SEM Roughness Metrology Using Variable-Frame Asymptotic Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current metrology protocols for microfabrication patterns, especially those using Extreme Ultraviolet Lithography (EUVL) with high numerical aperture (NA), face challenges in accurately estimating roughness parameters due to low signal-to-noise ratio (SNR), leading to underestimation of line-edge roughness, linewidth roughness, and pattern placement roughness.
Innovation Solution
A method involving a scanning electron microscope (SEM) generates multiple sets of frames with varying numbers of frames, and a computing device estimates feature data, computes preliminary roughness parameters, fits a model equation to these estimates, and extrapolates to an asymptotic value for improved accuracy, independent of SNR.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of frames is increased to improve SNR, then measurement precision improves, but productivity decreases
Solution Approach 1:
The method performs preliminary measurements with varying frame numbers to establish a relationship model before final measurement. This preliminary action creates a calibration curve that maps frame number to SNR correction factor, enabling accurate roughness estimation without requiring excessive frames during actual production metrology.
Solution Approach 2:
The invention changes the parameter of frame number systematically to observe its effect on SNR and roughness measurement. By varying frame number and establishing a correction model, the system can compensate for low SNR conditions without actually increasing frame count during production, thus maintaining productivity while improving measurement accuracy.
2Manufacturing precision
If the resist film thickness is reduced to avoid pattern collapse, then manufacturing precision improves, but measurement precision deteriorates due to lower SNR
Solution Approach 1:
The invention introduces an intermediary correction model that mediates between the physical constraint (thin resist causing low SNR) and the measurement requirement (accurate roughness estimation). The model acts as a mathematical intermediary that translates low-SNR measurements into accurate roughness values without requiring physical changes to the resist film.
Solution Approach 2:
The invention replaces the mechanical approach of increasing frames (which would improve SNR but reduce productivity) with a computational approach using modeling and correction. This substitution uses software-based signal enhancement instead of hardware-based signal accumulation, maintaining both measurement accuracy and productivity.
3Measurement precision
If the number of frames is increased to improve SNR, then measurement precision improves, but use of energy increases
Solution Approach 1:
The method performs preliminary measurements with varying frame numbers to establish an energy-efficient correction model. This preliminary action identifies the optimal frame number that achieves sufficient measurement accuracy with minimal energy consumption, preventing wasteful energy use during production metrology.
Solution Approach 2:
The invention systematically varies the frame number parameter to establish its relationship with both SNR and energy consumption. The resulting correction model enables accurate roughness measurement at optimized frame numbers, reducing energy consumption while maintaining measurement precision through mathematical compensation rather than brute-force signal accumulation.
Data Source
Figure 1~3
Figure 2
Figure 4
AI summary
There is provided a method of performing metrology on a microfabrication pattern, the method comprising: generating, by a SEM, a first set of frames of a first region of the pattern, a second set of frames of a second region of the pattern, and a third set of frames of a third region of the pattern, wherein a number of frames of the first, second and third sets of frames are different; for each of the first, second and third set of frames, by a computing device: estimating feature data representing edge positions, linewidths or centerline positions of one or more features of the region of the pattern of the set of frames, and computing a preliminary estimate of a roughness parameter from the feature data, wherein the roughness parameter is indicative of a line edge roughness, a linewidth roughness, or a pattern placement roughness of the one or more features; wherein the method further comprises, by the computing device: fitting a model equation to the preliminary estimates of the roughness parameter and a model parameter dependent on the number of frames of the set of frames, the model equation relating the model parameter to the roughness parameter; and computing a final estimate of the roughness parameter as an asymptotic value of the fitted model equation.