Estimating primary vehicular emission contributions to PM2.5 using the Chemical Mass Balance model: Accounting for gas-particle partitioning of organic aerosols and oxidation degradation of hopanes

Highlights: • CMB model is modified to account for G-P partitioning of OA and hopane oxidation. • Modified CMB corrects bias in apportioning PM2.5 by diesel and gasoline vehicles. • The gasoline-diesel split by the new CMB notably differs from that by ordinary CMB. • Adopting static vehicular source profiles in CMB produces flawed PMvehicle estimates. Particulate matter emitted from vehicles (PM_vehicle) represents a major air pollution source in urban areas. Ambient measurements of hopanes and elemental carbon have traditionally been coupled with the Chemical Mass Balance (CMB) model to quantify the contributions to fine PM from diesel and gasoline vehicular emissions (VE). The organic carbon part of PM_vehicle, however, undergoes gas-particle partitioning and oxidation degradation as VE move from exhaust pipe to receptor sites. This creates an issue of deviation from mass conservation in the utility of CMB. The impact of this issue on quantifying PM_vehicle has remained largely uncharacterized. In this study, we incorporate in CMB the gas-particle partitioning of VE organic aerosols and hopane oxidation, which is equivalent to adopting dynamic VE source profiles. The modified version of CMB is applied to quantify primary PM_vehicle contributions at a roadside and a general urban site in Hong Kong. For the roadside site, the modified CMB reports predominant PM_vehicle by diesel VE, a result consistent with previous studies. For the general urban site, the apportioned gasoline contribution by the modified CMB is tripled (0.8 ± 0.5 vs. 2.7 ± 2.1 μg/m³) while the diesel contribution is reduced by one third (1.7 ± 1.2 vs. 1.1 ± 1.2 μg/m³), producing a gasoline-diesel split significantly different from that by traditional CMB (1:2 vs. 5:2). Our work strongly indicates that a static representation of VE source profiles in CMB modeling would create flawed PM_vehicle estimation and demonstrates the necessity of considering gas-particle partitioning of organic aerosol and hopane oxidation degradation.