Lithography Patterning Model Training for Accurate Printed Pattern Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithographic technologies face challenges in accurately reproducing patterns on substrates with dimensions smaller than the classical resolution limit, particularly due to the complexity of low-k1 lithography, where sophisticated fine-tuning steps are required to overcome difficulties in achieving desired electrical functionality and performance.
Innovation Solution
A patterning process model is trained using a deep learning convolutional neural network framework, combining a first model based on physical terms and a machine learning model to predict patterns on a substrate, iteratively adjusting parameters to minimize differences between measured and predicted patterns, and employing this model for optical proximity corrections and etch bias determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If low-k1 lithography is used to print features smaller than the classical resolution limit, then manufacturing precision is improved, but device complexity increases due to sophisticated fine-tuning steps
Solution Approach 1:
The patterning process is segmented into multiple controllable parameters including illumination conditions, projection optics settings, and resist processing parameters. Each parameter can be independently optimized to achieve desired pattern dimensions without requiring complex fine-tuning of the entire system.
Solution Approach 2:
The patent employs parameter changes in the lithographic projection apparatus, such as adjusting numerical aperture, illumination wavelength, and focus settings, to optimize pattern formation. These parameter adjustments enable precise control over printed feature dimensions while maintaining process simplicity.
2Measurement precision
If traditional patterning process models are used, then model simplicity is maintained, but measurement precision deteriorates due to inability to accurately predict printed patterns
Solution Approach 1:
The patterning process model incorporates feedback mechanisms that compare predicted patterns with actual measured patterns from substrates. The model iteratively adjusts its parameters based on this feedback to improve prediction accuracy, enabling precise measurement and optimization of lithographic processes.
Solution Approach 2:
The model performs preliminary predictions of pattern formation before actual lithography execution. By simulating and predicting pattern outcomes in advance, the system can identify and correct potential issues before they manifest in physical substrates, improving measurement precision without requiring complex post-processing analysis.
3Manufacturing precision
If sophisticated fine-tuning steps are applied to optimize patterning, then manufacturing precision is improved, but productivity decreases due to additional process steps
Solution Approach 1:
Optimization of patterning parameters is performed in advance through computational modeling and simulation. By determining optimal settings before actual substrate processing, the system eliminates the need for time-consuming fine-tuning steps during production, thereby maintaining high precision while preserving productivity.
Solution Approach 2:
The patterning process model automatically determines optimal process parameters and predicts pattern outcomes without requiring manual intervention or iterative fine-tuning. This self-service capability enables the system to achieve high manufacturing precision while maintaining rapid substrate processing speeds.
Data Source
AI summary
A method for training a patterning process model, the patterning process model configured to predict a pattern that will be formed by a patterning process. The method involves obtaining an image data associated with a desired pattern, a measured pattern of the substrate, a first model including a first set of parameters, and a machine learning model including a second set of parameters; and iteratively determining values of the first set of parameters and the second set of parameters to train the patterning process model. An iteration involves executing, using the image data, the first model and the machine learning model to cooperatively predict a printed pattern of the substrate; and modifying the values of the first set of parameters and the second set of parameters such that a difference between the measured pattern and the predicted pattern is reduced.