Sulfuric Acid-Driven\nNucleation Enhanced by Amines\nfrom Ethanol Gasoline Vehicle Emission: Machine Learning Model and\nMechanistic Study

The sulfuric acid (SA)-amine nucleation mechanism gained\nincreasing\nattention due to its important role in atmospheric secondary particle\nformation. However, the intrinsic enhancing potential (IEP) of various\namines remains largely unknown, restraining the assessment on the\nrole of the SA-amines mechanism at various locations. Herein, machine\nlearning (ML) models were constructed for high-throughput prediction\nof IEP of amines, and the nucleation mechanism of specific amines\nwith high IEP was investigated. The formation free energy (Δ<i>G</i>) of SA-amines dimer clusters, a key parameter for assessing\nIEP, was calculated for 58 amines. Based on the calculated Δ<i>G</i> values, seven ML models were constructed and the best\none was further utilized to predict the Δ<i>G</i> values\nof the remaining 153 amines. Diethylamine (DEA), mainly emitted from\nethanol gasoline vehicles, was found to be one of the amines with\nthe highest IEP for SA-driven nucleation. By studying larger SA-DEA\nclusters, it was found that the nucleation rate of DEA with SA is\n3–7 times higher than that of dimethylamine, a well-known key\nbase for SA-driven nucleation. The study provides a powerful tool\nfor evaluating the actual role of amines on SA-driven nucleation and\nrevealed that the mechanism could be particularly important in areas\nwhere ethanol gasoline vehicles are widely used.