Machine learning interatomic potentials with accurate long-range interactions for molecular dynamics collision simulations of atmospherically-relevant molecules

Abstract

Molecular collisions and subsequent clustering events are fundamental to atmospheric cluster formation. Accurately modeling these processes requires interatomic potentials that simultaneously capture the long-range forces governing collision kinetics and the short-range quantum effects driving reactivity. In this work, we evaluate the AIMNet2 and PaiNN machine learning architectures trained on GFN1-xTB and ωB97X-3c quantum chemical data for molecular collisions involving sulfuric acid. The models exhibit low mean absolute errors in energies and forces and accurately reproduce potentials of mean force relative to the GFN1-xTB reference. However, discrepancies are observed for the collision dynamics. While AIMNet2 accurately reproduces reference collision rate coefficients across all systems, PaiNN underestimates the rate coefficient for the charged sulfuric acid–bisulfate system by ∼ 50 %. This error originates from the model's local atomic environment approximation, which neglects the strong long-range attractive forces at large intermolecular distances. Simulations with the OPLS-AA classical force field demonstrate that simple fixed partial charges are sufficient to describe these interactions. Comparing models trained on GFN1-xTB and ωB97X-3c data reveals that while increasing the level of electronic structure theory significantly alters the potential energy surface in the short-range binding region, it generally has less impact on the long-range shoulder and the resulting collision rate coefficients. Our results highlight that while local equivariant models like PaiNN offer exceptional accuracy for thermodynamics, correctly simulating collision kinetics in systems with strong long-range interactions requires models that explicitly account for forces beyond the local environment, such as AIMNet2.