Mechanically-tunable bandgap closing in 2D graphene phononic crystals

The study addresses how to dynamically control phononic bandgaps in phononic crystals (PnCs), which are essential for applications such as waveguiding, phononic shielding, and quantum information processing. Conventional rigid PnCs (e.g., SiN membranes) have fixed, growth-defined tension, limiting dynamic tunability and coupling to external systems. In contrast, 2D materials allow electrostatic control of tension, enabling tunability of bandgap center frequency; however, prior work showed that tuning tension did not alter the band hierarchy and thus could not close existing gaps. The research question is whether uniaxial tension in a 2D-material PnC can change band hierarchy to close an out-of-plane phonon bandgap, effectively switching the device between mechanically insulating and conductive states. The context includes leveraging suspended graphene’s mechanical robustness and electrical conductivity, and using electrostatic gating to engineer tension distributions. The purpose is to demonstrate, via simulations and realistic device design, a mechanically-tunable bandgap closing and to establish transmission measurements as a probe in finite devices. The importance lies in enabling a mechanical transistor for MHz phonons, tunable coupling between mechanical elements (e.g., qubits), and a platform to emulate condensed-matter analogues such as a metal–insulator transition.

The paper situates its work within PnC research where periodic modulation of material properties yields band structures analogous to electronic Bloch waves. Prior studies demonstrated phenomena including quantum entanglement, topological states, negative refraction, ultracoherent resonators via phononic shielding, waveguiding, and thermal management. For 2D materials, electrostatic gating has been shown to tune biaxial tension and thereby the bandgap center frequency. However, prior systems retained their band hierarchy under tension; i.e., a gapped system remained gapped, limiting control over coupling strength. This work advances the literature by employing uniaxial (not biaxial) tension to produce different scaling of band edges at different Brillouin zone points, enabling closure of an indirect out-of-plane bandgap in a graphene PnC.

• Phononic lattice design: Honeycomb lattice of circular holes (lattice constant a, hole diameter d) patterned into suspended monolayer graphene to maximize robustness and preserve material while yielding a robust indirect out-of-plane bandgap. Nominal parameters used in examples: a = 1 μm, d/a = 0.5.
• Band structure simulations: Finite element method (FEM) with COMSOL Multiphysics (v5.5). Material parameters: graphene E2D = 1.0 TPa, Poisson’s ratio ν = 0.15, thickness h = 0.335 nm, density ρ = 2260 kg m⁻³. Calculated tension redistribution within the unit cell due to patterning; computed phononic bands along Brillouin zone high-symmetry lines for out-of-plane modes under specified tension states.
• Tension conditions and metrics: Examined biaxial (σxx = σyy) and uniaxial (σxx ≠ σyy) tension. Quantified uniaxiality by the ratio σxx/σyy and tracked valence-band maximum fVB and conduction-band minimum fCB (out-of-plane) at specific k-points (indirect bandgap: fCB at Γ, fVB along Γ–X).
• Finite-size device modeling and transmission probe: Designed a rectangular suspended device (9 μm × 28 μm; 7 × 17 unit cells). Defined transmission between an excitation area (A) and detection area (B) via spatially integrated out-of-plane displacement z(x, y, f, t) over illumination spots, using harmonic excitation to map transmission versus frequency. Identified stop bands, resonances, and broad band regions; extracted bandgap visibility in finite geometry.
• Engineering uniaxial tension via electrostatic pressure: Modeled clamped rectangular membrane over a gate electrode separated by d = 300 nm oxide. Applied gate voltage Vgate yields electrostatic pressure pel = ε0 Vgate² / (2 d²). Varied aspect ratio W/L to induce higher tension along the shorter dimension. Computed σxx/σyy versus pressure and aspect ratio, identifying conditions to reach σxx/σyy ≈ 1.7.
• Comparative geometry: Modeled a circular device as a reference for biaxial tension scaling (σxx ≈ σyy) under pressure to contrast with the rectangular device’s uniaxial response.
• Disorder studies: Simulated surface contamination by randomly placing PDMS islands (thickness 18 nm, diameter 4 μm) on the PnC to vary areal mass density; also tested a uniform PDMS film. Modeled random tension variations using superpositions of randomized plane waves with spatial scales down to a/4; parametrized disorder by standard deviation Var(σ)/σ0. Calculated transmission spectra versus disorder strength.
• Additional analyses: Extracted average strain under pressure (e.g., ε ≈ 0.24% at 10 kPa). Assessed multilayer graphene devices (up to ~200 layers) for persistence of bandgaps and required increased pressures for bandgap closing.
• Data and code availability: FEM simulations in COMSOL; data and any unreported code available from authors upon request.

• Existence and tunability of an indirect out-of-plane phononic bandgap: For a honeycomb graphene PnC under biaxial tension σxx = σyy = 0.01 Nm⁻¹, FEM band-structure simulations show a bandgap from 48.8 to 56.4 MHz.
• Uniaxial tension closes the bandgap: Increasing tension uniaxiality to σxx/σyy = 1.7 causes fVB to upshift more strongly than fCB, closing the indirect bandgap (fCB − fVB → 0). Under biaxial scaling, the gap widens in absolute terms and the center frequency follows √σ, with constant relative gap.
• Finite-size transmission probe: A 9 μm × 28 μm device exhibits an average 5 orders-of-magnitude transmission suppression within the bandgap (≈48.5–56.5 MHz), confirming a mechanically insulating state in finite geometry.
• Gate-induced uniaxiality: Electrostatic pressure pel = ε0 Vgate²/(2 d²) with d = 300 nm produces uniaxiality in rectangular membranes. For pel ≈ 3 kPa (≈8 V), σxx/σyy reaches ~1.7, matching the bandgap closing condition. Phase diagram indicates the lowest required pressure near aspect ratio W/L ≈ 0.32.
• Mechanical transistor behavior: Applying pressure (e.g., 5 kPa) to the rectangular device eliminates the bandgap in transmission, switching to a mechanically conductive state. On/off ratio ~10^5 (≈100 dB), ≈6 dB suppression per unit cell.
• Geometry dependence: Rectangular device (W/L = 0.32) shows bandgap closing around 3 kPa; circular device (biaxial scaling) maintains a clear bandgap up to 30 kPa.
• Strain levels: Average strain ε ≈ 0.24% at 10 kPa, below thresholds for phonon instabilities and graphene fracture.
• Disorder tolerance:
  • Surface contamination: Bandgap signatures persist up to total areal mass density ~4× graphene’s areal mass (ρ2D ≲ 4 ρgraphene); even with three PDMS islands (each mass ≈ entire clean resonator) a weak gap remains. A uniform PDMS film still yields a clear gap.
  • Tension disorder: Increasing Var(σ)/σ0 smears the bandgap; beyond Var(σ)/σ0 ≈ 0.40, the gap becomes indistinct. This threshold is comparable to spreads observed in experiments via Raman spectroscopy.
• Multilayer devices: Bandgaps persist up to ~200 layers; closing requires higher pressures, but standard 300 nm SiO2/Si backgates allow ≈100 V before breakdown (~50 kPa), sufficient for closing in multilayers.

The work demonstrates that uniaxial tension engineering in a graphene PnC can alter band hierarchy to close an indirect out-of-plane bandgap, addressing the limitation of prior biaxial tuning that left gaps intact. By using a rectangular suspended geometry and gate-induced electrostatic pressure, the device transitions from a mechanically insulating to a conductive state at experimentally accessible voltages (~8 V), providing a mechanical analogue of a metal–insulator transition for out-of-plane modes. Transmission measurements in finite-size devices faithfully reflect the band structure, enabling practical readout of the bandgap state. The large on/off transmission ratio (~10^5) positions the system as a mechanical transistor for MHz phonons, enabling phonon logic and tunable coupling between mechanical elements (e.g., qubits). Robustness analyses indicate that realistic contamination levels and moderate tension disorder preserve bandgap functionality, while identifying critical thresholds that inform fabrication tolerances. Geometry dependence further clarifies the roles of uniaxial versus biaxial tension, guiding device design. Overall, the results advance dynamic control over phononic band structures and coupling strengths in 2D-material PnCs.

This study introduces a tunable graphene phononic crystal in which uniaxial tension closes an indirect out-of-plane bandgap, switching the device between mechanically insulating and conductive states. Simulations show that a uniaxiality ratio σxx/σyy ≈ 1.7 closes the 48.8–56.4 MHz bandgap (at σ = 0.01 Nm⁻¹). A realistic rectangular device achieves the required uniaxiality using gate-induced electrostatic pressure (~3 kPa, ~8 V across 300 nm oxide), with transmission measurements confirming gap closure and yielding an on/off ratio ~10^5. The bandgap is robust to surface contamination up to ~4× graphene’s areal mass and to tension disorder up to Var(σ)/σ0 ≈ 0.40. The approach scales across frequencies (~10 MHz to ~1 GHz via lattice constant) and supports applications in phonon logic, tunable coupling between mechanical resonators (qubits), and dynamic phononic shielding studies. Future work could realize logic gate architectures, explore multilayer devices with higher operating pressures, and investigate analogues of more complex condensed-matter phenomena (e.g., quantum Hall, Mott transitions, topological phase changes) in phononic platforms.

• The metal–insulator analogy applies only to out-of-plane modes and lacks a direct electronic counterpart of chemical potential; thus, the analogy is qualitative.
• Finite-size devices exhibit boundary-related disorder not captured by infinite-lattice band structure models, shifting the pressure required for gap closing.
• Fabrication tolerances are critical: excessive surface mass loading (>~4× graphene areal mass) or high tension disorder (Var(σ)/σ0 ≳ 0.40) smear or remove the bandgap.
• Results are based on FEM simulations; experimental validation is implied but not presented in this paper.
• Achieving and controlling high uniaxiality depends on device geometry and uniform gating; multilayer implementations require higher voltages/pressures.
• Focus is on out-of-plane modes; in-plane modes and multimode coupling effects are not explored in detail.