One-port acoustic resonators

Imaging acoustic leakage and higher-order modes in gigahertz resonators, and designing and fabricating an AlN-on-silicon resonator.

← Back to research

TMIM · Device design · Fabrication

Optical image of a fabricated AlN-on-silicon one-port resonator with a central IDT, reflector gratings and contact pads
Optical image of the fabricated AlN-on-silicon one-port resonator.

This project has two parts that involve different devices. First, I imaged an existing X-cut LiNbO₃-on-SiC resonator, in collaboration with Prof. Ruochen Lu. Second, I designed and fabricated my own AlN-on-silicon resonator.

Imaging leakage in a gigahertz resonator

Measured and simulated Y11 spectrum of a one-port resonator, with TMIM maps inside and outside the acoustic resonator near 1.03 GHz and 1.13 GHz
Y₁₁ spectrum (experiment vs. simulation) and TMIM maps inside and outside the acoustic resonator (AR) near 1.03 GHz (±1 V scale) and 1.13 GHz (±0.1 V scale).

A resonator’s performance is limited by the energy that escapes it. An electrical spectrum (Y₁₁) shows that there are losses and spurious modes, but not where the energy goes.

Using TMIM on a one-port X-cut LiNbO₃-on-SiC resonator, I imaged the acoustic field both inside the resonator and just outside its reflectors. I did this at frequencies around the main resonance (about 1.03 GHz) and around the higher-frequency features (about 1.13 GHz) in the spectrum. This gives a direct, real-space picture of leakage.

The inside and outside maps use different color scales (±1 V and ±0.1 V), so their contrast should not be read as a calibrated leakage ratio.

TMIM MOD signal versus input power, and line scans near the noise floor
Noise floor: TMIM MOD signal vs. input power down to about −87 dBm, with line scans near the noise floor.

Before interpreting weak signals outside the device, I characterized the noise floor of the measurement by sweeping the input power down to about −87 dBm.

Higher-order modes inside the resonator

TMIM CH1 and CH2 maps across the resonator aperture, a simulated mode profile and a ray schematic
TMIM CH1/CH2 maps across the resonator aperture (red box on the optical image), compared with a simulated mode profile.

Inside the resonator cavity, the TMIM maps show field variation across the aperture, not only along the propagation direction. This is the fingerprint of transverse higher-order modes, a common source of spurious responses in SAW resonators. Comparing the measured CH1/CH2 patterns with the simulated mode shape identifies which mode is excited.

Designing an AlN-on-silicon resonator

AlN-on-silicon device stack, CAD mask, simulated dispersion and simulated admittance spectrum
AlN-on-silicon resonator design: device stack, CAD mask, simulated dispersion, and simulated admittance with a resonance near 1 GHz for λ = 3.6 µm.

I then designed my own one-port resonator on a 200 nm AlN film on silicon. Using COMSOL, I simulated the dispersion of the guided acoustic mode to choose the IDT period. A wavelength of 3.6 µm places the resonance near 1 GHz in the simulated admittance spectrum. From the simulation I drew the CAD mask: a central IDT flanked by reflector gratings.

Fabrication

Fabrication steps for the AlN-on-silicon resonator and an optical image of the finished device
Fabrication steps and an optical image of the finished AlN-on-silicon resonator.

I fabricated the device using:

  1. Electron-beam lithography on PMMA resist.
  2. Deposition of 10 nm Cr / 100 nm Au, followed by lift-off.

The optical image shows the finished resonator: a central IDT, flanking reflectors, and contact pads.

Project status — August 2024

Leakage and higher-order modes were imaged in the LiNbO₃ resonator, and the AlN-on-silicon resonator was designed and fabricated. Next steps were electrical characterization of the AlN device, remeasuring the sample, and preparing a standard sample for low-temperature measurements.