Sidewall defects in deep cryogenic Si etching in SF6/O2 plasma: a numerical simulation

A model for the simulation of two-dimensional profile evolution during cryogenic Si etching in SF6/O2 plasma is proposed and implemented. It employs Monte-Carlo method for particle fluxes computation and cell-based representation of the profile. The model is tuned specifically for studying various profile defects of stochastic nature. To this end the state of a model cell is represented as the combination of states of several subcells, stochastically chosen on each particle-surface interaction, thus allowing to represent profile phenomena with high- resolution without compromising simulation performance. The model is verified by matching the simulation results with experimental data; good qualitative agreement is observed. Then it is used to investigate surface defects specific to high aspect ratio cryogenic etching. Among them are depth-dependent wall roughness, cavern formation, trench splitting and black silicon formation.