Source Mask Optimization Pattern Selection via Diffraction Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional source mask optimization (SMO) techniques are computationally expensive and inefficient, particularly for complex designs, leading to challenges in guaranteeing optimal performance across all IC components, especially as feature sizes shrink below the classical resolution limit of optical projection systems.
Innovation Solution
A method and system for selecting a subset of critical patterns from a design layout by accessing diffraction order data, identifying representative peaks based on grouping criteria, and performing SMO on these selected patterns to optimize source and mask configurations, thereby reducing computational complexity and improving pattern coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional source mask optimization techniques are applied to all patterns in a design layout, then optimization performance is improved, but computational cost and runtime increase significantly
Solution Approach 1:
The patent segments the design layout into multiple regions of interest (ROIs) based on pattern complexity and criticality. Instead of optimizing all patterns uniformly, the system identifies and prioritizes critical patterns (such as sub-resolution patterns, patterns near edges, and high-frequency patterns) for dedicated optimization resources, while applying standard optimization to less critical regions. This segmentation enables focused computational effort where it matters most.
Solution Approach 2:
The patent applies different optimization strategies to different regions of the design layout based on their specific characteristics. Critical patterns receive enhanced optimization treatment including higher-resolution diffraction order calculations, while less critical patterns use standard optimization. This local differentiation maintains high optimization quality for critical features while reducing overall computational burden.
2Ease of manufacture
If source mask optimization is performed on simple repeating designs, then optimization is feasible, but it cannot guarantee optimal performance for complex designs
Solution Approach 1:
The patent performs preliminary analysis of the design layout to identify critical patterns and regions before applying optimization. This preliminary action includes analyzing pattern types, identifying sub-resolution features, and determining which patterns require special optimization attention. By preparing this information in advance, the system can efficiently apply appropriate optimization strategies to complex designs without trying to optimize everything uniformly.
Solution Approach 2:
Instead of applying a universal optimization approach and then filtering results, the patent inverts the approach by first identifying which patterns need optimization based on their criticality, then applying optimization only to those selected patterns. This inversion of the conventional approach enables the system to handle complex designs effectively by focusing computational resources on the most important patterns.
3Reliability
If all patterns in a design layout are included in SMO, then complete optimization coverage is achieved, but computational complexity increases
Solution Approach 1:
The patent applies partial optimization by focusing on a strategically selected subset of critical patterns rather than all patterns. This partial action approach targets the most important patterns (sub-resolution, edge-proximity, high-frequency) that have the greatest impact on manufacturing quality, achieving reliable optimization coverage for critical features while significantly reducing computational complexity compared to full optimization.
Data Source
AI summary
Apparatuses, systems, and methods for selecting a subset of critical patterns from a plurality of patterns of a design layout. In some embodiments, the method includes accessing diffraction order data based on the plurality of patterns that represent features to be formed on at least a portion of a substrate, the diffraction order data including a plurality of peaks corresponding to the plurality of patterns. The method also includes identifying a subset of representative peaks from the plurality of peaks according to one or more grouping criteria, the identifying including identifying a first representative peak that covers another peak colinear with the first representative peak, wherein the first representative peak is a discrete peak having a frequency that is an integer multiple of frequency of another discrete peak. The method further includes selecting the subset of critical patterns corresponding to the subset of representative peaks.