Design of Spurious-Free Surface Acoustic Wave Resonators on LNOI Platform Using Machine Learning

LiNbO<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> thin film on insulator (LNOl) platform has emerged as a promising solution for wideband and low-loss surface acoustic wave (SAW) applications in Sub-6G era. However, Rayleigh spurious mode is a very serious problem in LNOI platform. In this work, machine learning approach random forest developed by Bayesian Optimization is employed to predict the performance of SAW resonators based on LNOI platform. Additionally, an intelligent algorithm is formulated to quantify the number of spurious modes of LNOI platform, thereby enabling the selection of parameter combinations that eliminate spurious modes and possess electromechanical coupling coefficient (<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K² </i>) higher than 20%. Moreover, this work selected parameter combinations from predicted results and designed corresponding SAW resonators based on LNOl platform to verify accuracy of the prediction and spurious modes. [2025-0081]