Semiconductor Pattern Shape Measurement Using Directional Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of pattern shape measurement in semiconductor devices is compromised due to variations in captured image contour information, leading to decreased measurement precision.
Innovation Solution
A method involving the generation of first and second contour point group data, where location information of contour points is associated with orientation angles, and weights are determined based on standard deviation, allowing for high-accuracy shape measurement through a weighting table and random extraction process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contour information is extracted from captured images to measure pattern shape, then measurement can be performed, but measurement accuracy decreases due to image variations
Solution Approach 1:
The patent applies preliminary action by pre-calculating standard deviations of contour points in different directions and creating a weighting table before actual measurement. This allows the system to predict which directions have higher variability and pre-determine the weights for random extraction, rather than reacting to variations after they occur in the measurement process
Solution Approach 2:
The patent changes the parameter of contour point selection by introducing random extraction with direction-based weighting. Instead of uniformly selecting contour points or using fixed extraction rules, the system dynamically adjusts selection probability based on the standard deviation calculated for each direction, thereby adapting to image variations
2Quantity of substance
If all contour points are used for measurement, then more data is available, but measurement accuracy decreases due to inclusion of high-variation points
Solution Approach 1:
The patent applies local quality by treating different directions of contour points differently based on their variability characteristics. Directions with high standard deviation (high variability) are assigned lower weights and thus lower probability of selection, while directions with low standard deviation are assigned higher weights. This creates a non-uniform quality distribution in the selected contour points that optimizes measurement accuracy
3Adaptability or versatility
If contour points with high standard deviation are included, then more comprehensive coverage is achieved, but measurement reliability decreases
Solution Approach 1:
The patent applies partial action by selectively extracting only certain contour points based on direction-specific weights rather than using all contour points equally. The random extraction process with weighting ensures that points from directions with high variability are extracted less frequently, achieving sufficient coverage of the pattern while filtering out problematic high-variation points
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy of pattern shape measurement by selectively extracting contour points with low variation, resulting in precise dimension calculation and improved measurement reliability.
Implementation Method 1
acquiring image data of a target pattern by irradiating an observation region of a sample with charged particles or an electromagnetic wave
Data Source
AI summary
A pattern shape measurement method includes acquiring image data of a target pattern obtained by irradiating an observation region of a sample with a charged particle or an electromagnetic wave, generating first contour point group data in which location information of a contour point of the target pattern extracted based on the image data and an index related to an orientation angle of a line connecting a center of the target pattern and the contour point from a reference line passing through the center of the target pattern are associated with each other, and generating, based on a weighting table in which the index and weight determined based on a standard deviation of location information of the contour point of the target pattern present in a direction of the orientation angle are associated with each other and stored, and the first contour point group data, second contour point group data according to the weight.