Remote plasma etching reactors: Modeling and experiment

Abstract

A combined theoretical–experimental analysis of a remote plasma etching reactor was undertaken. Chlorine plasma etching of polysilicon was chosen as a case study to illustrate the approach. A model of the plasma (radical source) and of the transport and reaction of radicals through the downstream section was developed. The effect of tube dimensions, gas pressure, flow rate, and wall recombination probability was examined. In the parameter range investigated, low pressure, high flow rate, short downstream section, and low wall recombination probability favor higher radical fluxes at the substrate. A conflict arises in choosing the tube radius. A tubular reactor with different radii for the source and downstream sections is proposed to optimize the radical flux. Model predictions were tested in a radical beam reactor. Actinometry was used to monitor the Cl-atom density in the source and to verify the predictions of the plasma model. Using the wall recombination probability as the only adjustable parameter, reasonable agreement between predicted and measured atom flux at the reactor exit was also obtained.