Tube-sample coupling in XRF spectrometry

Abstract

The introduction of the latest generation of wavelength-dispersive XRF spectrometers, with closely coupled x-ray tube-to-sample optics, has led to significantly higher x-ray intensities, lower detection limits and shorter counting times for the same level of precision. The higher intensities result from the inverse square law relationship between decreasing distance and the exponential increase in x-ray flux. However, with decreasing distance, small errors (e.g. ±100 µm) in sample positioning with respect to the sample datum height will lead to intensity errors which should be similarly magnified by the inverse square law. Such small positioning errors can in theory lead to intensity errors exceeding ±1% in closely coupled spectrometers. However, intensity measurements on height-adjustable test samples show that for inclined tube spectrometers (side-window and end-window sequential), the inverse square law is not followed. Observed intensity variations with small changes in sample height were much less than expected from the inverse square law. Instead the measured intensities follow a second-order parabolic relationship. For some spectrometers the crest of this parabola is close to the sample datum height and intensity variations are greatly reduced (i.e. <0.1% per ±100 µm). The parabolic relationship is considered to reflect (i) the asymmetric distribution of the primary x-ray flux on the sample and (ii) the restricted collimator field-of-view (product of collimator mask size and take-off angle). The interaction between these factors leads to compensating effects that buffer the measured x-ray intensities against changes in sample height. By contrast, in end-window simultaneous spectrometers the measured intensities indicate conformance to the inverse square law with important implications for analytical errors. Copyright © 2000 John Wiley & Sons, Ltd.