SRAF Optimization via MRC-Integrated ILT Mask Shape Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mask manufacturing techniques for integrated circuits lack integration of mask-manufacturing rule checking (MRC) rules into inverse lithography techniques (ILT), leading to sub-optimal mask image optimization and increased Edge Placement Errors (EPEs.
Innovation Solution
The integration of MRC rules into the ILT mask optimization process using a fast marching method (FMM) for determining sub-resolution assist features (SRAFs), which involves receiving a pixelated SRAF bitmap image, calculating arrival times based on pixel brightness values, and resolving MRC rule violations to optimize mask shapes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If MRC rules are not integrated into the ILT mask optimization process, then the mask optimization can be performed more simply and quickly, but the mask quality deteriorates with increased Edge Placement Errors
Solution Approach 1:
The patent merges the MRC rule checking process with the ILT mask optimization process into a unified computational framework. The speed function in the fast marching method incorporates MRC rule constraints directly, allowing simultaneous optimization of mask patterns while ensuring compliance with manufacturing rules, thereby improving mask quality without proportionally increasing process complexity
Solution Approach 2:
The patent modifies the speed function parameters in the fast marching method to reflect MRC rule constraints. By changing the parameter definition from a simple distance-based speed to a compliance-aware speed function that considers MRC rules, the system achieves better mask quality while maintaining computational efficiency through parameter optimization rather than structural complexity increases
2Manufacturing precision
If traditional shape extraction methods are used without MRC integration, then the processing speed is faster, but the manufacturing precision decreases due to MRC rule violations
Solution Approach 1:
The patent performs preliminary action by integrating MRC rule checking into the shape extraction process itself rather than as a subsequent separate step. The fast marching method with compliance-aware speed function proactively ensures MRC rule adherence during SRAF generation, preventing rule violations before they occur and eliminating the need for time-consuming post-processing correction steps
Solution Approach 2:
The patent replaces traditional multi-step mechanical processing workflows (separate shape extraction followed by separate MRC checking and correction) with a unified computational field-based approach. The fast marching method uses a compliance-aware speed function to directly generate MRC-compliant shapes in a single computational pass, substituting iterative mechanical correction with a streamlined field-based optimization
Data Source
AI summary
Extracting shapes from a pixelated SRAF bitmap image of pixels for mask making is disclosed. A method includes receiving the pixelated SRAF bitmap image of pixels, each pixel having a respective brightness value; selecting a ridge point in the pixelated SRAF bitmap image; for each pixel of at least some of the pixels, determining a respective arrival time at the pixel; and determining a mask shape using the arrival times of the at least some of the pixels. The ridge point is one of the pixels and is selected based on the respective brightness value of the one of the pixels. An arrival time is based on a respective brightness value of the pixel and a Mask Rule Check (MRC) rule.