Application of Response Surface Methodology to the Optimizing of the Reactive Ion Etching of Polysilicon.

摘要

The goal of this research was to learn an empirical procedure for process optimizing and to apply it to the reactive ion etching (RIE) of polysilicon. The procedure chosen was Response Surface Methodology (RSM). RSM is explained in terms of the techniques of experimental design; factor evaluation using designed experiments; model building using experimentally obtained data and least squares regression; and model analysis using the method of steepest ascent, canonical analysis, and the method of ridge analysis. These techniques are combined into a generic flow chart that outlines the organization of a response surface study for process optimizing. This organization is then used to plan and begin the optimizing of the RIE of polysilicon in a chlorine and helium plasma. Two responses were critical to the optimizing of this process, polysilicon-to-oxide selectivity and anisotropy. The four factors: helium flow rate, chlorine flow rate, electrode power, and chamber pressure, were found to be significant in their effects on these responses and were used as the variable parameters for optimizing the process. But the study revealed that a fifth parameter, the initial photoresist profile, had a dominant effect on the anisotropy results due to the low polysilicon-to-photoresist selectivity. Keywords: Response surface methodology, Empirical model building, Theses, Process optimizing, Least squares regression, Design of experiments, Reactive ion etching, Integrated circuits. (jg)