Deep learning-based multimodal analysis for transition-metal dichalcogenides

IF 4.1 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY Mrs Bulletin Pub Date : 2024-06-25 DOI:10.1557/s43577-024-00741-6

Shivani Bhawsar, Mengqi Fang, Abdus Salam Sarkar, Siwei Chen, Eui-Hyeok Yang

{"title":"Deep learning-based multimodal analysis for transition-metal dichalcogenides","authors":"Shivani Bhawsar, Mengqi Fang, Abdus Salam Sarkar, Siwei Chen, Eui-Hyeok Yang","doi":"10.1557/s43577-024-00741-6","DOIUrl":null,"url":null,"abstract":"<h3 data-test=\"abstract-sub-heading\">Abstract</h3><p>In this study, we present a novel approach to enable high-throughput characterization of transition-metal dichalcogenides (TMDs) across various layers, including mono-, bi-, tri-, four, and multilayers, utilizing a generative deep learning-based image-to-image translation method. Graphical features, including contrast, color, shapes, flake sizes, and their distributions, were extracted using color-based segmentation of optical images, and Raman and photoluminescence spectra of chemical vapor deposition-grown and mechanically exfoliated TMDs. The labeled images to identify and characterize TMDs were generated using the pix2pix conditional generative adversarial network (cGAN), trained only on a limited data set. Furthermore, our model demonstrated versatility by successfully characterizing TMD heterostructures, showing adaptability across diverse material compositions.</p><h3 data-test=\"abstract-sub-heading\">Graphical abstract</h3><h3 data-test=\"abstract-sub-heading\">Impact Statement</h3><p>Deep learning has been used to identify and characterize transition-metal dichalcogenides (TMDs). Although studies leveraging convolutional neural networks have shown promise in analyzing the optical, physical, and electronic properties of TMDs, they need extensive data sets and show limited generalization capabilities with smaller data sets. This work introduces a transformative approach—a generative deep learning (DL)-based image-to-image translation method—for high-throughput TMD characterization. Our method, employing a DL-based pix2pix cGAN network, transcends traditional limitations by offering insights into the graphical features, layer numbers, and distributions of TMDs, even with limited data sets. Notably, we demonstrate the scalability of our model through successful characterization of different heterostructures, showcasing its adaptability across diverse material compositions.</p>","PeriodicalId":18828,"journal":{"name":"Mrs Bulletin","volume":"44 1","pages":""},"PeriodicalIF":4.1000,"publicationDate":"2024-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Mrs Bulletin","FirstCategoryId":"88","ListUrlMain":"https://doi.org/10.1557/s43577-024-00741-6","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MATERIALS SCIENCE, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

In this study, we present a novel approach to enable high-throughput characterization of transition-metal dichalcogenides (TMDs) across various layers, including mono-, bi-, tri-, four, and multilayers, utilizing a generative deep learning-based image-to-image translation method. Graphical features, including contrast, color, shapes, flake sizes, and their distributions, were extracted using color-based segmentation of optical images, and Raman and photoluminescence spectra of chemical vapor deposition-grown and mechanically exfoliated TMDs. The labeled images to identify and characterize TMDs were generated using the pix2pix conditional generative adversarial network (cGAN), trained only on a limited data set. Furthermore, our model demonstrated versatility by successfully characterizing TMD heterostructures, showing adaptability across diverse material compositions.

Graphical abstract

Impact Statement

Deep learning has been used to identify and characterize transition-metal dichalcogenides (TMDs). Although studies leveraging convolutional neural networks have shown promise in analyzing the optical, physical, and electronic properties of TMDs, they need extensive data sets and show limited generalization capabilities with smaller data sets. This work introduces a transformative approach—a generative deep learning (DL)-based image-to-image translation method—for high-throughput TMD characterization. Our method, employing a DL-based pix2pix cGAN network, transcends traditional limitations by offering insights into the graphical features, layer numbers, and distributions of TMDs, even with limited data sets. Notably, we demonstrate the scalability of our model through successful characterization of different heterostructures, showcasing its adaptability across diverse material compositions.

Abstract Image

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

基于深度学习的过渡金属二钴化物多模态分析

摘要在本研究中，我们提出了一种新方法，利用基于生成式深度学习的图像到图像转换方法，对过渡金属二钙化层（TMD）的各层（包括单层、双层、三层、四层和多层）进行高通量表征。利用基于颜色的光学图像分割，以及化学气相沉积生长和机械剥离 TMD 的拉曼光谱和光致发光光谱，提取了图形特征，包括对比度、颜色、形状、薄片尺寸及其分布。用于识别和表征 TMD 的标记图像是使用 pix2pix 条件生成式对抗网络 (cGAN) 生成的，该网络仅在有限的数据集上进行了训练。此外，我们的模型通过成功表征 TMD 异质结构而展示了其多功能性，显示出对不同材料组成的适应性。虽然利用卷积神经网络进行的研究在分析 TMD 的光学、物理和电子特性方面显示出了前景，但它们需要大量的数据集，而且在较小的数据集上显示出有限的泛化能力。这项工作引入了一种变革性方法--基于生成式深度学习（DL）的图像到图像转换方法，用于高通量 TMD 表征。我们的方法采用了基于 DL 的 pix2pix cGAN 网络，即使数据集有限，也能深入了解 TMD 的图形特征、层数和分布，从而超越了传统方法的局限性。值得注意的是，我们通过对不同异质结构的成功表征证明了我们模型的可扩展性，展示了它对不同材料组成的适应性。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

Mrs Bulletin 工程技术-材料科学：综合

CiteScore

7.40

自引率

2.00%

发文量

193

审稿时长

4-8 weeks

期刊介绍： MRS Bulletin is one of the most widely recognized and highly respected publications in advanced materials research. Each month, the Bulletin provides a comprehensive overview of a specific materials theme, along with industry and policy developments, and MRS and materials-community news and events. Written by leading experts, the overview articles are useful references for specialists, but are also presented at a level understandable to a broad scientific audience.