A statistical approach to optimization of alumina etching in a high density plasma

Abstract

Inductively coupled plasma (ICP) reactive ion etching of Al2O3 with fluorine-based gas chemistry in a high density plasma reactor was carried out in an initial investigation aimed at data storage applications. A statistical design of experiments was implemented to optimize etch performance with respect to process variables such as ICP power, platen power, direct current (dc) bias, and pressure. Both soft photoresist masks and hard metal masks were investigated in terms of etch selectivity and surface properties. The reverse power dependence of dc bias on the ratio of ICP to platen power was elucidated. Etch mechanisms in terms of physical and ion enhanced chemical etchings were discussed. The F-based chemistry greatly enhances the etch rate of alumina compared to purely physical processes such as ion milling. Etch rates as high as 150 nm/min were achieved using this process. A practical process window was developed for high etch rates, with reasonable selectivity to hard masks, with the desired profile, and with low substrate bias for minimal damage.