Gauge Pattern Measurement Using Unit Cell Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for pattern correction and verification in lithographic processes face challenges in accurately measuring geometric dimensions and compensating for distortions, particularly due to the reliance on shape-fitting algorithms that can introduce algorithm-dependent offsets and inaccuracies.
Innovation Solution
A method involving the synthesis of unit cells from SEM images to create a high-quality representation of gauge patterns, averaging pixel information to reduce noise and eliminate filtering distortions, and determining geometric dimensions without relying on shape-fitting algorithms, thereby enhancing signal-to-noise ratio and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If shape-fitting algorithms are used to measure geometric dimensions from SEM images, then measurement can be performed, but algorithm-dependent offsets and inaccuracies are introduced
Solution Approach 1:
The patent extracts and removes the problematic shape-fitting algorithm step from the measurement process. Instead of fitting shapes to measured features, the method directly measures geometric dimensions from the SEM image data, eliminating the source of algorithm-dependent offsets and inaccuracies.
Solution Approach 2:
The patent replaces the mechanical/algorithmic shape-fitting process with a direct measurement approach. By substituting the iterative fitting algorithm with direct geometric dimension extraction, the method eliminates the inherent inaccuracies and offsets introduced by the fitting process.
2Measurement precision
If filtering is applied to SEM images to reduce noise, then noise is reduced, but filtering distortions are introduced
Solution Approach 1:
The patent extracts and removes the filtering step from the image processing chain. By directly measuring geometric dimensions from the original SEM images without applying filters, the method eliminates the introduction of filtering distortions while maintaining adequate signal-to-noise ratio through the measurement approach itself.
Solution Approach 2:
The patent replaces the filtering operation with a direct measurement approach. Instead of modifying the image data through filtering to improve signal-to-noise ratio, the method achieves accurate measurements directly from the original images, avoiding the trade-off between noise reduction and distortion introduction.
3Measurement precision
If unit cell synthesis is performed by averaging pixel information, then noise is reduced and accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent merges multiple unit cell measurements by averaging their pixel information to create a synthesized representation. This combining approach reduces noise and improves measurement accuracy by leveraging the statistical power of multiple measurements while maintaining manageable processing complexity through efficient averaging algorithms.
Solution Approach 2:
The patent creates a synthesized copy of the unit cell by averaging pixel information from multiple measurements. This synthesized representation serves as an improved model that captures the essential geometric features while reducing noise, enabling more accurate geometric dimension extraction without requiring complex processing of individual noisy measurements.
Data Source
AI summary
A method including providing a plurality of unit cells for a plurality of gauge patterns appearing in one or more images of one or more patterning process substrates, each unit cell representing an instance of a gauge pattern of the plurality of gauge patterns, averaging together image information of each unit cell to arrive at a synthesized representation of the gauge pattern, and determining a geometric dimension of the gauge pattern based on the synthesized representation.