Fig. 1 shows the sketch of the experimental setup. A commercial unleaded gasoline was used to produce an exhaust gas. The engine was kept running at an idle condition. The exhaust gas sampled from the engine was first passed successively through a condensator and a dryer filled with CuSO4 to remove its moisture content. It was noted that the cooling and drying process could influence the exhaust gas contents to some extent due to particle deposition, HC-condensation, etc. However, this step is inevitable, so all of the experiments were carried out under similar cooling and drying conditions so as to reduce the potential error caused by this step. Second, a fraction of the dried exhaust gas was forced to flow into a NTP reactor at a certain flow rate, which was located in a constant temperature electric oven to maintain constant reaction temperature and was treated by NTP in the reactor. Third, the HC concentration of the treated exhaust gas was measured by an exhaust gas analyzer (model FGA- 4100) at the outlet of the reactor. Fourth, to accommodate the requirement of a particle counter (handheld laser particle counter (model 3886 GEO-), Kanomax Japan Inc., 2-1Shimizu, Suita, Osaka 565-0805, Japan), the treated exhaust gas was diluted in a mixer by N2 with a content of 98%. Finally, the diluted exhaust gas was passed through the particle counter to measure its PM concentration and then the gas was vented. The measurable particle sizes ranged from 0.3 to 5.0m. The data from the exhaust gas analyzer and the particle counter were collected by an online data acquisition system.