Share Email Print

Journal of Micro/Nanolithography, MEMS, and MOEMS

Implementing a fully automatic macro defect detection and classification system in a high-production semiconductor fab
Author(s): Juanita Miller; Lloyd Lee; Michael Pham; David Pham; Manyam Khaja; Kathleen A. Hennessey
Format Member Price Non-Member Price
PDF $20.00 $25.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The process of shifting from human after develop inspection (ADI) to automatic detection, classification, and review of macro defects in the photolithography process is overviewed. The experiences and improvements from integration of automated macro defect detection and classification systems in a full-volume, state-of-the-art production fab is described. As the semiconductor industry moves toward adoption of integrated metrology solutions, the increased use of commercially available macro automated defect detection and classification systems provide multiple benefits for today's sub-half micron chip manufacturers. Such systems impact the yield by achieving substantial defect detection, classification accuracy, and reporting consistency. The system not only minimizes the time to detect and to fix manufacturing and processing problems but also simplifies process flow. Identification of process problems is increased considerably by providing 100% detection and review of all wafers in a lot using both stand-alone and inline integrated defect metrology. In order for the system to be utilized with the full capacity, the system has to be set up easily and quickly. Furthermore, programming and optimizing recipes used by the system should be flexible enough to identify for new defect types as they appear or when a different set of classification criteria is needed to focus on specific process problems. These systems are more effectively utilized in the production environment when tool-to-tool matching not only among inline-integrated units, but also between inline integrated and standalone units is available. Defect management methods that help increase process yield at a lower cost are becoming more important than ever in today's competitive semiconductor market. The most important task in defect management is to identify defects and their possible causes as early as possible. Early detection and diagnosis eliminates engineering time focused on existing bad wafers/dies. Another benefit is to correct the faulty processes to avoid producing additional bad wafers/dies. Early detection of photo problems can significantly increase process yield because the wafers can be reworked.

Paper Details

Date Published: 1 January 2003
PDF: 5 pages
J. Micro/Nanolith. MEMS MOEMS 2(1) doi: 10.1117/1.1528947
Published in: Journal of Micro/Nanolithography, MEMS, and MOEMS Volume 2, Issue 1
Show Author Affiliations
Juanita Miller, Texas Instruments Inc. (United States)
Lloyd Lee, Tokyo Electron Ltd. (United States)
Michael Pham, Tokyo Electron America, Inc. (United States)
David Pham, ISOA Inc. (United States)
Manyam Khaja, ISOA Inc. (United States)
Kathleen A. Hennessey, ISOA Inc. (United States)

© SPIE. Terms of Use
Back to Top