Study of normalization techniques using low and high energy x-ray fluorescence excitation beams with standard samples

Abstract

X-Ray fluorescence is gaining popularity in the scientific community as it is portable, non-invasive, and inexpensive to operate. Rigorous method analysis is needed to accurately assess health and mineral conditions in people or crops through concentrations of various elements. The current study examines the role of the 6.4keV X-ray fluorescence (XRF) excitation beam in the analysis of low atomic number (Z) elements and its comparison to the 17.48keV excitation beam for these elements. Fifty okra samples were analyzed using XRF and results were compared to inductively coupled plasma mass spectrometry (ICP-MS) values through linear correlation plots. The correlations for calcium, potassium, and phosphorus, revealed that the 6.4keV excitation beam, was better suited for elements up to but not including chromium. The low energy beam had an average R2 value of 0.74 whereas the high energy beam had an R2 of 0.52 for the okra samples. The same procedure was used to examine standard caprine horn samples where only calcium was examined. These supported the use of the low energy beam for low Z elements with an average R2 value of 0.45 compared to the high energy R2 value of 0.26. Normalization of the data was required and four methods, Region of Interest (ROI), Molybdenum or Iron, Scatter Peak, and total area ratio (TAR) were created and tested to determine which produced normalized counts with the best correlation to ICP-MS. Linear correlations of the okra samples indicated the ROI method was most promising for studying low Z elements while using the low energy X-ray fluorescence excitation beam. The goat horn samples indicated the Fe method provided the best correlations though all R2 were similar and the relative uncertainties were twice the okra. More research is needed to determine the potential of the ROI normalization method with higher atomic number elements.