In 1965 when Gordon Moore made his famous observation regarding the exponential growth of semiconductor device capacity, little consideration was given to the fact that the volume of data required to manage the manufacturing process would follow suit. To stay the course predicted by Moore's Law, it will be required that a rapid reduction in process data be achieved through its conversion to useful process control information. This can be partially accomplished through the introduction of new automation technologies that can assimilate manufacturing data from inspection equipment to assist the engineer in the rapid root-cause diagnosis of defect generating mechanisms. These analysis tools will be required to achieve cost-effective yield learning in the next generation fab. In this article, we describe an emerging technology known as Spatial Signature Analysis (SSA) which automates the interpretation of product wafer defect data. SSA is an artificial intelligence method that has been developed in partnership between SEMATECH, Austin, Texas, and the Oak Ridge National Laboratory, Oak Ridge, Tennessee. The method relies on capturing operator experience through a teaching method to emulate the human response to various manufacturing situations. This has been successfully accomplished through the development and application of an image processing-based, fuzzy classifier system. The technique uses data collected from current in-line inspection tools to interpret and rapidly identify characteristic patterns, or "signatures", that are uniquely associated with the manufacturing process. The SSA system then alerts fabrication engineers to probable yield-limiting conditions that require attention. The system has been validated at three major manufacturing sites around the U.S. and is now available as a product through several commercial suppliers. We conclude by discussing future directions required for this and similar technologies if next generation productivity goals are to be achieved.
[1]
C. H. Stapper,et al.
Fact and fiction in yield modeling
,
1989
.
[2]
Richard O. Duda,et al.
Pattern classification and scene analysis
,
1974,
A Wiley-Interscience publication.
[3]
Kenneth W. Tobin,et al.
Image paradigm for semiconductor defect data reduction
,
1996,
Advanced Lithography.
[4]
U. Kaempf,et al.
The binomial test: a simple tool to identify process problems
,
1995
.
[5]
Allan Y. Wong.
A statistical parametric and probe yield analysis methodology [IC manufacture]
,
1996,
Proceedings. 1996 IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems.
[6]
Keinosuke Fukunaga,et al.
Introduction to statistical pattern recognition (2nd ed.)
,
1990
.
[7]
Costas J. Spanos,et al.
Semiconductor yield improvement: results and best practices
,
1995
.
[8]
Kenneth W. Tobin,et al.
Automatic classification of spatial signatures on semiconductor wafer maps
,
1997,
Advanced Lithography.
[9]
James M. Keller,et al.
A fuzzy K-nearest neighbor algorithm
,
1985,
IEEE Transactions on Systems, Man, and Cybernetics.
[10]
S. Hyakin,et al.
Neural Networks: A Comprehensive Foundation
,
1994
.