Wafer Contour Extraction Using Simulated SEM Image Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor metrology techniques face limitations in accurately extracting contours of structures on wafers due to imaging properties like beam size and noise, which impact resolution and throughput, especially when measuring buried layers.
Innovation Solution
A system and method that utilize a particle beam source, detector, and processor to generate and analyze images, incorporating image modeling algorithms to account for noise and imaging properties, enabling accurate contour extraction with lower resolution imaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution imaging is used to accurately extract contours, then measurement precision is improved, but throughput decreases due to longer measurement times
Solution Approach 1:
The patent creates a simulated image that copies the appearance of the actual structure as it would be imaged by the metrology tool, including all imaging properties and noise characteristics. This simulated copy is then compared to the actual measured image to extract the true contour, eliminating the need for high-resolution imaging while maintaining accuracy.
Solution Approach 2:
The patent introduces a simulated image as an intermediary between the actual measured image and the true contour. This simulated image acts as a mediator that accounts for imaging properties and noise, allowing accurate contour extraction from lower-resolution measurements without direct high-resolution imaging.
2Measurement precision
If low beam current is used to achieve higher resolution, then measurement precision is improved, but throughput decreases due to reduced signal intensity
Solution Approach 1:
The patent generates a simulated image that replicates what the structure would look like under the actual imaging conditions (including low beam current effects), allowing accurate contour extraction without requiring high beam current or high resolution imaging.
Solution Approach 2:
The patent changes the approach from improving physical imaging parameters (beam current, resolution) to modifying the data processing parameters. By using image modeling algorithms that account for imaging properties and noise, the system achieves accurate measurements without changing the physical imaging parameters to higher values.
3Measurement precision
If image modeling algorithms are used to account for noise and imaging properties, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent creates a simulated image that copies the expected appearance of the structure including all imaging properties and noise characteristics. This simulated copy is generated using image modeling algorithms and then compared to the actual measured image to extract accurate contour information.
Solution Approach 2:
The patent uses a feedback approach where the simulated image (based on modeling algorithms) is compared to the actual measured image, and the difference is used to refine the contour extraction. This iterative feedback process improves measurement precision while managing processing complexity through efficient algorithm design.
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
Enables accurate contour extraction independent of imaging properties, allowing faster measurements of both surface and buried layers, and improves the precision of edge placement error metrology.
Implementation Method 1
a particle beam source that generates a particle beam
Data Source
AI summary
A wafer metrology tool, such as a scanning electron microscope, can generate an image of a structure on a wafer. A simulated image of the structure also is determined from a design of the wafer. A contour of the structure in the image and a contour of the structure in the simulated image are determined. These contours are compared.