Wafer Yield Graph Matching for Process Parameter Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual analysis of yield problems in semiconductor wafer manufacturing is inefficient, time-consuming, and lacks accuracy, making it difficult to identify and address process parameters causing yield issues effectively.
Innovation Solution
A data analysis method that involves generating stacked graphs for yield problems and measurement data, performing graph matching and correlation calculations, and conducting weighted calculations to automatically determine the target measurement parameters causing yield issues, thereby improving analysis efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual analysis methods are used to identify yield problems in wafer manufacturing, then the analysis process can be performed with simple tools, but the analysis efficiency is low and time consumption is high
Solution Approach 1:
The patent replaces manual mechanical analysis with an automated computer-based system that performs graph matching and correlation calculations. The system automatically compares stacked graphs of yield data with measurement data graphs, eliminating the need for manual inspection and significantly improving analysis efficiency while reducing time consumption.
Solution Approach 2:
The system enables self-service analysis by automatically performing graph generation, matching, and correlation analysis without requiring manual intervention. The computer executes algorithms that autonomously identify yield problems and their causes, allowing the system to serve itself in the analysis process.
2Measurement precision
If manual analysis methods are used to identify yield problems, then the system complexity is low, but the measurement precision and accuracy are insufficient
Solution Approach 1:
The patent replaces simple manual analysis with a sophisticated computer-based system that uses graph matching algorithms and correlation calculations. This substitution enables precise automated comparison of yield data with measurement data, significantly improving analysis accuracy despite the increased system complexity.
Solution Approach 2:
The system creates graphical representations (stacked graphs) of yield data and measurement data, allowing for visual and computational comparison. These graph copies enable precise analysis by transforming raw data into comparable visual formats that can be systematically evaluated.
3Reliability
If automated graph matching and correlation analysis are implemented, then analysis accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent segments the analysis process into distinct modules: graph generation from yield data, graph generation from measurement data, graph matching comparison, and correlation calculation. This segmentation allows each function to be independently implemented and optimized, improving reliability while managing system complexity through modular design.
Solution Approach 2:
The computer-based system performs multiple functions including data processing, graph generation, graph matching, and correlation analysis within a single integrated platform. This multi-functionality improves analysis reliability by ensuring consistent processing across all stages while avoiding the need for multiple separate systems.
Data Source
AI summary
A data analysis method includes: a target yield problem stacked graph corresponding to a wafer list is obtained, and measurement data stacked graphs of the wafer list under different types of tests are obtained; graph matching is performed on the target yield problem stacked graph and each of the measurement data stacked graphs to obtain matching degree data corresponding to the target yield problem stacked graph and each of the measurement data stacked graphs; correlation data corresponding to each of the measurement data stacked graphs and the target yield problem stacked graph is calculated; and weighted calculation is performed on the matching degree data and the correlation data, and a target measurement parameter causing a target yield problem is determined according to a result of the weighted calculation.