| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 4439063 | Atmospheric Environment | 2012 | 9 Pages |
Twenty one organic compounds including three monocarboxylic and ten dicarboxylic acids, four aldehydes, three polyols and one amine were determined in 61 atmospheric aerosol particle samples with different sizes (30 ± 4-nm, 40 ± 5-nm, 50 ± 5-nm and total suspended particles) collected at two sampling sites, the SMEAR II and SMEAR III stations during different seasons of the year. Non supervised pattern recognition techniques, such as hierarchical cluster analysis and principal component analysis were used to study the influence of the collection place, the season of the year and the particle size on the concentration and behavior of the target compounds. The reliability of these results was proved using a supervised pattern recognition technique such as soft independent modeling of class analogy. The results achieved demonstrate that the chemical composition of the atmospheric aerosol particles is affected by the potential emission sources and the reactivity of the studied compounds under certain atmospheric conditions. In addition, from quantitative analysis methodologies partial least squares regression and principal component regression models were successfully used to clarify the influence of the number of nucleation events on the chemical composition of the particles.
► 61 aerosol particle samples (30 nm, 40 nm, 50 nm, TSP) were studied. ► Twenty one compounds studied are important in atmospheric aerosol chemistry. ► Chemical composition in ultrafine and TSP samples differ in place and in time. ► Correlations were found between analyte profiles and nucleation event number. ► Statistical models used as data analysis tools successfully completed the study.
