TEM Metrology Using 3D Model Simulation and Weighted Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional TEM-based metrology techniques for measuring patterned semiconductor structures are slow, inaccurate, and unable to provide robust full geometrical and material information due to their reliance on 2D image processing and limited robustness in handling low contrast features and noise.
Innovation Solution
A novel model-based interpretation method using 3D parametrized models and weighting approaches to process TEM image data, enabling accurate retrieval of geometric and material parameters by simulating TEM images and adjusting weights based on feature importance, thereby improving the measurement of both strong and weak parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional 2D image processing methods are used for TEM-based metrology, then the measurement process is simple, but the measurement precision and robustness are insufficient for complex patterned structures
Solution Approach 1:
The patent transitions from conventional 2D image processing to 3D model-based interpretation. A parametrized 3D model of the patterned structure is created, and simulated TEM images are generated from this 3D model. The measured 2D TEM image is then compared with the simulated images to extract parameters, effectively using 3D modeling to enhance 2D measurement capabilities and achieve higher precision for complex structures.
Solution Approach 2:
The patent performs preliminary simulation of TEM images based on a parametrized 3D model before actual measurement interpretation. By pre-generating simulated images with known parameters and comparing them against measured images, the system establishes a reference framework that guides the interpretation process and improves measurement accuracy without requiring complex real-time processing during actual measurement.
2Measurement precision
If model-based interpretation with reference system data is used, then measurement accuracy improves, but processing time and computational complexity increase
Solution Approach 1:
The patent creates simulated copies of TEM images from a parametrized 3D model that replicate the appearance and characteristics of actual measured images. These simulated images serve as reference patterns for comparison with measured images, enabling accurate parameter extraction through pattern matching without requiring time-consuming iterative reconstruction or complex reference system measurements.
3Productivity
If conventional TEM measurement methods are used, then the equipment and process are simple, but the productivity and measurement speed are slow
Solution Approach 1:
The patent replaces conventional mechanical/image-processing-based TEM analysis with a computational approach using 3D model simulation and digital image comparison. Instead of relying on complex mechanical scanning and manual or algorithmic 2D image processing, the system uses computational simulation to generate reference images and automatically compares them with measured images, significantly accelerating the measurement process while maintaining simplicity in the physical TEM instrument.
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 and robustness of parameter measurement, particularly for weak parameters, and allows for automatic, efficient retrieval of geometric and material information from TEM images, improving the performance of TEM-based metrology in semiconductor manufacturing.
Implementation Method 1
TEM/STEM is a microscopy technique in which a beam of electrons is transmitted through a very thin section or slice (also termed 'Lamellae') of a structure, and interaction of the electrons with features of the structure within the Lamellae, forms an image of said Lamellae
Implementation Method 2
interaction of the electrons with features of the structure within the Lamellae, forms an image
Data Source
AI summary
A metrology method for use in determining one or more parameters of a patterned structure, the method including providing raw measured TEM image data, TEMmeas, data indicative of a TEM measurement mode, and predetermined simulated TEM image data including data indicative of one or more simulated TEM images of a structure similar to the patterned structure under measurements and a simulated weight map including weights assigned to different regions in the simulated TEM image corresponding to different features of the patterned structure, performing a fitting procedure between the raw measured TEM image data and the predetermined simulated TEM image data and determining one or more parameters of the structure from the simulated TEM image data corresponding to a best fit condition.