Pith. sign in

REVIEW 3 major objections 4 minor 21 references

Photonic Crystal Microring Resonators on a Hybrid Silicon Nitride-on-Lithium Niobate Platform

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A periodic corrugation on a silicon-nitride microring over lithium niobate creates a 14.6 GHz split between two optical supermodes, tunable by voltage at 0.85 pm/V.

desk verdict A plausible engineering first for PhCRs on SiN-on-LNOI, but the central mode-splitting claim rests on an unresolved mode identification and the paper overstates its frequency-conversion readiness. read the letter →

arxiv 2505.00678 v1 pith:DYP4ZW5W submitted 2025-05-01 physics.optics physics.app-ph

Zhongdi Peng , Rakesh Krishna , Xi Wu , Amir H. Hosseinnia , Tianren Fan , Ali Adibi This is my paper · ORCID
classification physics.opticsphysics.app-ph
keywords photoniccrystalresonatormicroringlithiumniobatesiliconnitrideelectro-opticeffectmodesplittingfrequencyconversionintegratedphotonics
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes a single, corrugated microring as the working element of a microwave-assisted optical frequency converter. It claims that a periodic ripple on the inner edge of a silicon-nitride ring sitting on lithium niobate splits each resonance into two supermodes whose frequency separation is set by the ripple depth, and that an applied voltage shifts the pair together through the lithium-niobate electro-optic effect. The measured device shows a 14.6 GHz splitting with an intrinsic quality factor of 1.47e5, a voltage tuning rate of 0.85 pm/V, and a splitting-versus-corrugation slope of 93.4 MHz/nm, all achieved without etching the lithium niobate. If the claims hold, this is a simpler, CMOS-compatible route to bidirectional frequency conversion than paired coupled resonators, because the corrugation replaces the mode-matching and gap-control needed in a two-resonator converter.

What carries the argument

The central object is the photonic-crystal resonator (PhCR): a microring whose radius is modulated as $r = r_0 + A_n \cos(n\phi)$, where $n$ is the number of corrugation periods. This ripple couples the otherwise degenerate clockwise and counterclockwise modes of the ring, splitting them into two supermodes separated by $\beta_m = k\,\omega_m A_n$, with $k$ an overlap integral over the corrugation boundary. That linear dependence is what lets the device designer choose the conversion frequency by picking $A_n$. The active mechanism is the X-cut lithium-niobate layer underneath the unetched silicon nitride: electrodes spaced along the crystal $c$-axis use the largest electro-optic coefficient ($r_{33} = 30$ pm/V) to shift both supermodes together at 0.85 pm/V while leaving $\beta_m$ essentially unchanged.

What would settle it

Drive the fabricated $A = 150$ nm PhCR with an over-coupled waveguide, inject light at one split resonance, apply a microwave tone at the 14.6 GHz splitting frequency, and look for converted light at the other split resonance. If no bidirectional sideband appears under the matching conditions the paper cites from its reference [16], or if the two dips do not remain a single constant-separation pair across the full spectrum, then the mode-splitting claim is not sufficient for the intended frequency converter.

Watch

Extended reading notes

Core claim

The central discovery is that the hybrid silicon-nitride-on-lithium-niobate platform can support a high-quality photonic-crystal resonator whose mode splitting is both large and precisely controlled by geometry. The authors report the first PhCR on this platform that lifts the degeneracy of clockwise and counterclockwise modes to form a pair of supermodes, demonstrating a splitting bandwidth of 14.6 GHz while keeping the intrinsic quality factor at 1.47e5, close to the 1.69e5 of an unperturbed ring on the same chip. They further show that the split resonances shift together under an applied DC voltage at 0.85 pm/V, giving voltage control of the converter operating frequency, and that the splitting grows linearly with corrugation amplitude at 93.4 MHz/nm so the intended frequency shift can be designed in advance. The platform avoids etching lithium niobate altogether, patterning only the deposited silicon-nitride layer, which sidesteps the poor sidewall angles of etched LNOI waveguides.

Load-bearing premise

The load-bearing assumption is that the two dips seen near 1620 nm are the two split directions of one and the same ring mode; if they are instead two unrelated modes or a polarization artefact, then the reported splitting does not describe a usable supermode pair.

Editorial extensions

If this is right

  • A designer can set the converter frequency shift in advance by choosing the corrugation amplitude, using the measured 93.4 MHz/nm slope.
  • Voltage tuning at 0.85 pm/V lets the same device be re-centred onto different input wavelengths without changing the ring geometry.
  • Because only silicon nitride is etched, the fabrication stays CMOS-compatible and avoids the sloped-sidewall loss of etched lithium-niobate waveguides.
  • The splitting of 14.6 GHz with an intrinsic quality factor near 1.47e5 is large enough for practical bidirectional frequency conversion, and the near-unity efficiency condition identified for coupled-resonator converters transfers to a single PhCR.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the linear scaling in $\beta_m = k\,\omega_m A_n$ holds beyond $A = 150$ nm, narrower waveguide widths or deeper corrugations should push the splitting well past 14.6 GHz; the paper hints at this route but does not measure it.
  • Because the X-cut lithium niobate's anisotropy produces a range of split modes at large $A$, a practical broadband converter will likely need the gradient-compensated PhCR design described in the supplement, not the plain periodic device measured in the main text.
  • The same corrugation-on-silicon-nitride recipe could be transferred to other electro-optic films beneath the silicon nitride, such as thin-film aluminum nitride or barium titanate, to trade the 0.85 pm/V tuning rate for different material properties; this is an extension the paper does not pursue.
  • A direct microwave-driven conversion experiment at $\beta_m$ would be the decisive test of the bidirectional frequency-converter claim, since the reported measurements demonstrate the split modes and their voltage tuning but stop short of showing converted output light.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The manuscript reports a hybrid SiN-on-LNOI microring resonator with a periodic sidewall corrugation (a photonic-crystal resonator, PhCR) that is intended to lift the degeneracy of clockwise and counterclockwise modes into a pair of supermodes. The authors claim a supermode splitting bandwidth of 14.6 GHz with an intrinsic quality factor of 1.47×10^5, a voltage-driven resonance shift of 0.85 pm/V, and a linear dependence of the splitting bandwidth on corrugation amplitude with a slope of 93.4 MHz/nm. They position this as the first demonstration of precise mode-splitting control for bidirectional electro-optic frequency conversion on a hybrid SiN-on-LNOI platform, achieved without etching lithium niobate. The paper presents transmission spectra of a corrugated and an uncorrugated ring, DC electro-optic tuning of the split resonances, and a set of transmission measurements at different corrugation amplitudes to support the linear slope. It also points to Supplement 1 for additional spectra and gradient-design simulations. No direct frequency-conversion experiment (e.g., with an applied RF tone and detection of the converted optical frequency) is reported.

Significance. If the mode-splitting identification is sound, this work offers a practically appealing route to integrated electro-optic frequency conversion: it avoids etching lithium niobate, uses CMOS-compatible SiN patterning, and provides a geometric knob (corrugation amplitude) to set the conversion shift. The comparison between the Q factor of the corrugated ring (1.47×10^5) and an uncorrugated reference ring (1.69×10^5) is a useful and reassuring data point, showing that the corrugation introduces little extra loss. The EO tuning demonstration is also concrete and directly relevant to the frequency-conversion concept. However, the verification of the central physical picture is currently incomplete: the paper does not positively prove that the observed doublet is the CW/CCW supermode pair of a single azimuthal mode, and it does not demonstrate the frequency-conversion process itself. The reported four-point slope without uncertainties further weakens the quantitative control claim. The platform concept is valuable, but the current evidence is not yet sufficient to establish the advertised level of control and the converter demonstration.

major comments (3)
  1. [Section 3 (Fig. 3) and Section 5 (Fig. 5)] The identification of the two transmission dips in Fig. 3(a) as the clockwise/counterclockwise supermodes of a single fundamental TE azimuthal mode is load-bearing for all subsequent claims, but it is not directly supported. The paper does not provide polarization-resolved transmission, a full free-spectral-range mode map, or an eigenmode simulation that includes the X-cut LN anisotropy. The manuscript itself acknowledges in the discussion of Fig. 5 that the broad transmission spectrum of the A = 150 nm device shows a range of split modes due to the anisotropic LN layer (Supplement 1, Fig. S1), which means the clean two-dip reading near 1620 nm may be mode-selective rather than representative of the entire resonator. If the two dips are adjacent azimuthal modes or TE/TM hybrid modes, then the quoted splitting bandwidth and the linear slope do not arise from corrugation-induced CW/CCW coupling, and Eqs. (1)–(2) do not describe the data. I ask the authors to either provide a full FSR trace with the azimuthal order of the split pair identified, a polarization-resolved measurement, or a simulation of the anisotropic structure that reproduces the observed doublet.
  2. [Section 5 (Fig. 5)] The linear slope of 93.4 MHz/nm is fitted to only four points (A = 0, 50, 100, and 150 nm, based on the spectra in Figs. 5(d)–(g)) and is shown without error bars. The same azimuthal index m = 2106 is asserted for all devices, but no method for determining m is given. A four-point fit without uncertainties and without verification that the same mode is tracked across devices does not establish the claimed linear relation, especially since the mode assignment is itself the concern. Please include additional corrugation amplitudes, error bars or confidence intervals from repeated measurements, and an explicit procedure for confirming the azimuthal mode index.
  3. [Title, Abstract, and Conclusion] The paper claims to demonstrate bidirectional frequency conversion, but no frequency-conversion experiment is shown. The reported experiments are passive transmission spectra (Fig. 3), DC electro-optic tuning of the resonance wavelengths (Fig. 4), and corrugation-dependent splitting (Fig. 5). There is no application of a microwave/RF signal at the splitting frequency and no measurement of output power at the converted frequency. As written, the manuscript demonstrates the components of a converter (split modes, EO tunability, splitting control) but not the conversion process. The claim should be softened to a statement about enabling such conversion, or a direct conversion measurement with efficiency should be added.
minor comments (4)
  1. [Fig. 5] The corrugation amplitudes are labeled inconsistently: the SEM images in Figs. 5(a)–(c) are described as A = 30, 60, and 150 nm, while the spectra in Figs. 5(e)–(g) are described as A = 50, 100, and 150 nm. Please harmonize the notation.
  2. [Eq. (2)] The notation in Eq. (2) is not fully defined: the meaning of the integration variable (presumably the angle along the ring), the boundary dS, and the permittivity contrast terms should be stated explicitly for readers who do not have access to ref. [18].
  3. [Section 4 (Fig. 4)] The reported EO tuning rate of 0.85 pm/V is the average shift of the two split resonances. Please report the shift of each resonance separately and indicate whether the response is linear over the full voltage range; this is relevant for the later claim of controllable frequency-conversion operation.
  4. [General] The phrase 'photonic-crystal resonator' is used for a sidewall-corrugated ring, which is a specific subclass; a brief comment on how this relates to conventional 2D photonic-crystal cavities would improve clarity for a general optics readership.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported values are direct measurements and empirical fits; the cited perturbation formula is background, not used to generate the claimed results.

full rationale

The paper's central quantitative claims—supermode splitting of 14.6 GHz, intrinsic Q of 1.47e5, voltage tuning of 0.85 pm/V, and the 93.4 MHz/nm splitting-versus-amplitude slope—are obtained from measured transmission spectra and DC voltage sweeps, not from a model that was itself fitted to those same quantities. Equation (1)-(2) is cited from prior work (including the authors' own ref. [18]) as the theoretical motivation for linear splitting versus corrugation amplitude, but the paper does not use it to predict the measured values; Fig. 5(h) explicitly reports a 'fitted slope.' Thus there is no constructed equivalence between inputs and outputs, and no fitted parameter is renamed as a prediction. The only self-citation, ref. [18], is not load-bearing for the empirical demonstration, and the cited formula is also attributed to external ref. [17]. The assumption that the two transmission dips correspond to the CW/CCW supermodes of a single azimuthal mode is an interpretive step, and the paper itself notes that the broad spectrum contains a range of split modes due to the X-cut LN anisotropy, but that is a correctness/robustness concern, not a circularity. Overall, the derivation chain is self-contained as an experimental characterization.

Assumptions & free parameters 2 free parameters · 2 assumptions · 0 invented entities

The central claims rest on measured device performance and on a prior perturbative formula for mode splitting. No new physical entities are postulated. The main free parameters are the empirical slopes extracted from the authors' own data, not hidden parameters of a new theory.

free parameters (2)
  • Mode-splitting slope k = 93.4 MHz/nm
    Obtained from a linear fit to three nonzero corrugation amplitudes plus A=0 in Fig. 5(h). No error bars or independent prediction from Eq. (2) are provided.
  • EO tuning rate = 0.85 pm/V
    Average resonance shift of the two split modes versus applied DC voltage, measured on a separate 300 um radius ring with a 1.4 um wide waveguide (Fig. 4(c)). This is an empirical device response, not a material constant.
assumptions (2)
  • domain assumption Mode splitting follows Eq. (1) with k given by Eq. (2) from refs [17,18]
    Invoked without derivation in the paragraph after Eq. (1). Assumes the corrugation is a weak perturbation on an otherwise uniform resonator, and that the unperturbed fundamental TE fields dominate the overlap integral. The formula originates from silicon/silicon-nitride rings, and its validity for the hybrid SiN-LN cross-section is not independently checked.
  • domain assumption The two resonances in Fig. 3(a) are the CW and CCW supermodes of the same azimuthal mode
    The identification is inferred from the observation that an uncorrugated ring shows a single resonance while the corrugated ring shows a pair. The paper later notes a broad range of split modes due to X-cut LN anisotropy, which casts some doubt on whether every measured pair is a single azimuthal mode split rather than a selection from a complex spectrum.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Photonic Crystal Microring Resonators on a Hybrid Silicon Nitride-on-Lithium Niobate Platform." pith.science (2026). https://pith.science/paper/DYP4ZW5W

@misc{pith2026250500678,
  author       = {Pith},
  title        = {Pith review of: Photonic Crystal Microring Resonators on a Hybrid Silicon Nitride-on-Lithium Niobate Platform},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DYP4ZW5W}},
  note         = {Machine review of arXiv:2505.00678}
}
read the original abstract

Photonic-crystal resonators (PhCRs) have been widely used in nonlinear integrated photonics for frequency engineering applications. A microwave-assisted frequency converter based on PhCRs highlights its precise control of frequency (enabled by creation of a pair of supermodes by a corrugated PhCR) and bidirectional frequency conversion. In this paper, we demonstrate a high-quality PhCR on a hybrid silicon nitride-on-lithium niobate-on-insulator (SiN-on-LNOI) platform for the first time for voltage-driven flexible frequency conversion using the electro-optic effect (0.85 pm/V). The fabricated PhCR has a large supermode splitting bandwidth = 14.6 GHz and an intrinsic quality factor (Q) = 147,000. Using different periodic corrugation amplitudes in the fabricated PhCRs enables the precise control of mode splitting with a ratio of 93.5 MHz/nm between the mode splitting bandwidth and the corrugation amplitude.

Figures

Figures reproduced from arXiv: 2505.00678 by the authors.

Figure 1
Figure 1. Schematic of the PhCR and working principle of the PhCR-based frequency converter. (a) A PhCR with periodic geometric corrugation on the inner surface of the circumference of a microring resonator, leading to mode splitting with splitting bandwidth βm. (b) The PhCR in (a) can convert the frequency of light from one of the split resonance modes to the other with the coherent microwave excitation at the frequency diff… view at source ↗
Figure 2
Figure 2. (a) The SEM image of the hybrid platform, and (b) the axis-symmetric simulation of the fundamental TE mode using COMSOL Multiphysics. (c) The SEM image of the waveguide cross-section showing a sidewall angle of 85 ̊. (d) A top-view microscopy image of the PhCR with waveguide width = 1.8 µm, radius = 400 µm, corrugation amplitude A = 150 nm, and the number of corrugation period n = 5678; the highlighted regions refer… view at source ↗
Figure 4
Figure 4. (a) Microscopy image of the PhCR with radius = 300 µmand width of SiN waveguide = 1.4 µm. The zoom-in image shows two parallel bent electrodes with metal spacing = 7.2 µm. (b-c) Characterization resultsshow a shift of split modes with tunability of 0.85 pm/V and no significant change of βm. One advantage of our PhCRs is precise controllability of the splitting bandwidth using A. Figures 5(a-c) show the SEM images of… view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: The precise control of splitting bandwidth in PhCR resonators with waveguide width = 1.4 µm and bending radius = 250 µm. (a-c) The SEM images of the bent waveguides in PhCR with A= 30 nm, 60 nm, and 150 nm, respectively. Normalized transmission spectra of the resonant …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

  1. [1]

    H. Feng, T. Ge, X. Guo, et al., Nature 627, 80 (2024)

    work page 2024

  2. [2]

    Marpaung, J

    D. Marpaung, J. Yao, and J. Capmany, Nature photonics 13, 80 (2019)

    work page 2019

  3. [3]

    J. M. Lukens, H.-H . Lu, B. Qi, et al., Journal of Lightwave Technology 38, 1678 (2019)

    work page 2019

  4. [4]

    A. Yi, C. Wang, L. Zhou, et al., Applied Physics Reviews 9(2022)

    work page 2022

  5. [5]

    Kobayashi, R

    T. Kobayashi, R. Ikuta, S. Yasui, et al., Nature photonics 10, 441 (2016)

    work page 2016

  6. [6]

    Q. Li, M. Davanço, and K. Srinivasan, Nature Photonics 10, 406 (2016)

    work page 2016

  7. [7]

    L. Shao, N. Sinclair, J. Leatham, et al., Optics Express 28, 23728 (2020)

    work page 2020

  8. [8]

    D. B. Sohn, S. Kim, and G. Bahl, Nature Photonics 12, 91 (2018)

    work page 2018

Show all 21 references
  1. [9]

    L. J. Wright, M. Karpiński, C. Söller, et al., Physical review letters 118, 023601 (2017)

    2017

  2. [10]

    Fan, C.-L

    L. Fan, C.-L. Zou, M. Poot, et al., Nature Photonics 10, 766 (2016)

    2016

  3. [11]

    X. He, L. Cortes-Herrera, K. Opong-Mensah, et al., Optics Letters 47, 5849 (2022)

    2022

  4. [12]

    S. F. Preble, Q. Xu, and M. Lipson, Nature Photonics 1, 293 (2007)

    2007

  5. [13]

    Lo and H

    H.-P. Lo and H. Takesue, Optica 4, 919 (2017)

    2017

  6. [14]

    Savchenkov, W

    A. Savchenkov, W. Liang, A. Matsko, et al., Optics Letters 34, 1300 (2009)

    2009

  7. [15]

    Zheng, A

    Y. Zheng, A. Yi, C. Ye, et al., in 2022 Conference on Lasers and Electro-Optics (CLEO), (IEEE, 2022)

    2022

  8. [16]

    Y. Hu, M. Yu, D. Zhu, et al., Nature 599, 587 (2021)

    2021

  9. [17]

    Lu, Silicon and Silicon Carbide Photonics and the Applications (University of Rochester, 2016)

    X. Lu, Silicon and Silicon Carbide Photonics and the Applications (University of Rochester, 2016)

    2016

  10. [18]

    Z. Peng, R. Krishna, X. Wu, et al., in 2024 Conference on Lasers and Electro-Optics (CLEO), (IEEE, 2024)

    2024

  11. [19]

    D. Zhu, L. Shao, M. Yu, et al., Advances in Optics and Photonics 13, 242 (2021)

    2021

  12. [20]

    Kozlov, D

    A. Kozlov, D. Moskalev, U. Salgaeva, et al., Applied Sciences 13, 2097 (2023)

    2023

  13. [21]

    X. Lu, A. Rao, G. Moille, et al., Photonics research 8, 1676 (2020)

    2020

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.