Combined FEA-Matlab Optimization of Capacitive Micromachined Ultrasound Transducer Cell

Abstract

A novel approach to the design of a Capacitive Micromachined Ultrasound Transducer (CMUT) cells is presented, using a Genetic Algorithm (GA) based optimization that couples finite element simulations with parameter changes tuned through Matlab scripts. The optimization goal is a maximization of the CMUT efficiency in coupling the electrical, mechanical and acoustic energy domains. Global constraints are related to the pull-in actuation voltage, resonant frequency (set to 5–6MHz, suitable for biomedical applications) and limits on the achievable geometric dimensions. A CMUT is a relatively new MEMS device used to generate and sense ultrasound waves through electrical actuation. Our CMUT design aims to generate ultrasound waves in the 5–6 MHz range, suitable for breast tumor detection. The electro-mechano-acoustical nonlinear coupling makes the design and optimization process difficult, requiring several cycles of finite element analysis. Therefore, a genetic constrained optimization algorithm implemented in Matlab is coupled with finite element simulations of a parameterized structure modeled in COMSOL Multiphysics, in order to maximize the efficiency of the CMUT cell. The main goal is to achieve maximum transducer efficiency, reflected in a better operating device with lower power consumption. The Matlab-FEA optimization loop is driven by the transducer electro-mechanical efficiency, used as a main performance measure to vary the CMUT geometry parameters from one generation to another.