The inherent chemical properties of eight different dissolved organic matters (DOMs) originating from soil, surface and groundwater have been analysed. The samples consist of isolated fulvic acids (FA), humic acids (HA), and humic substances (HS), i.e. natural mixtures containing a humic and a fulvic fraction. The humic substances have been characterised by elemental analysis, size exclusion chromatography, E2/E3 and E4/E6 UV absorption ratios, and liquid-state 13C-NMR spectroscopy. The information contents of the different analytical methods have been investigated by pattern recognition, i.e. cluster analysis and principal component analysis (PCA). A comparative study of the information contents of DOM descriptors derived from different analytical methods is presented. Through extraction of information content of the individual analytical methods the inherent properties of DOM are quantified. Pattern recognition revealed significant quantitative differences in the inherent properties of DOM of different origin and type. PCA based on the NMR descriptors showed highest explained variance. However, all models showed low robustness due to the limited number of samples. The supervised pattern recognition, i.e. PCA, indicates a classification of DOMs into groups of similar properties by an increase in the number of samples. Furthermore that the number of groups may be higher and more continuously distributed than the conventional classification into fulvic acids, humic acids or humic substances.
Read full abstract