Wafer Array Edge Identification via Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wafer inspection systems fail to accurately identify the edges of array areas, leading to reduced sensitivity in defect detection due to positioning uncertainty, which results in critical defects near the edges being missed.
Innovation Solution
A computer-implemented method that determines the edge of a care area by analyzing difference images and expanding the original care area in both x and y directions using image processing techniques, such as array edge search algorithms and pattern recognition, to accurately identify the 'true' edges and increase the inspected area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the care area border is reduced to increase inspected area, then defect detection sensitivity is improved, but positioning uncertainty causes edge accuracy to deteriorate
Solution Approach 1:
The patent applies preliminary action by performing edge identification algorithms before the actual inspection to determine the true array area edges. The system pre-processes images to locate edges, then uses this information to expand the care area border outward from the true edges, ensuring that the expanded area still maintains accurate edge positioning while increasing the inspected region.
Solution Approach 2:
The patent replaces mechanical positioning methods with optical/image processing methods for edge identification. Instead of relying solely on stage positioning accuracy, the system uses image analysis algorithms (such as gradient-based edge detection or pattern recognition) to optically determine the true array area edges, thereby achieving higher edge identification accuracy without being constrained by mechanical positioning uncertainty.
2Measurement precision
If array mode inspection is used to improve defect detection sensitivity, then sensitivity is improved, but the care area border reduces the inspected area
Solution Approach 1:
The system performs preliminary edge identification to determine the true array area edges before setting the care area border. By knowing the true edges from image analysis, the system can expand the care area border outward from these true edges, thereby increasing the inspected area while maintaining the benefits of array mode inspection for defect detection sensitivity.
Solution Approach 2:
The patent makes the care area border dynamic by allowing it to extend beyond the originally defined array area edges based on the identified true edges. The care area border is no longer a fixed shrinkage from the array area edges but is instead dynamically positioned to maximize the inspected area while maintaining accurate edge identification through image processing algorithms.
Data Source
AI summary
Methods for identifying an edge of a care area for an array area formed on a wafer and/or for binning defects detected in the array area are provided. One method for identifying an edge of a care area for an array area formed on a wafer includes determining a value for a difference image as a function of position from a position known to be inside the array area to a position known to be outside of the array area. The method also includes identifying the position that is located closest to the inside of the array area and that has the value greater than a threshold as a position of the edge of the care area.