Staff scientist and director, EG&G ORTEC, Materials Analysis Division, Oak Ridge, TN
X-ray fluorescence techniques have been applied for the analysis of portland cement for the past 20 or more years. During this time, applications have been developed primarily with wavelength spectrometers. In the past several years, development of solid state detectors has allowed energy dispersive X-ray fluorescence to detect and measure all the elements of interest in portland cement.
Determination by energy dispersive X-ray fluorescence displays excellent precision, but a linear calibration of intensity versus concentration for many elements exhibits significant deviation from a straight line. Particle size effects that could cause errors in the calibration were minimized through fine grinding. The remaining factors responsible for nonlinear calibrations relate to interelement effects, which are usually considered as absorption and enhancement.
The purpose of this investigation is to determine the interelement effects in portland cement and demonstrate the quantitative capabilities of energy dispersive X-ray fluorescence using certified National Bureau of Standards (NBS) portland cements as standard. An additional purpose is to demonstrate the application of empirical calculations for interelement effects in obtaining quantitative results utilizing Portland Cement Association check specimens as unknowns.
Interelement effects in X-ray fluorescence are well documented with various techniques utilized. Many of these programs are complex and correct all elements for the presence of all other elements and require a large number of standards. The approach used by the author is an exponential model based on intensities of the interferring elements.
Linear least squares fit of uncorrected intensities displayed absolute errors of 0.6% for alumina and silicon oxide to 1.0% for calcium oxide. Average deviations obtained following interelement corrections ranged from a low of 0.007% for potassium oxide to a high of 0.19% for calcium oxide. Analysis of each specimen consumed 200 s. The accuracy obtained by this method is well within the limits expected by the industry and the precision possible by alternate procedures.
Paper ID: CCA10262J