We analyzed the size-dependent volatility of nanoparticles in a diameter range of 30-70 nm in diesel exhaust emissions. The test system included a medium-duty diesel truck on a chassis dynamometer, a single-stage dilution tunnel, a tandem differential mobility analyzer (TDMA) equipped with an electric furnace, and a condensation particle counter. The size shifts of monodispersed diesel nanoparticles under changing furnace temperatures were measured by TDMA in the gas phase. Together with the reduction of average particle size and volume, we observed the development of bimodal size distributions resulting from the separation between semivolatile and nonvolatile species as the furnace temperature was increased. While 91-98% of the particles were found to be semivolatile species by total volume during the idling engine condition, only 6-9% were semivolatile during the one-half engine load condition. We also found that smaller particles contained a larger fraction of semivolatile species.

This paper presents the numerical simulation of the flow patterns within three cyclones of different cone dimensions. The simulation was performed with commercial CFD package FLUENT V5.5. The accuracy of the simulation was tested by comparison with the available experimental data. Simulation results show that the cone dimension significantly influences the flow pattern within cyclones. The comparison of the flow patterns within the cyclones of different cone dimensions explained well the different separation behavior of these cyclones. The results suggest that the CFD simulation is capable of predicting changes in flow patterns resulting from cyclone configuration modification.

A semi‐continuous measurement system was constructed to investigate ionic species of fine particles less than 2.5 μm in aerodynamic diameter (PM2.5) in ambient air and at ca. 30 min time resolution. An intercomparison study was also performed with a typical integrated measurement technique (12‐h filter measurements) at the Gwangju Institute of Science and Technology (GIST), Gwangju, Korea. The system was composed of three main parts, i.e., a PM2.5 cyclone inlet connected to annular coating denuders, particle collection components and an ion chromatograph. Particle collection efficiencies in the range of 0.05 – 0.15 μm and 0.15 – 2.5 μm were 80 – 90 % and ca. 100 %, respectively, under the following conditions: distilled water (DW) flow rates, injected into the rectangular coil type air/liquid separator, over 0.23 mL min–1, with aerosol sampling and DW flow rates for steam generation of 2 L min–1 and 0.17 mL min–1, respectively. The limits of detection (3s of the blank value) of the system under the field experimental conditions were as follows: Na+ and NH4+: 0.20 μg m–3, K+: 0.25 μg m–3, Ca2+ and Mg2+: 0.50 μg m–3, and Cl–, NO3– and SO42–: 0.10 μg m–3. The intercomparison studies showed that the results of the semi‐continuous measurement system generally correlated with those of the 12‐h filter measurements for NH4+, SO42–, NO3– and Cl– (r = 0.82 ∼ 0.96, P < 0.0001, 95 % confidence limit). The correlation was best for SO42– (r = 0.96), but was relatively low for NO3– (r = 0.87) and Cl– (r = 0.82). The results indicated that the correlations between the two methods were mainly dependent upon the thermodynamic stability of the species in the fine particles and the sampling artifacts associated with the filter method.