constants and the predicted Rabi splitting.","rationale":"The paper's formal machinery is coherent, and the honest anchoring to previous work (Eq. 18 formally equivalent to Jiang et al.) plus the genuine non-fitted predictions of approach curves and Au/Si contrast provide real support. Those checks are valuable but all pass through the same idealized probe model: a PEC hyperboloid with a single calibrated radius. The paper itself flags the axisymmetric approximation and the uncertainty in half-angle θ, and it defers simultaneous calibration of θ and other shape parameters to future work. Appendix G(iv) addresses Ohmic losses only for 'reasonably metallic' probes, not for PtSi at THz, and the experimental methods appendix is marked TODO, which hampers replication. The most load-bearing risk is not the formalism but the sufficiency of the geometric idealization: if the real probe differs in ways not absorbable by a single radius, the extracted optical constants for SiC, STO, and kapton, and particularly the predicted STO Rabi splitting, are systematically biased. The reader's conditional verdict already captures this risk, so our stress-test does not change the verdict. The proposed full-wave comparison with the real probe geometry is a direct, decisive check: if it passes within tolerance, the concern is retired; if it fails, the quantitative claim needs re-evaluation.","tokens_in":53323,"tokens_out":6575,"duration_ms":104301,"concrete_test":"Run an independent full-wave boundary-element/finite-element simulation of the actual PtSi-FM probe (geometry from SEM; finite PtSi conductivity) and compute the same observables—the self-normalized ξ on Au/Si and the s3 approach curve over SiC—without any fitted parameter except a. If matching the measured ξ requires a hyperboloid-equivalent a that differs by more than ~20% from the axisymmetric fit, or if the predicted SiC resonance position shifts by more than one experimental linewidth, the single-parameter idealization is falsified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—Eq. 18 and its use for metrology—inherits the assumption that a real PtSi-FM probe is adequately represented by an axisymmetric PEC hyperboloid whose only adjustable parameter is the apex radius a. Sec. IV A states 'we consider exclusively PEC probes with axisymmetric geometry' with half-angle θ fixed at 10°/25° 'resembling' commercial probes; Sec. VI A admits the axisymmetric approximation is 'admittedly unlike the directional illumination and pyramidal shape of most realistic probes' and defers simultaneous calibration of θ to future work. The calibration (Fig. 7b) fits only a via ξ=s3/s2 on Au/Si; nothing checks that the fitted a in the idealized model reproduces the actual probe's near-field spectrum. Since the same forward model is used for the inversion in Sec. VI D, any geometry/conductivity error (pyramidal facet currents, finite PtSi conductivity, wear not captured by radius growth) is absorbed into the extracted ε(ω). The approach-curve (Fig. 5g-h) and Au/Si contrast (Fig. 7c) comparisons are genuine non-fitted predictions, but they share the same idealized geometry and may be insensitive to shape errors. Appendix G(iv) argues Ohmic losses are negligible for 'reasonably metallic' probes, but does not validate PtSi at THz nor compare against a full-wave simulation of the actual tip. Thus the quantitative accuracy claim is conditional on an unverified single-parameter geometric sufficiency.","agreement_with_reader":"agree"},"referee_report":null,"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — Here's the short version: this is a serious paper that does real work, but its quantitative metrology claim is only as solid as the single-parameter probe model it calibrates, and the manuscript is unfinished in places that matter.\n\nThe genuinely new pieces are the ω-dependent eigenmodes for realistic finite-length probes (antenna resonances in Fig. 3), the Chebyshev-demodulation formula Eq. 26, the self-normalized calibration scheme, and the experimental confrontations. The SiC approach-curve maxima (Fig. 5e-h) and the Au/Si contrast growth over two octaves (Fig. 7c) are true non-fitted predictions, and they match. That is real evidence, and it gives the central claim independent grounding. The extraction of SiC, STO, and kapton optical constants consistent with literature, at under a minute per inversion, is a meaningful advance if the forward model is trustworthy.\n\nThe soft spots are in proportion. The formal eigenmode expansion is explicitly acknowledged as equivalent to Jiang et al.'s Eq. (3); the novelty is the electrodynamic extension and the experimental validation, not the mathematical kernel. The bigger issue is geometric: the entire pipeline assumes an axisymmetric PEC hyperboloid whose only adjustable parameter is apex radius a, calibrated via ξ on Au/Si and then used without an independent full-wave check against the real pyramidal PtSi probe. The paper says as much in Sec. IV A and VI A. If pyramidal facet currents, finite PtSi conductivity, or wear not captured by radius growth matter, every extracted optical constant and the predicted Rabi splitting in Sec. VII inherit the bias. The ad hoc (1+0.3i·r_p)^2 factor for far-normalized spectra is a fit, not a prediction, and the STO/kapton 'predictions' are inversions anchored only to literature constants — genuine but weaker support. Finally, the manuscript is not finished: Appendix A is literally 'TODO', Appendices H and I are 'To be completed', and no code URL or commit hash is given despite the promise of public software. Reviewers need those details to assess reproducibility.\n\nNone of this kills the central argument, but it does mean the paper is not yet at the level of precision metrology it claims. The reader's conditional verdict is about right. My recommendation: send it to a serious referee, but with the expectation of major revision — complete the appendices, provide code and data, and add at least one independent full-wave validation of the idealized probe model against a realistic commercial probe geometry. That last step would convert a promising framework into one that can be trusted for quantitative inversion.","headline":"Genuinely useful eigenmode framework with real non-fitted experimental matches; the load-bearing single-parameter probe idealization is not independently validated and the manuscript is unfinished in ways that matter.","tokens_in":54318,"tokens_out":3103,"would_cite":true,"duration_ms":35215,"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":"A sparse sum over probe-cavity eigenmodes quantitatively predicts near-field infrared nanoscopy, retrieves optical constants, and forecasts strong coupling.","keywords":["near-field optical microscopy","s-SNOM","nano-gap polariton","probe-cavity eigenmodes","optical constant retrieval","strong coupling","infrared nanoscopy","EigenProbe"],"falsifier":"Take a commercial PtSi s-SNOM probe, characterize its tip shape by electron microscopy, and simulate (or measure) the probe-scattered field with the actual pyramidal geometry and finite conductivity; if the EigenProbe expression with a single fitted apex radius systematically fails to reproduce the Au/Si contrast and SiC approach curves in a way that correlates with asymmetric or non-PEC features, the idealized hyperboloid model is falsified.","tokens_in":53029,"feed_emoji":"🔬","tokens_out":8350,"duration_ms":391756,"temperature":0.7,"pith_summary":"The paper aims to show that scattering-type near-field optical microscopy (s-SNOM) can be made quantitative by expanding the probe-sample interaction in a sparse basis of 'probe-cavity eigenmodes' — current distributions on the probe that are simultaneously orthogonal under the probe's electrodynamic self-impedance and its quasi-electrostatic mirror interaction with the sample. The central result is a closed-form expression, Eq. 18, in which the composite probe-sample scattered field is a sum over these modes with poles at surface reflectivity equal to mode eigenvalues; those poles are the 'nano-gap polaritons.' The authors argue that, computed for an axisymmetric perfect-conductor hyperboloid probe calibrated by a single apex-radius parameter, this expansion predicts measured approach curves on SiC, gold/silicon contrast spectra, and — after a Kramers-Kronig-constrained Lorentz-oscillator fit — optical constants of SiC, SrTiO3, and kapton consistent with literature. They further use the same eigenmode framework to predict a strong-coupling avoided crossing between a THz antenna-resonant probe and the soft phonon polariton of thin SrTiO3. If correct, this turns near-field nanoscopy into a fast quantitative metrology of local optical constants and a design tool for nano-gap cavity quantum optics. The manuscript as supplied marks Appendix A (experimental methods) as 'TODO' and Appendices H and I as 'To be completed,' so some promised measurement and derivation details are not yet present.","feed_headline":"Probe eigenmodes turn near-field nanoscopy into quantitative metrology","feed_subtitle":"A 20-mode expansion predicts s-SNOM spectra on SiC, gold, and silicon, and retrieves optical constants of crystals and polymers.","key_machinery":"The central object is the set of probe-cavity eigenmodes |j_ν): surface current distributions on the probe that solve (1/iω)(Ê_P − ρ_ν Ê_S^QS)|j_ν)=0 (Eq. 16), where Ê_P is the probe's electrodynamic self-impedance and Ê_S^QS is the quasi-electrostatic mirror interaction with a planar surface. The eigenvalues ρ_ν, ordered by increasing |ρ|, quantify field confinement (via Reρ) and radiative loss (via −Imρ). By construction these modes are orthogonal under both operators, so they nearly diagonalize the probe-sample scattering matrix and reduce the composite response to the sparse rational expression Eq. 18, with poles at β(ω)=ρ_ν. This machinery supplies forward predictions (approach curves,","core_discovery":"The paper claims that the composite probe-sample response function of an s-SNOM experiment can be written, formally exactly and without a perturbative assumption, as G_PS = −Σ_ν |E_ν)(j_ν|/(ρ_ν − β(ω)) (Eq. 18), where |E_ν) are fields generated by probe-cavity eigencurrents |j_ν), ρ_ν are dimensionless 'eigenreflectivities' of a generalized eigenproblem combining the probe's electrodynamic self-impedance with a quasi-electrostatic mirror operator, and β(ω) is the local surface reflectivity. When β approaches an eigenvalue ρ_ν, the response is dominated by a self-sustaining nano-gap polariton — a collective excitation of the probe-sample cavity. The paper reports that retaining about 20 eigen","pith_inferences":["The same sparse eigenmode expansion should extend to other active nanoscopies (e.g., THz-STM and photo-induced force microscopy) by replacing the far-field scattering observable with tunneling current or mechanical force; the paper gestures at this but does not demonstrate it.","A testable extension is to compare the eigenmode prediction against an independent full-wave simulation of a realistic pyramidal PtSi probe with finite conductivity, isolating whether residual discrepancies come from the axisymmetric-PEC idealization or from the single-radius calibration.","The explicit role of −Imρ_ν (radiative loss) and Imβ (surface absorption) suggests the formalism could be adapted to predict photothermal-expansion nanoscopy lineshapes; the paper notes the ambiguity but leaves a concrete prediction for future work.","Because the forward model is fast, one could train a machine-learned surrogate on the eigenmode expansion for real-time optical-constant mapping, an avenue the paper mentions only in passing."],"forward_implications":["If Eq. 18 holds, near-field experiments previously interpreted qualitatively become quantitatively predictable from a small (~20) eigenmode set, including in the non-perturbative strong-coupling regime.","The formalism yields an inversion scheme that retrieves local optical constants of polar crystals (SiC, SrTiO3) and polymers (kapton) from demodulated scattering spectra in under a minute on down-sampled data.","The same eigenvalues predict gap-dependent approach curves on phonon-resonant surfaces, identifying nano-gap polaritons as real, gap-tunable excitations that dominate the measured signal.","The strong-coupling analysis predicts an experimentally observable avoided crossing and Rabi splitting near 0.4 THz for a THz antenna probe over thin SrTiO3 at gap d/a≈0.2, with g/g_crit≈2.","EigenProbe encoding (Eq. 27) reuses one modal-reflectivity calculation to compute scattering at all probe-sample gaps, making multi-harmonic demodulation and nano-imaging computationally cheap."],"fun_headline_variants":["EigenProbe: exact non-perturbative model for near-field nanoscopy","Nano-gap polaritons explain and quantify near-field nanoscopy","Probe-cavity eigenmodes make near-field nanoscopy quantitative","20 eigenmodes turn probe-sample coupling into metrology","Quantitative near-field via nano-gap polaritons and eigenmodes"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"All quantitative predictions inherit the assumption that a real pyramidal, finitely conducting PtSi probe is adequately represented by an axisymmetric, perfectly conducting hyperboloid whose only adjustable parameter is the apex radius a; if real probes deviate in ways not absorbable by changing a — facet currents, finite THz conductivity, or wear beyond radius growth — the extracted optical constants and predicted splittings are systematically biased.","fun_headline_variants_meta":{"raw":{"variants":["EigenProbe: exact non-perturbative model for near-field nanoscopy","Nano-gap polaritons explain and quantify near-field nanoscopy","Probe-cavity eigenmodes make near-field nanoscopy quantitative","20 eigenmodes turn probe-sample coupling into metrology","Quantitative near-field via nano-gap polaritons and eigenmodes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00018,"raw_usage":{"total_tokens":1201,"prompt_tokens":867,"completion_tokens":334,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":611,"completion_tokens_details":{"reasoning_tokens":237}},"tokens_in":611,"tokens_out":334,"duration_ms":4009,"temperature":1.0,"reasoning_tokens":237,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T03:29:40.774544+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a commercial PtSi s-SNOM probe, characterize its tip shape by electron microscopy, and simulate (or measure) the probe-scattered field with the actual pyramidal geometry and finite conductivity; if the EigenProbe expression with a single fitted apex radius systematically fails to reproduce the Au/Si contrast and SiC approach curves in a way that correlates with asymmetric or non-PEC features, the idealized hyperboloid model is falsified.","supporting_citations":[],"review_version":2}