GEOSCAN Search Results: Fastlink


TitlePortable x-ray fluorescence analysis: as an assessment of the significance of sample thickness
AuthorKnight, R DORCID logo; Valiquette, L J; Russell, HORCID logo
SourceCANQUA 2015: conference program and abstracts; 2015 p. 51-52 Open Access logo Open Access
LinksOnline - En ligne (complete volume - volume complet, PDF, 2.08 MB)
Alt SeriesNatural Resources Canada, Contribution Series 20170162
MeetingCANQUA 2015: Canadian Quaternary Association (CANQUA) Meeting; St. John's, NL; CA; August 16-19, 2015
Mediapaper; on-line; digital
File formatpdf (Adobe® Reader®)
ProgramGroundwater Geoscience, Aquifer Assessment & support to mapping
Released2015 08 01
AbstractPortable X-ray fluorescence (pXRF) spectrometry is increasing in popularity for the geochemical determination of unconsolidated sediments due to real time results and reduced analytical cost compared to traditional methods. This non-destructive surface technique is predicated on a programming assumption of infinite sample thickness as the depth of X-ray penetration and the ability of returning X-rays to encounter the detector are influenced by this parameter. Portable XRF analysis has been systematically completed on 1,045 core samples from 10 boreholes located in multiple glaciated basins across Canada (Champlain Sea, Oak Ridges Moraine, Spiritwood Valley, Nanaimo Lowlands). A grain size fraction of <63 micrometres is used to minimize nugget effect and to facilitate analytical comparisons between datasets. Situations may arise, however, where the <63 micrometre size fraction of coarse sand and gravel units result in samples of less than infinite thickness. To address the question of how much sample is required to obtain stable results, a comparative study on the analyses of variable sample thickness was undertaken using Canmet certified reference materials, Till-1 and Till-3. Analysis was completed on samples from 1 mm in thickness to 20 mm in thickness, incremented in 1 mm intervals. The pXRF was operated in Soil mode, which is optimized for elemental concentrations of <1%. A dwell time of 60 seconds each for the Low, Main and High filters was used. Sixteen of the 18 elements detected returned valid results (As, Ba, Ca, Cr, Cu, Fe, K, Mn, Ni, Pb, Rb, Sr, Ti, V, Zn, Zr).
Results indicate that detected elements fall into three categories that correlate with one of the three detectors, i) Elements detected using the Low filter for which concentrations remain fairly consistent with increasing sample thickness includes Ca, Cr, K, Ti and V. Of these elements Cr has the highest x-ray intensity and is determined using the Kalpha1 5.24 keV. ii) Elements detected using the Main filter for which concentration decreases significantly with increasing thickness until infinite thickness is achieved (generally >6 mm), include As, Cu, Fe, Mn, Ni, Pb, Rb, Sr, Zn, and Zr. Of this group Mn has the lowest X-ray intensity and is determined using the Kalpha1 15.98 keV. iii) Ba was the only element detected where concentrations continue to increase significantly over the full 20 mm of sample thickness indicating that infinite thickness has not been achieved. Barium is detected using the High filter and the Kalpha1 line of 32.70 keV.
Divergence from infinite sample thickness influences the analytical results depending upon the X-ray intensity used to detect a specific element. For elements determined at low energy levels sample thickness is not a significant factor. Elements determined using the Main filter, the Compton normalization algorithms over report concentration for samples of <6 mm in thickness. For elements with high X-ray intensities, such as Ba, the Compton normalization algorithms under report concentrations until infinite thickness is obtained.
Summary(Plain Language Summary, not published)
Documents a test of how much sediment is required for reliable analysis with a Portable X-ray fluorescence (pXRF) spectrometry. Sample thickness is assumed to be infinite by the processing algorithm. In the absence of infinite thickness, which varies by element, how do results differ with changing thickness.

Date modified: