Discovering new photovoltaics using optimal transport theory
- 1Department of Chemistry, University College London London WC1H 0AJ UK
- 2Department of Engineering, University of Cambridge UK
Abstract
Searching by chemical and structural analogy is one of the most commonly used and successful approaches to materials discovery. However, formulating this task for algorithmic implementation raises the question of how we define similar materials. Methods have been proposed for searching materials space using vectors based on chemical composition and functional fragments in the material. Descriptors for structural similarity have also been proposed. However, the question of how to incorporate and balance structural and compositional similarity measures in a single metric remains open. Here, we adapt methods developed for calculating distances between undirected graphs and apply them to crystalline materials similarity. The Fused Gromov–Wasserstein (FGW) metric uses optimal transport theory to map between two graphs considering a balance of the graph structure and the information present at the nodes of the graph (atoms in crystals). We apply the method to explore new photovoltaic materials. We demonstrate that FGW is competitive with embeddings from an equivariant graph neural network, trained on >106 materials, despite minimal training. We then apply FGW to a discovery campaign to identify materials from the Materials Project database that have not previously been explored as photovoltaics, but have similarities to known high-efficiency materials. After validating predictions with hybrid density functional theory, we identify seven previously unexplored high-efficiency photovoltaic absorber candidates, including Cs5Sb8, which is found to have a predicted SLME of >30% and to be thermodynamically stable. The FGW approach demonstrates the power of strong inductive bias for developing metrics for materials exploration with minimal training data.
Related
- Molecular Diffusion-Driven Graded Polymer Hole-Transport Layers Enabling Efficient and Durable Organic Photovoltaics
- Bis-aryl substituted dioxaborines as electron-transport materials: a comparative density functional theory investigation with oxadiazoles and siloles
- The Kinetics of Polymer Adsorption in Transport Through Porous Media: Basic Theory
- Novel Lead-free double perovskite Cs2LiFeBr6 as photovoltaics: a DFT study of structural, mechanical, spin-resolved electronic and optical properties
- Discovering and designing multistate protein assemblies
BibTeX
@article{walker2026discovering,
title = {{Discovering new photovoltaics using optimal transport theory}},
author = {Matthew A. H. Walker and Zibo Zhou and Junayd Ul Islam and Keith T. Butler},
journal = {Faraday Discussions},
year = {2026},
month = jul,
doi = {10.1039/d6fd00016a},
url = {https://doi.org/10.1039/d6fd00016a},
note = {CC BY 3.0},
}
RIS
TY - JOUR
TI - Discovering new photovoltaics using optimal transport theory
AU - Matthew A. H. Walker
AU - Zibo Zhou
AU - Junayd Ul Islam
AU - Keith T. Butler
PY - 2026
DA - 2026/07/21
JO - Faraday Discussions
DO - 10.1039/d6fd00016a
UR - https://doi.org/10.1039/d6fd00016a
AB - Searching by chemical and structural analogy is one of the most commonly used and successful approaches to materials discovery. However, formulating this task for algorithmic implementation raises the question of how we define similar materials. Methods have been proposed for searching materials space using vectors based on chemical composition and functional fragments in the material. Descriptors for structural similarity have also been proposed. However, the question of how to incorporate and balance structural and compositional similarity measures in a single metric remains open. Here, we adapt methods developed for calculating distances between undirected graphs and apply them to crystalline materials similarity. The Fused Gromov–Wasserstein (FGW) metric uses optimal transport theory to map between two graphs considering a balance of the graph structure and the information present at the nodes of the graph (atoms in crystals). We apply the method to explore new photovoltaic materials. We demonstrate that FGW is competitive with embeddings from an equivariant graph neural network, trained on >106 materials, despite minimal training. We then apply FGW to a discovery campaign to identify materials from the Materials Project database that have not previously been explored as photovoltaics, but have similarities to known high-efficiency materials. After validating predictions with hybrid density functional theory, we identify seven previously unexplored high-efficiency photovoltaic absorber candidates, including Cs5Sb8, which is found to have a predicted SLME of >30% and to be thermodynamically stable. The FGW approach demonstrates the power of strong inductive bias for developing metrics for materials exploration with minimal training data.
SN - 1359-6640
SN - 1364-5498
ER -