Defect Quantification Using Dynamic Measurement Regions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for defect inspection on semiconductor wafers fail to set appropriate measurement locations based on defect type, limiting the ability to quantify defects effectively and support various types of defects.
Innovation Solution
A defect quantification method and device that classify defect images, set measurement regions and locations based on classification results, and calculate evaluation values using measurement values from both defect and reference images to quantify defects accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If measurement regions and directions are specialized for photoresist application failure (radial form from center), then measurement efficiency for that specific defect type is improved, but adaptability to various types of defects deteriorates
Solution Approach 1:
The measurement region setting is made dynamic by automatically determining regions based on defect classification results. Instead of fixed radial regions from wafer center, the system adapts measurement regions to each defect's location and type, allowing the measurement configuration to change dynamically according to the specific defect being measured.
Solution Approach 2:
The system changes measurement parameters (region, direction, location) based on defect classification. Different defect types trigger different measurement parameter configurations, enabling the system to adapt to various defect types while maintaining measurement efficiency through automated parameter selection.
2Device complexity
If measurement is performed without basis on defect classification results, then measurement process is simplified, but ability to set appropriate measurement locations for each defect type deteriorates
Solution Approach 1:
Defect classification is performed as a preliminary action before measurement. The classification results are used to pre-determine appropriate measurement regions, directions, and locations, ensuring that measurements are tailored to each defect type while the entire process remains automated and integrated.
Solution Approach 2:
The system uses feedback from defect classification results to adjust measurement parameters. The classification outcome feeds into the measurement setup process, creating a closed-loop system where measurement configuration is continuously optimized based on defect characteristics.
3Loss of time
If only defect classification is performed without quantification, then processing time is reduced, but ability to provide detailed evaluation information for yield management deteriorates
Solution Approach 1:
The system performs continuous useful action by seamlessly integrating defect classification with quantification in an automated workflow. Instead of stopping at classification, the system continuously proceeds to measurement and quantification, providing comprehensive evaluation information without significant time penalty due to automation.
Solution Approach 2:
The system performs self-service by automatically executing the entire process from defect detection through classification to quantification without manual intervention. This automation eliminates time loss while ensuring comprehensive defect evaluation information is generated and provided for yield management.
Data Source
AI summary
To quantify the degree of a defect, and provide information useful for yield management. Disclosed is a defect quantification method wherein: a defect image is classified; a measurement region and a measurement area are set to each of the defect image and a reference image on the basis of defect image classification results, said reference image corresponding to the defect image; and an evaluation value of a defect is calculated using each of the measurement values obtained from each of the measurement areas of the defect image and the reference image, and the defect is quantified.