Hybrid Photonic-Phononic Waveguide

Updated 10 July 2026

Hybrid photonic-phononic waveguides are integrated structures that simultaneously guide optical and acoustic modes to enable microwave-to-optical conversion and on-chip signal processing.
They leverage mode hybridization via symmetry breaking, elastic anisotropy, and dispersion engineering to achieve polarization conversion, routing, and precise Brillouin metrology.
Applications span thin-film lithium niobate, silicon–chalcogenide, and hBN/silicon platforms, demonstrating efficient multi-channel signal transduction and quantum interfacing capabilities.

A hybrid photonic-phononic waveguide is an integrated waveguide structure in which photonic and phononic degrees of freedom are either co-guided in the same physical region or deliberately interfaced on the same chip so that optical, microwave, and mechanical signals can be routed, converted, or spectrally processed through piezoelectric, photoelastic, Brillouin, or phonon-polaritonic interactions. In the cited literature, the term covers traveling-wave microwave-to-optics converters on thin-film lithium niobate, stimulated-Brillouin platforms in hybrid silicon-chalcogenide waveguides, and dual-band silicon structures that guide hyperbolic phonon polaritons in hexagonal boron nitride while preserving conventional near-infrared photonics (Yang et al., 12 Sep 2025, Zarifi et al., 2018, He et al., 2020).

1. Definition, scope, and terminological boundaries

Within integrated-waveguide research, “hybrid photonic-phononic waveguide” does not denote a single canonical geometry. One established meaning is a waveguide that simultaneously supports a guided photonic mode and a guided phononic mode, with explicit conversion pathways among microwave photons, traveling phonons, and optical photons. A clear example is the thin-film lithium niobate traveling-wave converter that guides telecom light near $1550\,\text{nm}$ , GHz acoustic waves around $9\,\text{GHz}$ , and couples them through an interdigital transducer and backward Brillouin scattering (Yang et al., 12 Sep 2025).

A second meaning emphasizes a hybrid photonic structure that supports conventional photonics in one spectral band and phonon-polaritonic propagation in another. In the hBN/silicon heterostructure, the same patterned silicon waveguide supports ordinary silicon photonics at about $1.55\,\mu\text{m}$ while the transferred hBN layer supports mid-IR hyperbolic phonon polaritons around $6.5\text{–}7.0\,\mu\text{m}$ through substrate-induced index contrast, without etching the hBN itself (He et al., 2020).

A third usage is broader and circuit-level: a phononic waveguide may be treated as the acoustic half of a future hybrid photonic-phononic chip even when the reported device itself is primarily phononic. That role is explicit in programmable phononic integrated circuits and in cryogenic lithium-niobate-based phononic devices aimed at later co-integration with optical circuitry (Xu et al., 30 Oct 2025, Malik et al., 29 May 2025).

The terminology is not uniform. A “hybrid” terahertz waveguide formed from silicon photonic crystals and gold parallel plates is a hybrid metal/dielectric photonic-crystal waveguide rather than a photonic-phononic one (Li et al., 2019). Likewise, a “hybrid silicon waveguide scheme” that combines strip and shallow-ridge silicon waveguides for spontaneous four-wave mixing control is hybrid in waveguide geometry, not in coupled photon-phonon guidance (Yu et al., 2019). Any encyclopedic treatment therefore requires distinguishing genuine photonic-phononic guidance from broader hybrid-waveguide nomenclature.

A recurrent principle in hybrid photonic-phononic waveguides is mode hybridization. In GaN-on-sapphire phononic wire waveguides with subwavelength cross-section, two orthogonal guided transverse acoustic modes appear: a Rayleigh-like mode dominated by out-of-plane displacement $z$ , and a Love-like mode dominated by in-plane shear-horizontal displacement $x$ . These behave analogously to TE/TM polarizations in photonic waveguides. Because sapphire is elastically anisotropic, the substrate coefficient $c_{14}$ couples in-plane and out-of-plane motion for certain propagation directions, so crystallographic orientation controls whether the two polarizations mix (Shen et al., 2019).

The hybridized phononic eigenfrequencies are written as

$\Omega_{\pm}=\frac{1}{2}\left(\omega_{L,0}+\omega_{R,0}\pm\sqrt{\left(\omega_{L,0}-\omega_{R,0}\right)^{2}+4g^{2}}\right),$

with coupling strength $g\propto c_{14}$ . In the $[11\bar{2}0]$ orientation, this produces strong mode hybridization, two experimentally observed resonant peaks, and avoided crossings as waveguide width varies; in the $9\,\text{GHz}$ 0 orientation, the modes remain essentially decoupled (Shen et al., 2019). The practical consequence is that a standard interdigital transducer designed for Rayleigh-like motion can indirectly excite a Love-like mode that is otherwise piezoelectrically inactive.

The same logic reappears at circuit scale in large-scale programmable phononic integrated circuits. There, phononic waveguides support dual polarization in the form of quasi-Rayleigh and quasi-Love modes, and an adiabatic width taper at the sapphire X-crystal orientation converts between them through an avoided crossing in the dispersion curves (Xu et al., 30 Oct 2025). This makes polarization conversion a design primitive rather than a by-product of anisotropy.

A related but purely photonic example sharpens the analogy. In a one-dimensional silicon antislot photonic crystal waveguide, breaking the in-plane mirror symmetry of the unit cell couples a longitudinal electric field component and a transverse electric field component, generating two hybrid LE-TE modes and opening a geometry-induced bandgap (Zhang et al., 4 Mar 2026). That device is not phononic, but it shows that hybridization in guided-wave systems can be driven by symmetry breaking, avoided crossings, and full-vectorial mode mixing rather than by material composition alone. This suggests a shared design vocabulary across photonic and phononic waveguides.

3. Material systems and representative architectures

The literature spans several material stacks, each selected for a distinct combination of optical confinement, acoustic confinement, piezoelectricity, Brillouin response, or quantum-material compatibility.

Platform	Guided degrees of freedom	Representative function
GaN-on-sapphire phononic wire (Shen et al., 2019)	Rayleigh-like and Love-like guided phonons	Polarization hybridization and on-chip polarization conversion
GaN-on-sapphire programmable PnIC (Xu et al., 30 Oct 2025)	quasi-Rayleigh and quasi-Love phonons	Splitters, MZIs, microrings, gratings, AAWGs, synthesizers
Silicon–chalcogenide hybrid waveguide (Zarifi et al., 2018)	Optical pump/Stokes waves and traveling acoustic wave in SBS	Distributed Brillouin metrology of width variations
hBN/silicon heterostructure (He et al., 2020)	Near-IR silicon photonics and mid-IR hyperbolic phonon polaritons	Dual-band waveguiding without etching hBN
TFLN on sapphire traveling-wave converter (Yang et al., 12 Sep 2025)	TE optical mode, quasi-Rayleigh acoustic mode, microwave IDT coupling	Multi-channel microwave-to-optics conversion
Thin-film LN on diamond (Malik et al., 29 May 2025)	Quasi-Love phononic waveguide with piezoelectric launch	Phonon-mediated interfaces to strain-sensitive color centers
Lithium niobate-on-sapphire giant-atom circuit (Xiao et al., 18 Dec 2025)	Fundamental quasi-Love phononic waveguide coupled to transmon qubit	Non-Markovian waveguide quantum acoustodynamics

These platforms are not interchangeable. The silicon–chalcogenide architecture is motivated by retaining silicon’s integration benefits while exploiting the strong SBS performance of $9\,\text{GHz}$ 1 (Zarifi et al., 2018). TFLN on sapphire is chosen because the material must confine optical modes, confine acoustic modes, and remain piezoelectric in the same waveguide (Yang et al., 12 Sep 2025). LN on diamond combines the strong piezoelectricity of LN with the high acoustic velocity and color-center compatibility of diamond (Malik et al., 29 May 2025). The hBN/silicon platform, by contrast, uses hybridization across spectral domains: the silicon waveguide remains a near-IR dielectric guide, while the hBN layer carries deeply subwavelength mid-IR hyperbolic phonon polaritons (He et al., 2020).

4. Core functionalities: guidance, conversion, routing, and metrology

In stimulated-Brillouin devices, the hybrid waveguide serves simultaneously as an interaction medium and a diagnostic object. The silicon–chalcogenide platform contains a $9\,\text{GHz}$ 2-thick chalcogenide section whose width changes from $9\,\text{GHz}$ 3 to $9\,\text{GHz}$ 4. Distributed Brillouin optical correlation domain analysis resolves this local variation: the $9\,\text{GHz}$ 5 section supports a dominant Brillouin frequency shift near $9\,\text{GHz}$ 6, whereas the $9\,\text{GHz}$ 7 section shows peaks near $9\,\text{GHz}$ 8 and $9\,\text{GHz}$ 9. With $1.55\,\mu\text{m}$ 0 ASE bandwidth, the reported spatial resolution is $1.55\,\mu\text{m}$ 1, sufficient to isolate the middle section and verify that the Brillouin response follows the designed geometry (Zarifi et al., 2018).

In programmable phononic integrated circuits, the waveguide is elevated from a single interaction region to a full routed architecture. The reported library includes directional couplers, Y-splitters, multimode interferometers, microrings, phononic gratings, thermoacoustic modulators, Mach-Zehnder interferometers, and an acoustic arrayed-waveguide grating. Using these elements, the platform demonstrates a $1.55\,\mu\text{m}$ 2 on-chip acoustic power splitter with integration density of $1.55\,\mu\text{m}$ 3 in the abstract and about $1.55\,\mu\text{m}$ 4 in the detailed description, a 21-port acoustic frequency demultiplexer with $1.55\,\mu\text{m}$ 5 resolution, and a four-channel reconfigurable frequency synthesizer (Xu et al., 30 Oct 2025). Although this work is phononic rather than optically hybridized, it establishes the waveguide-circuit paradigm that a hybrid photonic-phononic chip would require.

The traveling-wave TFLN converter adds direct photon-phonon-microwave transduction. Its operating principle is continuous phase matching rather than discrete cavity resonance, with backward Brillouin conditions

$1.55\,\mu\text{m}$ 6

Because the interaction takes place in a straight $1.55\,\mu\text{m}$ 7 waveguide that supports a fundamental TE optical mode and a quasi-Rayleigh acoustic mode, the device avoids the single-frequency bottleneck of cavity-enhanced converters. The reported result is coherent conversion between $1.55\,\mu\text{m}$ 8 microwave photons and $1.55\,\mu\text{m}$ 9 telecom photons, with operational bandwidths exceeding $6.5\text{–}7.0\,\mu\text{m}$ 0 in the optical domain and $6.5\text{–}7.0\,\mu\text{m}$ 1 in the microwave domain, an internal efficiency of $6.5\text{–}7.0\,\mu\text{m}$ 2, a system efficiency of $6.5\text{–}7.0\,\mu\text{m}$ 3, and simultaneous operation of nine conversion channels in a single device (Yang et al., 12 Sep 2025).

Taken together, these demonstrations show that the hybrid photonic-phononic waveguide is not merely a co-localized interaction zone. It can be a precision metrology platform, a routed signal-processing fabric, and a traveling-wave transducer whose selectivity arises from dispersion and phase matching rather than from isolated resonances.

5. Cryogenic and quantum implementations

Hybrid photonic-phononic waveguides are increasingly motivated by quantum interfaces. In thin-film lithium niobate on diamond, a $6.5\text{–}7.0\,\mu\text{m}$ 4-long rib phononic waveguide with $6.5\text{–}7.0\,\mu\text{m}$ 5 width is transfer-printed onto bulk diamond and driven by compact interdigital transducers. The selected guided mode is a quasi-Love mode around $6.5\text{–}7.0\,\mu\text{m}$ 6, chosen because it is better separated in wavevector space and therefore less likely to hybridize with other acoustic branches. At $6.5\text{–}7.0\,\mu\text{m}$ 7, the total insertion loss through the device is $6.5\text{–}7.0\,\mu\text{m}$ 8 at $6.5\text{–}7.0\,\mu\text{m}$ 9, corresponding to $z$ 0 transducer efficiency (Malik et al., 29 May 2025). The platform is explicitly oriented toward phonon-mediated quantum systems involving strain-sensitive color centers in diamond.

A different quantum direction appears in hybrid superconducting-phononic integrated circuits. There, a frequency-tunable superconducting transmon is coupled to a lithium niobate phononic waveguide at two points separated by $z$ 1, corresponding to about 600 acoustic wavelengths, with a propagation delay of $z$ 2. The guided acoustic mode is the fundamental quasi-Love mode, and the delayed self-interference of emitted phonons produces non-Markovian relaxation dynamics characterized by phonon backflow and a frequency-dependent effective decay rate that varies four-fold over merely $z$ 3. The reported Purcell factor exceeds 40, and the same dissipation landscape is used to prepare steady states with purity around $z$ 4 (Xiao et al., 18 Dec 2025).

These quantum implementations are not yet fully photonic-phononic in the sense of simultaneous optical and acoustic routing on the same demonstrated device. However, both papers state a broader architectural ambition: phononic waveguides are being developed as compact, slow-wave, GHz interconnects that can later interface with photonic circuits, superconducting circuits, or both (Malik et al., 29 May 2025, Xiao et al., 18 Dec 2025). A plausible implication is that cryogenic hybrid waveguides will be evaluated not only by insertion loss or electromechanical efficiency, but also by their ability to mediate coherent transduction among optical emitters, qubits, and guided phonons.

6. Conceptual issues, misconceptions, and research trajectory

One common misconception is that “hybrid” always refers to the same physical mechanism. In practice, the cited literature uses the term in several non-equivalent ways. In the GaN-on-sapphire phononic wire, hybridization refers to coupling between Rayleigh-like and Love-like acoustic polarizations set by elastic anisotropy and propagation direction (Shen et al., 2019). In the antislot photonic crystal waveguide, it refers to coupling between longitudinal and transverse electric field components created by broken in-plane mirror symmetry (Zhang et al., 4 Mar 2026). In the terahertz photonic-crystal device, it refers to metallic and dielectric confinement rather than to any photonic-phononic interaction (Li et al., 2019). In the silicon quantum-photonics scheme, it refers to combining strip and shallow-ridge waveguides to control spontaneous four-wave mixing noise (Yu et al., 2019).

A second misconception is that hybrid photonic-phononic functionality must be cavity-based. The TFLN traveling-wave converter demonstrates the opposite design philosophy: conversion is governed by continuous phase matching in a straight waveguide and can support over 70 independent channels in principle (Yang et al., 12 Sep 2025). Similarly, the programmable phononic integrated-circuit work argues for a monolithic chip that combines phononic, electronic, and photonic circuits, with Brillouin photon-phonon interaction and piezoelectric phonon-electron transduction as the relevant coupling channels (Xu et al., 30 Oct 2025). This suggests that the field is moving from isolated resonant devices toward waveguide-native multiplexing, routing, and reconfigurability.

A third issue concerns what counts as a photonic-phononic waveguide when only one half of the hybrid pair is directly demonstrated. The LN-on-diamond platform is primarily phononic, but it is explicitly positioned as compatible with diamond photonics and with phonon-mediated quantum interfaces to color centers (Malik et al., 29 May 2025). The same is true of superconducting-phononic giant-atom circuits, which are presented as a route toward hybrid superconducting photonic-phononic chips (Xiao et al., 18 Dec 2025). This suggests that the modern concept of a hybrid photonic-phononic waveguide increasingly includes not just co-guidance of light and sound, but also any waveguide platform engineered so that optical, microwave, and mechanical subsystems can be monolithically composed on the same chip.

Across these strands, the unifying theme is mode engineering in a guided geometry: subwavelength confinement, anisotropy, symmetry breaking, dispersion control, and piezoelectric or photoelastic coupling are used to make phonons routable and convertible with a level of architectural intentionality long associated with photonics. In that sense, the hybrid photonic-phononic waveguide has become a general integrated-waveguide concept rather than a single device type.