A comparison of backscattered electron and optical images for submicron defect detection

A rapid and automated inspection system is a necessity for the detection of defects in x ray and optical lithography masks. Monte Carlo techniques and optical simulation are used to examine the relative qualities of backscattered electron signals and transmitted optical images for use in a mask inspection system. The backscattered electron signals from one and two dimensional submicron structures are examined for different take‐off angle ranges and beam sizes and compared to the corresponding optical images for different wavelengths and numerical apertures. It is shown that signal quality in electron‐beam inspection systems can improve as defect size decreases while the optical images tend to degrade. An examination of shot noise signal‐noise ratio characteristics shows that shot noise alone should not be a fundamental limit in a well designed inspection system.