Highlights • In-situ estimation of unregulated compounds emission factors. • Characterization of the fleet composition and the driving conditions. • Statistical analysis to study impacts of fleet composition on pollutant emissions. • Comparison between emission factors estimated in this work and in other studies.

Abstract The purpose of this study is to estimate the in-situ emission factors of several pollutants (particle number [PN], black carbon [BC] and several volatile and semi-volatile organic compounds [VOCs and SVOCs]) in an urban area of Nantes, France, with real-world traffic conditions and characterization of the fleet composition. The fleet composition and driving conditions are characterized by the number of vehicles, their speeds and their types (passenger cars [PCs], light commercial vehicles [LCVs], heavy-duty vehicles [HDVs]) as well as their characteristics (make, model, fuel, engine, EURO emission standard, etc.). The number of vehicles passing on the boulevard is around 20,000 per day with about 44% of Euro 5 and Euro 6 vehicles. The impacts of fleet composition on emission were analyzed by ANOVA. The results show that the fleet composition has a significant impact on emissions for different pollutants. Higher percentage of gasoline PCs between Euro 4 to Euro 6 and Euro 4 diesel PCs induces more BC emission. Higher percentage of old gasoline and diesel vehicles (≤ Euro 3) induces higher emission of toluene, ethylbenzene and m + p- and o-xylene. Furthermore, emission factors estimated in this work were compared to those calculated in other in-situ studies that show a good agreement. For the chassis bench comparison, the in-situ PN and BC emission factors are in the same range as those measured for diesel vehicles without particle filter and gasoline vehicles with direct injection system. These EFs are also comparable with old heavy duty vehicles without particle filter (5 × 10¹³–2 × 10¹⁴ #/km).