LAUSR

Abstract The aim of the article is to show that determination of the main chemical composition with Wavelength Dispersive X-ray Fluorescence Spectrometry method does not require the use of separate analytical applications and procedures for the tests of materials of different origin. This is however conditioned upon calcination of the tested initial sample at the temperature of 815 °C to remove organic matter and its preparation for X-ray measurements by fusing into a borate bead. Within the framework of the conducted research a special analytical application measuring 10 oxides: SiO2, Al2O3, Fe2O3, CaO, MgO, Na2O, K2O, SO3, TiO2 and P2O5 was developed. Standards and certified reference materials (CRMs) were applied for calibration. Synthetic standards with the use of oxides and salts of purity of 99.99% were also prepared to complete the measuring points on the calibration curves The curves determined based on the measuring points have high correlation coefficients and accuracy. The α empirical correction coefficients were applied to correct matrix effects. The range of the method was confirmed based on the measurements of 13 synthetic standards. The average values of the standard deviations in the lower range of the calibration curves were between 0.0008% for K2O and 0.0056% for Na2O and 0.0054% for MgO. The average values of the relative standard deviation were in the range between 1.81% for K2O and 13.92% for MgO and 11.21% for Na2O, respectively. In the upper range of the calibration curves, the average values of standard deviation were between 0.079% for TiO2 and 0.962% for P2O5, and the relative standard deviation was between 0.05% for SiO2 and 2.39% for P2O5. The accuracy of the developed WDXRF method was assessed based on the measurement of 10 certified reference materials representing all the tested materials (cement, coal ash, slag, fire clay, sediment and soil). Out of 98 conducted measurements (two CRMs did not contain the declared content of P2O5 and SO3 respectively), in 68 measurements, relative errors were below 5%, and, in next 27 measurements, they were between 5% and 10%. The errors were comparable for all the tested materials which confirms that in WDXRF analyses of borax beads the type, kind and origin of a sample do not affect accuracy of the obtained measurements.