3D Mask Corner Rounding Simulation Using Feature Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lithography simulations struggle with accurately modeling mask corner rounding effects due to limitations in mask making processes, leading to inefficient and computationally intensive simulations, especially for complex shapes with arbitrary edge angles and small jogs, which increase runtime and reduce accuracy.
Innovation Solution
The approach involves partitioning the layout geometry into feature images, applying mask corner rounding (MCR) corrections to these images, and using precomputed M3D filters to calculate the mask function, reducing the need for additional layout vertices and improving simulation efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional three-dimensional mask models are used to accurately model mask corner rounding effects, then manufacturing precision is improved, but computational time increases significantly
Solution Approach 1:
The patent segments the mask layout into discrete feature images representing different corner rounding scenarios. Each feature image is pre-processed and stored in a library, allowing the simulation to quickly retrieve and combine relevant features rather than computing three-dimensional models from scratch for each corner rounding case. This segmentation enables accurate modeling of mask corner rounding effects while significantly reducing computational runtime.
2Manufacturing precision
If complex shapes with arbitrary edge angles and small jogs are simulated using traditional methods, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The patent decomposes complex shapes with arbitrary edge angles and small jogs into simpler feature images that can be retrieved from a pre-computed library. Instead of using a single complex three-dimensional model for the entire shape, the method segments the geometry into manageable feature components (such as corner features, edge features, and junction features) that are processed independently and then combined. This reduces simulation model complexity while maintaining manufacturing precision.
Solution Approach 2:
The patent applies different levels of modeling detail to different parts of the layout geometry based on their specific characteristics. Simple regions use basic feature images, while complex regions with arbitrary edge angles or small jogs use more detailed feature combinations. This local quality approach allows the simulation to maintain high manufacturing precision where needed without unnecessarily increasing device complexity across the entire layout.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enhances computational efficiency and accuracy in lithography simulations by reducing runtime and maintaining precision, making it suitable for both machine learning frameworks and graphics processing units, while providing accurate results for Hopkins and Abbe imaging models.
Implementation Method 1
The feature images have corresponding mask 3D (M3D) filters, which represent the electromagnetic scattering effect of that feature image for a given source illumination
Data Source
AI summary
A layout geometry of a lithographic mask is received. The layout geometry includes at least one shape having one or more rounded corners. The layout geometry is partitioned into a plurality of feature images, for example as selected from a library. The feature images include at least one mask corner rounding (MCR)-corrected feature image that accounts for the rounded corners of the shape. The feature images have corresponding mask 3D (M3D) filters, which represent the electromagnetic scattering effect of that feature image for a given source illumination. The mask function contribution from each of the feature images is calculated by convolving the feature image with its corresponding M3D filter. The mask function contributions are combined to determine a mask function for the mask illuminated by the source illumination.