{"id":"88ce0359-2c0d-4109-bf04-4bc7886290d5","arxiv_id":"2505.14974","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Grating-induced loss in photonic crystal microrings is characterized as a function of grating-to-mode ratio, revealing broad OAM radiation loss and additional peaks that can degrade nonlinear frequency conversion.","lead":"Photonic crystal microrings with tiny wall gratings lose light in wavelength-dependent ways, and this paper maps those losses across the full grating-to-wavelength ratio range. The map gives designers a practical guide for choosing grating periods that avoid radiation loss when building frequency converters, microcombs, and other nonlinear photonic devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The full-spectrum map in Fig. 3(b) assumes the 1550 nm loss-vs-N/m curve is a universal function of λ0/(neff Λ); this scale invariance is unverified at other wavelengths, so the broadband loss prediction is the least-supported link in the central claim.","rationale":"The reader's weakest assumption identifies exactly the load-bearing concern: the transformation from a fixed-wavelength, variable-period measurement to a fixed-period, variable-wavelength spectrum assumes scale invariance beyond the neff conversion. My stress-test confirms this is the critical gap. The experimental N/m loss map at 1550 nm is credible on its own, and the FDTD supports the main loss channels at that wavelength. However, the headline 'full spectral response' is a prediction about other wavelengths, and the paper provides no direct test at any second wavelength. The appendix's demonstration that cladding symmetry changes the amplitudes of channels (v) and (vi) strengthens the worry that loss magnitudes and even relative channel strengths depend on details that vary with wavelength, such as cladding index contrast and mode field distribution. The phase-matching and radiation-angle analysis explains peak locations at 1550 nm but does not establish their persistence across an octave or more. Because this concern is sufficient to keep the reader's conditional verdict — the paper should be accepted only after validating or explicitly bounding the scale-invariance assumption — I recommend UNCHANGED rather than moving to acceptance or rejection. The concrete FDTD test at a second wavelength would settle whether the broadband map is physically valid or only a convenient interpolation.","tokens_in":14655,"tokens_out":4904,"duration_ms":46261,"concrete_test":"Run the same half-PhCR 3D FDTD setup at a second free-space wavelength, e.g., λ0 = 1064 nm, sweeping the grating period to cover N/m ≈ 0.1-2.5, and compare the resulting loss curve with Fig. 2(d) after converting the x-axis via N/m = λ0/(neff(λ0)Λ). If the positions of peaks (ii)-(vi) shift by more than 0.05 in N/m, or if the relative amplitudes change by more than a factor of two, the fixed-Λ spectrum in Fig. 3(b) is not a reliable full spectral response. A direct alternative is to simulate the fixed-Λ = 730 nm device at λ0 = 600, 1000, 1400, 1800, and 2200 nm and check whether each simulated κ/2π matches the mapped curve at the corresponding N/m value.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a single fixed-period PhCR exhibits the loss spectrum of Fig. 3(b) rests entirely on a scaling collapse: the experimentally measured loss κ(N/m) at λ0 = 1550 nm, m = 165 is treated as a universal function of N/m = λ/Λ = λ0/(neff(λ0)Λ). The mapping in Fig. 3(a) accounts only for the wavelength dependence of neff in converting the x-axis. It does not account for wavelength-dependent changes in mode confinement and its overlap with the grating tooth, the refractive-index contrast between air and SiO2 claddings over 400-2470 nm, the effective indices of TE1 and OAM modes that enter the phase-mismatch and radiation-angle analysis, or material dispersion. These can shift the positions and alter the amplitudes of the identified loss channels. For example, the θr = ±90° surface-radiation condition at N/m ≈ 0.33 depends on neff and n_r through Snell's law; both are dispersive, so that peak could move or change strength at shorter wavelengths. The appendix itself shows that changing the cladding environment from asymmetric to symmetric redistributes loss between channels (ii)/(v) and (iii)/(vi), demonstrating that loss amplitudes are not fixed solely by N/m. No experimental or FDTD data at any second wavelength is presented to validate the broadband prediction. The phrase 'full spectral response' therefore overstates what has been established: only the 1550 nm N/m loss map is measured, and the fixed-Λ spectrum is a plausible but unverified extrapolation. Secondary issues, such as the undocumented FDTD amplitude rescaling from A = 20 nm to A = 6 nm and the absence of released data, affect loss magnitudes but do not by themselves threaten the spectral positions; the scale-invariance assumption is the load-bearing gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an experimental and numerical study of grating-induced loss in photonic crystal microrings (PhCRs). The authors measure 37 devices with a fixed microring circumference and azimuthal mode number (m=165) at a fixed free-space wavelength (λ0≈1550 nm), varying the grating period Λ (or number of periods N) from 0 to about 600. From transmission fits they extract the loaded Q and total loss κ, and plot κ versus N/m, where N/m = λ/Λ = λ0/(neffΛ). They identify distinct loss channels: a broad excess-loss region centered at N/m≈1 attributed to vertical out-coupling into OAM states, a pronounced radiation peak at N/m≈0.33, a symmetric peak near N/m≈1.67, TE0–TE1 coupling peaks near N/m≈0.13 and 1.87, and negligible excess loss for N/m≥2 including the Bragg condition at N/m=2. These observations are compared with 3D FDTD simulations and interpreted through OAM radiation-angle and phase-mismatch calculations. In the latter part of the paper, the authors transform the measured N/m-dependent loss into a wavelength-dependent loss spectrum for a single fixed-period device (Λ=730 nm) spanning roughly 400–2470 nm, and they discuss implications for OPO and four-wave-mixing Bragg scattering.","tokens_in":15006,"tokens_out":4659,"duration_ms":40497,"significance":"If the N/m-universality assumption underlying the wavelength mapping is correct, the paper provides a useful design rule for PhCR-based nonlinear photonics, showing where grating-induced loss degrades Q for far-detuned pump, signal, and idler wavelengths. The strength of the work is the systematic 37-device experimental dataset at 1550 nm and the generally good qualitative agreement between experiment and FDTD for the positions of the main loss channels. The phase-mismatch and radiation-angle analysis is physically transparent and helps assign mechanisms to the observed peaks. However, the central 'full spectral response' claim for a fixed device is not independently validated: it is obtained by a change of variables from single-wavelength data, and the paper does not present experimental or FDTD results at any second wavelength. The appendix itself shows that loss amplitudes redistribute when the cladding environment changes, indicating that N/m alone does not fix the loss spectrum.","major_comments":[{"comment":"The fixed-Λ spectrum over 400–2470 nm is generated by reparametrizing the 1550 nm loss-versus-N/m curve via N/m = λ0/(neff(λ0)Λ). This assumes that grating-induced loss is a universal function of N/m only, with absolute wavelength entering only through the TE0 effective index. No experimental or FDTD data at any other wavelength is presented to test this assumption. The effective indices of the TE1, TM0, and OAM modes that determine the phase-matching conditions and radiation angles in Fig. 2(d) are dispersive, so the positions and amplitudes of the identified loss channels (e.g., the N/m≈0.33 peak and the broad OAM band) could shift or change strength at shorter or longer wavelengths. I recommend either validating the mapping with a second-wavelength measurement or FDTD simulation at one or two probe wavelengths for a fixed Λ, or explicitly presenting Fig. 3(b) as a model extrapolation rather than an experimentally established 'full spectral response.'","section":"Fig. 3(a)–(b) and accompanying text"},{"comment":"The appendix shows that changing the cladding environment from asymmetric (air/SiO2) to symmetric (SiO2/SiO2) substantially redistributes loss between channels (ii)/(v) and (iii)/(vi), and changes the total loss by 56–97% depending on grating configuration. This demonstrates that loss amplitudes are not determined by N/m alone but also depend on the refractive-index environment and modal field overlap, both of which vary with absolute wavelength. Consequently, the universal curve used to generate Fig. 3(b) carries an unquantified error. The manuscript should state this limitation explicitly and, if possible, estimate how much the loss-channel positions and strengths move over the 400–2470 nm range when the dispersions of neff for all relevant modes are included.","section":"Appendix, Figs. 6–7"},{"comment":"The 3D FDTD results are computed with a grating modulation amplitude of A=20 nm and then 'rescaled' to the experimental A=6 nm, but the rescaling law is not specified or justified. Because the simulated loss magnitudes are compared directly with the measured κ values, this rescaling factor acts as a free parameter that can absorb discrepancies in amplitude. The authors should state whether the rescaling is a fixed power-law scaling, a best fit to the experimental data, or derived from a separate set of simulations, and they should report the resulting uncertainty in the loss magnitudes.","section":"Section II, Fig. 2(d) and simulation text"}],"minor_comments":[{"comment":"The caption for panel (d) says the experimental loss values are 'extracted from the transmission spectra in (a)', but panel (a) contains SEM images; the transmission spectra appear in panel (c).","section":"Fig. 2 caption"},{"comment":"The caption states 'λ = 2π(RR−RW)/m = λ0/(neffΛ)', which is dimensionally incorrect: λ = λ0/neff, while N/m = λ/Λ = λ0/(neffΛ). This should be corrected to avoid confusing the variable transformation.","section":"Fig. 3 caption"},{"comment":"The abstract and conclusion state that the paper establishes the 'full spectral response' of grating-induced loss. Given that the experimental evidence is obtained at a single free-space wavelength, the broadband part of the claim should be qualified as a predicted or extrapolated response unless additional validation is added.","section":"Abstract and Conclusion"},{"comment":"The TE0–TE1 coupling peaks are predicted by FDTD and phase-mismatch analysis but are not observed in the experimental transmission measurements. The text should state more prominently that these channels are numerical predictions rather than directly measured features.","section":"Section II, regions ii and vi"}],"recommendation":"major_revision","confidential_remarks":"The 1550 nm dataset is solid and the paper is within scope for the journal. The main concern is whether the 'full spectral response' claim is supported: Fig. 3(b) rests on an unverified scale-invariance assumption, and the appendix reveals sensitivity of loss amplitudes to the cladding environment. I would encourage the editor to ask the authors to either add a validating measurement or FDTD run at a second wavelength for a fixed grating period, or to rewrite the title, abstract, and conclusion to frame the broadband spectrum as a predicted design tool rather than an experimentally measured response."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this paper gives a genuinely useful experimental map of grating-induced loss in photonic crystal microrings at 1550 nm, and the physics attributions are mostly convincing. The headline claim of a \"full spectral response\" for a fixed device, however, is a variable transformation from that single-wavelength data, and the underlying scale-invariance assumption is not tested. The soft spot is real but addressable, not disqualifying.\n\nWhat is new: earlier work focused on N≈m and N=2m; this paper systematically fills in the intermediate N/m regions and identifies loss channels nobody had mapped experimentally—the pronounced peak at N/m≈0.33, the symmetric surface-radiation peak at 1.67, and the TE0–TE1 coupling peaks at 0.13 and 1.87. The vary-N-fix-m strategy is a sensible way to get a broadband loss spec without multiple tunable lasers, and the 37-device dataset is internally consistent. The FDTD reproduces the main features, and the OAM radiation-angle and phase-mismatch analyses use independently computed effective indices, so there is no fit-to-claim circularity. The appendix's cladding-dependence study is honest and actually adds useful design information.\n\nThe soft spots, in order of importance. First, the mapping in Fig. 3(b) treats the measured κ(N/m) at 1550 nm as a universal function of λ/Λ = λ₀/(n_eff Λ), adjusting only n_eff(λ) on the x-axis. It does not account for wavelength-dependent mode confinement, cladding index contrast, or the effective indices of TE1 and OAM modes that enter the phase-mismatch and radiation-angle analysis. Those can shift or rescale the peaks. The appendix itself shows that changing the cladding environment redistributes loss between channels (ii)/(v) and (iii)/(vi), which demonstrates that loss amplitudes are not fixed by N/m alone. No experimental or FDTD data at any second wavelength is presented. That is the load-bearing gap. Second, the FDTD loss rescaling from A=20 nm to A=6 nm is undocumented; it affects magnitudes but not the spectral positions. Third, no data or code is released, which makes it hard to reuse the map.\n\nWho gets value: integrated photonics designers working on OPO, FWM, or microcombs where a fixed PhCR must operate across widely separated wavelengths. The 1550-nm N/m map alone is worth having, and the design guidelines in Fig. 3(c–d) are practical.\n\nMy recommendation: send it to peer review. A serious referee should ask for validation or justification of the scale-invariance assumption—ideally a measurement or FDTD at a second wavelength—and should require documentation of the amplitude rescaling. Those are reasonable conditions, not reasons to desk-reject.","headline":"Useful measured loss map for PhCRs at 1550 nm with plausible physics, but the broadband 'full spectral response' for a fixed device is an unverified extrapolation that needs a second-wavelength check.","tokens_in":15565,"tokens_out":1372,"would_cite":true,"duration_ms":13633,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.79.Dj","42.60.Da","42.65.-k"],"model":"deepseek-v4-flash","headline":"Grating-induced loss in photonic crystal microrings is governed by one ratio, N/m, which maps out a full spectrum of radiation and mode-coupling loss channels.","keywords":["photonic crystal microring","grating-induced loss","orbital angular momentum","quality factor","nonlinear photonics","radiation loss","mode coupling","broadband spectral mapping"],"falsifier":"Measure or simulate the same fixed-$\\Lambda$ PhCR (for example $\\Lambda = 730$ nm) at several different free-space wavelengths across the 400–2470 nm range and check whether the loss peaks appear at the $N/m$ positions predicted from the 1550 nm data; if the peaks shift, split, or change relative strength with wavelength, the claim that loss depends only on $N/m$ fails.","tokens_in":14450,"feed_emoji":"🌀","tokens_out":12390,"duration_ms":87230,"temperature":0.7,"pith_summary":"This paper establishes that grating-induced loss in photonic crystal microrings is controlled by a single dimensionless ratio, $N/m = \\lambda/\\Lambda$, where $N$ is the number of grating periods, $m$ the azimuthal mode number, $\\lambda$ the modal wavelength, and $\\Lambda$ the grating period. By measuring 37 devices with different grating periods at a fixed wavelength, the authors map out the full loss spectrum and identify distinct loss channels: a broad vertical-emission region centered at $N/m = 1$, a sharp radiation peak near $N/m \\approx 0.33$, a symmetric surface-radiation peak near $1.67$, and TE$_0$–TE$_1$ mode-coupling peaks near $0.13$ and $1.87$, with negligible excess loss for $N/m \\geq 2$. The result matters because nonlinear PhCR devices operate across widely separated wavelengths, so a single device can land in a lossy region even when its design wavelength is safe. If correct, the spectrum gives a loss-aware design rule for avoiding degraded $Q$ and even for deliberately suppressing unwanted nonlinear processes.","feed_headline":"Grating-induced loss in photonic microrings collapses to one ratio","feed_subtitle":"A measured loss spectrum now predicts which wavelengths will degrade Q in nonlinear photonic devices.","key_machinery":"The load-bearing identity is $N/m = \\lambda/\\Lambda = \\lambda_0/(n_{\\mathrm{eff}}\\Lambda)$, which equates the grating-to-mode-number ratio with the ratio of modal wavelength to grating period. Because the experimental study varies $N$ at fixed $m$ and wavelength, this identity lets the authors reparametrize single-wavelength, many-device data into a broadband spectrum for a single device of fixed $\\Lambda$. The physical mechanisms are assigned using two supplementary tools: the OAM radiation angle $\\theta_r$ from Snell's law with internal ejection angle $\\theta_e = (l/m)(\\pi/2)$, and the phase mismatch $\\Delta\\beta = \\beta_1 - \\beta_2 - 2\\pi/\\Lambda$ among TE$_0$, TE$_1$, OAM, and counterpropagating TE$_0$ modes. Together these identify which loss features are radiation into claddings and which are mode conversions.","core_discovery":"The central claim is that grating-induced loss in a PhCR is a function of the ratio $N/m$ rather than of the grating period or wavelength separately. Experimentally, at $\\lambda_0 \\approx 1550$ nm with $m = 165$, varying $N$ from 0 to 600 produces a loss spectrum with a broad excess-loss band centered at $N/m = 1$ from vertical out-coupling into OAM-carrying states; a pronounced radiation peak at $N/m \\approx 0.33$; a symmetric surface-radiation peak at $N/m \\approx 1.67$; TE$_0$–TE$_1$ forward/backward coupling peaks at $N/m \\approx 0.13$ and $1.87$; and essentially no excess loss for $N/m \\geq 2$, including the Bragg backscattering condition at $N/m = 2$. 3D FDTD simulations reproduce the main features, and radiation-angle and phase-mismatch calculations assign each peak to a physical mechanism. The paper then transforms this $N/m$ spectrum into a wavelength-dependent loss spectrum for a fixed grating period $\\Lambda = 730$ nm, showing which wavelengths of a broadband nonlinear process will suffer degraded $Q$.","pith_inferences":["If the N/m scaling is universal, the same loss spectrum should reappear in other PhCR geometries and material systems once rescaled by their effective index, so the peak positions could serve as a design-rule table across platforms, a transfer the paper does not itself demonstrate.","The mapping from fixed-wavelength multi-device data to a broadband spectrum assumes loss depends only on N/m, with dispersion entering only through the effective index; measuring a single fixed-Λ device at several wavelengths would test this assumption directly.","The near-symmetric placement of peaks around N/m = 1 (0.13/1.87, 0.33/1.67) hints at a reciprocal phase-matching picture in which forward and backward couplings mirror each other, so a coupled-mode theory could likely predict the whole spectrum from one side.","In high-power comb generation, the broad OAM loss region could act as a built-in limiter on parasitic modes, potentially improving soliton stability by damping unwanted resonances, a consequence the paper suggests but does not quantify."],"forward_implications":["A fixed-Λ PhCR spanning a wide bandwidth will have degraded Q at wavelengths corresponding to N/m ≈ 0.33 and across the broad OAM loss region up to N/m ≈ 1.7, raising thresholds and lowering conversion efficiency in OPO and four-wave mixing Bragg scattering.","The Bragg mode-splitting condition at N/m = 2 carries negligible excess loss, so it remains a safe choice for phase matching long-wavelength modes.","Loss-aware design can avoid the loss channels by applying mode splitting to the shortest-wavelength mode instead, since no excess loss is seen for N/m > 2.","The broad excess-loss region can be used deliberately to suppress competing nonlinear processes by aligning unwanted wavelengths with it.","The positions of loss peaks in N/m are stable as the grating modulation amplitude increases; only the overall loss magnitude grows, and small TE0–TM0 coupling peaks appear at higher amplitudes."],"supporting_citations":[{"why":"Supplies the OAM-mode ejection-angle model and the vertical emission condition that identify the broad loss region centered at N/m = 1.","marker":"[29]"},{"why":"Supplies the phase-mismatch definition and Bragg-condition formalism used to assign peaks to TE0–TE1 and TE0–TE0 coupling.","marker":"[41]"},{"why":"Demonstrates the N/m = 2 wavenumber-selective mode splitting that the paper shows is loss-free and uses as a phase-matching anchor for nonlinear processes.","marker":"[14]"},{"why":"Provides the multiple-mode splitting engineering context that motivates the loss-aware design guidance.","marker":"[30]"},{"why":"Exemplifies optical parametric oscillation in PhCRs, the application whose signal wavelength can overlap the N/m ≈ 0.33 loss peak.","marker":"[12]"},{"why":"Supplies the Snell's-law relation used to compute the OAM radiation angle from the internal ejection angle.","marker":"[45]"},{"why":"Gives the threshold-power scaling with cavity quality factor used to quantify the performance impact of the identified loss channels.","marker":"[46]"}],"fun_headline_variants":["Grating loss in microrings collapses to N/m ratio","Full loss map for photonic crystal microrings","One ratio predicts all grating loss peaks in microrings","Microring loss spectrum reveals design trade-offs","N/m ratio governs loss in photonic microring gratings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The broadband spectrum for a fixed device is obtained by reparametrizing loss measured at a single wavelength (1550 nm) across many devices with different grating periods, assuming the loss depends only on the ratio $N/m$ and that wavelength enters only through the effective index $n_{\\mathrm{eff}}$; if grating-induced loss changes with absolute wavelength through dispersion, cladding contrast, or mode-field redistribution, the predicted spectrum would be inaccurate.","fun_headline_variants_meta":{"raw":{"variants":["Grating loss in microrings collapses to N/m ratio","Full loss map for photonic crystal microrings","One ratio predicts all grating loss peaks in microrings","Microring loss spectrum reveals design trade-offs","N/m ratio governs loss in photonic microring gratings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000649,"raw_usage":{"total_tokens":3044,"prompt_tokens":1074,"completion_tokens":1970,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":690,"completion_tokens_details":{"reasoning_tokens":1889}},"tokens_in":690,"tokens_out":1970,"duration_ms":15396,"temperature":1.0,"reasoning_tokens":1889,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:25:46.788508+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure or simulate the same fixed-$\\Lambda$ PhCR (for example $\\Lambda = 730$ nm) at several different free-space wavelengths across the 400–2470 nm range and check whether the loss peaks appear at the $N/m$ positions predicted from the 1550 nm data; if the peaks shift, split, or change relative strength with wavelength, the claim that loss depends only on $N/m$ fails.","supporting_citations":[{"cited_title":"Highly- twisted states of light from a high quality factor photonic crystal ring,","cited_arxiv_id":null,"evidence_quote":"Supplies the OAM-mode ejection-angle model and the vertical emission condition that identify the broad loss region centered at N/m = 1."},{"cited_title":"Yariv and P","cited_arxiv_id":null,"evidence_quote":"Supplies the phase-mismatch definition and Bragg-condition formalism used to assign peaks to TE0–TE1 and TE0–TE0 coupling."},{"cited_title":"Wavelength-accurate nonlinear con- version through wavenumber selectivity in photonic crys- tal resonators,","cited_arxiv_id":null,"evidence_quote":"Demonstrates the N/m = 2 wavenumber-selective mode splitting that the paper shows is loss-free and uses as a phase-matching anchor for nonlinear processes."},{"cited_title":"Universal frequency engineering tool for micro- cavity nonlinear optics: multiple selective mode splitting of whispering-gallery resonances,","cited_arxiv_id":null,"evidence_quote":"Provides the multiple-mode splitting engineering context that motivates the loss-aware design guidance."},{"cited_title":"Optical-parametric oscillation in photonic-crystal ring resonators,","cited_arxiv_id":null,"evidence_quote":"Exemplifies optical parametric oscillation in PhCRs, the application whose signal wavelength can overlap the N/m ≈ 0.33 loss peak."},{"cited_title":"Diffraction- less propagation beyond the sub-wavelength regime: a new type of nanophotonic waveguide,","cited_arxiv_id":null,"evidence_quote":"Supplies the Snell's-law relation used to compute the OAM radiation angle from the internal ejection angle."},{"cited_title":"Milliwatt-threshold visible–telecom optical parametric oscillation using silicon nanophotonics,","cited_arxiv_id":null,"evidence_quote":"Gives the threshold-power scaling with cavity quality factor used to quantify the performance impact of the identified loss channels."}],"review_version":1}