GSA Annual Meeting in Denver, Colorado, USA - 2016

Paper No. 241-4
Presentation Time: 9:00 AM-6:30 PM


TIBBITS, Tawny L.B., Department of Earth and Environmental Sciences, University of Iowa, 115 Trowbridge Hall, University of Iowa, Iowa City, IA 52240,

The use of portable x-ray fluorescence (pXRF) on homogeneous materials has been established in the literature. However, the use of pXRF on non-homogeneous materials is still being explored. This project used granite from the Maya Mountains of the Belize as the non-homogeneous material. The samples used during this project were cut to provide a flat surface. Only samples with minimal weathering were analyzed. A new sampling strategy was employed to manage problems associated with the differences between beam size (10 mm) and grain size (0.5 to 2 cm). Using Monte Carlo Simulation on a data set of 50 discreet points taken in a grid pattern on three granite samples. The simulation showed that no fewer than five data points were needed to develop an average geochemical signature that is statistically indistinguishable from using the full pool of 50 data points on the Maya Mountain granite. Lab-based XRF and pXRF results were compared to examine the accuracy of pXRF on heterogeneous materials. Thirteen whole rock samples were analyzed using the pXRF on Geochem mode for thirty seconds per beam, with two beams active. After the whole rock pXRF analysis, part of the sample was powdered using a ceramic ball mill. The powders were then analyzed using the same settings on the pXRF. The powders were then sent for analysis at the Illinois State Geological Survey XRF Lab. The results for whole rock and powdered rock analyzed by pXRF are statistically significant, indicating this method for using multiple points to generate an average bulk composition is effective. The data from the XRF are statistically indistinguishable from the whole rock and powdered pXRF results. These results indicate that the pXRF can replicate lab-based XRF results on heterogeneous materials when a sampling strategy is employed.