Characterization of profile dependency on nitride substrate thickness for a chemically amplified I-line negative resist

During implementation of a chemically amplified I-line negative resist (INR) into the eight- inch-wafer manufacturing line at IBM's facility in Essex Junction, Vermont we found that the resist profiles were being undercut during development on silicon-nitride surfaces. The undercut was attributed to a `poisoning' of the photoacid generated in the resist during exposure by the nucleophilic characteristics of the silicon-nitride surface. A second-generation negative resist has since been formulated by IBM, which greatly reduces these undercut effects but does not eliminate them. This newer version of INR is referred to as INR2. To further understand the effects of silicon nitride on the I-line negative-resist profiles, we used Prolith/2 (v3.05) 3 to model the underlying stack and evaluate the performance of INR2. Silicon monitors with oxide and different nitride thicknesses were patterned with INR2 using 4-Mb DRAM manufacturing ion-implant block masks and associated photo processing. The results indicated that the undercut observed by this chemically amplified negative resist on silicon nitride was not only chemical in nature, but related to the optical properties of the nitride substrate. Simulation results obtained using Prolith/2 are give and compared with resist profiles obtained from wafers processed with INR2; the accuracy of the Prolith/2 model to predict slope profiles for INR2 is also discussed.