Defect Inspection System Using Signal Intensity Recipes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect inspection systems require significant skill and time to set inspection conditions, especially when no sample is available, and existing methods do not assist in selecting optimal conditions efficiently, leading to lengthy evaluation times and high workloads.
Innovation Solution
A semi-automated defect inspection system that accumulates and compares inspection images with reference images, digitalizes defect signal intensities, and repeatedly adjusts conditions to generate an inspection condition recipe, enabling easy and rapid condition setting without a sample.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual inspection condition selection is performed by skilled workers through trial and error, then defect detection sensitivity can be optimized, but considerable skill requirement and heavy workload are incurred
Solution Approach 1:
The system performs self-evaluation of inspection conditions by automatically calculating defect signal intensities and misinformation ratios through simulation, eliminating the need for manual trial-and-error optimization by skilled workers. The condition evaluation unit autonomously assesses multiple inspection conditions and selects the optimal one based on quantitative criteria.
Solution Approach 2:
The system performs preliminary evaluation of inspection conditions through simulation before actual inspection, calculating defect signal intensities and misinformation ratios in advance. This allows the selection of optimal inspection conditions prior to sample inspection, reducing the need for iterative manual adjustment.
2Measurement precision
If entire wafer inspection is performed on multiple test conditions to evaluate inspection sensitivity, then optimal condition selection is enabled, but unrealistically long prior evaluation time is required
Solution Approach 1:
The system creates simulated copies of defect patterns and inspects these virtual replicas under various conditions through simulation, rather than performing actual inspections on physical samples. This allows rapid evaluation of multiple inspection conditions without the time-consuming process of inspecting entire wafers repeatedly.
Solution Approach 2:
The system performs preliminary evaluation of inspection conditions through simulation before actual inspection, calculating defect signal intensities and misinformation ratios in advance. This allows the selection of optimal inspection conditions prior to sample inspection, reducing the need for iterative manual adjustment.
3Measurement precision
If simulation-based defect classification is performed according to sizes and shapes, then defect characterization is improved, but inspection condition setting assistance is not provided
Solution Approach 1:
The system provides feedback by calculating and displaying defect signal intensities and misinformation ratios for different inspection conditions, enabling operators to select optimal conditions. The condition evaluation unit feeds back quantitative assessment results that directly guide inspection condition selection, bridging the gap between simulation and practical application.
Data Source
AI summary
A defect inspection system is disclosed for easily setting inspection conditions and providing an inspection condition and a defect signal intensity to an operator. The defect inspection system digitizes a defective image, and a reference image corresponding thereto and a mismatched portion of the defective image and the reference image as a defect signal intensity and accumulates them in association with the inspection condition. The inspection conditions are changed to repeat evaluations while repeating accumulating works until the evaluation of all the inspection conditions in a set range is completed. A recipe file including the accumulated conditions having the high defect signal intensity and an inspection condition item distribution as a inspection condition recipe is automatically outputted and provided to the operator.