Time-resolved sampling experiments were conducted on a single-cylinder direct-injection two-stroke engine, using an air-assisted and a pressure-swirl injector. The engine was operated in a stratified mode, at an engine speed of 2000 rpm with an overall A/F ratio of 30:1. Three injection timings were investigated for each injector. A fast-response solenoid valve was used to sample the engine exhaust at a position 1 cm downstream of the engine exhaust port. Measurements of the HC, NOx, CO2, CO and O2 concentrations were collected for 15º windows over the time that the exhaust port was open. Conventional concentration measurements were also performed downstream of an exhaust mixing tank, where the engine pollutants were assumed to be well-mixed.

The instantaneous mass flow through the engine exhaust port was calculated from the measured cylinder and exhaust pressures. This information was coupled with the time-resolved concentrations to determine the instantaneous mass flow rate of each species over the entire engine cycle. Good agreement was found between this instantaneous pollutant mass flow when integrated over the entire cycle, and the conventional measurement of each species flowrate downstream of the exhaust mixing tank.

The time-resolved concentrations of the combustion products (NOx, CO2, CO) were found to decrease throughout the scavenging event. It was believed that this was due to dilution of these bulk combustion gases by fresh intake air. This was confirmed by the increasing trend observed in the measured O2 concentration. The calculated dilution rate was seen to differ between the products of combustion. It was proposed that this was due to continuing chemical activity within the cylinder for differing species.

The hydrocarbon concentration was found to be the lowest at exhaust port opening, then increase during scavenging. Dilution, while present, was not sufficient to explain this effect. It is suggested that lean quench of the bulk combustion gases could account for the concentration of hydrocarbons at exhaust port opening, whereas other hydrocarbon production mechanisms must be present to account for increasing hydrocarbon concentrations throughout the rest of the scavenging event. Likely mechanisms include vaporization of liquid fuel from the combustion chamber surfaces, or hydrocarbons released from crevice volumes such as the injector tip.

Injection timing differences were seen to affect the time-resolved emissions for each injector. For the air-assisted injector, the time-resolved HC concentration at exhaust port opening was the same for both timings, however more hydrocarbons were released later in the cycle for the advanced condition. It was also observed that more oxidation of hydrocarbons to carbon monoxide was likely occurring for the advanced condition throughout scavenging.

For the pressure-swirl injector, the retarded condition exhibited the highest time-resolved HC and CO concentrations upon exhaust port opening. This may have been occurring due to undermixing of the fuel spray for the retarded condition. The advanced case exhibited the highest HC mass throughout the rest of the cycle, suggesting that early-injected fuel may have been released later during scavenging.

Few differences were seen in the trends of the time-resolved pollutant concentrations between injectors.