Deep generative model for the inverse design of Van der Waals heterostructures

Abstract This study proposes ConditionCDVAE+, a crystal diffusion variational autoencoder (CDVAE) based deep generative model for inverse design of van der Waals (vdW) heterostructures. To address the challenges of traditional experimental methods relying on trial-and-error and existing models strug...

Full description

Saved in:
Bibliographic Details
Main Authors: Shikun Gao, Qinyuan Huang, Chen Huang, Cheng Li, Kaihao Liu, Baisheng Sa, Yadong Yu, Dezhen Xue, Zhe Liu, Mengyan Dai
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-025-06432-9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849768709019664384
author Shikun Gao
Qinyuan Huang
Chen Huang
Cheng Li
Kaihao Liu
Baisheng Sa
Yadong Yu
Dezhen Xue
Zhe Liu
Mengyan Dai
author_facet Shikun Gao
Qinyuan Huang
Chen Huang
Cheng Li
Kaihao Liu
Baisheng Sa
Yadong Yu
Dezhen Xue
Zhe Liu
Mengyan Dai
author_sort Shikun Gao
collection DOAJ
description Abstract This study proposes ConditionCDVAE+, a crystal diffusion variational autoencoder (CDVAE) based deep generative model for inverse design of van der Waals (vdW) heterostructures. To address the challenges of traditional experimental methods relying on trial-and-error and existing models struggling to incorporate target property constraints, this work achieves breakthroughs through three innovative stages: (1) introduce the SE(3)-equivariant graph neural network EquiformerV2 as the encoder-decoder within the CDVAE framework to enhance the generation quality of the model; (2) design a module integrating Low-rank Multimodal Fusion and Generative Adversarial Networks to map properties and structures into a joint latent space; and (3) for the first time propose a generative model for the vdW heterostructures, by conducting experimental validation on the dataset constructed from Janus III–VI vdW heterostructures. Experiments demonstrate that ConditionCDVAE+ achieves optimal root mean square error for crystal reconstruction, with improved generation quality. Density Functional Theory calculations confirms 99.51% of generated samples converge to energy minima, indicating superior ground-state convergence. The effectiveness of the model under conditional guidance has also been extensively validated. This framework provides an efficient solution for target-oriented design of vdW heterostructures and holds promise for accelerating the development of novel optoelectronic devices.
format Article
id doaj-art-15949522db8b43cabed07336d074a3cb
institution DOAJ
issn 2045-2322
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-15949522db8b43cabed07336d074a3cb2025-08-20T03:03:42ZengNature PortfolioScientific Reports2045-23222025-07-0115111210.1038/s41598-025-06432-9Deep generative model for the inverse design of Van der Waals heterostructuresShikun Gao0Qinyuan Huang1Chen Huang2Cheng Li3Kaihao Liu4Baisheng Sa5Yadong Yu6Dezhen Xue7Zhe Liu8Mengyan Dai9School of Automation and Information Engineering, Sichuan University of Science and EngineeringSchool of Automation and Information Engineering, Sichuan University of Science and EngineeringChemical Defense Institute, Academy of Military SciencesState Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong UniversityChemical Defense Institute, Academy of Military SciencesMultiscale Computational Materials Facility & Materials Genome Institute, School of Materials Science and EngineeringChemical Defense Institute, Academy of Military SciencesState Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong UniversityChemical Defense Institute, Academy of Military SciencesChemical Defense Institute, Academy of Military SciencesAbstract This study proposes ConditionCDVAE+, a crystal diffusion variational autoencoder (CDVAE) based deep generative model for inverse design of van der Waals (vdW) heterostructures. To address the challenges of traditional experimental methods relying on trial-and-error and existing models struggling to incorporate target property constraints, this work achieves breakthroughs through three innovative stages: (1) introduce the SE(3)-equivariant graph neural network EquiformerV2 as the encoder-decoder within the CDVAE framework to enhance the generation quality of the model; (2) design a module integrating Low-rank Multimodal Fusion and Generative Adversarial Networks to map properties and structures into a joint latent space; and (3) for the first time propose a generative model for the vdW heterostructures, by conducting experimental validation on the dataset constructed from Janus III–VI vdW heterostructures. Experiments demonstrate that ConditionCDVAE+ achieves optimal root mean square error for crystal reconstruction, with improved generation quality. Density Functional Theory calculations confirms 99.51% of generated samples converge to energy minima, indicating superior ground-state convergence. The effectiveness of the model under conditional guidance has also been extensively validated. This framework provides an efficient solution for target-oriented design of vdW heterostructures and holds promise for accelerating the development of novel optoelectronic devices.https://doi.org/10.1038/s41598-025-06432-9
spellingShingle Shikun Gao
Qinyuan Huang
Chen Huang
Cheng Li
Kaihao Liu
Baisheng Sa
Yadong Yu
Dezhen Xue
Zhe Liu
Mengyan Dai
Deep generative model for the inverse design of Van der Waals heterostructures
Scientific Reports
title Deep generative model for the inverse design of Van der Waals heterostructures
title_full Deep generative model for the inverse design of Van der Waals heterostructures
title_fullStr Deep generative model for the inverse design of Van der Waals heterostructures
title_full_unstemmed Deep generative model for the inverse design of Van der Waals heterostructures
title_short Deep generative model for the inverse design of Van der Waals heterostructures
title_sort deep generative model for the inverse design of van der waals heterostructures
url https://doi.org/10.1038/s41598-025-06432-9
work_keys_str_mv AT shikungao deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT qinyuanhuang deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT chenhuang deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT chengli deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT kaihaoliu deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT baishengsa deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT yadongyu deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT dezhenxue deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT zheliu deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures
AT mengyandai deepgenerativemodelfortheinversedesignofvanderwaalsheterostructures