Characteristics of black carbon emissions from in-use light-duty passenger vehicles

Abstract

Mitigating black carbon (BC) emissions from various combustion sources has been considered an urgent policy issue to address the challenges of climate change, air pollution and health risks. Vehicles contribute considerably to total anthropogenic BC emissions and urban BC concentrations. Compared with heavy-duty diesel vehicles, there is much larger uncertainty in BC emission factors for light-duty passenger vehicles (LDPVs), in particular for gasoline LDPVs, which warrants further studies. In this study, we employed the dynamometer and the Aethalometer (AE-51) to measure second-by-second BC emissions from eight LDPVs by engine technology and driving cycle. The average BC emission factors under transient cycles (e.g., ECE-15, New European Driving Cycle, NEDC, Worldwide Harmonized Light Vehicles Test Cycle, WLTC) are 3.6–91.5 mg/km, 7.6 mg/km and 0.13–0.58 mg/km, respectively, for diesel (N = 3), gasoline direct injection (GDI) (N = 1) and gasoline port-fuel injection (PFI) engine categories (N = 4). For gasoline PFI LDPVs, the instantaneous emission profiles show a strong association of peak BC emissions with cold-start and high-speed aggressive driving. Such impacts lead to considerable BC emission contributions in cold-start periods (e.g., the first 47 s–94 s) over the entire cycle (e.g., 18–76% of the NEDC and 13–36% of the WLTC) and increased BC emission factors by 80–440% under the WLTC compared to the NEDC. For diesel BC emissions, the size distribution exhibits a typical unimodal pattern with one single peak appearing approximately from 120 to 150 nm, which is largely consistent with previous studies. Nevertheless, the average mass ratios of BC to particle mass (PM) range from 0.38 to 0.54 for three diesel samples, representing substantial impacts from both driving and engine conditions. The significant discrepancy between gasoline BC emission factors obtained from tailpipe exhaust versus ambient conditions suggest that more comparative measurements and fine-grained simulations should be designed and implemented to address this discrepancy.