Polarizable atomic multipoles for learning long-range electrostatics

2026-05-07

AI Summary

Current machine learning interatomic potentials struggle to accurately capture long-range electrostatic and polarization effects in ionic, polar, and interfacial systems. This work proposes a semi-local polarizable multipole framework that predicts environment-dependent monopole, dipole, and quadrupole moments using local equivariant descriptors, while incorporating non-self-consistent linear response theory to account for non-local charge transfer and polarization. The approach establishes a physically transparent and systematically improvable multipole hierarchy, requiring only standard energy and force training data to accurately predict polarization-sensitive observables. Evaluated across four diverse benchmarks, the model substantially improves potential energy surface accuracy, successfully reproduces experimental infrared spectra, achieves semi-quantitative agreement with Raman spectra, and yields precise Born effective charge tensors and emergent polarizabilities.

Abstract

Long-range electrostatics and polarization remain central obstacles to extending machine learning interatomic potentials (MLIPs) to ionic, polar, and interfacial systems. Here, we introduce a semi-local framework for learning electrostatics from energies and forces using polarizable atomic multipoles. Local equivariant descriptors predict environment-dependent latent monopoles, dipoles, and quadrupoles, while residual non-local charge transfer and polarization are captured by non-self-consistent linear response in induced charges and dipoles. Across four diverse benchmarks and four short-range MLIP architectures, the multipole hierarchy and response terms systematically improve potential energy surface accuracy, with the largest gains in systems where long-range effects are essential. More importantly, the learned latent variables recover physically meaningful electrical responses: accurate Born effective charge tensors, emergent polarizabilities, infrared spectra in close agreement with experiments, and semi-quantitative Raman spectra for bulk water and hybrid MAPbI$_3$ perovskite. This systematically improvable, physically transparent framework enables MLIPs trained on standard energy and force labels to predict polarization-sensitive observables.

Problem

• long-range electrostatics
• polarization
• machine learning interatomic potentials
• ionic systems
• polar systems

Innovation

• polarizable atomic multipoles
• long-range electrostatics
• machine learning interatomic potentials
• linear response
• equivariant descriptors