Volume 39, Issue 6 (November 1994)
A Possible Approach to Soil Discrimination Using X-ray Fluorescence Analysis
Soil samples collected from 110 different sites in the Kyoto district, Japan, were analyzed quantitatively using X-ray fluorescence spectroscopy to predict unknown soil origins. More than 250 mg soil samples, ground to −100 + 200 mesh and pressed, were analyzed for Si, K, Ca, Ti, Fe, Rb and Sr with good reproducibility. Analytical data were normalized to a standard rock sample (JG-1) and were subjected to multivariate analysis. Especially trace elements, such as Sr and Rb, as well as K and Fe were available to characterize soil samples, and these soils were classified into nine types that show good agreement with geological features. Probabilities of correct identity by comparing unknown soils with control data sets were about 71%, according to the systematic discrimination that was derived from multivariate analysis and a geochemical survey map of soils.