Genetic Algorithm-Based Approach for Mask Design Generation in Wet Etching Silicon Corrosion

Abstract

Microelectromechanical Systems (MEMS) embody interdisciplinary expertise across design, engineering, and manufacturing, integrating fields like integrated circuit fabrication, mechanical and electrical engineering, materials science, and chemistry. MEMS complexity extends to diverse markets and applications, with fabrication challenges exacerbated by the anisotropic nature of silicon substrates and intricate manufacturing processes. Computational simulations play a crucial role in mitigating risks and costs by predicting cleanroom outcomes, thus saving considerable time and money. With the complexity and multitude of processes inherent in microfabrication, advanced techniques and software have been innovated to simulate these processes prior to cleanroom execution. This proactive approach not only streamlines the fabrication process but also enhances cost-effectiveness. This study aims to address these challenges by focusing on inverse problem techniques for MEMS mask generation. Leveraging optimized devices, it employs the ViennaLS open-source simulator for anisotropic corrosion simulation using a level-set approach and C++ scripts with a genetic algorithm to propose masks, thereby ensuring robust MEMS device fabrication. The ultimate goal of this research is to find the optimal mask design that leads to the fabrication of the desired MEMS device.