Semiconductor Recipe Error Inference for Faster Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of recipe creation and increased number of recipes in semiconductor measurement and inspection apparatuses lead to lower availability rates due to recipe errors, and the inefficiency of manual correction processes, especially in situations where skilled engineers are scarce.
Innovation Solution
A system that uses a machine learning-based approach to learn the correspondence between recipe parameters and errors, inferring the cause of errors and providing correction candidates by analyzing apparatus data, measurement recipes, and measurement results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual recipe correction by engineers is used, then recipe errors can be corrected, but the process is time-consuming and depends heavily on engineer skill
Solution Approach 1:
The patent replaces the manual mechanical process of recipe correction with an automated information processing system. The system automatically collects apparatus data, performs error analysis, and generates correction candidates, eliminating the need for manual engineer intervention in the core correction process while maintaining high accuracy through systematic analysis methods.
Solution Approach 2:
The system enables self-service by automatically performing recipe error detection and correction without requiring external engineer intervention. The automated analysis system continuously monitors apparatus data, identifies potential errors, and generates correction recommendations, allowing the system to correct its own operational parameters independently.
2Adaptability or versatility
If the number of recipes is increased to handle diverse semiconductor products, then product versatility is improved, but the complexity of recipe management increases
Solution Approach 1:
The patent creates a universal recipe management system that can handle multiple semiconductor product types and recipe variations through a single unified platform. The system uses standardized data collection and analysis methods that work across different product types, reducing the need for separate management processes for each recipe while maintaining the ability to handle diverse product requirements.
Solution Approach 2:
The system manages recipe complexity by dynamically adjusting analysis parameters and correction thresholds based on the specific apparatus data and error patterns observed. This allows the system to adapt its analysis depth and correction strategies to match the complexity of each recipe, preventing uniform high-level analysis from becoming a bottleneck while maintaining thoroughness where needed.
3Ease of repair
If manual recipe correction processes are used, then correction can be performed, but the availability rate of the apparatus decreases
Solution Approach 1:
The system performs preliminary error detection and analysis continuously in the background, identifying potential recipe errors before they cause apparatus failures or production issues. By proactively detecting and correcting errors ahead of time, the system prevents downtime and maintains high apparatus availability while still providing comprehensive correction capabilities.
Solution Approach 2:
The system implements continuous feedback loops where apparatus operational data is constantly monitored, analyzed for errors, and used to generate real-time correction recommendations. This closed-loop feedback mechanism enables rapid detection and correction of recipe errors, minimizing their impact on apparatus availability and ensuring quick restoration of optimal performance.
Data Source
AI summary
An objective of the present invention is to provide a system which can infer the cause of a recipe error and present a correction candidate for the recipe error. A recipe information presentation system or recipe error inference system according to the present invention: causes a learner to learn a correspondence between a recipe and an error originating from the recipe; and acquires from the learner an inference result as to whether the error occurs when a new recipe is used (refer to FIG. 1).