Laser Surface Scanning Inspection Systems (SSISs) have been widely used to automate wafer inspection processes. Although SSISs are capable of detection of particles below 0.100μm, visual inspections continue to be used in silicon wafer manufacturing facilities. Visual inspections are required because conventional SSISs are not capable of identifying and classifying a variety of defects, such as ring defects, slips, stacking faults, voids, and mounds. As the line width decreasing, the ring defect detection is becoming one of the most important analysis functions for wafer inspection because these defects are yield killing defects of the chip production. This paper discusses a new ring defect detection method for wafer inspection using an improved SSIS, such as our CR8xE, equipped with both a dark channel and a new Quad-Cell Detection (QCD) light channel. Some sophisticated algorithms have been developed to detect and analyze ring defects on wafers. This new function provides the capability of successfully identifying and quantifying ring defects. Our experimental results show the efficacy and efficiency of our algorithm.
展开▼
