Off-Line Metrology on SEM Images Using Gray Scale Morphology

Discrimination and metrology results of microlithographic patterns from top-down SEM images are explored by means of morphological image analysis. The method relies on the use of various morphological filters on a top down SEM image. The resulted images are segmented in order to derive a quality factor which discriminates the candidate images as under- or fully-developed. Furthermore, the fully developed images are processed in order to extract useful measurements. The proposed image analysis methodology achieves for first time, to the authors’ knowledge, successful off-line discrimination between under-developed and fully-developed cases. For the latter case, the measuring method relies upon the evaluation of the connected regions in the SEM image after segmentation. This is expressed by the Useful Threshold Range (UTR), which corresponds to that specific value of connected regions obtained for the wider range of the threshold. The method is experimentally demonstrated by employing 72 test images from high resolution patterns. The evaluated critical pattern parameters are found in good agreement to those derived from on-line procedures.