Published online by Cambridge University Press: 06 March 2019
It will be shown that there are different types of influence coefficients to correct for matrix effects in XRF analysis: the empirical coefficients calculated from regression analysis and valid for the standards on hand only; the theoretical binary coefficients that are essentially constant in a given concentration range; then the theoretical multi-element coefficients calculated from one given composition and valid for one specimen only.
With theoretical coefficients, however, a special calibration process is needed, because theory cannot take into account all the instrumental parameters. In the proposed calibration method it will be shown that the slope of the calibration curves is the link to correlate the theoretical formalism to the experimental reality; that the standards used do not have to be similar to the unknowns and that a few only are required; that the calibration curves can be extrapolated by a factor of two or three; that pure analytes are not required.