{"id":"6c451d6b-9c2f-4a88-8f23-bba72c404dcc","arxiv_id":"2507.21272","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A new microresonator design generates efficient, flat, low-threshold normal-dispersion light combs on X-cut lithium niobate, plus a novel combined Kerr-Raman comb state spanning 33 THz.","lead":"Researchers made a tiny chip-based light source that produces a broad, flat frequency comb using lithium niobate, a material that also handles fast electrical signals. The comb is efficient, low-power, and could let a single chip both generate and process light for communications.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The interleaved-comb claim rests on a gLLE simulation whose Raman fraction is not independently constrained; no direct measurement of distinct CEO offsets is presented.","rationale":"The reader's weakest assumption (gLLE with a single scalar field and fitted f_R) is essentially correct, but I would locate the load-bearing link one step further: the defining property of the new state—distinct CEO offsets—is observable only in the simulation's temporal walk-off, not in the measured spectra. If the simulation's Raman parameters are free, the 'interleaved' interpretation is not independently established. This is not an accusation of fitting-to-match; it is a statement that the paper does not show how f_R was obtained, so the agreement in Fig. 4f cannot be distinguished from a parameterized fit. The concern does not invalidate the demonstrated high-efficiency normal-dispersion comb (Fig. 3), which stands on its own microwave beatnote and efficiency data. It does affect the second, more novel claim that is highlighted in the abstract. Since the reader's verdict is already CONDITIONAL, my concern supports that verdict but does not move it to rejection: the normal-dispersion comb results are credible, and the new state could be real, but it needs independent verification. Hence verdict_should_be = UNCHANGED. I partially agree with the reader: they flagged the same modelling concern, though I emphasize the unmeasured CEO offset as the specific quantity that must be verified.","tokens_in":12250,"tokens_out":5563,"duration_ms":74309,"concrete_test":"Measure the Raman response of the 25.7 GHz-FSR racetrack independently—e.g., by pump-probe stimulated Raman spectroscopy or spontaneous Raman scattering from the fabricated waveguide—to fix f_R and the Raman lineshape. Rerun the gLLE simulation of Figs. 4b/f with these independently obtained parameters and no free Raman fitting. If the simulated state no longer shows the interleaved temporal drift (Fig. 4h inset), the CEO-offset claim is unsupported; if it does, the mechanism is confirmed. As a cross-check, attempt an optical heterodyne between the pump comb and the Stokes comb: two interleaved combs should produce a stable RF beat at the CEO-offset difference, whereas a single comb with Raman sidebands will not.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's most novel claim—a new normal-dispersion Kerr–Raman state consisting of two interleaved combs with distinct carrier-envelope offset frequencies (Figs. 4f–i)—is not directly evidenced. The measured data (OSA spectra, clean beatnote) show a broad spectrum around ~1592 nm and a Stokes feature near ~1762 nm, but an OSA cannot resolve whether the Stokes lines belong to a separate comb with a different CEO offset or are phase-locked Raman sidebands of the same comb. The temporal walk-off in Fig. 4h inset, cited as evidence for distinct CEO offsets, comes from gLLE simulations. The Raman fraction f_R used in that model is described as chosen for each device, and no independent measurement of f_R or the Raman lineshape for the 25.7 GHz-FSR racetrack is presented. If f_R is adjusted to reproduce the observed spectrum, the simulated drift is not an independent confirmation. This matters because the claim of a microcomb 'mechanically distinct from Stokes solitons' is the central conceptual advance; without heterodyne or independent Raman characterization, the observed state could instead be a single frequency comb with strong Raman-shifted spectral content.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports the generation of normal-dispersion Kerr microcombs in racetrack microresonators on X-cut thin-film lithium niobate. The authors engineer avoided mode crossings through non-adiabatic waveguide tapers, achieving high intrinsic Q factors (up to ~15 million) and demonstrating a 102.4 GHz repetition-rate comb with broad bandwidth (>24 THz), pump-to-comb conversion efficiency up to 54%, and low on-chip turn-on power (~8 mW). They also present a 25.7 GHz-FSR device in which a normal-dispersion Kerr comb coexists with a strong Stokes-shifted feature; based on gLLE simulations with a fitted Raman fraction, they interpret this as a new state consisting of two interleaved frequency combs with distinct carrier-envelope offset frequencies. The paper includes measurements of optical linewidths showing 'quiet' modes with reduced Lorentzian linewidth relative to the pump.","tokens_in":12391,"tokens_out":12592,"duration_ms":143255,"significance":"The experimental results are potentially significant for integrated photonics: normal-dispersion combs on an EO-compatible platform with high efficiency and low threshold would enable monolithic comb sources with on-chip EO processing. The AMX engineering via non-adiabatic tapers appears robust and device-reproducible, and the linewidth noise scaling analysis (Fig. 3(e)) is a valuable addition to the literature on comb coherence. The claim of a new interleaved Kerr–Raman comb state, if validated, would be a novel nonlinear state. However, the current evidence for this state is incomplete, and the simulations used to support it rely on a free parameter (Raman fraction f_R) that is not independently constrained.","major_comments":[{"comment":"The central claim that the stable state in Fig. 4(f) consists of two interleaved frequency combs with distinct carrier-envelope offsets is not directly evidenced. The temporal walk-off shown in Fig. 4(h) is a simulation result, not a measurement, and walk-off between envelope and carrier oscillations can occur for a single comb with dispersion; it does not by itself imply two distinct CEO offsets. The optical spectra in Fig. 4(f) cannot resolve whether the Stokes feature belongs to a separate comb or is part of the same phase-locked comb. The clean f_FSR beatnote in Fig. 4(g) does not exclude a beatnote at a sub-FSR frequency (e.g., at the offset Δf = Ω_R mod f_FSR), which would be expected if two combs with different CEO were present; such sub-FSR measurements are not reported. To support the distinct-CEO claim, the authors should either (i) measure the RF spectrum below f_FSR to search for a modulation at the expected offset, or (ii) perform heterodyne measurements that directly reveal the CEO difference, or (iii) explicitly soften the claim to a single comb with Raman-shifted spectral content.","section":"Normal-dispersion Kerr and Raman Stokes microcombs; Figs. 4(f)–(i)"},{"comment":"The generalized Lugiato–Lefever simulations used to corroborate the interleaved-comb state rely on a Raman fraction f_R that is 'chosen for each device' without independent measurement. The agreement between the simulated spectra (red curves in Figs. 3(b), 4(b), 4(f)) and experiments is therefore not an independent confirmation of the underlying Raman dynamics or of the distinct-CEO interpretation. The authors should report the f_R values used, demonstrate the sensitivity of the predicted state to f_R, and, if possible, provide independent Raman spectroscopy of the same waveguides to constrain f_R. Without this, the simulation's prediction of a separate CEO-offset comb remains a hypothesis rather than a validated mechanism.","section":"Results; gLLE simulations (Figs. 3(b), 4(b), 4(f)) and Discussion"}],"minor_comments":[{"comment":"The paper should specify how the Lorentzian linewidths are extracted from the frequency noise PSDs and provide an error estimate for the quoted values; the current presentation gives no uncertainty on the linewidths or on the quadratic fit.","section":"Fig. 3(e) and linewidth measurements"},{"comment":"The abstract contains an extra space in 'e fficient', and the manuscript has several spacing and formatting artifacts (likely from typesetting); a careful proofread is needed.","section":"Abstract and general text"},{"comment":"The term 'quiet modes' is used without definition or a pointer to the reference; please clarify the meaning at first use.","section":"Notation, first use of 'quiet modes'"},{"comment":"The paper references a Supplementary note for details of the efficiency comparison and linewidth measurements, but this material is not included in the arXiv submission; it should be available to reviewers to assess the quantitative claims.","section":"Supplementary note availability"},{"comment":"The electro-optically divided beatnote frequency is given as 102.4509 GHz; the division factor and the RF drive frequency should be stated for reproducibility.","section":"Fig. 3(c) beatnote division"},{"comment":"The phrase 'mechanically distinct' from Stokes solitons is ambiguous; clarify whether it refers to the generation mechanism, phase-locking, or another physical distinction.","section":"Introduction, 'mechanically distinct'"}],"recommendation":"major_revision","confidential_remarks":"The paper is technically solid in its experimental core: high-Q X-cut TFLN resonators, controllable AMX via non-adiabatic tapers, and careful comb characterization. The main obstacle to acceptance is the over-interpretation of the Kerr–Raman state as two interleaved combs with distinct CEO offsets without direct experimental evidence. The required experiment—a search for RF beatnotes below f_FSR—is straightforward and decisive. If the authors cannot provide such evidence, the claim should be softened to a spectrally extended Raman-assisted comb; the paper would still be publishable as a strong demonstration of efficient normal-dispersion combs on an EO platform, but the novelty claim would be reduced. The simulation parameter f_R should be reported and constrained, otherwise the simulation corroboration carries little weight."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper earns its place. It reports the first normal-dispersion Kerr microcombs on X-cut TFLN, with Q factors around 5–7.5 million (up to ~15M), pump-to-comb efficiency above 50%, and turn-on below 10 mW. The 102.4 GHz comb spectra are flat across C-L bands, the beatnote is clean, and the linewidth data follow the expected parabolic quiet-mode pattern. For an experiment this demanding, the core claims are credible and mutually consistent. That alone makes it a serious advance.\n\nThe new physics claim—the interleaved Kerr-Raman two-comb state—is the part I'd flag. The evidence in the paper is an OSA spectrum showing a primary comb around 1592 nm and a Stokes feature around 1762 nm, a clean beatnote, and a gLLE simulation that shows temporal walk-off between the pulse envelope and the Raman oscillations. The walk-off is what implies different carrier-envelope offset frequencies. But that walk-off is simulated, not measured. The Raman fraction f_R is chosen per device, and no independent Raman lineshape or heterodyne measurement is presented. So the interpretation that this is a state mechanically distinct from Stokes solitons is plausible but not demonstrated. I agree with your skeptic on this point. The authors should provide either a two-comb interferometry measurement, a heterodyne of the CEO offset, or at minimum a sensitivity analysis showing the interleaved state is not an artifact of the particular f_R choice.\n\nThere's also the usual lack of error bars on Q, efficiency, and linewidth, and the state-of-the-art comparison is in the Supplementary. Minor, but a referee should ask for the comparison table in the main text or at least a clear statement of which numbers are single devices.\n\nThe paper is honest about its limitations: it says the simulations are tuned to salient parameters, and it shows the chaotic-to-stable transition as computational. The citation pattern is fine—they cite the relevant normal-dispersion and Raman literature, including their own prior soliton work, appropriately. Nothing looks circular.\n\nBottom line: I'd send this to peer review. The referee effort is worth it. The interleaved-comb interpretation needs more direct evidence, but the experimental core—normal-dispersion combs on X-cut TFLN with record performance—will stand regardless.","headline":"First normal-dispersion Kerr microcombs on X-cut TFLN with record Q and >50% efficiency; the new interleaved Kerr-Raman state is plausible but needs direct CEO-offset evidence.","tokens_in":13025,"tokens_out":2585,"would_cite":true,"duration_ms":30210,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper reports the first normal-dispersion Kerr microcombs on X-cut thin-film lithium niobate, achieving 54% pump-to-comb conversion efficiency and discovering a coherent two-interleaved-comb state from Kerr-Raman cooperation.","keywords":["microcombs","Kerr frequency combs","X-cut thin-film lithium niobate","normal dispersion","stimulated Raman scattering","avoided mode crossings","integrated photonics","electro-optic compatibility"],"falsifier":"Measure the Raman spectrum of each fabricated racetrack directly, extract $f_R$ independently, and then check whether the stable interleaved Kerr-Raman comb appears at the predicted pump detuning and threshold without re-fitting $f_R$; if the state appears independent of $f_R$, or if the simulated spectra only match after freely adjusting $f_R$ per device, the proposed Kerr-Raman synergy would not be confirmed.","tokens_in":11988,"feed_emoji":"💡","tokens_out":9196,"duration_ms":89760,"temperature":0.7,"pith_summary":"This paper reports the first normal-dispersion Kerr microcombs on X-cut thin-film lithium niobate (TFLN), the integrated platform prized for its electro-optic response. By orienting racetrack resonators to minimize the material's anisotropic Raman gain and inserting a non-adiabatic taper that supplies avoided mode crossings, the authors obtain combs with 54% pump-to-comb conversion efficiency, over 24 THz of spectral bandwidth, and roughly 8 mW turn-on power. They also identify and simulate a new kind of coherent microcomb in which normal-dispersion Kerr dynamics and stimulated Raman scattering cooperate, producing two interleaved combs centered on the pump and Stokes frequencies and spanning nearly 33 THz. If these results hold, a single X-cut TFLN chip could combine the comb source with high-speed electro-optic modulation, removing a major obstacle to monolithic microcomb-driven systems.","feed_headline":"Microcombs on X-cut lithium niobate reach 54% pump-to-comb efficiency","feed_subtitle":"Racetrack design tames Raman noise, a non-adiabatic taper seeds flat combs and a new interleaved Kerr-Raman state.","key_machinery":"The key machinery is a racetrack microresonator whose orientation relative to the X-cut TFLN crystal axes lowers the round-trip Raman fraction $f_R$, the ratio of Raman to Kerr nonlinear response, combined with a rapidly tapering non-adiabatic waveguide section that converts a controlled fraction of the TE0 mode into the TE1 mode. The taper creates a localized avoided mode crossing in the dispersion profile, which locally perturbs the resonance frequencies and seeds normal-dispersion Kerr comb formation under continuous-wave driving. A generalized Lugiato-Lefever equation with a Lorentzian Raman response and experimentally tuned parameters reproduces the observed spectra and identifies the temporal waveforms, including the picosecond quasi-rectangular pulse with rapid oscillations at the Stokes shift that underlies the interleaved Kerr-Raman comb.","core_discovery":"The central claim is that X-cut TFLN, despite its strong and complex Raman response, can support high-performance normal-dispersion Kerr microcombs when the resonator geometry is used to suppress Raman gain and to engineer avoided mode crossings. Specifically, a 102.4 GHz-FSR racetrack microresonator yields a comb with 19 lines in the C-band and 14 in the L-band with less than 2 dB variation, a measured pump-to-comb conversion efficiency of 54%, and an on-chip turn-on power of about 8 mW; linewidth measurements reveal quiet modes whose intrinsic linewidths fall below that of the pump, with the quietest mode blue-shifted from the pump. In a second, 25.7 GHz-FSR device, the paper reports a stable coherent state comprising two interleaved frequency combs, one around the pump at roughly 1592 nm and one around the Stokes-shifted wavelength at roughly 1762 nm, with the same repetition rate but different carrier-envelope offset frequencies and a total span of nearly 33 THz. The paper argues this state is mechanically distinct from Stokes solitons and from two-pumped cavity solitons, and that it emerges from the synergetic interplay of Kerr and stimulated Raman scattering in a single spatial mode.","pith_inferences":["If the racetrack orientation genuinely controls $f_R$ with the repeatability claimed, the same design rule should transfer to other anisotropic crystalline platforms such as lithium tantalate, giving a generic knob for balancing Raman and Kerr effects.","Because the two interleaved combs share a repetition rate but differ in carrier-envelope offset, the state is a single-resonator dual-comb source; it may be usable for dual-comb spectroscopy or as a self-referencing seed, though the paper does not demonstrate those uses.","The linewidth parabola's shift implies that engineering the collective dispersive-wave recoil from avoided mode crossings could be used as a design lever to place quiet comb lines at desired wavelengths, an extension the paper hints at but does not pursue.","A direct test of the proposed mechanism would be to vary the taper-induced avoided-mode-crossing strength and the racetrack orientation independently and check that the onset of the interleaved state tracks the simulated $f_R$ threshold."],"forward_implications":["A single X-cut TFLN chip can host both the microcomb source and high-speed electro-optic modulators, eliminating chip-to-chip coupling losses in comb-driven systems.","Normal-dispersion combs on this platform inherit the flat spectral profile, high conversion efficiency, and low turn-on power needed for WDM communications and signal processing.","The new interleaved Kerr-Raman comb state shows that stimulated Raman scattering can extend a microcomb's spectrum into widely separated wavelength bands rather than simply destroying it.","The measured quiet-mode linewidths indicate that pump noise transduction in normal-dispersion combs follows the same general scaling as in anomalous-dispersion soliton combs, with the quietest modes shifted from the pump.","Suppressing parasitic Raman effects may also enable other Kerr-nonlinearity-based devices, such as parametric oscillators and optical isolators, on the same platform."],"supporting_citations":[{"why":"Anisotropic Raman response measurements and prior X-cut soliton comb work that the racetrack orientation exploits to suppress Raman gain.","marker":"42,43"},{"why":"Demonstrates Raman lasing and soliton mode-locking in lithium niobate microresonators, establishing how parasitic SRS can suppress modulation-instability comb generation.","marker":"44"},{"why":"Supplies the avoided-mode-crossing mechanism that locally perturbs dispersion and is needed to seed normal-dispersion comb formation; the paper extends it with a non-adiabatic taper.","marker":"54–60"},{"why":"Defines Stokes solitons in the anomalous-dispersion regime, the closest known state to which the new interleaved Kerr-Raman comb is compared and distinguished.","marker":"49"},{"why":"Describes spectrally extended microcombs generated with more than one pump, the other comparison for the two interleaved combs.","marker":"50,51"},{"why":"Provides the linewidth and frequency-noise characterization method used to identify quiet modes and the parabolic noise distribution.","marker":"46"},{"why":"Gives the state-of-the-art electro-optic comb performance on lithium niobate that the new normal-dispersion Kerr combs surpass in efficiency, bandwidth, or turn-on power.","marker":"37–40"}],"fun_headline_variants":["X-cut lithium niobate microcombs hit 54% efficiency","Normal-dispersion microcombs on X-cut lithium niobate","Interleaved Kerr-Raman comb on X-cut TFLN","X-cut lithium niobate microresonator comb breaks 54%","High-efficiency microcombs from X-cut lithium niobate"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's interpretation rests on the assumption that a single-scalar-field generalized Lugiato-Lefever equation with a Lorentzian Raman response, whose Raman fraction is chosen for each device, faithfully captures the strongly multimode, anisotropic dynamics of the real X-cut TFLN racetrack resonators.","fun_headline_variants_meta":{"raw":{"variants":["X-cut lithium niobate microcombs hit 54% efficiency","Normal-dispersion microcombs on X-cut lithium niobate","Interleaved Kerr-Raman comb on X-cut TFLN","X-cut lithium niobate microresonator comb breaks 54%","High-efficiency microcombs from X-cut lithium niobate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000884,"raw_usage":{"total_tokens":3923,"prompt_tokens":1158,"completion_tokens":2765,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":774,"completion_tokens_details":{"reasoning_tokens":2670}},"tokens_in":774,"tokens_out":2765,"duration_ms":20195,"temperature":1.0,"reasoning_tokens":2670,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:56:31.397745+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the Raman spectrum of each fabricated racetrack directly, extract $f_R$ independently, and then check whether the stable interleaved Kerr-Raman comb appears at the predicted pump detuning and threshold without re-fitting $f_R$; if the state appears independent of $f_R$, or if the simulated spectra only match after freely adjusting $f_R$ per device, the proposed Kerr-Raman synergy would not be confirmed.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates Raman lasing and soliton mode-locking in lithium niobate microresonators, establishing how parasitic SRS can suppress modulation-instability comb generation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines Stokes solitons in the anomalous-dispersion regime, the closest known state to which the new interleaved Kerr-Raman comb is compared and distinguished."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the linewidth and frequency-noise characterization method used to identify quiet modes and the parabolic noise distribution."}],"review_version":1}