Raster Scan OPC Using Pixel Grayscale Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lithography systems face defects such as rounded corners and line end shortening due to boundary effects, which are compensated for by manipulating pattern data files with additional geometries, leading to increased data file size and processing time.
Innovation Solution
The method involves generating sub-pixel data to adjust grayscale values of corner pixels and neighboring pixels post-rasterization, allowing for corner correction without expanding the data file size, enabling flexible pixel configurations and distributed processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If geometry-based optical proximity correction is used to compensate for rounded corners by adding serifs to the pattern, then corner accuracy is improved, but data file size increases significantly
Solution Approach 1:
The patent transitions from geometry-based correction (modifying pattern shapes in 2D space) to pixel-value-based correction (modifying grayscale values in the rasterized image domain). This dimensional shift allows corner correction to be applied after rasterization by adjusting pixel intensity values rather than adding geometric features, thereby avoiding data file size expansion while achieving the same optical proximity correction effect
Solution Approach 2:
Instead of modifying the original vector pattern data with additional geometric elements (serifs), the patent creates a corrected version of the rasterized image by copying and adjusting pixel values. The correction is applied to the pixel map representation rather than the source geometry, allowing the original data file to remain unchanged while producing a corrected output image
2Manufacturing precision
If geometry-based optical proximity correction is used to add serifs to compensate for rounded corners, then corner sharpness is improved, but processing time increases
Solution Approach 1:
The patent performs corner correction as a post-rasterization operation using pre-computed correction kernels. By having the correction algorithm ready and the rasterization already complete, the system can apply corrections efficiently without the time-consuming process of geometric manipulation and re-rasterization that would be required by geometry-based methods
Solution Approach 2:
The patent replaces the mechanical/geometric approach of adding serifs (which requires complex geometric operations and re-rasterization) with a digital signal processing approach using convolution with correction kernels. This substitution of mechanical geometric manipulation with mathematical image processing operations significantly reduces processing time while achieving the same optical proximity correction
3Manufacturing precision
If additional geometries are added to the data file to compensate for rounded corners, then pattern fidelity is improved, but data transfer time increases
Solution Approach 1:
The patent creates a corrected copy of the rasterized image rather than modifying the original vector data file. The correction is applied to the pixel map representation, allowing the original compact vector data to remain unchanged for transfer, while the corrected image is generated locally at the lithography system
Solution Approach 2:
The patent shifts the correction operation from the vector geometry domain to the raster image domain. By applying corrections to the already-rasterized pixel map rather than modifying vector geometries, the system avoids increasing the size of the transferred data file while still achieving improved pattern fidelity through pixel value adjustments
Data Source
AI summary
Methods and apparatus for correcting defects, such as rounded corners and line end shortening, in patterns formed via lithography are provided. Such defects are compensated for “post-rasterization” by manipulating the grayscale values of pixel maps.