Registration Error Determination for Semiconductor Lithography Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining registration error in semiconductor lithography masks are prone to inaccuracies due to unknown rotations and irrelevant image intensity errors, leading to falsified results.
Innovation Solution
A method that generates an image of the mask, determines measuring contours, and matches the forms and registration of design and measuring contours to minimize mean lateral distances, thereby accounting for rotations and optimizing structure alignment without relying on intensity values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the threshold method or correlation method is used to determine structure position, then the measurement process is simplified, but unknown rotation of structures leads to falsified results
Solution Approach 1:
The patent extracts and isolates the rotation parameter as a separate variable to be determined. By specifically extracting rotational information from the aerial image and treating it independently, the method eliminates the confounding effect of rotation on position measurement, thereby resolving the contradiction between measurement simplicity and accuracy.
Solution Approach 2:
The patent transitions from two-dimensional position measurement to three-dimensional parameter space by adding rotation as an additional measurement dimension. This allows the system to account for rotational deviations separately, preventing them from corrupting the position measurement results while maintaining operational simplicity.
2Measurement precision
If the correlation method or symmetry-correlation method is used based on intensity values, then structure characterization is achieved, but errors in image intensity away from structures lead to falsified measurement results
Solution Approach 1:
The patent extracts only the essential contour information from the aerial image, discarding the problematic intensity values. By focusing exclusively on edge detection and contour extraction rather than using full intensity data, the method eliminates the harmful effect of intensity errors while maintaining the ability to characterize structure positions accurately.
Solution Approach 2:
The patent introduces contour extraction as an intermediary step between image acquisition and position measurement. This intermediary process filters out the harmful intensity variations by converting the image data into binary contour representations, thereby protecting the final measurement from intensity-related errors.
3Manufacturing precision
If mask production processes are made highly complex to achieve subnanometer registration accuracy, then registration error is reduced, but manufacturing cost and complexity increase significantly
Solution Approach 1:
The patent replaces complex mechanical mask production and alignment systems with a computational approach. By using image processing algorithms to measure and compensate for registration errors after mask fabrication, the method achieves high precision without requiring equally complex manufacturing processes, thereby reducing overall system complexity while maintaining subnanometer accuracy.
Data Source
AI summary
The invention relates to a method for determining a registration error of a structure on a mask for semiconductor lithography, comprising the following method steps: generating an image of at least one region of the mask, determining at least one measuring contour in the image, and matching the forms of a design contour and a measuring contour to one another while at the same time matching the registration of the two contours.