Semiconductor Virtual Metrology With DNN Trend Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity and number of processes in semiconductor manufacturing, coupled with long turn-around times, necessitate more immediate management and stable yield, requiring accurate measurement value estimation and identification of factors affecting these values.
Innovation Solution
A semiconductor manufacturing process measurement system utilizing a processor to collect data, preprocess it using DNN, normalize and handle missing data with autoencoders, select key data, and detect trends to determine data contributions, incorporating features like hybrid recursive feature selection and explainable AI for anomaly detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of processes in semiconductor manufacturing increases to handle design complexity, then manufacturing capability improves, but process complexity and management difficulty increase
Solution Approach 1:
The patent segments the complex manufacturing process into multiple sub-processes and monitors each separately using dedicated sensors. The measurement system divides process parameters into different categories (temperature, pressure, flow rate, etc.) and applies specific processing methods to each, making the complex system manageable through modular decomposition
Solution Approach 2:
The patent introduces an AI-based measurement value estimation system as an intermediary between raw sensor data and final quality decisions. This intermediary process synthesizes data from multiple sensors, estimates critical measurement values, and provides actionable insights, simplifying the management of complex processes
2Measurement precision
If traditional measurement methods are used, then system simplicity is maintained, but measurement precision and reliability are insufficient
Solution Approach 1:
The patent replaces traditional mechanical measurement systems with AI-based virtual measurement. Instead of physical contact with products, the system uses sensor data and machine learning models to estimate measurement values, achieving higher precision without the complexity of sophisticated physical measurement equipment
Solution Approach 2:
The patent creates a universal measurement system that can handle multiple measurement types (dimensional measurements, material property measurements, etc.) through a single AI platform. This multi-functional approach achieves high measurement precision across different parameters without requiring separate complex systems for each measurement type
3Speed
If immediate management is implemented to respond to outliers, then response speed improves, but system complexity increases
Solution Approach 1:
The patent implements real-time feedback mechanisms where sensor data is continuously monitored, measurement values are estimated, and alerts are generated immediately when outliers are detected. This automated feedback loop enables rapid response to process deviations without requiring complex manual management systems
Solution Approach 2:
The measurement system performs self-monitoring and self-diagnosis, automatically detecting outliers and generating alerts without external intervention. This self-service capability enables immediate management response while keeping the system relatively simple through autonomous operation
Data Source
AI summary
A semiconductor manufacturing process measurement system and a semiconductor manufacturing process measurement method are provided. The semiconductor manufacturing process measurement system includes a memory; and a processor configured to execute a program stored in the memory, wherein the program is configured to be executed by the processor to cause the semiconductor manufacturing process measurement system: collect data from a semiconductor manufacturing apparatus; preprocess the data in consideration of characteristics of the semiconductor manufacturing apparatus; acquire an estimated measurement value using DNN (Deep Neural Network); and detect a trend of the estimated measurement value over time, and determine a contribution of the data based on the trend.