Virtual Metrology Using Multi-Network Time-Series Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual metrology techniques face challenges in achieving highly accurate inspection data estimation for manufacturing processes, particularly in semiconductor manufacturing, due to the complexity and variability of processing units.
Innovation Solution
A virtual metrology apparatus and method utilizing a configuration that includes an acquisition unit and a training unit with multiple network sections trained through machine learning to process time series data, allowing for detailed analysis of processing units by dividing and normalizing data across different aspects, thereby enhancing inference accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single network section is used for virtual metrology, then the device complexity is low, but the inference accuracy is insufficient
Solution Approach 1:
The patent divides the processing system into multiple network sections, where each network section processes different aspects or features of the manufacturing data independently. This segmentation allows each section to specialize in specific patterns, improving overall inference accuracy while maintaining manageable complexity through modular architecture
Solution Approach 2:
The patent combines the outputs from multiple network sections to produce the final virtual metrology result. By merging the specialized processing results from different network sections, the system achieves higher inference accuracy than any single network section could provide alone
2Measurement precision
If multiple network sections are used to analyze manufacturing processes from multiple angles, then the inference accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent segments the analysis task across multiple network sections, with each section focusing on specific manufacturing parameters or process aspects. This division enables comprehensive multi-angle analysis while keeping individual network sections relatively simple and manageable
Solution Approach 2:
The multiple network sections are designed to handle different aspects of manufacturing data processing, creating a multi-functional system that can analyze various parameters simultaneously. This universality allows the system to improve inference accuracy across different manufacturing scenarios without requiring entirely separate systems for each analysis type
Data Source
AI summary
A virtual metrology apparatus, a virtual metrology method, and a virtual metrology program that allow a highly accurate virtual metrology process to be performed is provided. A virtual metrology apparatus includes an acquisition unit configured to acquire a time series data group measured in association with processing of a target object in a predetermined processing unit of a manufacturing process, and a training unit configured to train a plurality of network sections by machine learning such that a result of consolidating output data produced by the plurality of network sections processing the acquired time series data group approaches inspection data of a resultant object obtained upon processing the target object in the predetermined processing unit of the manufacturing process.