Statistical Metrology - Measurement and Modeling of Variation for Advanced Process Development and Design Rule Generation

Statistical metrology is the body of methods for understanding variation in microfabricated structures, devices, and circuits. The goal of this paper is to describe key features of statistical metrology, to review the tools and methods developed to date, and present an application of statistical metrology to advanced technology and design rule development. A running application example will be used to illustrate the concepts and experience in statistical metrology. In this section, we present the variation problem being addressed, namely the variation of interlevel dielectric (ILD) thickness across the wafer and chip resulting from chemical mechanical polishing (CMP). The progression of our development of statistical metrology concepts and methods will then be presented. These phases are (I) variation assessment, (II) variation modeling, (III) semi-physical model calibration, (IV) circuit impact modeling, and (V) design rule generation. Finally, future work and a summary will be presented. In essence, this paper serves as a review of the MIT work on statistical metrology. While the references are drawn almost entirely from the MIT research, it is important to note that much work on statistical metrology has been contributed elsewhere (e.g. Spanos et al. (6) and many others) which is not reviewed or addressed in the present paper.