Weighted XOR Density for IC Pattern Matching Weak Spots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pattern matching methods in semiconductor integrated circuit (IC) designs are not sensitive to the placement of individual edges and struggle to identify weak spots effectively, leading to challenges in predicting patterning failures during defect inspection and failure analysis.
Innovation Solution
The method involves automatically selecting and aligning areas of an IC design with known patterns, performing an XOR comparison to identify unweighted shapes, and weighting these shapes based on their location relative to a weighting zone pattern to calculate a dissimilarity measure, thereby determining the match factor between the design area and the known pattern.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional pattern matching is used to identify design weak points in IC designs, then the overall pattern configuration can be analyzed, but the method is not sensitive to placement of individual edges and struggles to identify weak spots effectively
Solution Approach 1:
The patent applies local quality by introducing weighting zones that assign different importance levels to different spatial regions within a pattern. The center region receives higher weight while peripheral regions receive lower weight, allowing the system to focus sensitivity on critical areas where individual edge placement matters most for identifying weak spots in IC designs.
Solution Approach 2:
The patent segments the pattern matching process into multiple components: unweighted shape comparison to identify differences, weighting zone application to prioritize certain regions, and dissimilarity calculation to quantify matches. This segmentation allows the system to separately optimize for edge placement sensitivity while maintaining overall pattern configuration analysis.
2Reliability
If optical/photoresist/etch models are used for defect inspection, then patterning failures can be detected, but the models are not capable of predicting patterning failures by defect inspection alone
Solution Approach 1:
The patent merges multiple analysis approaches by combining unweighted shape difference identification with weighted dissimilarity measurement. This hybrid approach integrates the strengths of both conventional pattern matching (overall configuration) and edge-sensitive methods (individual placement), creating a more robust system that overcomes the limitations of using optical/photoresist/etch models alone.
3Extent of automation
If conventional pattern matching is used to find weak spot patterns, then the process can be automated, but it struggles to find the weak spot pattern that most closely matches the IC design shapes
Solution Approach 1:
The patent implements feedback through the dissimilarity calculation that quantifies how well a known pattern matches a design area. This numerical feedback (dissimilarity measure) allows the automated system to evaluate match quality and identify the weakest matching pattern, thereby improving the accuracy of weak spot identification while maintaining automation.
Data Source
AI summary
A known pattern area (of known shapes) is aligned with a first area of an IC design (by overlaying and centering the known pattern area on the first area). The first area is compared to the known pattern area to identify dissimilarities between shapes in the first area and the known shapes in the known pattern area as unweighted shapes. The unweighted shapes are weighted (based on the location of the unweighted shapes relative to a weighting zone pattern) to produce weighted shapes; and a dissimilarity measure between the first area and the known pattern area is calculated by summing the weighted shapes. A determination of whether the first area matches the known pattern area is made (based on whether the dissimilarity measure exceeds a threshold) and the same is output.