AI-assisted Material Property Extraction for Piezoelectric Resonators

New piezoelectric materials such as ScAlN, sputtered PZT and NaNbO
3
with much higher piezoelectric coefficients than AlN, have been reported in the literature. However, their exact material properties are not reliably known. We demonstrate, for the first time, an AI-assisted method which combines a fast forward solver neural network with a minimizing optimizer. We extracted the material properties which give the best match of the impedance spectra of a coupled bulk acoustic resonator (CBAR) between simulation and experiment.