Wafer Inspection Difference Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor wafer inspection methods face challenges in sensitivity due to variations in tool conditions and imaging parameters, leading to false positives and reduced signal-to-noise ratio, as differences in critical dimensions, edge profiles, and layer thicknesses across dies can be misinterpreted as defects.
Innovation Solution
A method and system that use a computed reference image derived from multiple dies to minimize the norm of the difference image, accounting for benign variations and focusing on defect detection by adjusting imaging conditions and using a linear combination of images with adjustable filters, thereby enhancing sensitivity and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a reference image from a single neighboring die or median image is used for defect detection, then the inspection process is simple and fast, but the sensitivity is reduced and false positives increase due to tool condition variations and wafer process variations
Solution Approach 1:
The patent combines multiple reference die images into a single computed reference image through linear combination and optimization. This merging process integrates information from multiple dies to create a more robust reference that represents typical variations, thereby improving defect detection sensitivity while filtering out false positives caused by individual die variations.
Solution Approach 2:
The patent changes the parameter of the reference image from a single static image to a computed composite image with optimized coefficients. By adjusting the linear combination coefficients through optimization algorithms, the system adapts the reference image to minimize differences caused by tool and process variations, thereby enhancing measurement precision.
2Measurement precision
If traditional difference image method is used, then the inspection process is fast, but the signal-to-noise ratio is reduced due to benign variations being flagged as defects
Solution Approach 1:
The patent performs preliminary computation of the optimized reference image before the actual defect inspection process. By pre-calculating the reference image with optimized coefficients that account for typical variations, the system prepares a high-quality reference that improves signal-to-noise ratio during inspection, while the computational overhead is incurred only once rather than for each inspection.
3Measurement precision
If focus variation and tool condition variations are not compensated, then the inspection system is simple to operate, but false positives increase and detection accuracy decreases
Solution Approach 1:
The patent applies parameter changes by transforming the image data through linear combination with optimized coefficients. This mathematical transformation adjusts the reference image parameters to compensate for focus variations and tool condition differences, enabling the system to distinguish between benign variations and actual defects with higher accuracy.
Solution Approach 2:
The computed reference image acts as an intermediary between the raw inspection images and the defect detection decision. This intermediate representation filters out variations caused by tool and process conditions, serving as a mediator that improves the reliability of defect detection by comparing test images against a optimized reference rather than raw reference images.
Data Source
AI summary
Systems and methods increase the signal to noise ratio of optical inspection of wafers to obtain higher inspection sensitivity. The computed reference image can minimize a norm of the difference of the test image and the computed reference image. A difference image between the test image and a computed reference image is determined. The computed reference image includes a linear combination of a second set of images.