A series of cold start experiments using a 2.0 liter gasoline turbocharged direct injection (GTDI) engine with custom controls and calibration were carried out using gasoline and iso-pentane fuels, to obtain the cold start emissions profiles for the first 5 firing cycles at an ambient temperature of 22°C. The exhaust gases, both emitted during the cold start firing and emitted during the cranking process right after the firing, were captured, and unburned hydrocarbon emissions (HC), CO, and CO2 on a cycle-by-cycle basis during an engine cold start were analyzed and quantified. The HCs emitted during gasoline-fueled cold starts was found to reduce significantly as the engine cycle increased, while CO and CO2 emissions were found to stay consistent for each cycle. Crankcase ventilation into the intake manifold through the positive-crankcase ventilation (PCV) valve system was found to have little effect on the emissions results. Cold start experiments fueled by highly volatile iso-pentane saw an overwhelming majority of the injected carbon captured in the exhaust gases, while a significant portion of the injected carbon during the gasoline-fueled cold starts was not captured. The comparative results not only validated the experimental methods, but also demonstrated that a significant fraction of the injected gasoline failed to evaporate during cold starts. During the first 5 firing cycles, 22% to 34% of the injected fuel mass was estimated to remain in the liquid phase and escaped capture. Because fuel could be carried over from one cycle to the next, in some cases, the actual unevaporated gasoline portion in a given cold start cycle could be even higher than that measured.
