Pith. sign in

REVIEW 3 major objections 4 minor 2 references

Hybrid SiO2/Si Pillar-Based Optomechanical Crystals for On-Chip Photonic Integration

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Hybrid silicon/silica nanopillar cavities achieve on-chip optomechanical coupling above 1 MHz.

desk verdict A solid integration demo for pillar optomechanics, but the >1 MHz coupling claim is simulated, not measured—worth refereeing with a required abstract fix. read the letter →

arxiv 2505.17822 v1 pith:NYPS4H6A submitted 2025-05-23 physics.optics

Martin Poblet , Christian Vinther Bertelsen , David Alonso-Tomas , Rahul Singh , Elena Lopez-Aymerich , Jens Goldschmidt , Katrin Schmitt , Maria Dimaki
show 3 more authors
Winnie E. Svendsen Albert Romano-Rodriguez Daniel Navarro-Urrios
This is my paper · ORCID
classification physics.optics
keywords optomechanicalcrystal1Dphotonicnanopillarcavitysilicon-on-insulatorevanescentwaveguidecouplingtransductionthermalmechanicalmodesCMOS-compatiblephotonics
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 tries to establish that a one-dimensional photonic crystal made of vertical nanopillars, with a silicon upper section and a silica lower section, can be integrated on a silicon-on-insulator chip and still function as an optomechanical cavity. The design is meant to solve a coupling problem: full-silicon pillar cavities need waveguides too narrow for practical fiber coupling, because wider waveguides leak light into the substrate. By resting the waveguide's lower part on silica, the authors show that wider waveguides are viable and that evanescent coupling from an adjacent waveguide can excite the cavity modes. They report optical quality factors around $4x10^{3}$ and optomechanical coupling rates exceeding 1 MHz for the fundamental cantilever-like mechanical modes, driven purely by thermal motion at room temperature. If correct, this makes pillar-based optomechanical sensing practical on a CMOS-compatible photonic platform.

What carries the argument

The central object is the 1D photonic crystal pillar cavity, a linear array of nanopillars with a central defect region where the pitch and radius shrink quadratically, creating a confined optical mode in the TM bandgap. The mechanical motion comes from the cantilever-like flexural modes of the individual nanopillars, which modulate the optical resonance partly through the moving-boundary mechanism. The coupling between the cavity and the adjacent waveguide is described by an evanescent decay model, kappa_e(d) = kappa_i exp((d_c - d)/d_0), where d_c is the critical coupling distance and d_0 the decay length. Using this model and the input-output formalism, the paper derives that maximum intra-cavity power occurs at d = d_c, while maximum direct-detection transduction sensitivity occurs at a different distance, d_OM = d_c - d_0 ln(sqrt(3)-1) for bidirectional coupling. This distinction is what lets the authors optimize the geometry for transduction rather than simply for critical coupling.

What would settle it

Measure the thermomechanical noise spectrum of the fundamental pillar mode at a known temperature and input power, calibrate the photodetection and optical insertion loss, and extract the vacuum optomechanical coupling rate g_OM/2pi from the peak area and known thermal phonon occupation. If the inferred value is significantly below 1 MHz (or the measured linewidths do not match the simulated displacement profiles), the central claim of >1 MHz coupling would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that a 1D pillar photonic crystal cavity, composed of nanopillars with a silicon top section and a silica bottom section, can be efficiently excited through an adjacent integrated waveguide and transduce thermally driven mechanical vibrations with vacuum optomechanical coupling rates above 1 MHz. The optical field is confined in the upper silicon portion of the pillars, preventing substrate leakage, while the silica lower portion allows the adjacent waveguide to be widened without vertical optical loss. By tuning the waveguide-to-cavity separation, the authors find an optimal distance for optomechanical transduction that is slightly larger than the critical coupling distance for maximum intra-cavity power. The experimental RF spectra show peaks attributed to the fundamental and second cantilever-like mechanical modes of the pillars, and the simulated coupling rates for these modes match the spectral dependence of the transduced signal. The work proposes this as a scalable, CMOS-compatible platform for on-chip optomechanical sensors.

Load-bearing premise

The experimental RF transduction peaks are assumed to come from the simulated pillar mechanical modes with the simulated vacuum optomechanical coupling rates, but the spectra are not calibrated against an independent displacement or coupling measurement, so the claim of measured coupling rates above 1 MHz rests on the FEM simulation being quantitatively accurate.

Editorial extensions

If this is right

  • If the platform works as claimed, photonic circuits can include wide waveguides, fiber butt-coupling, splitters, and modulators on the same chip without the substrate-leakage penalty that plagued full-silicon designs.
  • The transduction of thermally activated mechanical modes from tens to hundreds of MHz means no external actuation is needed for readout, simplifying sensor operation.
  • The design rule that the optimal transduction distance exceeds the critical coupling distance by d_0 ln(sqrt(3)-1) provides a quantitative guideline for future pillar-cavity waveguide placements.
  • The demonstrated coupling rate above 1 MHz suggests that these cavities could resolve small forces or masses, extending nanowire- and nanorod-based MEMS/NEMS sensing into an on-chip photonic format.
  • The hybrid SiO2/Si geometry is compatible with established silicon foundry processes, so the cavities could be co-fabricated with other photonic components.

Reading between the lines

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

  • If the uncalibrated RF spectra quantitatively match the FEM predictions, then the same platform could be pushed to higher-frequency mechanical modes by scaling down pillar dimensions, potentially reaching several GHz with coupling rates that scale favorably.
  • The multimode interference in the 200 nm wide waveguide, which currently complicates the transmission spectrum, could be suppressed by narrowing the waveguide to a single-mode condition, which would also make the transduction spectroscopy cleaner and easier to calibrate.
  • The optomechanical coupling of individual pillars within the cavity varies along the array, so the device effectively provides a spatially resolved readout of pillar motion; this could be exploited for parallel sensing or for mapping the mechanical mode shape with sub-pillar resolution.
  • The lack of independent calibration suggests a straightforward test: measure the thermomechanical noise floor with a known temperature and laser power, extract the absolute displacement noise, and compare the inferred coupling rate to the FEM value.
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 SiO2/Si one-dimensional photonic crystal (1D-PhC) pillar cavity fabricated on a silicon-on-insulator platform and coupled evanescently to an integrated Si waveguide. FEM simulations describe the photonic band structure, the optical cavity mode, the mechanical cantilever-like modes, and the waveguide-to-cavity coupling, while experiments characterize the isolated cavity by a reflection measurement (Q ~ 4e3) and the integrated system by RF spectra of thermally driven mechanical motion. The paper derives an optimal waveguide-to-cavity distance for optomechanical transduction from a coupled-mode model and reports simulated vacuum optomechanical coupling rates reaching about 1 MHz for the fundamental pillar modes, with the abstract and conclusions stating that both simulated and experimentally measured vibrations show coupling rates exceeding 1 MHz.

Significance. The device concept is genuinely useful: vertically oriented pillars with a SiO2 base prevent vertical optical leakage while allowing wide access waveguides, which is a practical step toward CMOS-compatible on-chip nanopillar optomechanics. The fabrication details are complete, the FEM and coupled-mode derivations are standard and transparent, and the experimental RF spectra do show peaks at frequencies consistent with the simulated mechanical modes. However, the headline quantitative claim that optomechanical coupling rates exceeding 1 MHz are experimentally measured is not supported by the data as presented: the RF spectrum is uncalibrated and the g_OM values are simulated. If the claims are revised to separate simulated predictions from uncalibrated transduction measurements, or if a calibrated thermomechanical-noise measurement is added, the paper would be a solid contribution; as written, the central claim overreaches the evidence.

major comments (3)
  1. [Abstract, §2.4, Fig. 4e, Conclusions] The abstract and conclusions state that 'both simulated and experimentally measured mechanical vibrational frequencies show large optomechanical coupling rates exceeding 1 MHz.' This conflates measured frequencies with measured coupling rates. In Fig. 4e, the experimental RF spectrum is plotted in arbitrary units on the left axis, while the g_OM/2π values are simulated and plotted on a separate right-hand axis. The agreement claimed is only spectral coincidence; the absolute amplitude is not calibrated, and it depends on input power, detuning, detection gain, mechanical quality factor, and thermal phonon population. The paper itself notes that the transduced signal should be normalized by the thermal phonon population for a direct comparison, but this normalization is not performed. Therefore, the >1 MHz coupling rate is a FEM prediction, not an experimentally measured quantity. Please revise the claims accordingly or provide a calibrated measurement (e.g., a thermomechanical noise spectrum with known temperature and detection gain).
  2. [§2.3 and Supplementary S3] The prediction d_OM = 266 nm is not independent of the fitting procedure: it is computed from d_c and d_0 that are extracted by fitting Eq. (3) to simulated Q(d) data, and d_c itself depends on adding an artificial imaginary part to the silicon refractive index so that the simulated intrinsic Q matches the measured value of 4e3. The experimental optimum is found at d = 225 nm, and the discrepancy is attributed to a reduced intrinsic Q in the integrated devices, but no measurement of Q as a function of d is provided for those devices. This makes the 'optimization' partly a fit of simulated quantities rather than a parameter-free prediction. Please state explicitly which parameters are fitted, and ideally measure the waveguide-coupled Q or the transduction amplitude as a function of d to validate the predicted optimum.
  3. [§2.4, Fig. 4e] The assignment of each experimental RF peak to a specific simulated mechanical mode is assumed rather than demonstrated. The text states that lower frequencies correspond to smaller pillars near the cavity center and higher frequencies to larger pillars at the edges, and the simulated g_OM values are then overlaid on the experimental spectrum. Because the vertical scale of the spectrum is arbitrary and the mode identities come from the same FEM model used for g_OM, the agreement in Fig. 4e is largely a consistency check of the frequency scale, not an independent validation of the coupling rates. An independent check, such as comparing measured frequency ratios with the simulated eigenfrequency ratios or varying the pillar dimensions and observing the predicted scaling, would significantly strengthen the claim.
minor comments (4)
  1. [§2.4 / Fig. 4] The figure layout is confusing: the text says 'Representative RF measurements are shown for each separation in Figure 4b,' but the caption describes Fig. 4b as an RF spectrum as a function of laser wavelength for a single separation d = 225 nm, while Fig. 4c is said to contain the measurements for several separations. Please reorganize the panels and their descriptions to make it clear which spectrum corresponds to which separation.
  2. [§2.4, Fig. 4e] The two vertical axes in Fig. 4e are not labeled with units for the experimental RF signal, and the simulated g_OM points are plotted on a right-hand axis without error bars. Please add units and state explicitly that the comparison is qualitative in the vertical direction.
  3. [Abstract and Conclusions] The phrase 'mechanical vibrational frequencies show ... optomechanical coupling rates' is grammatically misleading, since coupling rates are properties of the coupled optical and mechanical modes, not of the frequencies themselves. Rephrase to 'the simulated optomechanical coupling rates for the measured mechanical modes exceed 1 MHz.'
  4. [§2.2] The statement that the waveguide supports multiple guided modes even at its minimum width, and that intermodal interference complicates the transmission spectrum, is an important limitation that should be mentioned in the conclusions as well as in the methods, since it affects the practical usability of the platform for sensing.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claims rest on independent FEM simulation plus uncalibrated but independent RF transduction measurements, and cited prior work is background rather than load-bearing.

full rationale

The load-bearing derivation chain is self-contained rather than circular. The simulated Q(d) values are fit with the exponential model of Eq. 3 to obtain d_c and d_0, and d_OM = d_c − d_0 ln(√3−1) is a derived quantity from that fitted model; however, it is subsequently compared with an independent experimental optimum (225 nm versus 266 nm), so it functions as a model prediction tested against new data, not as a fitted parameter renamed as a prediction. The vacuum optomechanical coupling rates in Fig. 4e are obtained from FEM perturbation theory (moving-boundary with negligible photo-elastic contribution), not by fitting the experimental RF spectrum; the measured spectrum independently shows the mechanical mode frequencies, though the transduced amplitudes are not calibrated against a known thermal noise floor. The paper's own caveat that the transduced signal should be normalized by thermal phonon population is a validity limitation, not a circular reduction. Reference [16] supplies the measurement scheme and prior platform, but no load-bearing claim relies on an unverified uniqueness theorem or ansatz smuggled in through self-citation. The abstract's wording overstates the experimental status of the >1 MHz coupling rates, but that is an evidence-strength concern, not a circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central design predictions rest on three fitted or calibrated parameters (d_c, d_0, and the artificial loss) and on a defect scaling factor chosen by hand; no new physical entities are introduced. The main experimental claims are grounded in direct RF measurements, with the coupling rates inherited from simulation rather than independently measured.

free parameters (4)
  • critical coupling distance d_c = 251.80 +/- 0.15 nm
    Extracted by fitting Eq. 3 to the simulated Q(d) curve in Fig. 3d (Section 2.3).
  • evanescent decay length d_0 = 46.64 +/- 0.14 nm
    Extracted from the same fit; used in Eq. 5 and in the d_OM prediction.
  • imaginary part of silicon refractive index (loss calibration) = not stated
    Introduced ad hoc in Section 2.3 to match the simulated intrinsic Q to the experimental value of 4x10^3; the specific value is not reported.
  • defect scaling factor g = 0.85
    Design parameter that sets the cavity mode energy; the predicted resonance (1376 nm) differs from the measured one (1364 nm), so the design model is approximate.
assumptions (5)
  • standard math Standard Maxwell equations solved by FEM provide the optical mode frequencies, Q factors, and coupling rates.
    Used throughout Sections 2.1-2.3; assumes the FEM implementation and the material refractive indices are correct.
  • domain assumption The waveguide-cavity external coupling rate decays exponentially with separation (Eq. 2).
    Invoked in Section 2.3 and S2 to derive the optimal coupling and transduction distances; standard in coupled-mode theory but an approximation.
  • domain assumption The RF peaks in Fig. 4 correspond to the cantilever-like mechanical modes of the pillars, and the transduction signal amplitude scales with the simulated optomechanical coupling strength.
    Section 2.4; no independent calibration verifies the mode assignment or the absolute coupling rate.
  • domain assumption Moving-boundary optomechanical coupling dominates and photo-elastic contributions are negligible.
    Section 2.4, citing Refs. 35 and 36; standard for high-index-contrast structures.
  • domain assumption Material properties of SiO2 and Si (crystalline p-type, 1-5 Ohm-cm) used in simulations match the fabricated samples.
    S1 specifies resistivity and etch process; optical properties are assumed from literature.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid SiO2/Si Pillar-Based Optomechanical Crystals for On-Chip Photonic Integration." pith.science (2026). https://pith.science/paper/NYPS4H6A

@misc{pith2026250517822,
  author       = {Pith},
  title        = {Pith review of: Hybrid SiO2/Si Pillar-Based Optomechanical Crystals for On-Chip Photonic Integration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NYPS4H6A}},
  note         = {Machine review of arXiv:2505.17822}
}
read the original abstract

One-dimensional photonic crystal (1D-PhC) pillar cavities allow transducing mechanical pillar vibrations to the optical domain, thereby relaxing the requirements typically associated with mechanical motion detection. In this study, we integrate these geometries into a silicon-on-insulator photonics platform and explore their optical and mechanical properties. The 1D-PhC structures consist of a linear array of high aspect ratio nanopillars with nanometer-sized diameters, designed to enhance the interaction between transverse-magnetic (TM) polarized optical fields and mechanical vibrations and to minimize optical leaking to the substrate. Integrated waveguides are engineered to support TM-like modes, which enable optimized coupling to the 1D-PhC optical cavity modes via evanescent wave interaction. Finite element method simulations and experimental analyses reveal that these cavities achieve relatively high optical quality factors (Q = 4x10^3). In addition, both simulated and experimentally measured mechanical vibrational frequencies show large optomechanical coupling rates exceeding 1 MHz for the fundamental cantilever-like modes. By tuning the separation between the 1D-PhC and the waveguide, we achieve optimal optical coupling conditions that enable the transduction of thermally activated mechanical modes across a broad frequency range (from tens to several hundreds of MHz). This enhanced accessibility and efficiency in mechanical motion transduction significantly strengthens the viability of established microelectromechanical (MEMS) and nanoelectromechanical systems (NEMS) technologies based on nanowires, nanorods, and related structures, particularly in applications such as force sensing and biosensing.

Figures

Figures reproduced from arXiv: 2505.17822 by the authors.

Figure 1
Figure 1. a, which consists of a cylindrical pillar of lattice constant (a = 350 nm), diameter (d = 210 nm), with a lower section of SiO₂ (1100 nm in height), and an upper section of Si (1300 nm in height), resting on a Si substrate. This unit cell exhibits propagating modes well confined in the upper section of the nanopillar and a wide bandgap for light of TM-like polarization ranging approximately from 200 THz to 240 THz, … view at source ↗
Figure 2
Figure 2. Integrated waveguides. a) Waveguide’s effective refractive index calculations as a function of their widths for the first three supported optical modes. b) Tilted SEM image of a waveguide at one edge of the sample, where the SiO2/Si interface can be clearly distinguished. c) FEM simulation of the first TM optical cavity mode electric field distribution in a 200nm wide waveguide. d) Optical image from above of the ho… view at source ↗
Figure 3
Figure 3. Optimal coupling distance simulation. a) Schematic representation of the fields and interactions between a waveguide and a bidirectional cavity. b) Electric field distribution for the TM fundamental optical supported mode in the waveguide-cavity system, when coupling distance is the critical one, 𝑑𝑐 . c) Ratio between 𝐸𝑐𝑎𝑣 2 and 𝐸𝑤𝑎𝑣 2 , i.e., 𝑅, for several waveguide-to-cavity separations. d) Optical quality factor… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Experimental setup scheme and measurements. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages

  1. [1]

    Aspelmeyer M, Kippenberg T and Marquardt F. (2014). Cavity optomechanics. Rev.Mod.Phys. 86(4), 54

    work page 2014

  2. [2]

    Dissertation (Ph.D.), California Institute of Technology

    Hill, Jeffrey Thomas (2013) Nonlinear Optics and Wavelength Translation Via CavityOptomechanics. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/DKW6TF64. 5 Fig. S1: Scanning electron microscopy image of one of the fabricated samples. Fig. S2: Coupling distance simulations without considering additional losses. a) Ratio between 𝐸𝐸𝑐𝑐𝑐𝑐...

    work page doi:10.7907/dkw6tf64 2013

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.