Some problems in the analysis of steels by X-ray fluorescence

Abstract

The measurement of the proportion of elements of atomic numbers higher than 20 in steels by X-ray fluorescence using a Geiger-counter spectrometer has been investigated, primarily with a view to determining the suitability of the method as a rapid means of analysis during the steel-making process, particularly for high alloy steels. A monitoring system to compensate for fluctuations in X-ray tube output is described and shown to increase the accuracy, though gives higher variances than theoretical once obtained by assuming a Poissonian distribution of counting events, and shows "drift" over long periods. From measurements on some iron/chromium/nickel steels it appears that unless a very powerful X-ray tube is used the method gives quick results of high accuracy only when more than about 10% of the alloying element is present. It is also shown that in some cases standard diagrams would have to be drawn up to allow for "mutual fluorescence." Certain possible improvements are noted. An Appendix gives the theory of this X-ray fluorescence analysis.