Fullerenes, carbon-based nanomaterials with sp2-hybridized carbon atoms arranged in polyhedral cages, exhibit diverse isomeric structures with promising applications in optoelectronics, solar cells, and medicine. However, the vast number of possible fullerene isomers complicates efficient property prediction. In this study, we introduce FullereneNet, a graph neural network based model that predicts fundamental properties of fullerenes using topological features derived solely from unoptimized structures, eliminating the need for computationally expensive quantum chemistry optimizations. The model leverages topological representations based on the chemical environments of pentagon and hexagon rings, enabling efficient capture of local structural details. We show that this approach yields superior performance in predicting the C–C binding energy for a wide range of fullerene sizes, achieving mean absolute errors of 3 meV/atom for C60, 4 meV/atom for C70, and 6 meV/atom for C72–C100, surpassing state-of-the-art machine learning interatomic potentials GAP-20. Additionally, the FullereneNet model accurately predicts 11 other properties, including HOMO-LUMO gap and solvation free energy, demonstrating robustness and transferability across fullerene types. This work provides a computationally efficient framework for high-throughput screening of fullerene candidates and establishes a foundation for future data-driven studies in fullerene chemistry.
Fullerenes, as characteristic carbon nanomaterials, offer significant potential for diverse applications due to their structural diversity and tunable properties. Numerous isomers can exist for a specific fullerene size, yet a comprehensive understanding of their fundamental properties remains elusive. In this study, we construct an up-to-date computational database for C20-C60 fullerenes, consisting of 5770 structures, and calculate 12 fundamental properties using DFT, including stability (binding energy), electronic properties (HOMO-LUMO gap), and solubility (partition coefficient logP). Our findings reveal that the HOMO-LUMO gap weakly correlates with both binding energy and logP, indicating that electronic properties can be tailored for specific uses without affecting stability or solubility. In addition, we introduce a set of novel topological features and geometric measures to investigate structure-property relationships. For the first time, we apply atom, bond, and hexagon features to effectively predict the stability of C20-C60 fullerenes, surpassing the conventional qualitative isolated pentagon rule, and demonstrating their robust transferability to larger-size fullerenes beyond C60. Our work offers guidance for optimizing fullerenes as electron acceptors in organic solar cells and lays a foundational understanding of their functionalization and applications in energy conversion and nanomaterial sciences.
Perovskite materials, renowned for their versatility and remarkable properties, pose challenges in discovering optimal candidates due to the vast compositional space. Data-driven machine learning (ML) offers promise in expediting material discovery; however, the trade-off between accuracy and efficiency across different ML models for predicting perovskite properties is not well understood. In this study, we conducted a comprehensive assessment of various ML models for predicting the formation energy (Ef) and bandgap (Eg) of perovskites. We designed a protocol to extract perovskite structures from three databases based on the stoichiometry, octahedral lattice motif, and alignment with established perovskite prototype structures. Benchmarking conventional ML algorithms (CML) against graph neural network (GNN) models across three datasets, we identified the GATGNN model as the top performer, balancing exceptional prediction accuracy and computational efficiency. We further investigated the impact of data size on model performance, emphasizing the need for over 1000 data points for optimal prediction accuracy. Additionally, through SHAP analysis, we provided valuable insights into the interpretation of CML models in predicting Ef and Eg. Our study establishes a standardized benchmark for evaluating various ML models across diverse datasets of perovskite materials, facilitating future applications in materials science, particularly in model selection and the advancement of perovskite materials.