COMPARISON OF TRACE ELEMENT CONCENTRATIONS IN LOESS DETERMINED BY ICP AND PXRF

Trace elements in minerals can be used to determine provenance of sediments, including Quaternary loess deposits. Historically, determination of trace element concentrations required acid digestion and wet chemistry techniques or fused pellets for x-ray fluorescence (XRF). Portable XRF (pXRF) instruments are revolutionizing investigations by allowing rapid determination of elemental concentrations of soils and sediments, including in situ measurements from outcrop or cores. This research compares three methods of measurement of trace element concentrations in the 8-63 µm silt fraction of Peoria Loess collected in Wisconsin. The silt fraction of 29 samples was separated by settling and micronized to a fine powder. Trace element chemistry was determined by acid digestion (HF and aqua regia) and analysis by ICP-OES, and two methods in a Bruker Tracer IV-SD pXRF. The two pXRF methods include the GeoQuant TR and Soil methods analysis packages installed in the Bruker S1 software. Results were compared graphically and with Pearson correlation coefficients for a total of 8 elements (Ti, Cr, Rb, Sr, Zr, Nb, Ba, Ce) that were measured by ICP and both pXRF methods. Correlation coefficients between ICP and the pXRF methods ranged between 0.12 for Ba to 0.88 for Sr. With the exception of Sr, Zr, and Ti the two pXRF methods were not consistent predictors, and neither method appeared to be a better overall predictor of trace element concentration. Elemental concentrations measured by pXRF were not consistently lower, equal, or higher than measured by ICP. For both pXRF methods the greatest correlation values were consistently obtained for Sr, Zr, and Ti. These results support using pXRF for rapid determination of the long established ratio of Ti to Zr for provenance and sediment uniformity studies, although results for other trace elements should be used cautiously.