Process-Sequence Mining for Queue-Time Yield Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor fabrication processes face challenges in predicting the impact of queue times (QTs) on product quality and yield due to the complexity of analyzing large amounts of data with combinatorial dependence structures, limiting the effectiveness of traditional ad-hoc investigations and individual QT optimizations.
Innovation Solution
A novel process-step sequence mining technique that encodes the entire fabrication process-step sequence into symbols, applying sequence-based analysis to identify rules governing the influence of QTs and process steps on quality and yield, using machine learning algorithms and natural language processing to reduce data complexity and uncover quality-related and yield-related rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional ad-hoc investigations and individual QT optimizations are used, then implementation simplicity is maintained, but prediction effectiveness of impact on product quality and yield deteriorates due to complexity of analyzing large amounts of data with combinatorial dependence structures
Solution Approach 1:
The patent segments the complex fabrication process into discrete process steps and queue times, representing them as sequential symbols in a process-step sequence. This segmentation transforms the overwhelming combinatorial data structure into manageable discrete units that can be analyzed systematically through sequence mining, resolving the contradiction between maintaining analysis simplicity and achieving prediction effectiveness.
Solution Approach 2:
The patent introduces sequence mining algorithms and symbolic representations as intermediaries between the raw fabrication data and the quality/yield predictions. These intermediaries process the complex combinatorial data structures, extracting meaningful patterns and relationships without requiring direct analysis of the full data complexity, thus improving prediction effectiveness while managing analytical complexity.
2Measurement precision
If sequence-based analysis of entire process-step sequence is applied, then prediction accuracy of quality and yield is improved, but computational burden and data processing complexity increase
Solution Approach 1:
The patent applies preliminary encoding of process steps and queue times into symbolic representations before sequence mining analysis. This preliminary action transforms raw data into a standardized symbolic format that facilitates more efficient sequence-based analysis, reducing the computational burden and processing time required for accurate quality and yield predictions while maintaining comprehensive sequence analysis.
3Loss of information
If comprehensive process-step sequence analysis is performed, then identification of quality and yield influencing parameters is improved, but ease of operation deteriorates due to vast data sets
Solution Approach 1:
The patent extracts key quality and yield influencing parameters from the comprehensive process-step sequence through sequence mining operations. By taking out only the critical parameters and relationships from the vast data set, the system maintains complete parameter identification while presenting results in a manageable and operationally easy-to-use format, resolving the contradiction between comprehensive analysis and operational ease.
Data Source
AI summary
Embodiments of the invention are directed to a computer-implemented method. A non-limiting example of the computer-implemented method includes accessing, using a processor system, a process-step sequence that includes a plurality process-steps and a plurality of queue-times. A process-step sequence mining operation is applied to the process-step sequence, wherein the process-step sequence mining operation is operable to make a prediction of an impact of a portion of the process-step sequence on a characteristic of a product generated by the process-step sequence.