{"id":"88dbd015-22c1-4ab3-957f-e3b4faa6eff9","arxiv_id":"2506.20873","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Resonant q-BIC membrane metasurfaces produce odd harmonics with non-integer power-scaling exponents, attributed to higher-order susceptibilities activated by strong local fields.","lead":"A silicon membrane metasurface supporting a quasi-bound state in the continuum is shown to generate third through ninth harmonics whose power dependence deviates from the usual integer power laws. The work suggests that strong resonant field enhancement can change the effective nonlinearity of a metasurface, with implications for compact harmonic sources.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed non-integer power scaling may be an artifact of the broad femtosecond pump spectrum overlapping the narrow qBIC resonance, not the proposed higher-order susceptibility feedback.","rationale":"The reader's weakest assumption focused on the reliability of the susceptibility values and the neglect of harmonic-only nonlinear source terms in the simulation. While those are legitimate secondary concerns, the paper's central claim is an experimental one: the observation of non-integer intensity scaling. The reader's concern does not challenge the validity of that observation, only the theoretical explanation. My concern challenges the observation itself: the experimental conditions (broad femtosecond pump, narrow high-Q resonance) create a plausible mechanism for apparent non-integer slopes that does not require any new physics. This is more load-bearing because if the experimental slopes are an artifact, the 'break this principle' claim is unsupported; if only the susceptibility assumptions are wrong, the observation might still stand as a discovery even if the mechanism is misidentified. The paper's own model, by assuming a monochromatic pump, cannot rule out the spectral-overlap mechanism. I do not see this concern in the reader's analysis, hence 'disagree'. That said, the paper has clear strengths: the linear transmission is well characterized, the unpatterned membrane provides a control showing integer scaling, and the model reproduces the data within its assumptions. The appropriate resolution is to require the new test before accepting the non-integer scaling as a genuine effect. The verdict remains CONDITIONAL, but the condition should be the spectral-overlap check rather than only the susceptibility sensitivity.","tokens_in":74,"tokens_out":8017,"duration_ms":188450,"concrete_test":"Repeat the power-dependence measurement of the third, fifth, and seventh harmonics from the metasurface with a pump whose spectral width is narrower than the qBIC linewidth, e.g., by spectral filtering the 250 fs pulse to ~3 nm bandwidth (accepting some pulse lengthening) or by using a longer pulse (~5 ps) tuned to the same resonance. If the extracted slopes return to integer values (3, 5, 7), then the non-integer scaling is an artifact of the broad pump spectrum overlapping the narrow resonance and does not reflect the proposed higher-order susceptibility feedback. As a complementary check, record the transmitted pump spectrum at each input power to see whether spectral reshaping or resonance shifts occur over the power range used in Fig. 4d.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central observation is that harmonic powers from the qBIC metasurface scale with non-integer exponents (3ω: 2.3, 5ω: 3, 7ω: 4.6, 9ω: 4.5), whereas the unpatterned membrane gives integer exponents. The paper interprets this as evidence that the high-Q resonance boosts higher-order susceptibilities, modifying the effective nonlinearity. However, the experiment uses 250 fs pulses at 3.96 µm, which have a Fourier-limited spectral width of roughly 90 nm (FWHM), while the qBIC resonance has a measured Q of 480, corresponding to a linewidth of about 8 nm. The pump spectrum is therefore much broader than the resonance. The measured harmonic signal at a fixed pump wavelength is an integral over the pulse spectrum weighted by the resonance enhancement. Any power-dependent change in the pulse spectrum (e.g., self-phase modulation, spectral broadening) or in the resonance position (e.g., Kerr-induced index shift, free-carrier dispersion) will alter the spectral overlap and thus the apparent intensity scaling. The numerical model in Methods V.D solves the nonlinear wave equation in the frequency domain with the electric field expanded only over harmonic components, i.e., it assumes a monochromatic pump at the resonance frequency. It does not include the finite bandwidth of the pump or power-dependent spectral reshaping. Therefore, the non-integer slopes observed in the experiment could be reproduced by a mundane spectral-overlap effect that the model cannot capture, undermining the claim that the deviation is caused by resonant field-driven modification of effective susceptibilities. This is a load-bearing concern because if the slopes become integer when the spectral-overlap effect is removed, the central conclusion collapses.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports high-harmonic generation (HHG) from a free-standing silicon membrane metasurface supporting a quasi-bound-state-in-the-continuum (qBIC) resonance at 3.96 µm. Experiments show that resonant pumping enhances harmonic signals—by more than three orders of magnitude for the seventh harmonic—and enables observation of the ninth harmonic only on resonance. The authors also report non-integer intensity scaling exponents for the harmonic powers (3ω: 2.3, 5ω: 3, 7ω: 4.6, 9ω: 4.5), in contrast to the integer slopes observed for an unpatterned membrane. A frequency-domain nonlinear wave equation model, including susceptibilities up to ninth order and feedback terms at the pump frequency, is used to reproduce the trends. The central claimed result is that high-Q resonances modify the effective nonlinearity, producing unconventional power scaling.","tokens_in":10022,"tokens_out":3873,"duration_ms":48907,"significance":"If the observation is correct, this is a significant result for nonlinear nanophotonics: it would demonstrate that resonant field enhancement can substantially modify the effective order of a nonlinear process, breaking the conventional integer-power-law scaling of HHG. The experimental platform—free-standing membrane metasurfaces—is clean, and the comparison with unpatterned membranes is a useful control. The reported enhancements and the first observation of ninth-harmonic generation from this platform are valuable. However, the theoretical explanation relies on higher-order susceptibilities whose values are assumed from prior literature or an atomic scaling model, and the finite spectral bandwidth of the pump is not included in the model. These issues leave the mechanism open to alternative explanations, so the paper requires additional validation before the central claim can be accepted.","major_comments":[{"comment":"The non-integer slopes are the central experimental claim, but Figure 4d shows no error bars or confidence intervals, and the text does not describe the fitting procedure (fit range, weighting, number of points, or how the 'inflection at lower powers' is handled). Without uncertainties, the difference between slopes such as 2.3 and 3 for the third harmonic cannot be assessed as statistically meaningful. Please provide error bars, fit details, and the off-resonance slopes for the patterned membrane that are mentioned in the text but not shown in a figure.","section":"Section III.A, Figure 4d"},{"comment":"The numerical model expands the field only over harmonic components and therefore treats the pump as effectively monochromatic. This is a serious limitation because the 250 fs pump has a Fourier-limited bandwidth of roughly 90 nm, while the qBIC resonance has a measured Q of 480, corresponding to a linewidth of about 8 nm. The measured harmonic signal is an integral over the pump spectrum weighted by the resonance enhancement, so any power-dependent spectral reshaping (self-phase modulation, thermal effects, free-carrier dispersion) or power-dependent resonance shift could produce apparent non-integer scaling that is unrelated to higher-order susceptibilities. The model cannot capture this effect. Please include a quantitative estimate of the spectral-overlap contribution or provide a control experiment with a spectrally resolved pump and a measurement of the transmitted pump spectrum as a function of input power.","section":"Methods V.D and Section III.A"},{"comment":"The theoretical explanation depends on the values of χ(5), χ(7), and χ(9), which are taken from an atomic field scaling model rather than measured in this work. The paper states that reproducing the trends 'necessitated' these higher-order terms and the feedback effects, but it does not report the assumed values, their dispersion, or a sensitivity analysis. With several adjustable higher-order susceptibilities and cross-phase-modulation terms, it is not clear that the non-integer slopes are a robust prediction rather than a fit. Please provide the explicit susceptibility values used, show how the predicted slopes change when they are varied within a plausible range, and justify the truncation at ninth order.","section":"Methods V.D, Figure 5b"},{"comment":"The text states that 'we observe the same integer power scaling for the off-resonance excitation of the patterned membrane,' but no such data are presented in the main text or figures. This control is essential for ruling out a trivial explanation in which the resonance merely filters the pump spectrum or introduces a power-dependent coupling. Please show the off-resonance power dependencies for the patterned membrane, ideally alongside the on-resonance data, with the same fitting and error analysis.","section":"Section III.A, paragraphs after Figure 4"}],"minor_comments":[{"comment":"The caption text says the unpatterned membrane is pumped at '4.96 µm,' but the surrounding text and the experimental tuning range indicate this should be 3.95–3.96 µm. Please correct the typo.","section":"Section III.A, Figure 4 caption"},{"comment":"The abbreviation 'CdT' appears to be a typo for CdTe; please correct it.","section":"Introduction, reference [23]"},{"comment":"The description of the THG spectrum says the narrow peak 'aligns precisely with the tripled frequency of the qBIC mode,' but it would be useful to state explicitly how the spectrometer wavelength calibration and the qBIC resonance position were cross-checked, since the pump is tuned over a range much broader than the resonance linewidth.","section":"Section III.A, Figure 3a"},{"comment":"The phrase 'for the first time' in the conclusions should be qualified in light of previous reports of non-integer or non-standard power laws in resonant metasurfaces (e.g., even-harmonic generation in Ref. [21]). Please discuss how the present odd-harmonic observation differs from those prior results.","section":"Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The experimental data appear to support a real deviation from integer power scaling, but the manuscript's theoretical model is not yet convincing enough to rule out spectral-overlap or resonance-shift artifacts. The authors should be asked to provide error bars, off-resonance controls, a spectral-overlap estimate, and explicit susceptibility values. I do not see grounds for rejection, but the central mechanistic claim needs substantial additional evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper shows a clean experimental result: a free-standing q-BIC silicon metasurface produces odd harmonics (3ω to 9ω) with power-scaling slopes that are clearly non-integer (2.3, 3, 4.6, 4.5), while the unpatterned membrane gives the expected integer slopes. The controls are good—off-resonance pumping of the same metasurface restores integer scaling, and very high-Q samples that don't couple also show integer scaling. The enhancement of higher harmonics (up to 3 orders for 7ω) and the appearance of 9ω only on resonance are solid observations. The frequency-domain model is a serious attempt: it includes nonlinearities up to 9th order, pump depletion, and cross-phase modulation, and it reproduces both the absolute efficiencies and the non-integer slopes. That's genuinely useful, even if the model is monochromatic and relies on literature-based higher-order susceptibilities. The soft spots are in the interpretation and the completeness of the evidence. First, the claim that this 'breaks the conventional power scaling' is overstated—the same group reported non-integer scaling for even harmonics before (Ref [21]), and the mechanism (higher-order susceptibilities modifying effective nonlinearity) is standard nonlinear optics. What's new is the specific system and the odd-harmonic demonstration. Second, there are no error bars on the fitted slopes; the reader can't tell how robust the difference between 2.3 and 3 is. Third, the higher-order susceptibility values are assumed, not measured, so the model could match the data for the wrong reason. The paper acknowledges this by citing an atomic field scaling approach, but doesn't test sensitivity to those values. The stress-test concern about spectral overlap: the pump spectrum (90 nm FWHM) is much broader than the 8 nm resonance. But a broad pump alone doesn't change the power exponent—if the pump spectrum shape is fixed, the harmonic signal at a given wavelength still scales as P^n. To get non-integer slopes from spectral overlap, you'd need power-dependent spectral reshaping or resonance shifts, and the paper doesn't address those. So the concern is speculative rather than a demonstrated flaw. Still, a short discussion or a finite-bandwidth simulation would strengthen the paper considerably. Who is this for? People working on BIC-enhanced nonlinear nanophotonics and solid-state HHG. It's a worthwhile contribution that deserves serious peer review. I'd recommend sending it out, but requiring error bars on the slopes, a sensitivity analysis of the higher-order susceptibilities, and at least a discussion of the pump-spectrum/resonance-overlap issue before publication. The central observation is likely correct; the interpretation needs tempering.","headline":"Solid experiment with a real non-integer scaling observation, but the 'break the principle' framing is overstated and the theoretical mechanism is plausible but not nailed down.","tokens_in":798,"tokens_out":1447,"would_cite":true,"duration_ms":51417,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.65.Ky"],"model":"deepseek-v4-flash","headline":"A resonant membrane metasurface supporting a quasi-bound state in the continuum produces high harmonics whose intensity scales with non-integer powers of the pump intensity, breaking the conventional integer-power law of high-harmonic…","keywords":["high-harmonic generation","quasi-bound state in the continuum","membrane metasurface","non-integer power scaling","nonlinear susceptibility","field enhancement","resonant nonlinear optics"],"falsifier":"Measure the third-harmonic power slope over at least two decades of pump power while tuning the pump wavelength to, say, 100 nm away from the qBIC resonance in the same metasurface. The paper predicts the slope returns to the integer value 3 off-resonance and to about 2.3 on resonance; if the non-integer slope persists far from the resonance, the qBIC field-enhancement mechanism is not the cause. Alternatively, a comparative scan of the slope across metasurfaces with different Q factors would test the claim that the deviation tracks the resonant field enhancement.","tokens_in":9569,"feed_emoji":"⚡","tokens_out":7781,"duration_ms":77995,"temperature":0.7,"pith_summary":"This paper reports that high-harmonic generation in a free-standing silicon membrane metasurface does not follow the conventional rule that the nth harmonic's power scales as the nth power of the pump intensity. When the pump is tuned to a quasi-bound-state-in-the-continuum (qBIC) resonance, the measured slopes become non-integer: about 2.3 for the third harmonic, 3 for the fifth, and 4.6 for the seventh. The authors argue that the resonance's strong local field enhancement amplifies higher-order nonlinear susceptibilities and creates feedback among harmonics, changing the effective nonlinear order of the system. If true, this means resonant metasurfaces can reach extreme regimes of harmonic generation at moderate input powers, and the standard integer-power law is not universal for nanophotonic systems.","feed_headline":"Metasurface breaks high-harmonic generation's integer power law","feed_subtitle":"Experiment shows third and seventh harmonic slopes of 2.3 and 4.6 instead of 3 and 7 near the quasi-BIC.","key_machinery":"The central object is the quasi-bound state in the continuum (qBIC): a symmetry-protected dark mode of a hexagonal lattice of elliptical apertures in a 1 µm silicon membrane, made weakly radiating by breaking the circular symmetry of the apertures. It produces a Fano transmission resonance at 3.96 µm with a simulated Q factor of 580 (measured 480), and at the resonance it localises the pump field strongly inside the silicon. That local field is the mechanism that drives the unconventional scaling: it boosts higher-order nonlinear susceptibilities (up to ninth order) relative to lower-order ones, so the generated harmonic power becomes a mix of several integer power laws, yielding apparent non-integer slopes. The supporting numerical model solves the nonlinear wave equation in the frequency domain, retaining all nonlinear terms driven by the pump field, including pump depletion and cross-phase modulation among harmonics, while neglecting terms that depend only on harmonic fields.","core_discovery":"The central claim is that a qBIC resonance in a membrane metasurface drives high-harmonic generation into a regime where the generated harmonic power no longer scales as the integer power of the pump intensity, as it does in bulk solids and unpatterned membranes. Experimentally, the third, fifth, seventh, and ninth harmonics from the metasurface follow best-fit slopes of 2.3, 3, 4.6, and 4.5, respectively, while the unpatterned membrane shows slopes matching the harmonic orders. The same metasurface pumped off-resonance returns to integer scaling, so the deviation is tied to the qBIC excitation. The authors reproduce the behaviour numerically with a generalised perturbative model that keeps nonlinear susceptibilities up to ninth order and includes pump depletion, self- and cross-phase modulation, and feedback of harmonics onto the pump; the model captures the observed slopes and conversion efficiencies. The interpretation is that the enhanced local fields at the resonance make higher-order susceptibility contributions comparable to lower-order ones, so the effective nonlinearity of the system becomes a mixture of orders.","pith_inferences":["If the effective nonlinear order is set by the local field strength, the power-law slope could become a designable parameter, tunable through resonance Q, detuning, or aperture geometry.","The same mechanism should appear in other resonant platforms, such as dielectric resonators or epsilon-near-zero films, whenever local fields make high-order susceptibility terms compete with low-order ones; measuring the slope could serve as a generic probe of resonant field enhancement.","A testable extension is to pump with two beams or polarization states to see whether the non-integer slopes are accompanied by additional inter-order mixing, which the feedback model would predict."],"forward_implications":["Harmonic order no longer fixes the intensity scaling exponent in resonant metasurfaces; the effective exponent becomes a resonance-dependent quantity.","Resonant field enhancement lets high harmonics up to the ninth be observed at pump fluences around 2 mJ cm⁻², where an unpatterned membrane of the same material produces no detectable ninth harmonic.","The observed enhancement grows with harmonic order, exceeding three orders of magnitude for the seventh harmonic relative to the unpatterned membrane.","A full perturbative treatment including higher-order susceptibilities and pump feedback is sufficient to describe the non-integer scaling, so the effect does not require a non-perturbative strong-field mechanism.","Coupling efficiency into the qBIC matters: metasurfaces with Q factors above roughly 1000 show no enhanced HHG in this experiment, meaning the resonance must be matched to the pump linewidth."],"supporting_citations":[{"why":"Supplies the atomic-field scaling procedure and the subwavelength silicon film baseline that the metasurface result is compared against.","marker":"[26]"},{"why":"Provides the measured linear and nonlinear dispersion profile of silicon that the simulations integrate.","marker":"[31]"},{"why":"Extends harmonic generation in silicon membranes to visible and ultraviolet wavelengths and supports the perturbative susceptibility framework used here.","marker":"[33]"},{"why":"Demonstrates odd harmonics up to the 11th order from BIC-enabled dielectric metasurfaces, the prior state of the art this paper refines.","marker":"[19]"},{"why":"Establishes high-harmonic generation from a single subwavelength dielectric resonator, the platform class the metasurface builds on.","marker":"[5]"},{"why":"Introduces the symmetry-breaking route that converts a symmetry-protected BIC into a quasi-BIC with a high-Q resonance.","marker":"[12]"}],"fun_headline_variants":["Metasurface qBIC yields non-integer harmonic power slopes","Resonant metasurface breaks harmonic scaling law","Fractional harmonic orders from quasi-BIC metasurface","Harmonic power law broken by quasi-BIC resonance","Integer scaling fails for qBIC harmonic generation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The numerical story depends on assumed values for the material's higher-order nonlinear susceptibilities and on neglecting nonlinear terms that involve only harmonic fields, so if those assumptions are wrong the simulated slopes could match the data for the wrong reason.","fun_headline_variants_meta":{"raw":{"variants":["Metasurface qBIC yields non-integer harmonic power slopes","Resonant metasurface breaks harmonic scaling law","Fractional harmonic orders from quasi-BIC metasurface","Harmonic power law broken by quasi-BIC resonance","Integer scaling fails for qBIC harmonic generation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000303,"raw_usage":{"total_tokens":1740,"prompt_tokens":942,"completion_tokens":798,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":558,"completion_tokens_details":{"reasoning_tokens":731}},"tokens_in":558,"tokens_out":798,"duration_ms":8552,"temperature":1.0,"reasoning_tokens":731,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:39:13.108698+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the third-harmonic power slope over at least two decades of pump power while tuning the pump wavelength to, say, 100 nm away from the qBIC resonance in the same metasurface. The paper predicts the slope returns to the integer value 3 off-resonance and to about 2.3 on resonance; if the non-integer slope persists far from the resonance, the qBIC field-enhancement mechanism is not the cause. Alternatively, a comparative scan of the slope across metasurfaces with different Q factors would test the claim that the deviation tracks the resonant field enhancement.","supporting_citations":[{"cited_title":"Rodr ´ ıguez-Sun´ e, J","cited_arxiv_id":null,"evidence_quote":"Provides the measured linear and nonlinear dispersion profile of silicon that the simulations integrate."},{"cited_title":"Hallman, L","cited_arxiv_id":null,"evidence_quote":"Extends harmonic generation in silicon membranes to visible and ultraviolet wavelengths and supports the perturbative susceptibility framework used here."},{"cited_title":"Zograf, K","cited_arxiv_id":null,"evidence_quote":"Demonstrates odd harmonics up to the 11th order from BIC-enabled dielectric metasurfaces, the prior state of the art this paper refines."},{"cited_title":"Zalogina, L","cited_arxiv_id":null,"evidence_quote":"Establishes high-harmonic generation from a single subwavelength dielectric resonator, the platform class the metasurface builds on."},{"cited_title":"Koshelev, S","cited_arxiv_id":null,"evidence_quote":"Introduces the symmetry-breaking route that converts a symmetry-protected BIC into a quasi-BIC with a high-Q resonance."}],"review_version":1}