| Summary: | 碩士 === 國立中央大學 === 環境工程研究所 === 89 === Toward a developed country, Taiwan’s air quality is degraded by a fast development of industry and a steady increase of traffic flow. Particulate matter is still a predominant pollutant for bad air quality (PSI>100) (Taiwan EPA, 1998, 1999, 2000). An effective air pollution control strategy specific to pollution source becomes a focus point of research activities.
This study used honeycomb denuders to collect PM2.5 (particles with aerodynamic diameter smaller than 2.5mm) at Sin-chun site in Taipei County and Hsiao-kun site in Kaohsiung City from December in 1999 to November in 2000. In total, 40 samples collected in Taipei and 41 in Kaohsiung for a comparison in aerosol characteristics between a light and a heavy industrial site. Both single-particle analysis (using computer controlled scanning electron microscope, CCSEM) and bulk chemical analysis techniques were adopted for particle chemical compositions. Nineteen elements, namely C, O, Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, Cr, Mn, Fe, Ni, Cu, Zn, and Pb, were identified from single particles selected randomly by CCSEM. Statistical techniques such as cluster analysis, factor analysis, and absolute principal component analysis (APCA) were applied to the elemental data from single particles to apportion their source contributions. In contrast, the chemical species of particles resolved from bulk analysis were converted into elemental composition for a comparison with CCSEM data and were apportioned their source contributions using APCA.
Among the detected elements, in terms of signal intensity, carbon is the most abundance element with an average of 74%, oxygen is second to carbon with an average of 12%, silicon is the third most abundance element with an average of 4%, those followed were aluminum 2.5%, sulfur 1.7%, and sodium 1.3%. Six of twenty source types resolved from cluster analysis were with significant particle numbers. Notably, the fraction of particle numbers for the source type with carbon and oxygen only is around 35% in Taipei and 40% in Kaohsiung, respectively. In addition, factor analysis shows that Taipei and Kaohsiung are similar in source contributions with industrial sources, sea-salt spraying, cement and fertilization production, dust and mixed burning sources, and vehicle emissions in common. The apportioned source types are consistent with sources shown in local emission inventory. In general, for elemental data from CCSEM, cluster analysis has the best resolution in source identification than factor analysis and APCA.
The bulk analysis shows spring PM2.5 at 46.6 mg/m3 is the highest among the four seasons at Sin-chun site. In contrast, the highest PM2.5 for Hsiao-kun site is at 94.5 mg/m3 for winter season. Major compositions of PM2.5 are OC, sulfate, and ammonium ion. The APCA indicates a mixed source of secondary reactions, sea-salt spraying, and motor vehicle emissions is predominant at Sin-chun site; whereas the mixed source from motor vehicle emissions, agricultural burning, secondary reactions, industrial activities, and resuspended dusts is the largest source type.
In summary, although single-particle analysis and bulk analysis are two different techniques, both methods are comparable in the rank of the resolved elemental abundance. These two methods are complimentary in source apportionment of atmospheric aerosols.
|
|---|
