Wafer Peripheral Image Analysis for Dark-Field Defect Sensitivity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In wafer production, dark-field scanning detection is hindered by light interference superposition in repeated image regions, leading to reduced sensitivity and potential missed detection of critical defects due to manual errors in selecting these regions.
Innovation Solution
A method and apparatus that analyze gray-scale images of wafer peripheral circuit regions, dividing them into pixel regions and identifying repeated image regions with consistent gray-scale values to determine interference superposition positions, thereby improving detection sensitivity by accurately determining and blocking light interference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of repeated image regions is used in dark-field scanning detection, then the detection process can be performed, but detection sensitivity is reduced due to light interference superposition and manual errors
Solution Approach 1:
The patent replaces manual selection of repeated image regions with an automated image processing system that uses gray-scale analysis and pattern recognition algorithms. The system automatically identifies repeated regions by comparing gray-scale values across different areas of the wafer image, eliminating manual intervention and its associated errors. This substitution of mechanical/manual operations with automated computational methods resolves the contradiction by improving both detection sensitivity and reliability simultaneously.
Solution Approach 2:
The detection system performs self-identification of repeated image regions through automated analysis of the wafer's own image data. The algorithm extracts features from the captured image, identifies patterns with consistent gray-scale values, and determines interference regions without external manual input. This self-service capability allows the system to autonomously resolve the detection sensitivity issue while maintaining high accuracy.
2Measurement precision
If automated identification of repeated image regions is implemented, then detection sensitivity is improved, but device complexity increases
Solution Approach 1:
The patent segments the image processing task into distinct modular stages: gray-scale conversion, region division, repeated pattern identification, and interference determination. Each stage processes a specific aspect of the image data independently, allowing the complex overall task to be managed through simpler, specialized sub-processes. This segmentation reduces system complexity by breaking down the automated identification process into manageable, well-defined components.
Solution Approach 2:
The system uses parameter changes in the image data (specifically gray-scale values) as the basis for automated identification. By transforming the color image into gray-scale representation and analyzing intensity variations, the system simplifies the complexity of identifying repeated regions. This parameter-based approach converts a complex visual recognition problem into a simpler numerical analysis task, improving detection sensitivity without proportionally increasing system complexity.
Data Source
AI summary
A method and apparatus for improving sensitivity of wafer detection, and a storage medium are provided. The method including: acquiring a gray-scale image of a peripheral circuit region of a wafer; dividing the gray-scale image into at least one pixel region; determining a repeated image region in each pixel region, in which the repeated image region is a region with N*M pixels where M pixels on a same column in the repeated image region have a same gray-scale value, and N and M are integers greater than 0; and determining position information where interference superposition occurs in the peripheral circuit region of the wafer according to all repeated image regions in the gray-scale image.