Automatic Optimization of Reaction Mechanisms in Simulations of High Aspect Ratio Plasma Etching

Developing reaction mechanisms for feature profile simulation in plasma etching is challenged by the lack of fundamental data. In this regard, we have developed an optimization scheme which aims to aid in accelerating the development of such reaction mechanisms. An initial step in the process was to perform a series of plasma etches of high aspect ratio (HAR) SiO₂ features using a C₄F₆ / C₄F₈ / Ar /O₂ capacitively coupled plasma. Etch depth, etch rate, aspect ratio as and feature widths at several depths were determined using scanning electron microscopy. Using reactor scale plasma and feature scale simulation tools, HPEM (Hybrid Plasma Equipment Model) and MCFPM (Monte Carlo Feature Profile Model), the same processes were reproduced. The surface reaction mechanism used in the MCFPM to predict profiles consists of processes that are defined by parameters representing physical quantities such as chemical reaction probabilities, angular dependencies and energy dependence of processes having thresholds. A subset of these processes were coupled to a multi-variate 2-step optimizer that varies these physical parameters to achieve the best match between predicted profiles and experiments. The difference between etch profiles produced by simulation and experiment effectively acts as an objective function on which the optimization is based. The outcome and development of this optimization process will be discussed. The resulting reaction mechanism was used to simulate features using similar but quantitatively different processing conditions to demonstrate how broadly the mechanism can be applied.