Subresolution Assist Feature Placement Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithographic technologies face challenges in reproducing patterns with dimensions smaller than the classical resolution limit, necessitating sophisticated methods like optical proximity correction and customized illumination, which are inefficient in achieving precise feature placement and electrical functionality.
Innovation Solution
A machine learning model is trained to generate subresolution assist features by analyzing reference images, determining the presence and optimal placement of assist features using training data and a scoring function, to enhance pattern fidelity on substrates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If optical proximity correction and customized illumination are used to reproduce patterns smaller than the classical resolution limit, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The machine learning model is trained in advance on a comprehensive dataset containing design layouts, simulated aerial images, and corresponding SRAF placement patterns. This preliminary training phase enables the model to learn optimal SRAF placement strategies for various pattern configurations, allowing rapid prediction during actual manufacturing without requiring complex real-time calculations or multiple iterative correction steps
Solution Approach 2:
The patent replaces traditional optical proximity correction methods (which involve complex iterative calculations and multiple processing steps) with a machine learning-based prediction system. The trained model directly predicts optimal SRAF placement patterns from design layouts and aerial images, substituting the mechanical/optical correction process with an intelligent prediction system that achieves similar or superior precision with reduced complexity
2Manufacturing precision
If sophisticated methods like optical proximity correction are used to achieve precise feature placement, then manufacturing precision is improved, but productivity decreases
Solution Approach 1:
The patent replaces traditional optical proximity correction methods (which involve complex iterative calculations and multiple processing steps) with a machine learning-based prediction system. The trained model directly predicts optimal SRAF placement patterns from design layouts and aerial images, substituting the mechanical/optical correction process with an intelligent prediction system that achieves similar or superior precision with reduced complexity
Solution Approach 2:
The machine learning model is trained in advance on a comprehensive dataset containing design layouts, simulated aerial images, and corresponding SRAF placement patterns. This preliminary training phase enables the model to learn optimal SRAF placement strategies for various pattern configurations, allowing rapid prediction during actual manufacturing without requiring complex real-time calculations or multiple iterative correction steps
Data Source
AI summary
A method for training a machine learning model to generate a characteristic pattern, the method includes obtaining training data associated with a reference feature in a reference image. The training data includes (i) location data of each portion of the reference feature, and (ii) a presence value indicating whether the portion of the reference feature is located within a reference assist feature generated for the reference feature. The method includes training the machine learning model to predict a presence value based on the actual presence value in the training data. The predicted presence value indicates whether a portion of a feature (e.g., a skeleton point on a skeleton of a contour of the feature) is to be covered by an assist feature. The training is performed based on the training data such that a metric between a predicted presence value and the presence value is minimized.