Identification and Classification of Contaminations on Wafers Using Hyperspectral Imaging

We present a new characterization technique based on hyperspectral imaging applied to silicon wafers. It combines the measurement of spatially and spectrally resolved reflection features and a dedicated subsequent data analysis. This method allows for a rapid localization and classification of defects and contaminations on wafers. Thus, it complements standard imaging techniques such as infra-red or luminescence imaging. In our work, we show that hyperspectral imaging is capable of separating clean and contaminated regions including a classification of the contamination type and a quantification of the coverage.