Multi-Ellipse Shape Model for Semiconductor Contour Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor processing technologies face challenges in accurately and efficiently extracting dimensions of complex shapes from semiconductor samples, particularly with the increasing complexity of device structures and the need for high-speed, high-accuracy dimension measurement without manual operation.
Innovation Solution
A shape model using multiple ellipses is employed to describe contours, allowing for the fitting of a shape model to the intended structure's contour data, enabling accurate extraction of dimensions through analytical methods rather than relying on large databases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a simple shape model is used for database verification, then the estimation process is fast, but the recognition of complicated shapes is difficult
Solution Approach 1:
The shape model is segmented into multiple ellipse components that can be independently adjusted. Each ellipse represents a specific portion of the shape (e.g., top, bottom, sides), allowing the model to capture complex geometries by combining multiple simple geometric primitives. This segmentation enables accurate representation of complicated shapes while maintaining computational efficiency through parameterized ellipse definitions.
Solution Approach 2:
The shape model transitions from a static simple geometry to a dynamic multi-ellipse configuration. The ellipses can be independently positioned, scaled, and rotated to adapt to various shape complexities. This dynamic structure allows the same model framework to represent both simple and complicated shapes by adjusting the number and parameters of ellipse components, resolving the trade-off between model simplicity and shape recognition capability.
2Adaptability or versatility
If a very large-scale database is used to avoid difficulty of recognizing unknown shapes, then shape coverage is improved, but the estimation time becomes unreasonably long
Solution Approach 1:
Instead of expanding database size to improve shape coverage, the invention changes the parameters of a compact multi-ellipse model to adapt to different shapes. By varying ellipse parameters (position, size, orientation, number of ellipses), the model can represent a wide range of shape complexities without requiring a large database. This parameter-based adaptation achieves high versatility while maintaining fast estimation through direct model fitting rather than database search.
3Measurement precision
If manual extraction of dimensions is performed, then measurement accuracy can be ensured, but the process becomes increasingly difficult with finer structures
Solution Approach 1:
The invention creates a parametric copy of the shape using multiple ellipses that mirrors the actual semiconductor structure. This multi-ellipse model copy can be automatically fitted to measured data points, enabling automated dimension extraction without manual intervention. The parametric nature of the ellipse-based model allows automatic calculation of dimensions (width, length, curvature) from the fitted parameters, replacing complex manual extraction processes while maintaining accuracy even for fine structures.
4Manufacturing precision
If a large number of dimension measurements are made for process development, then optimization of control parameters is improved, but the process development time increases
Solution Approach 1:
The invention replaces the mechanical/manual process of repeated dimension measurements and manual analysis with an automated multi-ellipse fitting system. The system automatically fits multi-ellipse models to SEM images, extracts dimensions, and provides feedback for process optimization. This substitution of automated computational mechanics for manual measurement processes maintains the ability to optimize control parameters through comprehensive data collection while dramatically improving process development efficiency by eliminating manual extraction bottlenecks.
Data Source
AI summary
High-accuracy dimension extraction for a complicated shape possibly appearing through semiconductor processing. A starting point and an endpoint are placed on the periphery of a graphic shape including a combination of multiple ellipses, and a curve unicursally drawn on the periphery between the two points is used as a shape mode, thereby describing a contour of an intended structure.