Source-linked AI summary
Minimization of phonon-tunneling dissipation in mechanical resonators
Garrett D. Cole, Ignacio Wilson-Rae, Katharina Werbach, Michael R. Vanner, Markus Aspelmeyer
TL;DR
Mechanical-resonator measurements are limited by dissipation mechanisms such as thermoelastic damping, motivating methods to calculate support-induced losses. The paper develops a phonon-tunneling-based numerical approach and tests it against device calculations and measurements, finding agreement within reported geometric-model discrepancies.
Problem
Thermoelastic damping can limit mechanical-resonator measurements, while support-induced losses require a framework valid beyond the simplest weak-coupling geometry.
Method
The approach uses a phonon-tunneling master formula with symmetry-aware finite-element calculations, convergence checks, and experimentally characterized resonator geometries.
Results
4% discrepancy was obtained at the largest tested aspect ratio, decreasing monotonously from 20% to 4% for clamped-clamped beam calculations compared with the analytical approximation.
Takeaways & Limitations
The method applies beyond k_Rd ≪ 1 and can accommodate systems with exponentially suppressed contact stresses, including phononic-band-gap structures.
Abstract
from arXiv · showhide
Micro- and nanoscale mechanical resonators have recently emerged as ubiquitous devices for use in advanced technological applications, for example in mobile communications and inertial sensors, and as novel tools for fundamental scientific endeavors. Their performance is in many cases limited by the deleterious effects of mechanical damping. Here, we report a significant advancement towards understanding and controlling support-induced losses in generic mechanical resonators. We begin by introducing an efficient numerical solver, based on the "phonon-tunneling" approach, capable of predicting the design-limited damping of high-quality mechanical resonators. Further, through careful device engineering, we isolate support-induced losses and perform the first rigorous experimental test of the strong geometric dependence of this loss mechanism. Our results are in excellent agreement with theory, demonstrating the predictive power of our approach. In combination with recent progress on complementary dissipation mechanisms, our phonon-tunneling solver represents a major step towards accurate prediction of the mechanical quality factor.
I. METHODS
The study uses cryogenic optical interferometry to characterize mechanical resonators and extract their quality factors while reducing thermoelastic damping.
- I. METHODS: Cryogenic testing is used because thermoelastic damping limits the devices at room temperature.At 1.8 MHz and 300 K, the estimated thermoelastic-damping-limited Q is approximately 4000; at 20 K, it is estimated as 9.9 × 10^8.
- I. METHODS: An optical fiber interferometer with optical homodyne detection measures the resonators’ frequency response.
- I. METHODS: Resonance measurements are performed in a continuous-flow 4He cryostat.
Appendix A: Numerical calculation of Q-values
The numerical framework computes support-induced dissipation from resonator-mode stresses and elastic-wave propagation in a substrate half-space, under controlled geometric and weak-coupling approximations.
- Appendix A: Numerical calculation of Q-values: The method derives support-induced dissipation from stress-displacement overlaps between the resonator and substrate scattering modes.
- Appendix A: Numerical calculation of Q-values: The substrate is modeled as a free elastic half-space decomposed into cylindrical elastic eigenmodes, including longitudinal, transverse SV, and surface acoustic waves.
- Appendix A: Numerical calculation of Q-values: For thin plates, parallel contact stresses and bending-moment terms can be neglected under the stated geometric and low-frequency assumptions.
- Appendix A: Numerical calculation of Q-values: The leading symmetry- and mode-dependent corrections scale as (k_R t)^2 or k_R t, while the weak-coupling frequency-shift error scales as (k_R t)^3.
- Appendix A: Numerical calculation of Q-values: Numerical evaluation uses FEM-derived mode stresses and normalization, symmetry-reduced domains, and convergence checks estimating approximately 5% numerical error.
- Appendix A: Numerical calculation of Q-values: For clamped-clamped square beams with L/t from 15 to 40, numerical and analytical Q-values differ by 20% to 4%.
Appendix B: Analysis of Completed Devices
Completed-device analysis combines measured geometry with systematic design variations to supply realistic inputs for resonator simulations and support-loss comparisons.
- Appendix B: Analysis of Completed Devices: The chip contains 16 devices with identical nominal central-resonator dimensions of 130 × 40 µm^2 and two auxiliary-beam outer radii.
- Appendix B: Analysis of Completed Devices: Each subset contains eight auxiliary-beam contact positions spanning 13 to 62.5 µm, with two nominally identical chips measured.
- Appendix B: Analysis of Completed Devices: Measured devices have a 6.67 µm DBR thickness, dimensions enlarged by 1 µm at each free edge, and an average undercut distance of 27 µm.
- Appendix B: Analysis of Completed Devices: Reflectance-spectrum fitting determines physical thickness, while electron microscopy determines lateral dimensions and feeds the reconstructed geometry into CAD simulations.
- Appendix B: Analysis of Completed Devices: Dual-beam SEM/FIB measurements determine the lateral etch distance beneath the supports.
Appendix C: Q-value and frequency measurements
The study extracts Q from spectral and ringdown measurements, identifies the relevant free-free mode, and separates support-induced behavior from material-related background effects.
- Appendix C: Q-value and frequency measurements: Q is extracted either from resonance linewidths or from the 1/e ringdown decay time τ through τ = Q/πf.
- Appendix C: Q-value and frequency measurements: Spectral data are fitted with Lorentzian functions, while ringdown envelopes are fitted with decaying exponentials.
- Appendix C: Q-value and frequency measurements: The neighboring antisymmetric mode exhibits hardening Duffing nonlinearity and bistability, so it is driven below the bistability threshold during dissipation measurements.
- Appendix C: Q-value and frequency measurements: The symmetric and antisymmetric modes differ in surface-to-volume ratio, frequency shift, and background dissipation.
- Appendix C: Q-value and frequency measurements: The observed shifts and background losses are interpreted using bulk and surface material-related dissipation, while phonon-tunneling frequency shifts are predicted to be negligible.