Data processing: artificial intelligence – Knowledge processing system – Knowledge representation and reasoning technique
Reexamination Certificate
2000-12-14
2003-10-21
Gordon, Paul P. (Department: 2121)
Data processing: artificial intelligence
Knowledge processing system
Knowledge representation and reasoning technique
C706S012000, C706S934000
Reexamination Certificate
active
06636843
ABSTRACT:
BACKGROUND OF INVENTION
1. Field of Invention
Invention relates to measurements and evaluation of grating profile dimensions and more particularly to the characterization, classification, and use of grating profile data.
2. Description of Related Art
Spectroscopic reflectometry and ellipsometry are used to beam light on a grating and measure the spectra of reflected signals. Current practices basically use an empirical approach where the spectra of reflected light is measured for a known width of features in a grating, a time consuming and expensive even for a limited library of profiles. If such a library were built using either empirical or calculated spectrum data, the library would be very useful to have in a real-time environment. A grating profile library of feature dimensions and spectrum data could be used for matching against input grating spectrum data from metrology devices and get the grating dimensions associated with the matching library profile.
However, manufacturers are not only interested in the basic measurements of the grating dimensions but also in the shape of the grating features. If the shape of the grating features were correlated to the fabrication variables, manufacturers would have information upon which to base corrective actions. Knowledge of the grating shapes and other characterizations of the grating profile are also important in evaluating the acceptability of gratings. Thus, there is a need to associate profile shape data to a profile. Although possible, manual labeling of a typical run-time profile library of about 400,000 to 500,000 profiles is a time consuming and expensive task. The characterization and classification of library profiles should be done in a way that solves the laborious tasks of inputting profile shape data. Once an efficient way of characterizing the profiles is available, profile library data and statistical measurements may be used to define acceptable ranges and tolerances in the grating profile dimensions and shapes.
SUMMARY OF INVENTION
The invention resides in a method and a system of characterizing a data space of grating profile data such as a grating profile library. In one embodiment, the method for characterizing grating profile data comprises requesting classification of a grating profile data space consisting of grating profile data points, clustering the grating profile data points into a number of clusters, and associating profile shape data to each cluster. Profile shape data comprise a profile shape label, a profile shape graphical image, and a profile shape description.
One system for classifying grating profile data comprises a grating profile data space containing grating profile data points; a computer; and a cluster generator for generating a plurality of clusters of grating profile data points from the data space containing grating profile data points by utilizing a clustering algorithm, for associating profile shape data to each cluster, and for linking the associated profile shape data to the grating profile data points belonging to each cluster.
In one embodiment, the present invention includes a method for evaluating grating profile data by accessing profile data from a grating profile library instance selected as the closest match to an input grating profile; comparing the profile data from the grating profile library instance with a set of acceptable ranges of profile data for the application; flagging the input grating profile if the profile data is outside the set of acceptable ranges of profile data for the application; and presenting the profile data and flags associated with the grating profile. Presentation of the profile data and flags includes a display of a profile shape image and or a two-dimensional graph of key profile dimensions, and an alert identifying the profile shape data as exceeding the acceptable parameter ranges.
The present invention further includes in one embodiment a system for evaluating grating profile shapes. The system comprises a computer; a profile cluster database for storing the profile shape data for a cluster of profiles; a library of grating profiles; and a profile shape evaluator; where the profile shape evaluator accesses the profile shape data from a grating profile library instance selected as the closest match to an input grating profile, compares the profile shape data with a set of acceptable ranges of profile data for the application, and flags the input grating profile if the profile shape data is outside the set of acceptable ranges of profile data for the application.
REFERENCES:
patent: 5982891 (1999-11-01), Ginter et al.
T. Feder, et al, Optimal Algorithms for Approximate Clustering, 1988, ACM, Proceedings of the 20thACM Symposium on Theory of Computing, 434-444.*
T. Feder, et al., “Optimal Algorithms for Approximate Clustering”, Proceedings of the 20th ACM Symposium on Theory of Computing, pp. 434-444, 1988.
T. Gonzalez, “Clustering to Minimize the Maximum Intercluster Distance”, Theoretical Computer Science, vol. 38, pp. 293-306, 1985.
S. Kirkpatrick, et al., “Optimization by Simulated Annealing”, Science 220, 4598 (May), pp. 671-680, 1983.
Doddi Srinivas
Jakatdar Nickhil
Niu Xinhui
Gordon Paul P.
Hirl Joseph P.
Morrison & Foerster / LLP
Timbre Technologies, Inc.
LandOfFree
System and method for grating profile classification does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for grating profile classification, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for grating profile classification will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3124088