SEM ADC (Auto Defect Classification): How it improves the cost of ownership without risk of yield loss

Abstract

To be productive and profitable in a modern semiconductor manufacturing environment, large amounts of defect data must be collected, analyzed, and summarized. SEM Automatic Defect Classification (ADC) is becoming a viable technology in the Yield Enhancement field. Most of the past work focused on optical inline review. Now that defect size is beyond the optical resolution, SEM is the essential review tool and SEM ADC is inevitable as the consequence. This paper reports the result of the ADC experiment and concludes that the state-of-the-art technology of ADC improves greatly the Cost of Ownership (CoO) without risk of yield loss. © 2007 IEEE.