No Cover Image

Journal article 161 views 20 downloads

High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning

Xiangwen Wang, Muyang Chen, Gengyi Bao, Yan Lai, Jinghe Cao, Xinhua Liu, Rui Tan Orcid Logo

Advanced Science, Start page: e21575

Swansea University Authors: Muyang Chen, Rui Tan Orcid Logo

  • 71157.VOR.pdf

    PDF | Version of Record

    © 2026 The Author(s). This is an open access article under the terms of the Creative Commons Attribution License.

    Download (3.15MB)

Check full text

DOI (Published version): 10.1002/advs.202521575

Abstract

Rising energy generation from renewables (e.g., wind, solar power) will drive global demand for >1.0 TWh of long-duration energy storage by 2030 to stabilise grids and balance supply. Rechargeable batteries are central to this transition, with their performance critically governed by the properti...

Full description

Published in: Advanced Science
ISSN: 2198-3844 2198-3844
Published: Wiley 2026
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa71157
first_indexed 2025-12-18T11:14:54Z
last_indexed 2026-01-23T06:52:01Z
id cronfa71157
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2026-01-22T14:37:16.9813421</datestamp><bib-version>v2</bib-version><id>71157</id><entry>2025-12-18</entry><title>High&#x2010;Conductivity Electrolytes Screened Using Fragment&#x2010; and Composition&#x2010;Aware Deep Learning</title><swanseaauthors><author><sid>7c009156a96911675293ed37c70709eb</sid><firstname>Muyang</firstname><surname>Chen</surname><name>Muyang Chen</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>774c33a0a76a9152ca86a156b5ae26ff</sid><ORCID>0009-0001-9278-7327</ORCID><firstname>Rui</firstname><surname>Tan</surname><name>Rui Tan</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2025-12-18</date><deptcode>MACS</deptcode><abstract>Rising energy generation from renewables (e.g., wind, solar power) will drive global demand for &gt;1.0 TWh of long-duration energy storage by 2030 to stabilise grids and balance supply. Rechargeable batteries are central to this transition, with their performance critically governed by the properties of active materials and supporting electrolytes. However, designing electrolyte formulations remains a major challenge, as their performance arises from complex, non-additive interactions among lithium salts and organic solvents, requiring elegant molecular design and selection. Conventional trial-and-error strategies still dominate electrolyte design, but they are slow and resource-intensive. Recent machine learning approaches have improved electrolyte screening, yet many rely on coarse molecular representations that neglect fragment-level chemistry and explicit ratios, limiting interpretability and their utility for guiding experiments. Here we introduce a deep learning framework that integrates intermolecular attributions across solvents with intramolecular attributions from functional units. The framework builds a hierarchical representation, decomposing formulations into molecules and their functional units, while integrating ratios, physicochemical descriptors, and salt identity to generate mixture-invariant embeddings for accurate and interpretable conductivity prediction. Applied to benchmark datasets of lithium battery electrolytes, the framework achieves high accuracy in predicting ionic conductivity and enables large-scale virtual screening. Crucially, it provides chemically interpretable insights: fragment-level attentions align with functional units; composition-aware attention reveals the impact of mixing ratios; and counterfactual perturbations confirm causal roles of key motifs. This framework paves the way for data-driven, interpretable electrolyte design and can be generalized to broader formulation challenges in materials science.</abstract><type>Journal Article</type><journal>Advanced Science</journal><volume>0</volume><journalNumber/><paginationStart>e21575</paginationStart><paginationEnd/><publisher>Wiley</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>2198-3844</issnPrint><issnElectronic>2198-3844</issnElectronic><keywords>battery electrolyte, data-driven design, graph neural networks, ionic conductivity, machine learning</keywords><publishedDay>4</publishedDay><publishedMonth>1</publishedMonth><publishedYear>2026</publishedYear><publishedDate>2026-01-04</publishedDate><doi>10.1002/advs.202521575</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm>SU Library paid the OA fee (TA Institutional Deal)</apcterm><funders>Royal Society Research Grant (RGS\R2\252134), EPSRC Royce Industrial Collaboration Grant (RICP-R4-100029; ICP5334) and RSC Collaboration Grant (C25-1820146588).</funders><projectreference/><lastEdited>2026-01-22T14:37:16.9813421</lastEdited><Created>2025-12-18T11:10:49.4092782</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Engineering and Applied Sciences - Chemical Engineering</level></path><authors><author><firstname>Xiangwen</firstname><surname>Wang</surname><order>1</order></author><author><firstname>Muyang</firstname><surname>Chen</surname><order>2</order></author><author><firstname>Gengyi</firstname><surname>Bao</surname><order>3</order></author><author><firstname>Yan</firstname><surname>Lai</surname><order>4</order></author><author><firstname>Jinghe</firstname><surname>Cao</surname><order>5</order></author><author><firstname>Xinhua</firstname><surname>Liu</surname><order>6</order></author><author><firstname>Rui</firstname><surname>Tan</surname><orcid>0009-0001-9278-7327</orcid><order>7</order></author></authors><documents><document><filename>71157__36011__12f2ad1f5bce45c38c25906c3e31899e.pdf</filename><originalFilename>71157.VOR.pdf</originalFilename><uploaded>2026-01-15T16:33:11.9004918</uploaded><type>Output</type><contentLength>3305120</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>&#xA9; 2026 The Author(s). This is an open access article under the terms of the Creative Commons Attribution License.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>http://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling 2026-01-22T14:37:16.9813421 v2 71157 2025-12-18 High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning 7c009156a96911675293ed37c70709eb Muyang Chen Muyang Chen true false 774c33a0a76a9152ca86a156b5ae26ff 0009-0001-9278-7327 Rui Tan Rui Tan true false 2025-12-18 MACS Rising energy generation from renewables (e.g., wind, solar power) will drive global demand for >1.0 TWh of long-duration energy storage by 2030 to stabilise grids and balance supply. Rechargeable batteries are central to this transition, with their performance critically governed by the properties of active materials and supporting electrolytes. However, designing electrolyte formulations remains a major challenge, as their performance arises from complex, non-additive interactions among lithium salts and organic solvents, requiring elegant molecular design and selection. Conventional trial-and-error strategies still dominate electrolyte design, but they are slow and resource-intensive. Recent machine learning approaches have improved electrolyte screening, yet many rely on coarse molecular representations that neglect fragment-level chemistry and explicit ratios, limiting interpretability and their utility for guiding experiments. Here we introduce a deep learning framework that integrates intermolecular attributions across solvents with intramolecular attributions from functional units. The framework builds a hierarchical representation, decomposing formulations into molecules and their functional units, while integrating ratios, physicochemical descriptors, and salt identity to generate mixture-invariant embeddings for accurate and interpretable conductivity prediction. Applied to benchmark datasets of lithium battery electrolytes, the framework achieves high accuracy in predicting ionic conductivity and enables large-scale virtual screening. Crucially, it provides chemically interpretable insights: fragment-level attentions align with functional units; composition-aware attention reveals the impact of mixing ratios; and counterfactual perturbations confirm causal roles of key motifs. This framework paves the way for data-driven, interpretable electrolyte design and can be generalized to broader formulation challenges in materials science. Journal Article Advanced Science 0 e21575 Wiley 2198-3844 2198-3844 battery electrolyte, data-driven design, graph neural networks, ionic conductivity, machine learning 4 1 2026 2026-01-04 10.1002/advs.202521575 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University SU Library paid the OA fee (TA Institutional Deal) Royal Society Research Grant (RGS\R2\252134), EPSRC Royce Industrial Collaboration Grant (RICP-R4-100029; ICP5334) and RSC Collaboration Grant (C25-1820146588). 2026-01-22T14:37:16.9813421 2025-12-18T11:10:49.4092782 Faculty of Science and Engineering School of Engineering and Applied Sciences - Chemical Engineering Xiangwen Wang 1 Muyang Chen 2 Gengyi Bao 3 Yan Lai 4 Jinghe Cao 5 Xinhua Liu 6 Rui Tan 0009-0001-9278-7327 7 71157__36011__12f2ad1f5bce45c38c25906c3e31899e.pdf 71157.VOR.pdf 2026-01-15T16:33:11.9004918 Output 3305120 application/pdf Version of Record true © 2026 The Author(s). This is an open access article under the terms of the Creative Commons Attribution License. true eng http://creativecommons.org/licenses/by/4.0/
title High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
spellingShingle High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
Muyang Chen
Rui Tan
title_short High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
title_full High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
title_fullStr High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
title_full_unstemmed High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
title_sort High‐Conductivity Electrolytes Screened Using Fragment‐ and Composition‐Aware Deep Learning
author_id_str_mv 7c009156a96911675293ed37c70709eb
774c33a0a76a9152ca86a156b5ae26ff
author_id_fullname_str_mv 7c009156a96911675293ed37c70709eb_***_Muyang Chen
774c33a0a76a9152ca86a156b5ae26ff_***_Rui Tan
author Muyang Chen
Rui Tan
author2 Xiangwen Wang
Muyang Chen
Gengyi Bao
Yan Lai
Jinghe Cao
Xinhua Liu
Rui Tan
format Journal article
container_title Advanced Science
container_volume 0
container_start_page e21575
publishDate 2026
institution Swansea University
issn 2198-3844
2198-3844
doi_str_mv 10.1002/advs.202521575
publisher Wiley
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Engineering and Applied Sciences - Chemical Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Engineering and Applied Sciences - Chemical Engineering
document_store_str 1
active_str 0
description Rising energy generation from renewables (e.g., wind, solar power) will drive global demand for >1.0 TWh of long-duration energy storage by 2030 to stabilise grids and balance supply. Rechargeable batteries are central to this transition, with their performance critically governed by the properties of active materials and supporting electrolytes. However, designing electrolyte formulations remains a major challenge, as their performance arises from complex, non-additive interactions among lithium salts and organic solvents, requiring elegant molecular design and selection. Conventional trial-and-error strategies still dominate electrolyte design, but they are slow and resource-intensive. Recent machine learning approaches have improved electrolyte screening, yet many rely on coarse molecular representations that neglect fragment-level chemistry and explicit ratios, limiting interpretability and their utility for guiding experiments. Here we introduce a deep learning framework that integrates intermolecular attributions across solvents with intramolecular attributions from functional units. The framework builds a hierarchical representation, decomposing formulations into molecules and their functional units, while integrating ratios, physicochemical descriptors, and salt identity to generate mixture-invariant embeddings for accurate and interpretable conductivity prediction. Applied to benchmark datasets of lithium battery electrolytes, the framework achieves high accuracy in predicting ionic conductivity and enables large-scale virtual screening. Crucially, it provides chemically interpretable insights: fragment-level attentions align with functional units; composition-aware attention reveals the impact of mixing ratios; and counterfactual perturbations confirm causal roles of key motifs. This framework paves the way for data-driven, interpretable electrolyte design and can be generalized to broader formulation challenges in materials science.
published_date 2026-01-04T07:00:11Z
_version_ 1861427933078355968
score 11.1007185

Similar Items