Programmable Phononic Devices: Mechanisms & Applications

• Programmable phononic devices are reconfigurable platforms that manipulate vibrational modes using optical, electrical, magnetic, and strain fields.
• They employ engineered structures such as phononic crystals, guided-wave resonators, and moiré heterostructures to achieve tunable mode localization and controllable signal routing.
• Applications span quantum information processing, reconfigurable acoustics, and phononic logic computation, offering scalable integration and versatile on-chip programmability.

Programmable phononic devices are engineered platforms in which the generation, propagation, and interaction of phonons—quanta of mechanical vibrations—can be dynamically configured via external controls. Enabled by advances in nanofabrication, optomechanics, piezoelectrics, and hybrid quantum interfaces, these systems encompass solid-state, optically, electrically, magnetically, and strain-programmed architectures at scales from atomically thin layers to macroscale bulk phononic crystals. They operate across frequencies spanning the kilohertz to gigahertz regimes and underpin emerging paradigms in acoustic signal processing, quantum information, hybrid opto/electromechanical systems, and even phononic logic computation.

1. Fundamental Concepts and Programming Mechanisms

Programmable phononic devices manipulate vibrational modes, spatial and spectral confinement, and coherence properties by tuning local physical parameters through optical, electrical, magnetic, or strain fields. The general design builds on periodic or quasi-periodic structures (phononic crystals, PnCs), networks of guided acoustic waveguides, or arrays of interacting resonators.

A canonical programming method relies on altering local elastic moduli, effective masses, or boundary conditions. In optomechanical platforms, an optical spring is applied localized to a crystal site, dynamically shifting the frequency of targeted modes and localizing vibrational energy (Clark et al., 2024). Electric-field-driven architectures employ piezo-actuated modulation of phononic resonators or waveguides, implementing programmable coupling between phonon modes, gate-tunable interference, and dynamic mode hybridization (Ji et al., 31 Oct 2025, Taylor et al., 2021). Magneto-acoustic approaches exploit the acoustic Faraday effect in magnetic materials to realize on-demand polarization rotation and logic operations (Sklan et al., 2013). Twistronic and strain engineering in van der Waals heterostructures offer additional degrees of control via moiré potential engineering and atomistic strain fields (Chakraborty et al., 2 May 2025).

2. Material Platforms and Fabrication Approaches

Several material systems have been demonstrated for programmable phononic architectures:

• Nanomembrane phononic crystals: High-tensile Si₃N₄ membranes patterned into 2D hexagonal PnCs enable full in-plane bandgaps and reconfigurable mode localization via cavity-mediated optomechanical control (Clark et al., 2024).
• Lithium niobate and gallium nitride ridge platforms: Single-crystal LiNbO₃ and GaN on sapphire, with sub-micron waveguide geometries and high-index contrast, provide robust, suspension-free phononic integrated circuits (PnICs) and facilitate integration of superconducting circuits, piezoelectric transducers, and photonic elements (Wang et al., 4 Dec 2025, Zhang et al., 2 Mar 2025, Xu et al., 30 Oct 2025).
• Thin-film lithium niobate PnC resonators: PECVD Si₃N₄ patterned on x-cut LiNbO₃ supports high-Q, bandgap-confined phononic modes, programmable via nonlinear piezoelectric modulation for mode coupling, frequency conversion, and nonreciprocal routing (Ji et al., 31 Oct 2025).
• Trapped-ion arrays: Linear chains of trapped Yb⁺ ions utilize collective vibrational modes as phononic qubits, with programmable beam-splitter and phase-shifter gates realized via Raman coupling (Chen et al., 2022).
• Moiré heterostructures: CVD-grown monolayer transition-metal dichalcogenide (TMD) bilayers with programmable twist and strain yield strong moiré-induced tunability in phonon energies, lifetimes, and symmetries (Chakraborty et al., 2 May 2025).
• Magneto-acoustic materials: YIG and related ferrites enable magnetic-field programmable phononic logic and polarization control via magnon–phonon coupling (Sklan et al., 2013).

Key fabrication processes include e-beam lithography and reactive ion etching for sub-μm features; lift-off and metallization for electrodes and heaters; dry transfer or CVD growth for van der Waals stacks; and standard CMOS-compatible steps for integrated photonic-phononic platforms (Xu et al., 30 Oct 2025, Zhang et al., 2 Mar 2025, Taylor et al., 2021).

3. Building-Block Functionality: Local Control and Gate Operations

Programmable phononic functionalities are realized through a library of building blocks:

• Defect writing and localization: Optically defined defect modes are programmed in situ, with the optical spring constant $K_{\text{opt}}$ providing spatially selective stiffening and enabling a single collective mode to split off from the band edge and become exponentially localized (Clark et al., 2024). The participating mass reduces by up to 37× upon localization, as quantified by $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$.
• Electro-acoustic coupling and mode hybridization: Applying time-dependent electric fields across piezoelectric resonators yields Hamiltonians such as

$$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$

where $g_{01}(t)$ is modulated by the applied voltage, allowing realization of Autler-Townes splitting, ac Stark shifts, Rabi oscillations, and programmable nonreciprocal conversion between multiple modes (up to 20 dB isolation) (Ji et al., 31 Oct 2025).

• Piezo-acoustomechanical phase shifters: Strain-induced shifts in phase velocity $v_p$ provide $\pm \pi$ phase control of GHz phonons over tens of micron lengths with tens of volts (Taylor et al., 2021).
• Programmable interferometers and switching: Networks of directional couplers, Y-splitters, multimode interferometers (MMIs), and combiner trees enable arbitrary routing, multiplexing, and demultiplexing, configured dynamically by applied voltages, optical power, or heaters (Xu et al., 30 Oct 2025, Zhang et al., 2 Mar 2025, Taylor et al., 2021).
• Logic and parallel architectures: Magneto-acoustic gyrators implement polarization logic gates (NOT, NAND, OR, XOR) and enable “rainbow parallelization,” i.e., frequency-multiplexed parallel logic streams in a single circuit (Sklan et al., 2013).

4. Control Protocols, Reversibility, and Reconfiguration Dynamics

The temporal and spatial flexibility of programmable phononic devices is governed by the physical control channel:

• Optical programming: Actuation via radiation pressure features $\lesssim \mu$s response, reversible programming, and single-site spatial selectivity scalable to 2D arrays. Because the mechanism is non-dissipative, mass, defect location, and coupling can be reprogrammed in situ without device degradation (Clark et al., 2024).
• Electrostatic and piezoelectric tuning: Programmable phase shifters and modulated cavities respond on sub-μs timescales limited by mechanical resonance, with insertion loss dominated by coupler and IDT design (Taylor et al., 2021, Ji et al., 31 Oct 2025).
• Thermoacoustic phase shifters: Resistive heaters tune local acoustic phases in Mach–Zehnder topologies with reconfiguration times $\sim$10–100 μs, set by the thermal RC constant (Xu et al., 30 Oct 2025).
• Qubit-based switching: Coupling superconducting qubits to phononic cavities allows rapid ($\lesssim$375 ns) on-off emission or routing of individual phonons, controlled by flux-biasing or capacitive tuning (Wang et al., 4 Dec 2025).
• Magneto-acoustic logic circuit reconfiguration: Coil currents and local magnetic bias fields switch polarization logic within $\sim$1–10 ns, enabling GHz-band frequency-multiplexed circuits (Sklan et al., 2013).

Reprogramming is generally reversible and non-volatile as long as control signals are appropriately managed and no permanent material change is induced. Multimodal programmability—simultaneous adjustment of spatial, spectral, and polarization degrees of freedom—enables highly diverse signal processing and quantum information protocols.

5. Performance Metrics and Exemplary Devices

Table 1 summarizes select metrics from recent implementations:

Platform/Device	Key Metric(s)	Reference
Si₃N₄ PnC, optically defined defect	50 kHz mode shift, $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$0, 37× mass reduction	(Clark et al., 2024)
GaN-on-sapphire photonic-phononic IC	$$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$1, $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$2, SNR>35 dB, 3x3 RF-optical mapping	(Zhang et al., 2 Mar 2025)
LiNbO₃ microring PnIC	Single-phonon emission $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$3, $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$4, Purcell factor 19×	(Wang et al., 4 Dec 2025)
Phononic memory (Si/ScAlN)	$$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$5 read/write fidelity, switch time $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$6 ns	(Taylor et al., 2021)
Trapped-ion network	Gate fidelity 99.1%, state prep/detect fidelity $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$7, scaling to >100 modes	(Chen et al., 2022)
YIG magneto-acoustic logic	$$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$8 logic channels, $$m_\text{eff} = \int \rho |u(\mathbf{r})|^2 dV / \max|u(\mathbf{r})|^2$$9, crosstalk $$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$040 dB	(Sklan et al., 2013)
LiNbO₃ cavity electroacoustic	Cooperativity $$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$1, mode isolation 20 dB, $$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$2	(Ji et al., 31 Oct 2025)
Large-scale PnIC (GaN)	$$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$3 splitter ($$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$4), 21-channel demux with 3.8 MHz resolution	(Xu et al., 30 Oct 2025)

Performance is determined by acoustic mode confinement (bandgap engineering), coupling efficiency (piezoelectric, optomechanical, or electrostatic), reconfiguration speed, coherence times (phonon quality factors), and scalability of integration.

6. Applications in Quantum, Classical, and Hybrid Systems

Programmable phononic devices are being deployed in:

• Quantum information processing: On-chip phononic buses for qubit–qubit connectivity, entanglement distribution, and high-fidelity phonon storage or routing (Wang et al., 4 Dec 2025, Xu et al., 30 Oct 2025, Ji et al., 31 Oct 2025, Chen et al., 2022).
• Reconfigurable signal processing: On-chip acoustic frequency synthesizers, low-loss delay lines, spectral multiplexers, programmable demultiplexers (e.g., 1×128 splitters, 21-port demuxes), and photonic-phononic transducers bridging RF and optical domains (Zhang et al., 2 Mar 2025, Xu et al., 30 Oct 2025).
• Programmable logic and computation: Acoustic logic gates, polarization-based logic, and frequency-multiplexed parallel computing in magneto-acoustic platforms (Sklan et al., 2013).
• Moiré opto-straintronics: Tunable valley-phonon interactions and quantum sensing via twist/strain in TMD heterostructures, programmable SHG sources, valleytronic logic, and acousto-optical modulators (Chakraborty et al., 2 May 2025).
• Hybrid integration: Convergence of phononic, photonic, and electronic domains for signal transduction, metrology, and AI acceleration ("Zhengfu" chips) (Xu et al., 30 Oct 2025).

7. Prospects, Challenges, and Outlook

Programmable phononic devices continue to expand in architectural complexity, integration density, and controllability. Ongoing challenges include reduction of insertion loss (especially at IDT and waveguide interfaces), mitigation of materials disorder to maintain high $$H_{\rm eff}(t) = \hbar \omega_0 a_0^\dagger a_0 + \hbar \omega_1 a_1^\dagger a_1 + \hbar g_{01}(t)(a_0 a_1^\dagger + a_0^\dagger a_1),$$5 at scale, efficient scaling of control circuitry, and hybridization with superconducing, optoelectronic, and quantum photonic platforms. Trends include extending coherent control to the single-phonon regime via enhanced optomechanical or piezoelectric coupling (Clark et al., 2024, Wang et al., 4 Dec 2025); enabling dense on-chip superlattice and hybrid quantum-phononic architectures; and utilizing chiral, nonreciprocal, and topological effects for robust routing and protection (Chakraborty et al., 2 May 2025, Taylor et al., 2021).

As foundational building blocks for quantum acoustics, signal processing, nonlinear optics, and unconventional computing, programmable phononic devices are positioned as a third technologically significant information carrier alongside electrons and photons (Xu et al., 30 Oct 2025).