|
The cover illustrates scanning electron microscope (SEM) images of a heavily gold-coated oxide test sample, on the left, and their two-dimensional Fourier frequency magnitude distributions, on the right. The sharper image, at the bottom, has larger high spatial frequency components compared to the top image. A statistical measure of SEM image sharpness has been developed in terms of the multivariate kurtosis of the 2D spatial spectrum of the image. This statistical algorithm has been demonstrated to be useful for monitoring the performance of manual, semi-automated and automated SEM instrumentation, and it has been implemented in a user-friendly stand-alone analysis system that was recently developed by NIST in collaboration with Hewlett-Packard and Spectel. This work was recognized by a 1999 Department of Commerce Silver Medal, which was awarded jointly to Michael Postek (Manufacturing Engineering Laboratory) and Nien Fan Zhang (SED).
Date created: 7/20/2001 |