Chemical characterization of individual particles by electron probe X-ray microanalysis and electron energy-loss spectrometry
Xhoffer, C. (1993). Chemical characterization of individual particles by electron probe X-ray microanalysis and electron energy-loss spectrometry. PhD Thesis. Universiteit Antwerpen: Antwerpen. 248 pp. |  |
| Available in | Author | | Document type: Dissertation
|
| Abstract | Among the many techniques available at the Micro and Trace Analysis Center (MiTAC) for the observation and analysis of individual particles, electron probe X-ray microanalysis (EPXMA) in combination with multivariate classification techniques allows the most advanced characterization of individual particles in a fast and efficient way. All facilities of EPXMA were used for the source apportionment of more than 25,000 individual aerosol particles detected above the North Sea, the English Channel and the Celtic Sea. The North Sea aerosol is predominantly composed of sea salt, sulfur-rich particles, silicates and calcium sulfate particles. Their abundance is dependent on meteorological conditions and sample location. Differences were studied on the basis of the abundance variations by using principal component analysis. The first component represents the marine-derived aerosol fraction and is more important as wind speed increases or at more remote locations. The second component differentiates anthropogenically derived CaSO4-rich samples. Their relative abundance is more pronounced as the sampled air masses spend longer residence times over the South of England. The samples of the third cluster are related to high silicate and sulfur abundances. Two different clusters separate mixed marine/continental samples from pure continental-derived silicate and sulfur-rich particulate samples. On the average, of all samples taken during various locations and periods, about 60 % have a marine source. The other 40 % of the particles have a continental or mixed marine/continental origin. Five pairs of North Sea bulk waters and their corresponding surface micro layer samples were investigated for their particulate matter content and chemically analyzed and characterized by EPXMA and laser microprobe mass analysis (LAMMS). Multivariate techniques as hierarchical and non-hierarchical clustering in association with principal component analysis were performed on a data set containing more than 3000 individual particles. The classification procedure yielded eight different particle types. Differences in particle type abundances for the micro layer and bulk water were observed. They were apportioned to their most probable sources. The results were compared with atmospheric and riverine particle data. The domination of the major particle groups (aluminosilicates, Ca-rich aluminosilicates, CaCO3, silicates and organic particulate matter) suppresses the relative abundance of the metal-rich particles (Fe-rich, Ti-rich and others). A conditioned classification was performed using selection criteria derived from riverine suspensions and atmospheric particulate North Sea data. Similar particle abundances were observed for both the multivariate and the conditional particle classification procedures. Chemical characterization of very small substances is of increasing interest in many domains of research. The ability of electron energy-Ioss spectrometry (EELS) to perform qualitative and quantitative information of light elements combined with a high spatial resolution, is one of the innovations of the late seventies. Beside spectrum analysis, an energy filter integrated in a transmission electron microscope offers the possibility of elemental specific imaging (ESI). In this work, an attempt is made to explore the characteristics of EELS for single particle analysis. Optimization of the instrumental parameters is discussed and the advantage of energy filtering is demonstrated for the localization and sizing of particles. ESI gives direct evidence of the localization of specific elements in a matrix and has major advantages over an X-ray mapping. Despite of this, the ESI-technique is not fully exploited and still suffers from shortcomings. To obtain reliable element specific images, the net intensity of the elemental edge must be determined. The net intensity is obtained by subtracting a background image. This background image needs to be |
|
