Boundary-dominated optomechanics in silicon metamaterial membranes

Abstract: Stimulated Brillouin scattering in integrated photonic waveguides enables coherent coupling between optical photons and gigahertz acoustic phonons, providing a powerful mechanism for on-chip microwave photonics and opto-acoustic signal processing. Despite theoretical predictions of ultra-strong Brillouin interactions arising from enhanced light-sound coupling at device boundaries, most state-of-the-art integrated demonstrations remain governed by bulk photoelastic effects. This limitation stems from trade-offs between optical loss, interaction with waveguide boundaries and accessible phonon frequencies associated with the use of transverse-electric optical modes coupled to horizontally breathing mechanical modes. Here we demonstrate a new approach based on transverse-magnetic optical modes coupled to vertically breathing mechanical modes in suspended silicon membranes engineered with subwavelength metamaterial claddings. In this geometry, the interaction is dominated by the moving-boundary effect occurring at smooth top and bottom interfaces, while the phonon frequency is set primarily by the membrane thickness rather than its width. We observe forward Brillouin interactions at a record frequency of 12 GHz with a gain of 7200 W$^{-1}$ m$^{-1}$ and a mechanical quality factor of 620, yielding the highest Brillouin gain-to-quality-factor ratio reported in silicon waveguides. The devices exhibit net Brillouin amplification in millimeter-scale waveguides with pump powers below 15 mW, establishing a scalable platform for high-frequency integrated opto-acoustic signal processing.

• The paper demonstrates that boundary-dominated optomechanical coupling via the moving-boundary effect achieves unprecedented Brillouin gain (>7000 W⁻¹ m⁻¹) in silicon membranes.
• It employs comprehensive FEM simulations and experimental setups to optimize geometric designs, photon-phonon overlap, and fabrication tolerance.
• Experimental results validate high mechanical Q factors (600–700) and scalable on-chip integration for Brillouin amplifiers and related photonic applications.

Boundary-Dominated Optomechanics in Silicon Metamaterial Membranes

Introduction

The paper "Boundary-dominated optomechanics in silicon metamaterial membranes" (2605.07567) investigates the design, simulation, fabrication, and optomechanical characterization of subwavelength acoustic (SWA) membrane waveguides in silicon platforms. The focus is on harnessing boundary-dominated optomechanical coupling—primarily through the moving-boundary (MB) effect, as opposed to traditional photoelastic (PE) mechanisms—to achieve unprecedented Brillouin gain and robust mechanical resonance. The authors provide comprehensive theoretical, numerical, and experimental analysis, highlighting the interplay between geometric optimization, photon-phonon overlap, and practical fabrication tolerance.

Figure 1: Model of the suspended Brillouin waveguide with lateral metamaterial membranes. SWG cladding is modeled as homogeneous dielectric slabs for tractable simulation.

Optomechanical Modal Simulations and Design Rationale

Intramodal forward stimulated Brillouin scattering (FSBS) is leveraged, where synchronous co-propagation of pump, Stokes, and anti-Stokes optical modes induces an acoustic phonon via phase-matched interaction. The dominant mechanical mode is a vertically breathing resonance, realized in a core-membrane geometry. Finite-element (FEM) simulations include both the MB and PE contributions to the optomechanical overlap, but MB effects overwhelmingly dominate due to strong vertical displacement at membrane interfaces and concentration of the optical field at boundaries.

The optical SWG cladding is approximated as a synthetic dielectric (homogeneous slab) via effective medium theory, facilitating efficient calculation of leakage losses and modal profiles.

Figure 2: Simulated Brillouin gain coefficient $G_\mathrm{B}$ as a function of membrane periods and core widths.

Figure 3: Simulated mechanical frequency $\Omega_\mathrm{m}/2\pi$ for geometries.

Figure 4: Simulated Brillouin gain coefficient for optimal membrane period, partitioned into MB and PE contributions.

Numerical optimization yields maximal $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$ for narrow core ($W_\mathrm{c}=240$ nm), but this is counterbalanced by pronounced leakage losses, rendering such geometry impractical. A compromise at $W_\mathrm{c}=280$ nm maintains $$G_\mathrm{B}>7000\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$ while reducing optical leakage by a factor of 30 relative to the extremal configuration.

Fabrication Tolerance and Robustness

COMSOL simulations systematically evaluate the impact of typical fabrication deviations: etching offsets, layer thickness fluctuations, and longitudinal hole length changes. Mechanical frequency is primarily set by the silicon layer thickness, while Brillouin gain is more sensitive to tether/air-hole size adjustments due to altered rigidity and acoustic confinement.

Figure 5: Simulated mechanical frequency and Brillouin gain under etching, thickness, and hole-length variations.

Gain variation across realistic fabrication errors remains above $$G_\mathrm{B}>5900\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$, demonstrating substantial tolerance. This property significantly eases implementation of robust high-gain optomechanical waveguides.

Optical and Nonlinear Loss Analysis

The authors consider both linear and nonlinear loss mechanisms, including two-photon absorption (TPA) and free-carrier absorption (FCA), using power-dependent loss models. Optical frequency domain reflectometry (OFDR) measurements yield $\alpha \sim 2$–$2.6$ dB/cm. Nonlinear loss coefficients are calculated experimentally and via established analytical relations, with numerical values summarized and corroborated through measurement.

Figure 6: Optical loss measured across four waveguides using OFDR.

Experimental Configurations: Three-Tone and Two-Tone Brillouin Gain

Two experimental setups are deployed: the three-tone scheme (pump + Stokes + anti-Stokes injected together) and the two-tone scheme (pump + one sideband). The three-tone regime enables doubly resonant enhancement and detailed tracking of both Stokes amplification and anti-Stokes depletion/amplification while simplifying filtering requirements.

Figure 7: Three-tone gain setup: simultaneous energy transfers between anti-Stokes–pump and pump–Stokes throughout the chip.

Figure 8: Single-sideband gain setup: energy transfer from pump to sideband as modulation frequency traverses resonance.

Theoretical Modeling and Nonlinear Dynamics

A full three-mode coupled differential equation system models the FSBS process, incorporating mechanical susceptibility and loss coefficients. The model predicts non-exponential signal growth/depletion and non-Lorentzian spectral shapes—contradicting conventional backward Brillouin scattering assumptions. Output dependence on the Brillouin gain coefficient is quantified; strong numerical results include:

• Amplification regime for Stokes and anti-Stokes sidebands at high $G_\mathrm{B}$ and pump power.
• Asymmetric depletion/amplification dynamics between Stokes and anti-Stokes sidebands.
• Complete anti-Stokes depletion followed by regeneration with $\Omega_\mathrm{m}/2\pi$0 phase shift—observable only at sufficiently high gain and appropriate waveguide length.

  Figure 9: On/off relative power for Stokes and anti-Stokes versus input power for variable $\Omega_\mathrm{m}/2\pi$1.

  

  Figure 10: Amplitude, phase, and power evolution for pump, Stokes, and anti-Stokes along waveguide length.

  

  Figure 11: Signal evolution as function of modulation frequency for varying $\Omega_\mathrm{m}/2\pi$2. Notably, anti-Stokes regime switches from depletion to gain.

In most experimental cases, the undepleted pump and small-signal limits hold, but at high $\Omega_\mathrm{m}/2\pi$3 or extended waveguides, full FSBS modeling is required; small-signal approximations systematically underestimate sideband generation.

Figure 12: Comparison of full theoretical model vs. small-signal limit for sideband coupling—anti-Stokes sideband is regenerated from noise at high gain.

Experimental Results and Numerical Fitting

The study investigates two core width variants ($\Omega_\mathrm{m}/2\pi$4 nm and $\Omega_\mathrm{m}/2\pi$5 nm) and two lengths (4 mm, 6 mm). Brillouin gain coefficients extracted from fits reach $\Omega_\mathrm{m}/2\pi$67,200–8,000 $\Omega_\mathrm{m}/2\pi$7 for $\Omega_\mathrm{m}/2\pi$8 nm, and $\Omega_\mathrm{m}/2\pi$96,000–7,000 $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$0 for $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$1 nm, with mechanical frequencies matched to simulation ($$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$2 GHz). Mechanical Q factors are $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$3–700.

Figure 13: Three-tone pump results for $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$4 nm: RF beat notes and on/off gain curves for varying powers and lengths.

Figure 14: Three-tone pump results for $$G_\mathrm{B}>7300\, \mathrm{W}^{-1}\,\mathrm{m}^{-1}$$5 nm: analogous characterizations and gain curves.

Two-tone tests confirm asymmetric Stokes/anti-Stokes dynamics and full theoretical model predictions. Sideband regeneration is not observed at low power, requiring higher gain or extended length.

Figure 15: Two-tone pump experiment: normalized RF beat for Stokes/anti-Stokes and on/off gain.

Implications and Future Directions

From a practical perspective, the demonstrated boundary-dominated optomechanical interaction in SWA membrane waveguides enables ultra-high Brillouin gain and robust mechanical confinement—key for on-chip Brillouin amplifiers, lasers, and signal processing devices. The intrinsic fabrication tolerance allows reproducible, scalable manufacturing for silicon photonic circuits.

Theoretically, the established regime contradicts several assumptions in prior FSBS literature (e.g., neglect of MB contributions, oversimplistic small-signal models). The paper asserts that for high-gain, boundary-driven systems, the full nonlinear and modal-coupling dynamics must be considered.

Going forward, this paradigm opens new research directions in engineered photonic-phononic interactions, boundary-modulated metamaterial photonics, and nonlinear device physics. Applications include chip-scale microwave photonics, nonreciprocal devices, and noise-free Brillouin-based filtering, as well as fundamental exploration of non-Lorentzian and multi-sideband dynamics.

Conclusion

This study provides a rigorous account of boundary-dominated optomechanics in silicon metamaterial membranes, combining high-fidelity theoretical modeling, robust geometric optimization, and experimental validation. The resulting SWA membrane waveguides achieve record Brillouin gains, mechanically robust operation, and tolerance to fabrication imperfections, marking a significant advance in integrated optomechanics. The boundary-driven FSBS regime and its unique nonlinear dynamics have substantial implications for next-generation photonic circuits and theoretical foundations of optomechanical interactions.