# US 2026/0301156 A1 · Method for measuring overlay using scanning electron microscope device

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

- Number: US 2026/0301156 A1
- Document ID: US20260301156A1
- Status: Published application
- Kind: A1
- Application number: 19/353,945
- Filed: 2025-10-09
- Published: 2026-10-01
- Earliest claimed: 2025-03-28
- Assignees: Samsung Electronics Co., Ltd.
- Applicants: Samsung Electronics Co., Ltd.
- Inventors: Yeny Yim; Jhongkwon Kim; Wonhee Lee; Chulmoo Kang; Hyeju Kang; Woongkyu Son; Sungeun Eom; Taehyun Yun; Sanghyun Joo
- Web page: https://patentgrep.com/patent/US20260301156A1

## CPC

- G06T 7/001: {using an image reference approach}
- G06T 7/60: Analysis of geometric attributes
- G06T 7/74: {involving reference images or patches}
- H01J 37/222: {Image processing arrangements associated with the tube}
- G06T 2207/10061: from scanning electron microscope
- G06T 2207/20081: Training; Learning
- G06T 2207/30148: Semiconductor; IC; Wafer
- H01J 2237/221: Image processing
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H01J: ELECTRIC DISCHARGE TUBES OR DISCHARGE LAMPS (spark-gaps H01T; electric arc lamps H05B31/00; particle accelerators H05H)

## Abstract

A method of measuring an overlay includes obtaining a plurality of first images, each of the plurality of first images being generated by integrating N first frames generated by a scanning electron microscope (SEM) device, generating a plurality of training images by respectively applying noise to the plurality of first images, inputting the plurality of training images into a machine learning model, obtaining, from the machine learning model, predicted noise by determining, by the machine learning model, noise in each of the plurality of training images, training the machine learning model by respectively comparing the predicted noise of the plurality of training images with the noise applied to the plurality of first images, obtaining an original image in which M second frames generated by the SEM device are integrated, obtaining a result image by inputting the original image into the machine learning model for which training is completed.

## Claims

### Claim 1

A method of measuring an overlay, comprising:
obtaining a plurality of first images, each of the plurality of first images being generated by integrating N first frames generated by a scanning electron microscope (SEM) device, wherein Nis a natural number equal to or greater than 2;
generating a plurality of training images by respectively applying noise to the plurality of first images;
inputting the plurality of training images into a machine learning model;
obtaining, from the machine learning model, predicted noise by determining, by the machine learning model, noise in each of the plurality of training images;
training the machine learning model by respectively comparing the predicted noise of the plurality of training images with the noise applied to the plurality of first images;
obtaining an original image in which M second frames generated by the SEM device are integrated, wherein M is a natural number less than N;
obtaining a result image by inputting the original image into the machine learning model for which training is completed; and
measuring an overlay based on the result image.

### Claim 2 (depends on 1)

The method of claim 1, wherein Nis at least twice M.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
determining a boundary of at least one structure in the result image; and
determining a center position of the at least one structure based on the boundary; and
wherein measuring the overlay comprises measuring the overlay based on the center position.

### Claim 4 (depends on 3)

The method of claim 3, wherein determining the center position of the at least one structure comprises:
determining a first center position of the at least one structure from a first result image corresponding to a first level at a first height; and
determining a second center position of the at least one structure from a second result image corresponding to a second level at a second height that is different from the first height, and
wherein measuring the overlay based on the center position comprises comparing the first center position and the second center position.

### Claim 5 (depends on 1)

The method of claim 1, further comprising:
obtaining a first profile represented by at least one structure in the result image in a first direction; and
determining at least one first center position of the at least one structure in the first direction based on the first profile, and
wherein measuring the overlay comprises measuring the overlay in the first direction based on the at least one first center position.

### Claim 6 (depends on 5)

The method of claim 5, further comprising:
obtaining a first result image of a predetermined target region captured at a first height, and a second result image of the predetermined target region captured at a second height that is different from the first height;
acquiring a filter image by filtering the second result image from the first result image; and
generating the first profile represented by the at least one structure in the filter image in the first direction.

### Claim 7 (depends on 5)

The method of claim 5, further comprising:
obtaining a second profile represented by the at least one structure in the result image in a second direction that is perpendicular to the first direction; and
determining at least one second center position of the at least one structure in the second direction based on the second profile, and
wherein measuring the overlay comprises measuring the overlay in the second direction based on the at least one second center position.

### Claim 8 (depends on 5)

The method of claim 5, further comprising:
acquiring a filtered profile by deleting an asymmetric region corresponding to at least a portion of the first profile,
wherein determining the at least one first center position comprises determining the at least one first center position of the at least one structure in the first direction based on the filtered profile, and
wherein measuring the overlay comprises measuring the overlay in the first direction based on the at least one first center position.

### Claim 9 (depends on 8)

The method of claim 8, further comprising:
generating a reference profile based on the filtered profile; and
determining the at least one first center position by generating a correlation coefficient between the reference profile and the filtered profile.

### Claim 10 (depends on 1)

The method of claim 1, further comprising:
determining original image predicted noise that estimates original noise in the original image by inputting the original image into the machine learning model, and
wherein obtaining the result image comprises processing the original image based on the original image predicted noise.

### Claim 11 (depends on 10)

The method of claim 10, wherein the original image predicted noise comprises a predicted average value and a predicted standard deviation value that estimate the original noise.

### Claim 12 (depends on 1)

The method of claim 1, wherein the result image comprises an image in which the SEM device captures a target region in at least one chip region among a plurality of chip regions, and
wherein the method further comprises compensating an error of an overlay measured from an overlay key pattern formed in a scribing region between the plurality of chip regions based on the overlay measured based on the result image.

### Claim 13

A method of measuring an overlay, comprising:
inputting a training image into a machine learning model, the training image comprising an image captured by a scanning electron microscope (SEM) device to which noise is applied;
determining a loss function based on a predicted average value and a predicted standard deviation value that are output by the machine learning model, and an actual average value and an actual standard deviation value in the noise applied to the training image;
training the machine learning model; and
based on the training being completed, storing the machine learning model in a server, the machine learning model being configured to control the SEM device.

### Claim 14 (depends on 13)

The method of claim 13, wherein the training the machine learning model is executed in a training server that is separate from the server.

### Claim 15 (depends on 13)

The method of claim 13, wherein the machine learning model comprises a generative model.

### Claim 16

A method of measuring an overlay, comprising:
obtaining an original image in which M frames generated by a scanning electron microscope (SEM) device are integrated, wherein M is a natural number equal to or greater than 2;
obtaining a predicted average value and a predicted standard deviation value that estimate noise in the original image by inputting the original image into a machine learning model;
generating a result image by processing the original image using the predicted average value and the predicted standard deviation value;
determining a center position of each of structures in the original image based on the result image; and
measuring an overlay based on the center position of each of the structures.

### Claim 17 (depends on 16)

The method of claim 16, wherein integrating is 2 or more and less than 10.

### Claim 18 (depends on 16)

The method of claim 16, wherein the machine learning model is configured to predict the noise in the original image as a Gaussian distribution.

### Claim 19 (depends on 16)

The method of claim 16, wherein the SEM device is configured to generate the M frame by capturing at least one target region in a plurality of chip regions in a wafer.

### Claim 20 (depends on 19)

The method of claim 19, wherein measuring the overlay comprises:
measuring a first overlay based on an overlay key pattern in a scribing region between the plurality of chip regions on the wafer; measuring a second overlay based on the center position of each of the structures; and
compensating for an error of the first overlay based on the second overlay.
