Data analysis methods for evaluating lithographic performance

Lithographic improvements obtained from process modifications or application of optical enhancement techniques can often be obscured within experimental noise and inconsistent data analysis techniques. At IBM, the need for a common platform for the accurate analysis of lithographic data led to the development of a new software analysis package entitled LEOPOLD. In this article, the exposure-defocus methodology taken by LEOPOLD for accurate calculation of the total lithographic process window in the presence of experimental noise is outlined; the extension of this method to the evaluation of the common process window through applications in the areas of optical proximity effects, analysis of optical enhancement techniques, and across-field exposure tool characterization is also demonstrated. (C) 1997 American Vacuum Society.