Lithography Process Window Detection With Measurement Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing edge detection methods in noisy SEM images, such as those used in scanning electron microscopes, suffer from biased roughness measurements due to measurement noise, which overestimates the true roughness and is dependent on metrology tool settings, making it difficult to achieve accurate and precise unbiased roughness estimates of pattern features.
Innovation Solution
A physics-based inverse linescan model is employed to separate noise from actual roughness in SEM images, allowing for unbiased and precise measurement of roughness parameters without relying on image filtering that alters the feature roughness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional edge detection methods are used on noisy SEM images, then the detection process is simple, but the roughness measurement is biased and overestimates the true roughness
Solution Approach 1:
The patent segments the roughness measurement process into two distinct components: signal (true roughness) and noise (measurement uncertainty). By separating these components through probabilistic modeling, the method achieves unbiased roughness estimates while accounting for noise effects, resolving the contradiction between measurement accuracy and method simplicity.
Solution Approach 2:
The patent introduces measurement uncertainty as an intermediary parameter that mediates between the raw noisy measurements and the final roughness assessment. This intermediary allows the system to quantify and correct for noise effects, improving measurement accuracy without requiring complex filtering procedures that would alter the actual feature roughness.
2Measurement precision
If image filtering is applied to reduce noise, then noise is reduced, but the feature roughness is altered and measurements become biased
Solution Approach 1:
The patent extracts and quantifies the noise component separately from the signal through probabilistic modeling. Instead of removing noise through filtering, the method extracts the measurement uncertainty and uses it to correct the roughness measurement, thereby reducing noise effects while preserving the integrity of the actual feature roughness.
Solution Approach 2:
The patent converts the harmful effect of measurement noise into a beneficial by quantifying it as measurement uncertainty. This uncertainty information is then used to adjust and correct the roughness measurement, transforming the previously harmful noise into a useful parameter that improves measurement accuracy without altering the feature itself.
3Measurement precision
If measurement uncertainty is accounted for, then unbiased roughness estimates are achieved, but the analysis complexity increases
Solution Approach 1:
The patent changes the parameter representation from deterministic roughness values to probabilistic distributions that incorporate measurement uncertainty. By modeling roughness as a probability distribution rather than a single value, the method achieves unbiased estimates while managing complexity through standardized statistical frameworks.
Solution Approach 2:
The patent implements feedback by using the calculated measurement uncertainty to adjust and correct the roughness measurement. The uncertainty information feeds back into the measurement process, allowing iterative refinement of the roughness estimate until convergence is achieved, thereby resolving the contradiction between accuracy and complexity.
Data Source
Figure 1A~1B
Figure 2~3
Figure 4~5
AI summary
Methods, systems, and computer-readable mediums for configuring a lithography tool to manufacture a semiconductor device. The method includes selecting a first variable, selecting a second variable, selecting at least one response variable that is a function of the first variable and second variable, determining a measurement uncertainty for each response variable, determining, based on a measurement of the response variable, and the measurement uncertainty for the response variable, a plurality of probabilities representing a plurality of indications of whether a plurality of points associated with a lithography process meet a specification requirement for each response variable, wherein the plurality of probabilities represent a process window, and configuring, based on the process window, a lithography tool to manufacture a semiconductor device.