Sequential Sub-Model Defect Prediction for Semiconductor Production
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor manufacturing processes face challenges in efficiently managing and predicting defects due to the complexity of multivariate processes and the need for precise inspections, which can decrease productivity and reliability.
Innovation Solution
A defect prediction method and device that model the production process into multiple sub-models, map production data into each sub-model, and predict defect information by determining output data for each sub-model based on previous output data and production process data, while considering the sequence between sub-models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If precision inspection is performed on all wafers to detect defects, then measurement precision and reliability are improved, but productivity decreases due to the great amount of time required
Solution Approach 1:
The patent segments the defect detection task by dividing wafers into different inspection categories. Instead of inspecting all wafers with the same precision level, the system performs precision inspection only on selected wafers (e.g., first wafers in a sequence or wafers with higher risk profiles) while using less intensive methods for others. This segmentation allows maintaining high defect detection precision for critical cases while preserving overall productivity by reducing the number of full precision inspections required.
2Manufacturing precision
If multiple process variables are tracked through image form results to monitor production quality, then manufacturing precision is improved, but device complexity increases due to the need to process and analyze numerous variables
Solution Approach 1:
The patent extracts and focuses on specific critical process variables from the complete set of available variables. Rather than processing and analyzing all dozens of process variables through image form results, the system identifies and extracts the most influential variables (such as resolution, depth of focus, and overlay in photolithography) for targeted monitoring. This extraction approach maintains manufacturing precision by concentrating on key quality drivers while reducing device complexity by eliminating the need to process the entire variable set.
3Ease of operation
If univariate management method is used to manage defects in each unit process, then ease of operation is improved, but reliability decreases because only unit processes are managed without considering correlations between sequential processes
Solution Approach 1:
The patent merges univariate management of individual unit processes with multivariate analysis of sequential processes. The system maintains the operational simplicity of univariate management for each individual process step while simultaneously integrating correlations between sequential processes through multivariate statistical analysis. This merging approach preserves ease of operation at the unit level while enhancing reliability by considering inter-process relationships, allowing the system to capture defect patterns that span multiple sequential processes.
Data Source
AI summary
A method and device for predicting a defect. The method includes determining a sequence between a plurality of sub-models by modeling a production process into the plurality of sub-models, mapping production process data into each of the plurality of sub-models, determining, by a corresponding sub-model, output data comprising defect information on a potential defect occurring in a corresponding step, for each of the plurality of sub-models, predicting information associated with a defect in the production process based on the output data corresponding to each of the plurality of sub-models, and inputting the output data of each of the sub-models to a subsequent sub-model of the corresponding sub-model, based on the sequence.