부산대학교 도서관

In this paper, we discuss a multiscale method for process simulations with atomic resolution applied to plasma etching. We demonstrate that the accurate prediction of microstructural modifications, as a function of the equipment parameters, can be achieved by coupling two simulation approaches that model phenomena at different length scales. Focusing on the etching processing of nanopatterned silicon samples in HBr/O $_2$-type plasma, we highlight the main ingredients of the numerical method: 1) the comprehensive model of plasma reactions to determine the particle distribution of the active plasma components, and 2) a coupled Kinetic Monte-Carlo method simulating all the events concurring to the surface erosion at the atomic level. The technique predicts the variation of the process results when the macroscopic parameters related to equipment settings (e.g., power, potential, and injected gas composition) are modified. A comparison between the microscopic analysis of real structures and the etched profiles predicted by the feature scale simulations validates the coupled numerical approach.