Feature-Based Cell Extraction for Oblique Pattern Regions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computational lithography techniques face challenges in accurately identifying and processing pattern features with oblique angles or no vertices, which are common in IC designs, leading to increased computational complexity and reduced efficiency in correcting IC layouts.
Innovation Solution
A method for feature-based cell extraction that identifies unit cells with oblique angle features, determines their slope and pitch, and constructs a hierarchy for pattern regions, optimizing unit cell distribution using linear optimization to maintain block level symmetry and reduce duplicate cells, thereby improving processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If current computational lithography techniques are used to identify and process pattern features with oblique angles or no vertices, then the IC layout correction can be performed, but the computational complexity increases and processing efficiency decreases
Solution Approach 1:
The layout is divided into multiple pattern regions, each characterized by specific features (oblique angles, no vertices, line-space features). Unit cells are extracted from each region independently, allowing the complex layout processing to be segmented into manageable parts that can be handled more efficiently
Solution Approach 2:
The patent transforms the representation of pattern features by extracting unit cells with specific parameters (slope, pitch, line-space features) instead of processing raw geometric data. This parameter transformation simplifies the computational representation and enables more efficient processing while maintaining accuracy
2Manufacturing precision
If current computational lithography techniques are used to identify and process pattern features with oblique angles or no vertices, then the IC layout correction can be performed, but the processing efficiency is reduced
Solution Approach 1:
Unit cells are extracted as reusable templates that can be copied and applied across multiple pattern regions. Once a unit cell is extracted from one region, it can be matched and applied to similar regions, avoiding redundant processing and significantly improving efficiency
Solution Approach 2:
Unit cells are extracted and characterized in advance before the main layout correction process. This preliminary extraction and characterization of pattern features allows the subsequent matching and correction stages to proceed more quickly by working with pre-processed unit cell data rather than raw layout data
3Area of stationary object
If unit cells are extracted and distributed across pattern regions without optimization, then coverage is achieved, but duplicate unit cells increase computational overhead
Solution Approach 1:
Multiple extracted unit cells that represent the same pattern feature are merged into a single canonical unit cell. The patent identifies duplicate unit cells through matching operations and consolidates them, reducing the total number of unit cells that need to be processed and stored while maintaining complete pattern region coverage
Data Source
AI summary
Systems and methods of feature-based cell extraction. The methods include obtaining data representative of a layout, wherein the layout includes a pattern region having no vertices, extracting unit cells from the pattern region having no vertices, identifying, using the unit cells, a set of regions of the layout matching the unit cells, and generating, using the unit cells, a hierarchy for the set of regions. In some embodiments the pattern regions have oblique angle features or have no vertices of features. The pattern regions can have a feature including a feature slope, a horizontal or a vertical pitch, or a line-space feature. In some embodiments the hierarchy is optimized using a linear optimization and can be provided for use in modeling, OPC, defect inspection, defect prediction, or SMO.