Substrate Processing Recipe Prediction for Eco-Efficient Resource Use
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The semiconductor industry faces challenges in reducing its environmental impact due to increasing resource consumption and complexity in substrate processing, making it difficult to determine eco-efficient processing recipes without physical testing or empirical results.
Innovation Solution
A system and method that utilize machine learning models to predict environmental resource usage by inputting process recipes into trained models, providing recommendations for optimizing eco-efficiency in substrate processing, allowing for the selection of the most eco-efficient recipes before implementation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If physical testing and empirical results are used to determine eco-efficient processing recipes, then manufacturing precision and reliability are improved, but loss of time and productivity deteriorate due to extensive testing requirements
Solution Approach 1:
The patent applies preliminary action by training machine learning models in advance using historical process data and empirical results. Once trained, these models can predict eco-efficiency outcomes for new process recipes without requiring physical testing, thus resolving the contradiction between accurate eco-efficiency determination and time consumption.
Solution Approach 2:
The patent uses machine learning models as virtual copies that simulate the complex relationship between process parameters and environmental resource usage. These digital twins replicate the behavior of physical systems, allowing eco-efficiency prediction without actual substrate processing tests, thereby reducing time loss while maintaining prediction accuracy.
2Ease of manufacture
If traditional substrate processing methods are used, then manufacturing simplicity is maintained, but object-generated harmful factors increase due to high energy consumption and waste generation
Solution Approach 1:
The system performs preliminary prediction of environmental impact using trained machine learning models before actual substrate processing. This allows identification and selection of low-impact process recipes in advance, enabling eco-efficient manufacturing without complicating the actual processing workflow.
Solution Approach 2:
The patent implements feedback mechanisms where predicted environmental resource usage data is fed back into the process recipe selection system. This continuous feedback loop enables dynamic optimization of manufacturing processes to minimize harmful factors while maintaining operational simplicity.
3Productivity
If chip complexity is increased to meet growing demand, then productivity is improved, but object-generated harmful factors worsen due to increased resource consumption
Solution Approach 1:
The patent applies parameter changes by using machine learning models to identify and select process recipes with optimized parameter combinations that achieve high productivity while minimizing environmental impact. The models analyze multiple process parameters simultaneously to find optimal settings that decouple productivity gains from increased resource consumption.
Data Source
AI summary
In some embodiments, a method includes receiving a process recipe including process recipe setpoint data. The method further includes inputting the process recipe into one or more trained machine learning models that output predicted environmental resource usage data indicative of an environmental resource consumption associated with processing a substrate in a process chamber according to the process recipe. The method further includes outputting a recommendation associated with the process recipe based at least in part on the predicted environmental resource usage data.