{"id":"72213de5-b7ce-4613-8c32-c3fbc20490a7","arxiv_id":"2412.03422","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A lithium niobate metasurface with sharp quasi-BIC resonances demonstrates GHz-speed electro-optic modulation of both reflected light and second-harmonic signal.","lead":"Researchers built a lithium niobate metasurface that changes how much light it reflects and how much second-harmonic light it generates when an electrical voltage is applied. It works at gigahertz speeds and could make free-space optical switches and modulators faster and more compact.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SHG 'order-of-magnitude' claim conflicts with reported factor-5 maximum and 110% slope change; headline should be corrected.","rationale":"The reader identified the Fano-fit background as the weakest assumption, but I view that as less load-bearing: the linear modulation efficiency is measured directly as a lock-in AC/DC ratio and the bandwidth is a direct VNA measurement, so neither would be invalidated by a voltage-dependent background in the reflectance spectra used for the shift extraction. The SHG magnitude inconsistency is more central because the abstract and conclusion assert an order-of-magnitude nonlinear modulation, while the body reports a maximum factor-5 reduction on resonance and a factor-2.1 change on the slopes; these are internally inconsistent. Independent supports do exist: the sign-reversed SHG response under opposite bias confirms a Pockels origin, the linear lock-in and VNA traces are direct measurements, and the SHG slope change is within a factor of 2 of simulation. These supports make the underlying work likely sound, but the quantitative headline claim is overstated and needs correction. The appropriate verdict therefore remains CONDITIONAL, unchanged from the reader, pending the correction and a re-statement of the SHG modulation magnitude computed from the raw data.","tokens_in":15220,"tokens_out":14650,"duration_ms":144579,"concrete_test":"Recompute from the raw SHG traces behind Figure 4e,f the maximum intensity ratio P_SH(V)/P_SH(0) for V = 0, +9 V, and -9 V at every pump wavelength. If the largest ratio over this voltage range is below 10 (and no point shows a greater-than-10-fold decrease), revise the abstract and conclusion to state the observed factor (about 5 on resonance, about 2 on the slope) and quote the efficiency against that baseline before resubmission.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the device demonstrates 'more than one order of magnitude intensity modulation' of second-harmonic light is not supported by the paper's own numbers. In the SHG section, on-resonance pumping is reported to give 'a reduction up to a factor of 5' for ΔV = ±9 V, and on the slopes the relative variation is (P_SH(9V) - P_SH(0V))/P_SH(0V) > 1.1, i.e., an increase by a factor of about 2.1. Neither value reaches a factor of 10. The abstract and conclusion repeat the order-of-magnitude claim, and the reader's strongest_claim inherits it. The stated efficiency η > 0.12 V^-1 is computed from the 110% slope change, so it corresponds to a factor-2.1 effect, not a factor-10 effect. Since nonlinear modulation magnitude is one of the two headline demonstrations, this quantitative overstatement is load-bearing for the central claim as written. The underlying physics may be sound, but the headline metric must be corrected to the actual demonstrated factor before the claim is accepted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a monolithic lithium-niobate-on-insulator (LNOI) metasurface supporting a quasi-bound-state-in-the-continuum resonance, and demonstrates electro-optic modulation of the reflected fundamental and of the second-harmonic signal. The linear characterization yields Q > 8000, a resonance shift of about 0.05 nm for a 9 V bias, a relative reflectivity modulation of up to about 0.12 for Vpp = 20 V with a low-voltage efficiency near 0.015 V^-1, and a 3-dB bandwidth around 800 MHz. The SHG experiments show a factor-of-5 reduction on resonance and a relative variation greater than 1.1 on the resonance slopes under ±9 V DC bias. The abstract and conclusion claim 'more than one order of magnitude' SHG intensity modulation and fast modulation of both linear and nonlinear signals, but these statements are not supported by the numbers reported in the body.","tokens_in":15395,"tokens_out":8664,"duration_ms":76674,"significance":"If the quantitative claims are corrected, this is a valuable demonstration: it uses a standard LNOI platform, top-down fabrication, CMOS-compatible voltages, and independent COMSOL simulations based on literature values for electro-optic and nonlinear coefficients, with no fitting of the target effect to the data. The measured linear modulation depth, bandwidth, and quality factor are credible and represent a useful step toward free-space electro-optic modulators. The SHG modulation result is interesting even at the reduced factor-of-5 or factor-of-2 levels actually demonstrated, because it shows that Pockels control of nonlinear upconversion is possible with a CW pump. The main weakness is the overstatement of the SHG modulation depth and the unsupported claim of fast nonlinear modulation, not the underlying experiment.","major_comments":[{"comment":"The abstract states 'more than one order of magnitude intensity modulation of the second harmonic' and the conclusion states 'a SHG intensity modulation exceeding one order of magnitude by applying ΔVEO = 9 V bias'. These claims are contradicted by the body: on resonance the paper reports 'a reduction up to a factor of 5' for ΔVDC = ±9 V (Fig. 4e), and on the slopes it reports (P_SH(9V) − P_SH(0V))/P_SH(0V) > 1.1 (Fig. 4d,f), i.e., an increase by a factor of about 2.1. Neither value reaches a factor of 10. The stated efficiency η > 0.12 V^-1 is computed from the 110% slope change, so it corresponds to the factor-2.1 effect, not to a decade-scale modulation. This quantitative overstatement is load-bearing because the SHG modulation depth is one of the two headline results; the text should be corrected to report the actually demonstrated factors (about 5 on resonance, about 2.1 on the slopes).","section":"Abstract and Conclusion, SHG modulation (Fig. 4)"},{"comment":"The abstract claims that the metasurface 'achieve[s] fast electrical modulation of both linear and nonlinear optical properties', and the conclusion says 'high-speed, efficient EO modulation'. However, the linear modulation is the only quantity characterized as a function of frequency (Fig. 3f, up to about 1 GHz). The SHG modulation is demonstrated only under static DC bias (Fig. 4c-f); no AC or frequency-resolved measurement of the modulated SHG is presented anywhere in the paper. The claim of fast nonlinear modulation is therefore unsupported and should be removed or explicitly qualified as DC-only until such a measurement is provided.","section":"Abstract and Conclusion"},{"comment":"The tuning sensitivity ΔλEO/ΔV = 5.6 pm/V is extracted from Fano fits to reflectance spectra recorded at different DC biases. The text notes that the shift is 'superimposed to an unmodulated instrumental artifact' and that the measured linear modulation is about five times smaller than simulation due to a 'high level of unmodulated reflectance' (Fig. 3c,d). This assumes that the broad Fabry–Pérot background is strictly bias-independent, but no control experiment or error analysis is provided to support that assumption. If the background changes with bias, the extracted resonance shift and hence the reported tuning sensitivity and the comparison with the simulated 0.06 nm shift would be inaccurate. The direct lock-in modulation measurement at the derivative extrema is less affected, but the paper should either demonstrate background stability or quantify the sensitivity of the extracted shift to the background parameters.","section":"Design and linear characterization, Fig. 3a,d"},{"comment":"The conclusion states 'we report the first experimental demonstration of electrically modulated SHG in subwavelength devices using CW pumping [40]'. This claim is not adequately supported in the manuscript: Ref. [40] demonstrates continuous-wave SHG in a gallium-phosphide metasurface, but not electro-optic modulation, while Ref. [39], cited nearby, is titled 'Electro-optically Modulated Nonlinear Metasurfaces'. The authors should clarify what precisely is claimed as 'first' (e.g., first CW-pumped EO-modulated SHG in a lithium niobate metasurface) and explicitly distinguish their result from Ref. [39]; otherwise the novelty claim should be softened.","section":"SHG modulation, Conclusion"}],"minor_comments":[{"comment":"The sentence 'Our experimental results reveal a modulation efficiency exceeding 10%, driven by less than 10 V' is inconsistent with the paper's own definition of efficiency as relative modulation per applied volt, and with the later number of 0.07 modulation at Vpp = 10 V. Please clarify whether 10% refers to modulation depth or efficiency, and use consistent units.","section":"Introduction"},{"comment":"The caption of Fig. 4a says 'sample with FF = 0.3 and α = 0', but the text and the rest of the paper refer to the sample with α = 0.19. This appears to be a typo and should be corrected.","section":"Figure 4 caption"},{"comment":"The phrase 'good agreement with the nonlinear simulation under same excitation conditions (a factor 2 higher)' is internally contradictory: a factor-of-2 discrepancy is not 'good agreement'. Please report the simulated value separately and describe the discrepancy quantitatively.","section":"SHG modulation section"},{"comment":"The sentence 'Importantly, the low SHG signal outside the resonance leads to a high modulation efficiency, close to the simulated value' is confusing because the preceding sentence states that off-resonant excitation shows 'no electro-optic modulation thereof'. Please clarify which condition yields the high efficiency and how the off-resonant background affects it.","section":"SHG modulation section"},{"comment":"There are minor language issues: 'We models a plane-wave excitation' should be 'We model', and 'in herent' should be 'inherent'. The duplicated reference [32]/[34] (same Huang et al. paper) should also be merged.","section":"Methods"},{"comment":"The efficiency definition for the SHG modulation uses ΔVDC = 9 V while the linear efficiency uses Vpp/2. Please state explicitly whether the quoted efficiencies are referenced to the peak voltage excursion or to the peak-to-peak voltage, and keep the convention consistent across the paper.","section":"SHG modulation section"}],"recommendation":"major_revision","confidential_remarks":"The main experimental results appear sound and the linear modulation characterization is convincing. The revision hinges on aligning the abstract and conclusion with the body's numbers: the SHG modulation is factor-of-5 (or factor-of-2.1 on the slopes), not an order of magnitude, and is demonstrated under DC bias only. I also could not verify the 'first demonstration' claim against Ref. [39] from the text; if Ref. [39] already reports experimental EO modulation of nonlinear metasurfaces, the priority claim should be revised. The paper is otherwise within scope and worth reviewing after these corrections."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a real experimental step—first voltage-modulated CW second-harmonic generation from a subwavelength LNOI metasurface—but the paper's own numbers don't support the \"more than one order of magnitude\" SHG claim in the abstract and conclusion. That needs fixing before this is accepted.\n\nWhat's genuinely new: prior work did CW SHG in quasi-BIC metasurfaces (Anthur) and EO modulation of nonlinear metasurfaces (He), but not both in one monolithic LNOI device with a telecom CW pump. The linear modulation is credible: Q>8000, resonance shift ~5.6 pm/V at 9V close to the simulated 6 pm/V, ~0.015 V-1 efficiency at Vpp<10V, and -3dB at 800MHz. The simulations use literature electro-optic and nonlinear coefficients and measured geometry; no fitting of the target result. That's good practice.\n\nSoft spots, in order of severity. First, the SHG headline. On-resonance pumping gives a factor-of-5 reduction at ±9V; on the slopes the relative change is (P_SH(9V)-P_SH(0V))/P_SH(0V) > 1.1, i.e. a factor ~2.1 increase. Neither is tenfold. The \"η > 0.12 V-1\" is derived from the 110% slope change, so it corresponds to the factor-2.1 effect, not the order-of-magnitude one. The abstract and conclusion must be corrected to the demonstrated factors. Second, the linear modulation is five times weaker than simulation because of a large unmodulated reflectance; the Fano-fit extraction assumes that background is bias-independent. That's plausible but unverified, and there are no error bars on the key efficiencies or shifts. Third, the comparison with ref [39] is asserted rather than quantified; a table of modulation depth, voltage, and bandwidth against prior EO nonlinear metasurfaces would help.\n\nNone of this undermines the core demonstration. The data look internally consistent, the EO origin is supported by the sign-dependent slopes, and the overstatement is a fixable reporting error rather than a fatal flaw. It deserves a serious referee, but the referee should insist on corrected SHG claims and error bars.\n\nRecommendation: send to peer review with a request for major revision on the claims.","headline":"A credible first demonstration of CW-pumped electrically modulated SHG in an LNOI metasurface, but the abstract's order-of-magnitude SHG claim exceeds the data; fix the numbers and it's a solid paper.","tokens_in":15994,"tokens_out":2183,"would_cite":true,"duration_ms":19520,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["78.20.Jq","42.79.Hp","42.65.Ky"],"model":"deepseek-v4-flash","headline":"A single lithium-niobate metasurface modulates both reflected telecom light and its second harmonic at gigahertz speeds with less than 10 volts.","keywords":["lithium niobate","metasurface","electro-optic modulation","Pockels effect","bound states in the continuum","second-harmonic generation","guided-mode resonance","CMOS-compatible photonics"],"falsifier":"Record the full reflected spectrum while a DC bias is swept, and test whether the Fano-fit background parameters (offset and phase) remain constant; if they move with voltage, the reported 5.6 pm/V shift and the derived efficiencies collapse, whereas a stationary background with a linearly moving resonance peak confirms the Pockels mechanism.","tokens_in":15011,"feed_emoji":"⚡","tokens_out":9001,"duration_ms":80799,"temperature":0.7,"pith_summary":"This paper reports a flat, free-space electro-optic modulator made from a single 600 nm layer of lithium niobate patterned into an asymmetric array of nanowires. The device supports a nearly dark (quasi-bound-state-in-the-continuum) resonance with a quality factor above 8000, so that a few volts applied through CMOS-compatible electrodes shift the resonance enough to change the reflected power by about 10 per cent at gigahertz speeds. The same shift also modulates the intensity of the second harmonic generated by a continuous-wave pump, producing more than an order-of-magnitude change in the upconverted signal. Because lithium niobate is an established, transparent electro-optic material with a large Pockels coefficient, the result suggests that nonlocal metasurfaces can close the performance gap between bulk or on-chip modulators and free-space flat optics. If confirmed, this makes ultrafast electrical control of free-space light, both fundamental and frequency-upconverted, available in a compact, CMOS-compatible platform.","feed_headline":"Lithium niobate metasurface hits 800 MHz electro-optic modulation","feed_subtitle":"With <10 V it tunes a Q>8000 resonance, switching reflected and upconverted light at gigahertz rates.","key_machinery":"The load-bearing element is a quasi-bound-state-in-the-continuum guided-mode resonance (a dark guided mode made weakly radiating by breaking its symmetry) in a one-dimensional lithium-niobate-on-insulator grating. The grating is an asymmetric pair of nanowires with period 800 nm, fill factor 0.3, and asymmetry around 0.2 on a 450 nm lithium-niobate film over silicon dioxide on silicon; the asymmetry opens a controllable radiative channel, giving a sharp Fano line shape with Q near $10^{4}$. The Pockels effect shifts this resonance: the static field applied along the crystal z-axis changes the extraordinary index by half the cube of the index times the effective electro-optic coefficient times the field, displacing the resonance peak by a fraction of its linewidth per volt. In-plane interdigitated electrodes produce a uniform static field across the film, maximizing overlap with the confined optical mode. The same resonance shift is then read out twice: directly in reflectivity at a fixed wavelength, and squared in second-harmonic generation, where the quadratic field dependence sharpens the effective spectral linewidth and amplifies the modulation depth.","core_discovery":"The central claim is that a monolithically patterned lithium-niobate-on-insulator metasurface can act as a high-speed electro-optic modulator in both the linear and nonlinear regimes. The metasurface uses a guided-mode resonance whose symmetry-protected dark character is broken by a small asymmetry between neighbouring nanowires, turning it into a quasi-BIC with a theoretical Q of 11000 (measured Q greater than 8000) and a spectral width of about 0.18 nm near 1553 nm. An applied bias shifts the resonance through the Pockels effect, with a measured tuning sensitivity of 5.6 pm/V; this yields a relative reflectivity modulation of about 0.12 at 20 V peak-to-peak and a linear modulation efficiency of about 0.015 per volt, with a 3 dB bandwidth around 800 MHz and measurable response beyond 1.4 GHz. For the nonlinear response, pumping the same resonance with a narrow-linewidth continuous-wave laser at 12 kW/cm squared produces second-harmonic light whose excitation linewidth is about 0.14 nm, narrower than the fundamental resonance; applying a static bias of 9 V modulates the second-harmonic intensity by more than a factor of five on resonance and by more than 1.1 on the resonance slopes, corresponding to a modulation efficiency greater than 0.12 per volt. The paper therefore establishes that the same electrical signal can control both the fundamental reflection and the upconverted emission of a free-space metasurface.","pith_inferences":["Beyond the paper: suppressing the unmodulated reflectance background, for example with a buried reflector or a transmission configuration, should recover a large part of the five-fold gap between measured and simulated linear modulation.","Beyond the paper: the same resonance-shift mechanism should apply to sum-frequency and difference-frequency generation, since the nonlinear gain is set by the resonance linewidth and the Pockels coefficient rather than by the specific harmonic process.","Beyond the paper: the second-harmonic modulation depth should scale quadratically with the continuous-wave pump power at fixed bias; measuring that scaling would directly test the claim that the effect enters through the resonance-enhanced fundamental field.","Beyond the paper: shortening the 1.5 mm interdigitated electrodes should push the 3 dB bandwidth beyond 800 MHz, with the trade-off that the illuminated area must stay large enough to preserve the nonlocal resonance."],"forward_implications":["A single all-dielectric flat device can modulate a free-space telecom beam with a relative reflectivity swing of about 0.1 and a modulation efficiency above 0.01 per volt at CMOS-compatible voltages.","The same bias shifts the resonance enough to modulate continuous-wave second-harmonic emission by more than an order of magnitude, with efficiency around 0.12 per volt.","The quasi-BIC resonance, with Q above 8000 and a linewidth below 0.2 nm, makes the modulation sensitive to sub-nanometre refractive-index changes, so the required voltages stay below 10 V.","Because second-harmonic generation responds quadratically to the resonance field, the nonlinear modulation depth exceeds the linear one, and it can be driven by a low-power continuous-wave pump rather than pulsed lasers.","Electrode capacitance currently limits the 3 dB electrical bandwidth to about 800 MHz; reducing the 1.5 mm electrode length should extend the speed further."],"supporting_citations":[{"why":"Prior demonstration of gigahertz free-space electro-optic modulation in a different platform; supplies the performance baseline the paper must beat.","marker":"[8]"},{"why":"Defines the Pockels-effect index-change formula used throughout the modelling.","marker":"[12]"},{"why":"Supplies the lithium-niobate electro-optic coefficients r13 and r33 used in the static-field simulations.","marker":"[16]"},{"why":"Establishes the resonant-enhancement route for electro-optic modulation in lithium-niobate metasurfaces and the efficiency definition the paper adopts.","marker":"[25]"},{"why":"A thin-film lithium-niobate tunable metasurface whose efficiency is achieved at the cost of bandwidth; provides the comparison for the trade-off claimed here.","marker":"[26]"},{"why":"Reviews electro-optic lithium-niobate metasurfaces and the role of high-Q resonances in reaching large modulation efficiency.","marker":"[28]"},{"why":"Shows that frequency upconversion can amplify electro-optic modulation, motivating the second-harmonic modulation experiment.","marker":"[31]"},{"why":"Provides the Fano-profile model used to fit the measured reflection spectra and extract the resonance shift.","marker":"[38]"},{"why":"Demonstrates continuous-wave second-harmonic generation on quasi-BIC metasurfaces, enabling the CW-pumped second-harmonic modulation measured here.","marker":"[40]"}],"fun_headline_variants":["GHz-rate electro-optic modulation in a lithium niobate metasurface","Metasurface modulates reflection and second harmonic at 800 MHz","Nonlinear metasurface achieves fast electro-optic switching in free space","Lithium niobate metasurface: GHz modulation for linear and nonlinear light","High-Q metasurface enables efficient GHz electro-optic modulation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extracted tuning sensitivity of 5.6 pm/V assumes the broad Fabry-Perot background in the measured spectrum is completely unmodulated; if that background responds to the applied voltage, the reported modulation efficiencies would be inflated.","fun_headline_variants_meta":{"raw":{"variants":["GHz-rate electro-optic modulation in a lithium niobate metasurface","Metasurface modulates reflection and second harmonic at 800 MHz","Nonlinear metasurface achieves fast electro-optic switching in free space","Lithium niobate metasurface: GHz modulation for linear and nonlinear light","High-Q metasurface enables efficient GHz electro-optic modulation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000211,"raw_usage":{"total_tokens":1535,"prompt_tokens":1184,"completion_tokens":351,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":800,"completion_tokens_details":{"reasoning_tokens":260}},"tokens_in":800,"tokens_out":351,"duration_ms":3980,"temperature":1.0,"reasoning_tokens":260,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:24:37.064933+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Record the full reflected spectrum while a DC bias is swept, and test whether the Fano-fit background parameters (offset and phase) remain constant; if they move with voltage, the reported 5.6 pm/V shift and the derived efficiencies collapse, whereas a stationary background with a linearly moving resonance peak confirms the Pockels mechanism.","supporting_citations":[{"cited_title":"Chapter 11 - The Electrooptic and Photorefractive Effects,","cited_arxiv_id":null,"evidence_quote":"Defines the Pockels-effect index-change formula used throughout the modelling."},{"cited_title":"Evaluation of crystals of LiNbo3 doped with MgO or TiO2 for electrooptic devices,","cited_arxiv_id":null,"evidence_quote":"Supplies the lithium-niobate electro-optic coefficients r13 and r33 used in the static-field simulations."},{"cited_title":"Enhanced Electro-Optic Modulation in Resonant Metasurfaces of Lithium N i o b a t e ,","cited_arxiv_id":null,"evidence_quote":"Establishes the resonant-enhancement route for electro-optic modulation in lithium-niobate metasurfaces and the efficiency definition the paper adopts."},{"cited_title":"Tunable Metasurface Using Thin-Film Lithium Niobate in the Telecom Regime,","cited_arxiv_id":null,"evidence_quote":"A thin-film lithium-niobate tunable metasurface whose efficiency is achieved at the cost of bandwidth; provides the comparison for the trade-off claimed here."},{"cited_title":"Electro-optic lithium niobate metasurfaces,","cited_arxiv_id":null,"evidence_quote":"Reviews electro-optic lithium-niobate metasurfaces and the role of high-Q resonances in reaching large modulation efficiency."}],"review_version":1}