A workflow for exploration sampling in regolith-dominated terranes using portable X-ray fluorescence: comparison with laboratory data and a case study

This paper presents and validates a workflow that provides for the rapid collection of reliable and robust portable X-ray fluorescence (pXRF) data in a regolith-dominated exploration setting. The analysis of regolith material by pXRF is challenging because of its variable Fe content (e.g. 0 to >70 wt%) that results in large matrix effects, which can be mitigated with appropriate standards. Here, we present a dataset from the Western Mount Isa Inlier, Queensland, Australia, comprising soil, lag, rock and rotary air blast (RAB) samples. In the soil dataset, comparison of the laboratory and pXRF datasets for Cu, Pb and Zn have R² > 0.9, and Cu and Zn fall within 2% of the laboratory data, while Pb is 77% less than the corresponding laboratory analyses. Iron, Al, K and Ca by pXRF perform reasonably well when compared with the laboratory data (R² = 0.59 for Al, R² > 0.9 for Fe, K and Ca, <25% variation from the laboratory data), while Mn, Rb and Sr have very good correlations (<8% variation) with R² > 0.94. Titanium, Zr, Ni, Cr and As have poorer comparisons. Overall, the RAB dataset shows similar trends with Cu overestimated by 17%, Zn underestimated by 4% and Pb overestimated by 69%; R² for all elements is >0.92. Since no suitable standards are available, the solid rock dataset was uncorrected; despite this limitation, the dataset shows good correlations with the laboratory data for many elements, and Cu is overestimated by 9.8% with an R² = 0.87. The poor analytical performance of Pb in all datasets is associated with erroneous Bi concentrations being reported by the pXRF unit. When high Fe and Pb amounts are present in a sample, erroneous Pb and Bi concentrations are reported, owing to a pile-up of the Fe K_α peak (6.405 keV) at ～12.8 keV, which is proximal to the Pb L_β (12.614 keV) and Bi L_β (13.023 keV) peaks. Despite the care that is required in validating data, by using pXRF there is substantial opportunity for dynamic exploration campaigns in regolith-dominated terranes with rapid turnaround times, additional elements that may not otherwise be analysed for and low analytical costs. Decisions to stop, continue or infill drill holes while the drill rig is present can be made in near-real time, and not after laboratory results are available and the drill rig has left the area.