Transferable Machine Learning of Electronic Hamiltonians with Superposition-of-Atomic-Potentials Features
Journal of Chemical Theory and Computation · 11.11.2026
DOI: 10.1021/acs.jctc.6c01477 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Chemical Theory and Computation, online erschienen am 11.11.2026. Verfasst von Chaoqun Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 76 Literaturangaben. Gefördert von: Division of Chemistry, National Defense Science and Engineering Graduate.
Abstract (Originalsprache)
Machine learning of electronic Hamiltonians offers a unified route to electronic wave functions and physical observables. We introduce a Hamiltonian learning framework built on electronic features derived from the superposition-of-atomic-potentials (SAP) approximation, an efficient self-consistent-field initial guess that captures essential electron–electron screening. SAP quantities define a symmetry-adapted intrinsic atomic orbital learning basis and provide physics-informed inputs to an orbital-based graph neural network that predicts converged Kohn–Sham Fock matrices. To extend the approach to larger basis sets, we further develop a downfolding scheme that predicts large-basis electronic structure from minimal-basis features. On the QM9 data set, the model accurately reproduces frontier and core orbital energies, dipole moments, and the full density of states. For organic charge-transport materials, it yields accurate intermolecular transfer integrals for benzene, tetracyanoquinodimethane (TCNQ), and tetrathiafulvalene (TTF) dimers. The model also generalizes to unseen substituted-benzene heterodimers, with a mean absolute error of 4.8 meV. These results establish SAP-based Hamiltonian learning as a transferable and scalable tool for high-throughput electronic-structure prediction.
Förderung
- Division of Chemistry
- National Defense Science and Engineering Graduate
Zitieren
APA
Zhang, C., Venturella, C., Chen, E., & Zhu, T. (2026). Transferable Machine Learning of Electronic Hamiltonians with Superposition-of-Atomic-Potentials Features. Journal of Chemical Theory and Computation. https://doi.org/10.1021/acs.jctc.6c01477
Vancouver
Zhang C, Venturella C, Chen E, Zhu T. Transferable Machine Learning of Electronic Hamiltonians with Superposition-of-Atomic-Potentials Features. Journal of Chemical Theory and Computation. 2026. doi:10.1021/acs.jctc.6c01477
BibTeX
@article{Zhang2026,
author = {Zhang, Chaoqun and Venturella, Christian and Chen, Enzhi and Zhu, Tianyu},
title = {Transferable Machine Learning of Electronic Hamiltonians with Superposition-of-Atomic-Potentials Features},
journal = {Journal of Chemical Theory and Computation},
year = {2026},
doi = {10.1021/acs.jctc.6c01477}
}
EUVIMED