Application of ADC techniques to characterize yield-limiting defects identified with the overlay of E-test/inspection data on short loop process testers

Abstract

Automatic defect classification, using a combination of CRS IMPACT/sup TM/ ADC and 4300+ SEM ADC based approaches, can significantly improve data integrity and the rate of yield-limiting defect characterization. Implementation of ADC can enhance yield learning rates since it can improve the defect classification speed and improve the accuracy of the results. The data shows that both optical and SEM ADC approaches offer their own unique advantages. Optical ADC offers a faster significantly cheaper solution with less accuracy. Higher classification accuracy is obtained from the SEM based approach due to the higher magnifications coupled with the additional Z-information available only with tilted images. The ultimate approach would involve combination of the optical and SEM based approaches to take advantage of the strengths of each system.