Identification of environmental contaminants and their sources using various analytical spectroscopy methods, spectral libraries and multi-variate statistics

Two methods have been developed for the identification of the source of environmental contamination using various analytical techniques, multi-variate statistics and spectral matching. The first method was developed to identify the refining company responsible for chromium contamination of soil. Samples were collected from sites with known and unknown manufacturers and analyzed using X-ray fluorescence (XRF) spectroscopy. Cluster analysis was used to group the samples together, using various elemental combinations to isolate the elements present due to chromium waste. One unknown site was successfully identified as being produced by Diamond Shamrock; additional samples are necessary for further discrimination of the other unknown areas.; The second method was developed to identify the source of lead in manufacturers paint trade cards, which have a similar composition to the paint in cans, but are more useful for method development because a large number of samples are available from a manufacturer. The samples were analyzed using XRF, attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy and micro-Raman spectroscopy (MRS). The pigments present in the samples were identified using a combination of these spectral techniques. Spectral libraries for each technique were developed for the identification of unknown samples. Classification and regression trees (CART) determine the manufacturer using both quantitative and qualitative variables without the sample being present in the spectral library.; The final set of samples analyzed were house paints manufactured between the late 1800's and 1978. The method developed for the determination of the pigment composition was applied to paint in cans. Spectral libraries were developed for each technique and used to determine the manufacturer of unknown samples. Real world samples were simulated by applying several different layers of paint to a wooden block. The paint chips were then embedded into several different types of media and sectioned into thin layers. The individual paint layers were then analyzed using micro-XRF and micro-Raman spectroscopy.; The methods developed here, combining analytical techniques with multi-variate statistical analysis and spectral libraries have a broad range of applicability. They are appropriate for any type of analysis where a large number of factors are needed to characterize samples into specific groups.