AVS 50th International Symposium
    Plasma Science and Technology Tuesday Sessions
       Session PS-TuA

Paper PS-TuA3
Depth Dependent Spatial Frequency Analysis of Post-Etch Sidewall Roughness

Tuesday, November 4, 2003, 2:40 pm, Room 315

Session: Dielectric Etch
Presenter: S.A. Rasgon, Massachusetts Institute of Technology
Authors: S.A. Rasgon, Massachusetts Institute of Technology
H.H. Sawin, Massachusetts Institute of Technology
A.P. Mahorowala, IBM T.J. Watson Research Center
D. Goldfarb, IBM T.J. Watson Research Center
M. Angelopoulos, IBM T.J. Watson Research Center
S.D. Allen, IBM Microelectronics Division
Correspondent: Click to Email
For the patterning of sub-100 nm features, a clear understanding of the origin and control of line edge roughness (LER) is extremely desirable. Until recently, LER studies have focused on the analysis of top-down SEM micrographs of post-developed photoresist lines. However, plasma etching processes often roughen the feature sidewalls and might form striations. This post-etch substrate LER is probably more relevant from a manufacturing perspective than the post-developed LER. The depth dependence of the post-etch sidewall morphology cannot be captured adequately by top down SEM techniques. A novel atomic force microscopy (AFM) technique developed by Reynolds and Taylor (JVST B 17(2), p. 334-344, 1999) was used to examine sidewall roughness (SWR) transfer through photoresist, BARC/hardmask, and oxide layers simultaneously. Dense line-and-space structures were cleaved parallel to the line patterns, and turned 90 degrees to access the exposed sidewall with the AFM tip. The images vividly highlight the resulting SWR structure, allowing one to observe roughness transfer through materials and determine any potential correlations. This paper studies the effect of etch chemistry, BARC/hardmask material, and resist thickness/type on the morphology and structure of SWR striation spatial frequency and spatial correlation, under conditions typically encountered during oxide etch processes. Quantitative data on RMS roughness and striation spatial frequency as a function of feature depth is collected using AFM-based techniques. Power spectral density (PSD) and correlation analysis of the sidewall AFM images allows us to track the evolution of sidewall striations through the various feature layers after each process step (lithography, BARC/hardmask open, and oxide etch). Finally, the impact of these sidewall striations on future processing steps (for instance, conformal liner deposition) is discussed.