four-point technique

Since the neutrals were assumed to be thermally equilibrated with the feature surface, the trajectories of neutral species that are scattered were calculated assuming a cosine distribution about the surface normal. A four-point technique was used to determine the surface normal by the presence/ absence of the four adjacent neighboring cells.

by-product, sticking probability/coefficient

The incorporation probability of an incident particle was based on the computed sticking probability. For Cl and O neutrals, this quantity was dependent on the extent of surface chlorination and/or oxygenation of the surface layer where the probability of sticking was based on the sticking coefficient and the fraction of free sites in the surface layer. SiCl2, which is a byproduct of the electron impact dissociation of SiCl4 , and C were assumed to have a constant sticking probability of 0.5 independent of the chlorination/oxygenation. The etching products Cl2 , SiCl4 , and CCl4 were assumed to be volatile and immediately desorb. Ion incorporation probabilities were angle dependent where an ion that caused sputtering or ion-induced chemical reaction was incorporated.

Faceting of feature corners is commonly observed when they are physically sputtered by directional ions. The facet is formed at an angle corresponding to the peak yield θm for the governing angular distribution. The material surface that is inclined to the angle of incidence at θm will propagate faster than other planes, thereby overtaking them and forming the facet. Faceting occurs at corners, as the curved surface contains surface segments corresponding to the peak yield angle.

A. P. Mahorowala and H. H. Sawin: Etching of polysilicon in inductively coupled Cl2, J. Vac. Sci. Technol. B 20, 1064(2002)