Off-axis Beam Tilting for Automated Feature Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated measurement of dense lines in integrated circuit technology, particularly with critical dimension (CD) scatterometry tools, faces challenges in distinguishing between lines and trenches with similar aspect ratios, leading to erroneous data and reduced precision due to operator errors and limitations in conventional algorithms.
Innovation Solution
The method employs off-axis beam tilting to collect image data from a region of interest at multiple angles, analyzing edge width and intensity data to accurately identify features as lines or trenches, enabling automated microalignment and undercut detection without relying on pattern recognition, thus improving the precision of CD measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional pattern recognition with top-down view images is used, then automated measurement can be achieved, but measurement precision deteriorates when line-to-space aspect ratio is close to one
Solution Approach 1:
The patent transitions from conventional top-down 2D imaging to off-axis tilted beam imaging, introducing a new dimensional perspective. By tilting the electron beam at specific angles (e.g., 15-45 degrees), the system creates projected images that reveal depth information and side walls of features, enabling clear distinction between lines and trenches even when their top-down appearances are identical.
Solution Approach 2:
The patent applies different imaging conditions to different measurement needs. For features with aspect ratio close to one, off-axis tilted beam imaging is used to reveal side wall characteristics. For other cases, conventional top-down imaging suffices. This localized application of specialized imaging techniques optimizes measurement precision without unnecessarily complicating the overall measurement process.
2Adaptability or versatility
If operator intervention is used to set up measurement recipes, then measurement flexibility is maintained, but erroneous data increases due to operator errors
Solution Approach 1:
The patent implements automated feature identification algorithms that self-adjust measurement parameters based on off-axis image analysis. The system automatically determines whether a feature is a line or trench by analyzing the characteristic patterns in tilted beam images, eliminating the need for operator interpretation and reducing human error while maintaining measurement flexibility through programmable adaptability.
Solution Approach 2:
The system uses feedback from off-axis image analysis to automatically adjust and optimize measurement recipes. By comparing the observed image characteristics with expected patterns for lines and trenches, the system self-corrects and refines measurement parameters, ensuring high reliability without requiring operator intervention for each measurement setup.
3Measurement precision
If multiple imaging angles are used, then feature identification accuracy improves, but measurement time increases
Solution Approach 1:
The patent applies partial tilting rather than exhaustive multi-angle imaging. By using a single off-axis tilt angle (or a limited set of angles) sufficient to reveal the critical distinguishing features, the system achieves high identification accuracy without the time penalty of complete 360-degree imaging. This partial action approach captures only the essential information needed for accurate feature classification.
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 allows for fully automated metrology recipes with increased tolerance for error, reducing incorrect flagging and improving measurement accuracy in dense arrays, particularly for features with aspect ratios close to one, and can detect undercuts effectively.
Implementation Method 1
Image data is collected from a region of interest on a substrate at multiple beam tilts... The scattering of electrons by the feature produces an image that depends on the feature geometry and the beam incident angle
Implementation Method 2
The beam may be tilted electrostatically, or may alternately be achieved by other mechanisms
Data Source
AI summary
One embodiment relates to a method of automated microalignment using off-axis beam tilting. Image data is collected from a region of interest on a substrate at multiple beam tilts. Potential edges of a feature to be identified in the region are determined, and computational analysis of edge-related data is performed to positively identify the feature(s). Another embodiment relates to a method of automated detection of undercut on a feature using off-axis beam tilting. For each beam tilt, a determination is made of difference data between the edge measurement of one side and the edge measurement of the other side. An undercut on the feature is detected from the difference data. Other embodiments are also disclosed.