Semiconductor Process Variance Detection via Statistical Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor process monitoring systems are inadequate in detecting variations across multiple process steps, leading to inefficiencies in identifying the root causes of yield reduction and increased costs due to complex data analysis and resource wastage.
Innovation Solution
A method involving correlation analysis that collects and preprocesses raw data from semiconductor tools, using statistic analysis methods like Factor Analysis, Principal Component Analysis, and Analysis of Covariance to generate interrelationship indices, enabling the identification of essential causes of process variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional FDC systems are used to monitor semiconductor process tools, then real-time monitoring capability is achieved, but the systems cannot effectively detect variations across multiple process steps and locate root causes
Solution Approach 1:
The patent creates a unified correlation analysis framework that can handle multiple process steps and various data types simultaneously. The system processes tool process data, first raw data, and second raw data through a common statistical analysis pipeline, enabling the same system to detect variations across different process steps without requiring separate specialized systems for each step.
Solution Approach 2:
The patent segments the complex monitoring task into three distinct data categories (tool process data, first raw data, second raw data) that are processed separately through specific statistical methods before being integrated. This segmentation allows each data type to be analyzed with appropriate methods while maintaining the ability to detect cross-step variations through the interrelationship indices.
2Measurement precision
If complex rule-setting methods are used for flaw detection analysis, then detection capability is improved, but time consumption and resource waste increase significantly
Solution Approach 1:
The patent implements self-service through automated statistical analysis methods that automatically process the segmented data and generate interrelationship indices without requiring manual rule configuration. The system performs Factor_analysis, Principal_Component_analysis, and Analysis_of_Covariance automatically, eliminating the time-consuming manual rule-setting process while maintaining high detection accuracy.
Solution Approach 2:
The patent replaces the mechanical manual rule-setting process with automated statistical computing methods. Instead of engineers manually creating and adjusting detection rules, the system uses mathematical algorithms (Factor_analysis, PCA, ANCOVA) to automatically analyze the data and identify variations, significantly reducing time consumption and resource requirements.
3Measurement precision
If manual analysis methods are used to locate variation sources, then detailed investigation is possible, but productivity and efficiency are reduced
Solution Approach 1:
The patent replaces manual analysis methods with automated statistical computing systems that can process large datasets rapidly. The system uses Factor_analysis to identify underlying factors, Principal_Component_analysis to reduce dimensionality, and Analysis_of_Covariance to quantify relationships, all automatically generating interrelationship indices that locate variation sources without manual intervention, thereby maintaining precision while dramatically improving productivity.
Data Source
AI summary
A method of detecting variance by regression model has the following steps. Step 1 is preparing the FDC data and WAT data for analysis. Step 2 is figuring out what latent variable effect of WAT data by Factor Analysis Step 3 is utilizing Principal Component Analysis to reduce the number of FDC variables to a few independent principal components. Step 4 is demonstrating how the tools and FDC data affect WAT data by Analysis of covariance model, and constructing interrelationship among FDC, WAT and tools. The interrelationship can point out which parameter effect WAT significantly. By the method, when WAT abnormal situation happened, it is easier for engineers to trace where the problem is.