Wafer Defect Identification via Gray Scale and Slope Indices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inspection machines struggle to accurately distinguish between wafer scratches and grain boundaries, leading to unnecessary rework and increased costs.
Innovation Solution
A method that involves obtaining a marked frame from a wafer inspection picture, calculating a gray scale index, and using it in conjunction with a slope difference index to determine whether a trace pattern is a scratch or a grain boundary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a conventional inspection machine is used to detect wafer defects, then the inspection process is simple and fast, but the machine cannot accurately distinguish between scratches and grain boundaries, leading to misidentification
Solution Approach 1:
The patent applies parameter changes by introducing two new parameters: gray scale index (representing brightness distribution characteristics) and slope difference index (representing edge sharpness characteristics). These parameters transform the visual inspection problem into a quantitative analysis problem, enabling accurate distinction between scratches and grain boundaries through mathematical comparison rather than simple visual detection.
Solution Approach 2:
The patent uses an intermediary approach by introducing intermediate calculation steps (gray scale index and slope difference index) between the raw inspection image and the final defect classification. These intermediate parameters act as mediators that bridge the gap between simple visual detection and accurate defect identification, allowing the system to analyze subtle differences that conventional machines miss.
2Productivity
If all detected defects are treated as scratches, then no additional rework is needed for grain boundaries, but unnecessary rework time and cost are incurred on grain boundaries
Solution Approach 1:
The patent implements feedback by using the calculated gray scale index and slope difference index to provide information about the nature of each detected defect. This feedback mechanism allows the inspection system to automatically distinguish between scratches (which require rework) and grain boundaries (which do not), enabling real-time decision-making that prevents unnecessary rework and optimizes production efficiency.
3Measurement precision
If the gray scale index and slope difference index are both used for classification, then the identification accuracy is improved, but the calculation complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the defect analysis process into distinct stages: first calculating the gray scale index to assess brightness characteristics, then calculating the slope difference index to assess edge characteristics, and finally combining both indices for classification. This segmented approach breaks down a complex classification problem into manageable steps, making the calculation process more systematic and easier to implement despite the increased accuracy requirements.
Data Source
AI summary
A method for identifying a wafer is provided, which includes the following steps. A marked frame is obtained from a wafer inspection picture. A gray scale index corresponding to the marked frame is calculated based on a gray scale value corresponding to each of multiple pixels included in the marked frame. The gray scale index indicates a proportion of pixels whose gray scale values are greater than a specified value. Whether a trace pattern in the marked frame is a scratch or a grain boundary is determined based on the gray scale index.