Intra-Frame Defect Detection via Patch-to-Patch Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect detection methods in semiconductor manufacturing face challenges in distinguishing between defects and noise, particularly in Die-to-Die comparisons, where noise from process variations, mechanical, and electrical variations hinders sensitivity and integrity, especially as design rules shrink.
Innovation Solution
The implementation of an intra-frame defect detection method that compares microscopic object locations closer together within the same frame, reducing noise and improving sensitivity by using similar patches within an image for defect detection, and creating a new reference from other patches in the image, especially for scenarios without a traditional reference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If inter-die comparison method is used to detect defects, then defect detection coverage is improved, but noise from process variations and mechanical variations increases, reducing detection sensitivity
Solution Approach 1:
The patent segments the comparison process into two stages: first comparing patches within the same die (intra-die) to establish a local reference with minimal noise, then using this refined reference for inter-die comparison. This segmentation allows the system to maintain both broad defect coverage and high detection sensitivity by isolating the noise-prone inter-die comparison from the reference establishment process.
Solution Approach 2:
The patent introduces an intermediate reference frame created from intra-die patch comparisons as a mediator between the inspected die and the model. This intermediate reference absorbs the process variations and mechanical variations, allowing the final defect detection to focus on actual defects rather than noise. The intermediate reference acts as a buffer that preserves detection sensitivity while enabling comprehensive defect coverage.
2Productivity
If design rules are shrunk to increase device density, then productivity is improved, but noise from process variations increases relative to feature size, making defect detection more difficult
Solution Approach 1:
The patent applies local quality by performing defect detection at multiple scales: first at the patch level within dies, then at the die level, and finally at the wafer level. This multi-scale local analysis allows the system to detect defects in shrunk designs by comparing only locally relevant features, reducing the impact of global process variations that become more significant as feature sizes decrease.
Solution Approach 2:
The patent adds the dimension of spatial scale to defect detection by implementing a hierarchical comparison approach. Instead of a single-scale comparison, the system operates across multiple spatial dimensions (patch size, die size, wafer size), allowing it to distinguish between noise and actual defects even in shrunk designs where the signal-to-noise ratio is reduced.
3Measurement precision
If intra-frame patch-to-patch comparison is used, then noise is reduced and detection sensitivity is improved, but the scope of comparison is limited to locations closer together
Solution Approach 1:
The patent merges multiple comparison approaches by combining intra-die patch comparisons with inter-die comparisons in a hierarchical framework. The intra-die comparisons provide high-sensitivity local defect detection, while the inter-die comparisons extend the coverage to catch defects that may manifest differently across dies. This merging preserves both the sensitivity benefits of limited-scope comparison and the comprehensive coverage of broad comparison.
Solution Approach 2:
The patent performs preliminary intra-die patch comparisons to establish a refined reference before conducting the final defect detection. This preliminary action creates a high-quality local reference that minimizes noise, which then enables the system to safely expand the comparison scope to inter-die levels without sacrificing detection sensitivity. The preliminary comparison prepares the data structure for subsequent broader analysis.
Data Source
AI summary
A system receives, based on processing of an inspected frame of an inspected image generated by collecting signals indicative of a pattern on an article, at least one candidate defect location in the inspected frame. The system defines a candidate patch within the inspected frame. The candidate patch is associated with the candidate defect location. The system identifies at least one similar patch in the inspected frame using a predefined similarity criterion and determines whether a defect exists at the candidate defect location based on a comparison of at least a portion of the candidate patch with at least a corresponding portion of the at least one similar patch.