Deformation Pattern Recognition for Lithographic Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional lithographic apparatuses face challenges in accurately transferring subsequent patterns relative to previously transferred patterns on substrates due to the limitations of alignment marks, which can lead to reduced throughput and sensitivity to outliers, especially when dealing with substrate deformations caused by processing devices.
Innovation Solution
A method involving deformation pattern recognition, where alignment mark displacements are measured, deformation patterns are fitted to these displacements, and accuracy values are determined to assess the presence of specific deformation patterns, allowing for correction recipes to be applied to align subsequent patterns accurately, thereby monitoring processing device performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a large number of alignment marks are used to improve the accuracy of pattern transfer, then the manufacturing precision is improved, but the throughput of the lithographic apparatus deteriorates
Solution Approach 1:
The patent segments the alignment mark set into two distinct types: a first alignment mark positioned in a corner region and a second alignment mark positioned in a non-corner region. This segmentation allows the system to use a minimal set of alignment marks (reducing throughput impact) while still capturing sufficient deformation information through the specific spatial distribution of these segmented marks, thereby maintaining manufacturing precision without sacrificing productivity.
2Manufacturing precision
If advanced alignment models like higher-order alignment models are used to improve the accuracy of determining pattern position, then the manufacturing precision is improved, but the sensitivity to outliers increases
Solution Approach 1:
The patent applies local quality by assigning different functional roles to alignment marks based on their spatial locations. Corner region alignment marks and non-corner region alignment marks are treated differently in the deformation pattern recognition process. This local differentiation allows the system to robustly determine pattern positions using simplified models that are less sensitive to outliers, while still achieving high manufacturing precision through the strategic spatial distribution of marks with different qualities.
Data Source
AI summary
A deformation pattern recognition method including providing one or more deformation patterns, each deformation pattern being associated with a deformation of a substrate that may be caused by a processing device; transferring a first pattern to a substrate, the first pattern including at least N alignment marks, wherein each alignment mark is positioned at a respective predefined nominal position; processing the substrate; measuring a position of N alignment marks and determining an alignment mark displacement for the N alignment marks by comparing the respective nominal position with the respective measured position; fitting at least one deformation pattern to the measured alignment mark displacements; determining an accuracy value for each fitted deformation pattern, the accuracy value being representative of the accuracy of the corresponding fit; using the determined accuracy value, determining whether an associated deformation pattern is present.