Plasma development of a silylated bilayer resist: Effects of etch chemistry on critical dimension control and feature profiles

We have previously described a positive-tone, silylated, dry-developed bilayer resist process for which 0.2 mu m features with an aspect ration of 4.5 were obtained using deep-UV (248 nm) exposure. The plasma development of the resist is comprised of two steps: an initial etch with a mixture of Ar and Cl-2 to remove the thin layer of silylated resist in the exposed regions followed by pattern transfer using an oxidative plasma. In this process feature profiles with minimal undercut were obtained when CO2 was used instead of O-2 in the pattern transfer step. We have investigated the mechanism of these etching processes using x-ray photoelectron spectroscopy (XPS) and trilevel resist processing and found that there was little deposition and that the selectivity of the silylated resist decreased by a factor of 2 in the CO2 plasma compared to the O-2 plasma. The lower selectivity leads to increased erosion of the silicon-containing mask and reduced critical dimension (CD) control. in order to obtain good CD control and further improve the feature profiles we have evaluated other etching gases which are considered to undergo sidewall deposition during the pattern transfer step, including SO2 and mixtures of O-2 with N-2 and SO2. Using trilevel resist structures, the degree of lateral etching was determined for each plasma. The extent of deposition in each plasma was monitored through surface analysis using XPS. In the silylated bilayer resist, the feature profiles are significantly influenced by the etching rates, selectivities, and degree of overetch which were studied by varying the etcher power levels, and etching times. In this article we will discuss the results of these studies and show how 0.2 mu m features with nearly vertical sidewalls were obtained. (C) 1995 American Vacuum Society.