A comparison of XRFS and ICP-OES methods for soil trace metal analyses in a mining impacted agricultural watershed

Field portable X-ray fluorescence spectroscopy (XRFS) has become increasingly prevalent for in situ detection of trace metals, as it is both rapid and cost effective. The accuracy of in situ XRFS analyses has been questioned due to possible interferences from elevated soil moisture and organic content. In this study, three metal analysis protocols (Cd, Pb, Zn) were compared for surface soil samples collected near the Tar Creek Superfund Site in northeastern Oklahoma. The protocols included the use of a field portable XRF spectrometer for in situ analyses and on homogenized, pulverized, air-dried soil samples sieved to < 250-µm fraction in the laboratory. A subset of soil samples was also analyzed after microwave-assisted hot HNO₃ digestion followed by inductively coupled plasma-optical emission spectrometry (ICP-OES) analyses. Moisture content and loss-on-ignition (as a surrogate for organic matter) were determined for each sample. Soil moisture exceeding 10% in situ caused underreporting of field XRF readings when compared to the laboratory XRF readings. Relationships between concentrations determined by laboratory XRFS and ICP-OES for Pb (r² = 0.96) and Zn (r² = 0.91) were not statistically different (p < 0.025 for both analytes). A strong relationship between ICP-OES analyses of Zn and Cd (r² = 0.93) allowed prediction of Cd concentrations for additional samples not analyzed by ICP-OES. This study recommends that XRFS field readings be used for initial screening only and that samples analyzed via field portable XRFS be homogenized, air dried, sieved and re-analyzed in the laboratory to yield the most accurate results.