Industrial
Inspection
Automatic
Defect Classification for Semiconductor Wafer Inspection and Review
Rapid Yield
Improvement Through Automation
Semiconductor wafer manufacturers invest much of their time in isolating
the causes of yield-impacting defects during the lithographic printing
and processing of integrated circuits on wafers. Automatic Defect Classification (ADC) automates
the tedious manual process of defect review and classification during
optical microscopy and scanning electron microscopy (SEM); a key
step in the identification of the root cause of a manufacturing
problems. ORNL has been working with the semiconductor
industry since 1991 to provide new strategies, methods, and technologies
for defect detection, segmentation, and classification for the purposes
of process characterization, control, and rapid yield improvement.
The ORNL ADC technology provides flexible investigation
and development tools for front-end-of-line (FEOL) and back-end-of-line
(BEOL) inspection and review environments and is used for both optical
and SEM imaging modalities. The
technology currently is in commercial use for SEM defect review
and post-fabrication optical inspection systems.
Base
Technology
The
ORNL ADC methods use extensive monochrome and multi-spectral features;
principle component and linear discriminant feature analysis; and
pair-wise, multi-class, and fuzzy implementations of k-nearest neighbor
and radial basis function classifiers. A
unique classifier training mechanism and interface provides intuitive
visual and numeric feedback to assist in maintenance of defect libraries.
Specifications
and Features
- Windows
and UNIX libraries
- Standard
image I/O (JPG, Tiff, BMP, etc.)
- Flexible
feature selection and analysis
- Flexible
classifier hypothesis testing
Fact
Sheet available here in PDF format.
Point
of Contact:
Kenneth W. Tobin, Ph.D.
Group Leader, Image Science & Machine Vision
Engineering Science and Technology Division
Oak Ridge National Laboratory
P.O. Box 2008, MS-6010
Oak Ridge, Tennessee 37831-6010
Office: (865) 574-8521
E-mail: tobinkwjr@ornl.gov |