APPLICATION OF PXRF TO REGIONAL SURFICIAL TILL GEOCHEMISTRY STUDIES IN NORTHERN CANADA

The geochemistry of glacial till (diamicton) samples from two regional surficial studies in the area to the east of Great Slave Lake, Northwest Territories, Canada was determined by pXRF and by ‘traditional’ laboratory analyses. Sample preparation for the 1^st study (n = 241) consisted of drying at 60 °C, and then shaking the sample in a 15 ml container to induce granular convection, which brings the pebbles and coarse sand to the surface. This coarser material was removed, leaving a clay + silt + fine-medium sand fraction for subsequent pXRF analysis. pXRF analysis was also completed on the <0.063 mm fraction. Sample preparation for the 2nd study (n = 106) consisted of drying at 60 °C, then sieving to <2 mm followed by pXRF analysis, and also pXRF analysis on the original non-dried sample. For both sample sets, aqua regia, 4 acid, and fusion digestions followed by ICP-ES/MS analysis was determined on the <0.063 mm (clay+silt) and the <2 mm fractions. Comparison of the ‘shaken’ or sieved pXRF data with <0.063 mm laboratory data (fusion or 4-acid methods) from both studies demonstrates that pXRF analyses can provide highly valuable and cost effective information. While the pXRF data does not exactly replicate the laboratory data (in terms of precision and accuracy), the relationship between pXRF data and fusion or 4-acid data is highly systematic, and typically linear with high r² values. Coincident geochemical anomalies are observed on interpolated single element maps generated from the lab data and the pXRF data. Comparisons of aqua regia (partial digestion) and pXRF (a total analysis) data sets are poor and not systematic, as would be expected. pXRF data for the <0.063 mm size fraction, and the ‘shaken’ samples are an excellent proxy for the <0.063 mm laboratory data; the pXRF < 2 mm sieved data is also comparable to the laboratory data. pXRF data from till samples that have not been dried and sieved or shaken to remove pebbles and coarser material typically exhibit much greater scatter (lower r² values), and also lower concentration levels for a number of elements, and this practice is in general not recommended.