Imaging detection and classification of particulate contamination on structured surfaces

Abstract

We present new imaging techniques for the detection and classification of particulate contamination on structured surfaces. This allows for cleanliness inspection directly on the sample. Classical imaging techniques for particle detection, such as dark-field imaging, are typically limited to flat surfaces because structures, scratches, or rough surfaces will give similar signals as particles. This problem is overcome using stimulated differential imaging. Stimulation of the sample, e.g. by air blasts, results in displacement of only the particles while sample structures remain in place. Thus, the difference of images before and after stimulation reveals the particles with high contrast. Cleanliness inspection systems also need to distinguish (often harmful) metallic particles from (often harmless) nonmetallic particles. A recognized classification method is measuring gloss. When illuminated with directed light, the glossy surface of metallic particles directly reflects most parts of the light. Non-metallic particles, in contrast, typically scatter most of the light uniformly. Here, we demonstrate a new imaging technique to measure gloss. For this purpose, several images of the sample with different angles of illumination are taken and analyzed for similarity.