Skip to Main Content
This paper introduces the quality inspection of nanoscale patterns produced by the Laser Interference Lithography (LIL) technology using image analysis techniques. In this paper, patterns of two-beam and four-beam interferences are considered. Image analysis techniques based on the Hough transform (HT) and Maximum Likelihood Estimation (MLE) have been applied to detect and estimate various quality parameters for the two types of textures. The HT and a modified grey-scaled HT are introduced as a global approach for the analysis of the two-beam interference patterns. Surface parameters, such as the period, depth, and their uniformities, can be obtained directly without a priori knowledge of the textures. Due to different pattern structures and strong noise effects, the four-beam patterns are dealt with a different approach, using a statistical method based on the likelihood function to estimate each circle's center and shape. Taking into consideration of noises and defects, another further rejection step is introduced to filter out noises and defects. Results from experimental samples are presented.