Automated mapping method of crystalline structure and orientation of polycrystalline material with deep learning

Patent metadata

Field	Value
Publication number	US-12487196-B2
Application number	US-202217959793-A
Country	US
Kind code	B2
Filing date	Oct 4, 2022
Priority date	Oct 15, 2021
Publication date	Dec 2, 2025
Grant date	Dec 2, 2025

Abstract

A method for two-dimensional mapping of crystal information of a polycrystalline material may include acquiring a diffraction pattern acquired by scanning an electron beam to a polycrystalline material, generating a plurality of clusters by applying a clustering algorithm to the acquired diffraction pattern based on unsupervised learning, acquiring crystal information of the polycrystalline material by applying a parallel deep convolutional neural network (DCNN) algorithm to each of the plurality of generated clusters based on supervised learning, and generating a two-dimensional image in which the acquired crystal information is mapped.

First claim

What is claimed is:

1 . A method for two-dimensional mapping of crystal information of a polycrystalline material, the method comprising:
acquiring a diffraction pattern by scanning an electron beam to the polycrystalline material; generating a plurality of clusters by clustering the acquired diffraction pattern based on unsupervised learning; acquiring the crystal information of the polycrystalline material by distributing and processing the crystal information in a parallel deep convolutional neural network (DCNN) for each of the plurality of generated clusters based on supervised learning, wherein the crystal information includes a crystal symmetry, a crystal tilt and a crystal thickness, each acquired using the parallel DCNN; and generating a two-dimensional image in which the acquired crystal information is mapped.

2 . The method of claim 1 , wherein the generating of the plurality of clusters comprises:
classifying the acquired diffraction pattern based on crystal grains contained in the polycrystalline material; and generating a positional averaged diffraction pattern for the classified diffraction pattern.

3 . The method of claim 2 , wherein the number of the plurality of generated clusters is the same as the number of the crystal grains included in the polycrystalline material.

4 . The method of claim 2 , wherein the generating of the plurality of clusters further comprising:
performing rotation correction on a direction of the generated positional average diffraction pattern to correspond to a direction of a simulated positional averaged diffraction pattern.

5 . The method of claim 1 , wherein the generating of the plurality of clusters comprises applying at least one of a K-means algorithm, a mean shift algorithm, a Gaussian mixture model (GMM) algorithm, or a density based spatial clustering of applications with noise (DBSCAN) algorithm with respect to the acquired diffraction pattern.

6 . The method of claim 1 , wherein the parallel DCNN applied to each of the plurality of clusters comprises
a first DCNN algorithm associated with the crystal symmetry of the polycrystalline material, a second DCNN algorithm associated with the crystal tilt, and a third DCNN algorithm associated with the crystal thickness.

7 . The method of claim 6 , wherein the parallel DCNN applied to each of the plurality of generated clusters further comprises a fourth DCNN algorithm associated with strain of a crystal structure.

8 . The method of claim 1 , wherein the two-dimensional image comprises a first map image associated with the crystal symmetry, a second map image associated with the crystal tilt, and a third map image associated with the crystal thickness.

9 . The method of claim 1 , wherein the polycrystalline material has an ultra-thin film thickness of about 10 nm or less.

10 . The method of claim 1 , wherein the acquired diffraction pattern comprises convergent beam electron diffraction (CBED) data acquired through 4D-scanning transmission electron microscope (4D-STEM).

11 . An apparatus for two-dimensional mapping of crystal information of a polycrystalline material, the apparatus comprising:
an image acquisition unit configured to detect a captured image of a sample having the polycrystalline material by scanning an electron beam on the polycrystalline material; and an image processing unit configured to
acquire a diffraction pattern acquired by the scanning of the electron beam on the polycrystalline material,
generate a plurality of clusters by clustering the diffraction pattern acquired based on unsupervised learning,
acquire the crystal information of the polycrystalline material by distributing and processing the crystal information in a parallel deep convolutional neural network (DCNN) for each of the plurality of generated clusters based on supervised learning, wherein the crystal information includes a crystal symmetry, a crystal tilt and a crystal thickness, each acquired using the parallel DCNN, and
generate a two-dimensional image in which the acquired crystal information is mapped.

12 . The apparatus of claim 11 , wherein the image processing unit is configured to classify the acquired diffraction pattern based on crystal grains included in the polycrystalline material and to generate a positional average diffraction pattern with respect to the classified diffraction pattern.

13 . The apparatus of claim 12 , wherein the number of the plurality of generated clusters is the same as the number of the crystal grains included in the polycrystalline material.

14 . The apparatus of claim 12 , wherein the image processing unit is configured to perform rotation correction on a direction of the generated positional average diffraction pattern to correspond to a direction of a simulated positional averaged diffraction pattern.

15 . The apparatus of claim 11 , wherein the image processing unit is configured to generate the plurality of clusters by applying at least one of a K-means algorithm, a mean shift algorithm, a Gaussian mixture model (GMM) algorithm, or a density based spatial clustering of applications with noise (DBSCAN) algorithm with respect to the acquired diffraction pattern.

16 . The apparatus of claim 11 , wherein the parallel DCNN applied to each of the plurality of clusters comprises
a first DCNN algorithm associated with the crystal symmetry of the polycrystalline material, a second DCNN algorithm associated with the crystal tilt, and a third DCNN algorithm associated with the crystal thickness.

17 . The apparatus of claim 16 , wherein the parallel DCNN applied to each of the plurality of generated clusters further comprises a fourth DCNN algorithm associated with strain of a crystal structure.

18 . The apparatus of claim 11 , wherein the two-dimensional image comprises a first map image associated with the crystal symmetry, a second map image associated with the crystal tilt, and a third map image associated with the crystal thickness.

19 . The apparatus of claim 11 , wherein the apparatus is configured to generate the two-dimensional image for the polycrystalline material which has an ultra-thin film thickness of about 10 nm or less.

20 . The apparatus of claim 11 , wherein
the image acquisition unit includes a 4D-scanning transmission electron microscope (4D-STEM), and the acquired diffraction pattern comprises convergent beam electron diffraction (CBED) data acquired through the 4D-STEM.

Assignees

• Samsung Electronics Co Ltd
• Research & Business Found Sungkyunkwan Univ

Inventors

• Kim Young-Min
• Lee Eunha
• Jeong Myoungho
• Kim Young-Hoon
• Yang Sang-Hyeok

Classifications

• G06V10/763

  Non-hierarchical techniques, e.g. based on statistics of modelling distributions · CPC title

• G06T2207/10061

  from scanning electron microscope · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title

• G06V10/82

  using neural networks · CPC title