Wafer Defects Inspection Using Image Pattern Recognition Techniques

Abstract

This research project focuses on the development of wafer inspection tool for wafer fabrication process. In wafer fabrication process, physical defects become major challenge for the Engineers to monitor, detect, reduce and correct for any faulty process occurred in the production as quickly as possible. This exercise greatly minimizes the cost of production and ensures the quality of the products is maintained. The new developed tool provides a new approach of strategy which is capable to locate the defect and immediately analyze the parameters of the defect. Defects such as porosity, foreign inclusions and cracks are common and can be easily recognized using the new system. In the system first the image is captured by high resolution camera and immediately the captured image is translated into graphical pattern format. MATLAB engine is used to support the translation procedure. Finally the graphical pattern is compared with the good pattern which is stored in the system database to decide all the parameters of defects such as type of defect, size, area, perimeter and etc.