Machine-Learned Source Shapes for Faster Mask Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing source mask optimization processes in semiconductor manufacturing rely on arbitrary initial source shapes, leading to longer optimization runtimes and non-optimal solutions due to a lack of consideration for the lithographic mask layout.
Innovation Solution
A machine learning model is used to determine initial source shapes based on the actual mask layout, selecting clips that are challenging to manufacture or sensitive to process variations, and inferring corresponding source shapes for the source mask optimization process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If arbitrary initial source shapes are used in source mask optimization, then the process is simple to implement, but the optimization runtime is longer and the solutions are non-optimal
Solution Approach 1:
The machine learning model performs preliminary action by predicting optimized initial source shapes before the source mask optimization process begins. The model is trained on historical optimization data and uses the predicted initial source shapes as starting points for the optimization algorithm, which accelerates convergence and reduces runtime while maintaining simplicity of implementation.
2Ease of manufacture
If arbitrary initial source shapes are used in source mask optimization, then the implementation remains simple, but the quality of the final design is reduced
Solution Approach 1:
The machine learning model performs preliminary action by predicting optimized initial source shapes before the source mask optimization process begins. The model is trained on historical optimization data and uses the predicted initial source shapes as starting points for the optimization algorithm, which accelerates convergence and reduces runtime while maintaining simplicity of implementation.
3Loss of time
If mask layout-based initial source shapes are determined using machine learning, then the optimization runtime is reduced and convergence is faster, but the system complexity increases
Solution Approach 1:
The machine learning model acts as an intermediary between the mask layout and the source mask optimization process. It takes the mask layout as input and generates predicted initial source shapes that are optimized for the specific layout, serving as a bridge that connects layout features to optimization starting points without requiring direct complex interactions between the layout and optimization algorithm.
4Reliability
If machine learning model is used to infer source shapes from mask clips, then the lithographic process window is enlarged and manufacturing outcomes are improved, but the computational requirements increase
Solution Approach 1:
The machine learning model performs preliminary action by predicting optimized initial source shapes before the source mask optimization process begins. The model is trained on historical optimization data and uses the predicted initial source shapes as starting points for the optimization algorithm, which accelerates convergence and reduces runtime while maintaining simplicity of implementation.
Data Source
AI summary
Initial source shapes for source mask optimization are determined based on a layout of the lithographic mask. In one approach, a layout of a lithographic mask is received. Different sections of the lithographic mask, referred to as clips, are selected. These clips are applied to a machine learning model which infers source shapes from the clips. The inferred source shapes are used as the initial source shapes for source mask optimization.