Utilizing maximum likelihood deblurring algorithm to recover high frequency components of scanning electron microscopy images

摘要

Applying an image deblurring technique used in image processing fields, scanning electron microscopy (SEM) images were enhanced by recovering the high frequency components of the image, thereby reducing image blur. The accuracy of the deblurring method was tested by examining isolated trenches, dense lines:trenches, and unpatterned photoresist. The images were artificially blurred, then deblurred. They were also rotated and compared. Any periodic structures on the image must have a pitch larger than the expected image blur. When that qualification is satisfied, a statistical confidence level far greater than 99% demonstrated that the method does indeed deblur properly and as desired. To some degree, the deblurring technique can be used to determine stage drift, aperture misalignment, astigmation, or other problems with the SEM that affect image blur.