From tailpipe to the environment: how vehicles’ emissions drive secondary aerosol formation

• Study on secondary aerosol potential precursors from the exhaust of gasoline and diesel-fueled Euro 6D vehicles; • Role of aromatic hydrocarbons, NOx and particle number concentration on Secondary Organic Aerosol (SOA) formation; • Role of NOx and NH 3 on Secondary Inorganic Aerosol (SIA) formation; • Gaussian process regression as a cutting-edge method for data-driven prediction of SOA/SIA formation and mitigation design. Urban air pollution, which includes gaseous pollutants and particulate matter, is primarily caused by traffic emissions. This study explores the potential precursors of atmospheric aerosols emitted directly into the atmosphere from the exhaust of Euro 6D gasoline and diesel vehicles. Emissions, including NO x , aromatic hydrocarbons, ammonia and particulate matter from various gasoline and diesel fuels were analyzed at the exhaust vehicles during a test driving cycle. After dilution, primary emissions were fed into an oxidation flow reactor to simulate atmospheric oxidation conditions. The formed secondary aerosol was analyzed using a Scanning Mobility Particle Sizer and a High-Resolution Time-of-Flight Aerosol Mass Spectrometer. Gasoline fuels show a greater tendency to form secondary aerosols compared to diesel fuels, with observed organic mass concentrations typically 10–15 times higher than those measured for diesel. During the initial phases of the tests, i.e., cold start conditions, the organic fraction is the dominant component, while the inorganic fraction becomes more significant in the later stages. Ammonium nitrate is identified as the primary component of the inorganic fraction of the secondary aerosol, alongside organonitrates, peroxynitrates, and peroxyacyl nitrates. It was found that secondary inorganic aerosol formation from diesel vehicle exhaust is negligible. Correlations between the concentrations of secondary aerosols and their primary precursors have been found, and analytical expressions have been proposed. Furthermore, a data-driven Bayesian approach, namely Gaussian Process Regression, was employed and provided accurate predictions, achieving a robust coefficient of determination exceeding 0.95 when correlating measured secondary aerosol concentrations with primary precursors. The findings of this work offer a solid framework for characterizing and predicting secondary aerosol formation in the atmosphere from road transportation.