SEM Image Deconvolution for Sample Charging Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Scanning electron microscope (SEM) images of electrically insulating materials suffer from SEM-induced charging artifacts, leading to distorted representations of IC structures, which can result in false defect detection and reduced accuracy in inspection processes.
Innovation Solution
The method involves obtaining two SEM images using different scanning parameters, such as directions or velocities, and applying convolution equations to generate a reduced sample charging effect image by retrieving and utilizing the point spread functions corresponding to each image, thereby minimizing SEM-induced charging artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SEM imaging is performed on electrically insulating materials, then structural information can be obtained, but SEM-induced charging artifacts distort the image quality
Solution Approach 1:
The patent changes the scanning parameter (scan direction) to acquire multiple images with different charging artifact patterns. By rotating the scan direction by 90 degrees, the charging artifacts appear in different orientations in each image, allowing their mathematical separation from the true structural information through deconvolution operations.
Solution Approach 2:
The patent introduces a mathematical model (convolution equation) as an intermediary to separate the charging artifacts from the true image. The model represents the image formation process as I(x,y) = (O(x,y) ⊗ PSF(x,y)) + C(x,y), where the point spread function and charging artifact components can be independently solved and removed.
2Measurement precision
If multiple SEM images are acquired with different scan directions, then charging artifacts can be reduced, but inspection time increases
Solution Approach 1:
The patent uses a minimal set of two images acquired with 90-degree rotated scan directions, which is sufficient to solve the system of convolution equations and remove charging artifacts. This partial action approach avoids the need for acquiring multiple images in many different directions, thereby reducing inspection time while still achieving effective artifact removal.
3Object-affected harmful factors
If conventional averaging of multiple SEM images is used, then some charging artifacts are reduced, but image resolution and detail are lost
Solution Approach 1:
The patent replaces the simple mechanical averaging operation with a sophisticated mathematical deconvolution process. Instead of merely averaging pixel values, the system solves convolution equations to mathematically separate and remove charging artifacts while preserving the true structural information, thereby maintaining image resolution and detail that would be lost in simple averaging.
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 results in more accurate SEM images that faithfully represent the original structures, reducing false errors and improving defect detection accuracy by minimizing SEM-induced charging artifacts, thus enhancing the inspection process.
Implementation Method 1
scanning electron microscope (SEM) image... obtained from a first electron beam scan
Implementation Method 2
SEM-induced charging artifacts... sample charging effect
Data Source
AI summary
Disclosed herein is a method of reducing a sample charging effect in a scanning electron microscope (SEM) image, the method comprising: obtaining a first SEM image from a first electron beam scan with a parameter being a first quantity; obtaining a second SEM image from a second electron beam scan with the parameter being a second quantity different from the first quantity; and generating a reduced sample charging effect image based on convolution equations comprising a representation of the first SEM image, a representation of the second SEM image, a first point spread function corresponding to the first SEM image and a second point spread function corresponding to the second SEM image.