# US 2025/0209604 A1 · Crosstalk correction for images of semiconductor specimens

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0209604 A1
- Document ID: US20250209604A1
- Status: Published application
- Kind: A1
- Application number: 18/393,570
- Filed: 2023-12-21
- Published: 2025-06-26
- Earliest claimed: 2023-12-21
- Assignees: Applied Materials Israel Ltd.
- Applicants: Applied Materials Israel Ltd.
- Inventors: Lior AKERMAN
- Web page: https://patentgrep.com/patent/US20250209604A1

## CPC

- G06T 7/0004: {Industrial image inspection}
- G06T 7/0006: {using a design-rule based approach}
- G06T 5/50: using two or more images, e.g. averaging or subtraction
- G06T 5/90: Dynamic range modification of images or parts thereof
- G06T 7/001: {using an image reference approach}
- G06T 7/12: Edge-based segmentation
- G06T 7/143: involving probabilistic approaches, e.g. Markov random field [MRF] modelling
- G06T 2207/10061: from scanning electron microscope
- G06T 2207/20081: Training; Learning
- G06T 2207/30148: Semiconductor; IC; Wafer
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

There is provided a system and method of examination of a semiconductor specimen. The semiconductor specimen comprises at least a surface layer and a under layer. The method includes obtaining a secondary electron (SE) image and a backscattered electron (BSE) image of the specimen acquired by an electron beam tool, wherein the BSE image possesses one or more image artifacts caused by one or more structural features on the surface layer; processing the SE image to generate a feature mask comprising a set of segments representative of the one or more structural features; and generating a corrected BSE image based on the BSE image and the feature mask, wherein the corrected BSE image possesses suppressed image artifacts with respect to the one or more image artifacts in the BSE image.

## Claims

### Claim 1

A computerized system of examination of a semiconductor specimen, the semiconductor specimen comprising at least a surface layer and a under layer, the system comprising a processing circuitry configured to:
obtain a secondary electron (SE) image and a backscattered electron (BSE) image of the specimen acquired by an electron beam tool, wherein the BSE image possesses one or more image artifacts caused by one or more structural features on the surface layer;
process the SE image to generate a feature mask comprising a set of segments representative of the one or more structural features; and
generate a corrected BSE image based on the BSE image and the feature mask, wherein the corrected BSE image possesses suppressed image artifacts with respect to the one or more image artifacts in the BSE image.

### Claim 2 (depends on 1)

The computerized system according to claim 1, wherein the processing circuitry is configured to generate the corrected BSE image by applying a correction factor to the feature mask to obtain a corrected feature mask, and correcting the BSE image using the corrected feature mask.

### Claim 3 (depends on 2)

The computerized system according to claim 2, wherein the correction factor is used for minimizing effects of the one or more image artifacts in the BSE image, taking into consideration gray level variations between the SE image and the BSE image.

### Claim 4 (depends on 1)

The computerized system according to claim 1, wherein the SE image is a topographic image where contours of the structural features on the surface layer are enhanced, and wherein the feature mask is generated by performing image segmentation on the SE image so as to obtain the set of segments as geometrical representation of the structural features.

### Claim 5 (depends on 4)

The computerized system according to claim 4, wherein the image segmentation is performed based on one of the following: edge detection, machine learning, a segmentation threshold for the SE image, or design data of the specimen.

### Claim 6 (depends on 2)

The computerized system according to claim 2, wherein the correction factor is automatically selected by:
creating a temporary BSE image by applying an initial correction factor to the feature mask to get a temporary correction mask, and correcting the BSE image using the temporary correction mask;
calculating a gradient energy for pixels of the temporary BSE image that correspond to contour pixels of the SE image, and
iterating the creating and calculating with a different correction factor until an iteration condition is met, and selecting the correction factor having minimized gradient energy, wherein the selected correction factor, when being used, is expected to result in the suppressed image artifacts in the BSE image.

### Claim 7 (depends on 6)

The computerized system according to claim 6, wherein the calculating a gradient energy comprises:
generating a binary mask based on the SE image, the binary mask representing contours of the structural features on the surface layer;
calculating absolute gradients for pixels of the temporary BSE image that correspond to pixels of the contours in the binary mask; and
aggregating the absolute gradients to obtain a gradient energy corresponding to the initial correction factor.

### Claim 8 (depends on 2)

The computerized system according to claim 2, wherein the correction factor is automatically selected by:
selecting a region of interest (ROI) on the BSE image containing an image artifact and a surrounding area of the image artifact, the surrounding area not including any structural feature of the under layer;
creating a temporary BSE image by applying an initial correction factor to the feature mask to get a temporary correction mask, and correcting the BSE image using the temporary correction mask;
calculating, based on the temporary BSE image, a first average gray level (GL) value for pixels in the surrounding area and a second average GL value for pixels belonging to the image artifact;
computing an absolute difference value between the first and second average GL values, the absolute difference value corresponding to the initial correction factor; and
iterating the creating, calculating and computing with a different correction factor until an iteration condition is met, and selecting the correction factor that corresponds to a minimized absolute difference value, wherein the selected correction factor, when being used, is expected to result in the suppressed image artifacts in the BSE image.

### Claim 9 (depends on 1)

The computerized system according to claim 1, wherein the surface layer comprises multiple types of structural features, and wherein the processing circuitry is configured to process the SE image to generate multiple feature masks corresponding to the multiple types of structural features, and generate the corrected BSE image by applying respective correction factors to the multiple feature masks to obtain multiple corrected feature masks, and correcting the BSE image using the multiple corrected feature masks.

### Claim 10 (depends on 1)

The computerized system according to claim 1, wherein the specimen comprises multiple under layers at different depths, each under layer comprising respective structural features captured by the BSE image, and wherein the processing circuitry is configured to generate the corrected BSE image by dividing the BSE image into different regions based on the structural features from different under layers, applying different correction factors for regions in the feature mask corresponding to the different regions in the BSE image, so as to obtain the corrected feature mask; and correcting the BSE image using the corrected feature mask.

### Claim 11

A computerized method of examination of a semiconductor specimen, the semiconductor specimen comprising at least a surface layer and a under layer, the method comprising:
obtaining a secondary electron (SE) image and a backscattered electron (BSE) image of the specimen acquired by an electron beam tool, wherein the BSE image possesses one or more image artifacts caused by one or more structural features on the surface layer;
processing the SE image to generate a feature mask comprising a set of segments representative of the one or more structural features; and
generating a corrected BSE image based on the BSE image and the feature mask, wherein the corrected BSE image possesses suppressed image artifacts with respect to the one or more image artifacts in the BSE image.

### Claim 12 (depends on 11)

The computerized method according to claim 11, wherein the generating the corrected BSE image comprises applying a correction factor to the feature mask to obtain a corrected feature mask, and correcting the BSE image using the corrected feature mask.

### Claim 13 (depends on 12)

The computerized method according to claim 12, wherein the correction factor is used for minimizing effects of the one or more image artifacts in the BSE image, taking into consideration gray level variations between the SE image and the BSE image.

### Claim 14 (depends on 11)

The computerized method according to claim 11, wherein the SE image is a topographic image where contours of the structural features on the surface layer are enhanced, and wherein the feature mask is generated by performing image segmentation on the SE image so as to obtain the set of segments as geometrical representation of the structural features.

### Claim 15 (depends on 12)

The computerized method according to claim 12, wherein the correction factor is automatically selected by:
creating a temporary BSE image by applying an initial correction factor to the feature mask to get a temporary correction mask, and correcting the BSE image using the temporary correction mask;
calculating a gradient energy for pixels of the temporary BSE image that correspond to contour pixels of the SE image, and
iterating the creating and calculating with a different correction factor until an iteration condition is met, and selecting the correction factor having minimized gradient energy, wherein the selected correction factor, when being used, is expected to result in the suppressed image artifacts in the BSE image.

### Claim 16 (depends on 15)

The computerized method according to claim 15, wherein the calculating a gradient energy comprises:
generating a binary mask based on the SE image, the binary mask representing contours of the structural features on the surface layer;
calculating absolute gradients for pixels of the temporary BSE image that correspond to pixels of the contours in the binary mask; and
aggregating the absolute gradients to obtain a gradient energy corresponding to the initial correction factor.

### Claim 17 (depends on 12)

The computerized method according to claim 12, wherein the correction factor is automatically selected by:
selecting a region of interest (ROI) on the BSE image containing an image artifact and a surrounding area of the image artifact, the surrounding area not including any structural feature of the under layer;
creating a temporary BSE image by applying an initial correction factor to the feature mask to get a temporary correction mask, and correcting the BSE image using the temporary correction mask;
calculating, based on the temporary BSE image, a first average gray level (GL) value for pixels in the surrounding area and a second average GL value for pixels belonging to the image artifact;
computing an absolute difference value between the first and second average GL values, the absolute difference value corresponding to the initial correction factor; and
iterating the creating, calculating, and computing with a different correction factor until an iteration condition is met, and selecting the correction factor that corresponds to a minimized absolute difference value, wherein the selected correction factor, when being used, is expected to result in the suppressed image artifacts in the BSE image.

### Claim 18 (depends on 11)

The computerized method according to claim 11, wherein the surface layer comprises multiple types of structural features, and wherein the processing the SE image comprises processing the SE image to generate multiple feature masks corresponding to the multiple types of structural features, and generating the corrected BSE image by applying respective correction factors to the multiple feature masks to obtain multiple corrected feature masks, and correcting the BSE image using the multiple corrected feature masks.

### Claim 19 (depends on 11)

The computerized method according to claim 11, wherein the specimen comprises multiple under layers at different depths, each under layer comprising respective structural features captured by the BSE image, and wherein the generating the corrected BSE image comprises dividing the BSE image into different regions based on the structural features from different under layers, applying different correction factors for regions in the feature mask corresponding to the different regions in the BSE image, so as to obtain the corrected feature mask; and correcting the BSE image using the corrected feature mask.

### Claim 20

A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by a computer, cause the computer to perform a method of examination of a semiconductor specimen, the semiconductor specimen comprising at least a surface layer and a under layer, the method comprising:
obtaining a secondary electron (SE) image and a backscattered electron (BSE) image of the specimen acquired by an electron beam tool, wherein the BSE image possesses one or more image artifacts caused by one or more structural features on the surface layer;
processing the SE image to generate a feature mask comprising a set of segments representative of the one or more structural features; and
generating a corrected BSE image based on the BSE image and the feature mask, wherein the corrected BSE image possesses suppressed image artifacts with respect to the one or more image artifacts in the BSE image.
