Application of Energy Dispersive X-ray Fluorescence Spectrometry to the Determination of Copper, Manganese, Zinc, and Sulfur in Grass (Lolium perenne) in Grazed Agricultural Systems

Abstract

Conventional methods for the determination of major nutrients and trace elements in grass rely on acid digestion followed by analysis using inductively coupled plasma optical emission spectrometry (ICP-OES), which can be both time consuming and costly. Energy dispersive X-ray fluorescence (EDXRF) spectrometry offers a rapid alternative that can determine multiple elements in a single scan. Copper, Mn, Zn, and S in grass samples were determined using EDXRF with a number of different calibration approaches using both empirical standards and the theoretical relationships between concentrations and intensities. Using an existing archive of 467 grass samples of known concentrations, a suite of 30 samples was selected as empirical grass standards to build a calibration set between sample concentrations and EDXRF intensities. The theoretical or standardless approach used the fundamental parameters method to determine element concentrations. To validate the two calibration methods, 59 samples were randomly selected from the same archive and database and analyzed by EDXRF. The measurements of Cu, Mn, Zn, and S were compared with the ICP-OES values using agreement statistics. An excellent correlation was observed between the concentrations determined by EDXRF and ICP-OES ( R > 0.90) regardless of the calibration approach. However, agreement and closeness to the true value varied and were assessed using agreement statistics. Across all elements, the empirically calibrated samples were in excellent agreement with the values determined by ICP-OES. The theoretical calibrations provided excellent agreement for Mn and Zn, but a degree of fixed and proportional bias was observed in the Cu and S values. Fixed bias was corrected by subtracting the computed bias from the EDXRF concentrations and improved the overall agreement. Similarly, proportional bias was corrected using the linear regression model to predict the corrected EDXRF values. This improved the overall agreement with the ICP-OES values for both Cu and S using corrected fundamental parameters calibrations. This study provides a practical basis for the use of EDXRF to determine Cu, Mn, Zn, and S in grass samples to monitor forage quality in grazed systems without the need for sample digestion. The observed fixed and proportional bias in the theoretical calibrations can be corrected provided that a good correlation exists between EDXRF and conventional methods.