Towards a "virtual fabrication tool" for ultra-high precision semiconductor manufacturing

In this study, we conducted plasma simulations using the particle-in-cell (PIC) method in conjunction with Bayesian optimization to optimize the operating parameters of low-temperature plasma reactors used for microelectronics chip manufacturing. Our investigation focused on understanding how variations in pressure and temperature influence etch uniformity. We used Bayesian optimization to identify parameter configurations that maximize uniformity

The results demonstrated that integration of simulation-driven evaluation with Bayesian optimization enables efficient exploration of the parameter space. The results demonstrated that this approach can identify optimal conditions that improve plasma uniformity, providing valuable insights in advancing the development of manufacturing microelectronics