Aerial Image Feature Extraction for Lithography Pattern Coverage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for test pattern sampling and pattern coverage inspection in semiconductor manufacturing rely on engineering judgment and are limited to one-dimensional cutlines, leading to low reliability and failure to utilize two-dimensional contour measurements, which are essential for accurate pattern coverage and defect detection in photolithography processes.
Innovation Solution
The implementation of a computer-implemented method using artificial neural networks to automatically extract features from aerial images, allowing for clustering and selection of test samples, thereby improving pattern coverage inspection by leveraging deep learning for efficient data collection and reducing the number of required test patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional engineering judgment methods are used for test pattern sampling, then the process is simple and familiar, but the reliability and accuracy of pattern coverage inspection deteriorates
Solution Approach 1:
The patent replaces traditional engineering judgment methods (manual/mechanical process) with an artificial neural network-based automatic feature extraction system. The neural network automatically identifies and extracts critical features from aerial images, substituting human expert judgment with an automated intelligent system that provides more reliable and consistent pattern coverage inspection results.
Solution Approach 2:
The patent introduces an artificial neural network as an intermediary between the aerial images and the pattern coverage inspection process. This intermediary automatically extracts meaningful features from the images, serving as a bridge that transforms raw image data into actionable inspection information, thereby improving reliability without requiring direct human intervention.
2Measurement precision
If one-dimensional cutlines are used for measurement, then the measurement process is simple, but the measurement precision and defect detection capability deteriorates
Solution Approach 1:
The patent transitions from one-dimensional cutline measurements to two-dimensional contour measurements using aerial images. By utilizing the full two-dimensional information available in aerial images instead of reducing it to one-dimensional cutlines, the system achieves superior measurement precision and defect detection capability while capturing comprehensive pattern coverage information.
Solution Approach 2:
The patent uses aerial images as a comprehensive copy or representation of the actual patterns on the substrate. These aerial images contain complete two-dimensional pattern information that can be analyzed without physically measuring the substrate, enabling precise measurement and defect detection through image analysis rather than direct physical measurement.
3Productivity
If comprehensive test patterns are used for full coverage inspection, then the inspection coverage is complete, but the number of test samples and processing time increases
Solution Approach 1:
The patent extracts only the most critical and informative features from aerial images using an artificial neural network. Instead of analyzing all possible test patterns equally, the system identifies and extracts key features that are most relevant for pattern coverage inspection, thereby reducing the number of test samples needed while maintaining comprehensive inspection coverage.
Solution Approach 2:
The patent applies different levels of inspection focus to different regions and features. The neural network identifies areas and features that require detailed examination versus those that can be inspected more broadly. This localized quality approach ensures comprehensive coverage while concentrating resources on critical areas, improving efficiency without sacrificing inspection thoroughness.
Data Source
AI summary
According to one or more embodiments of the present invention a computer-implemented method for fabricating a chip includes generating, using an aerial image generation system, a set of aerial images for a chip layout, the set of aerial images including an aerial image corresponding to each region from the chip layout. The method further includes automatically determining, using an artificial neural network, a feature vector for each aerial image from the set of aerial images. The method further includes clustering the aerial images using their corresponding feature vectors. The method further includes selecting, as test samples, a predetermined number of aerial images from each cluster. The method further includes performing a pattern coverage inspection of the chip layout using the aerial images that are selected as test samples.