On-Stream X-Ray Analyzer And Digital Computer Simplify Ore Analysis

Considerable efforts have been expended to apply multiple regression techniques to quantitative determinations in x-ray fluorescence analysis. Multiple regression has been shown to be an effective method for correlating x-ray fluorescence intensities to elemental concentrations. Although many other techniques have been devised to solve the problem of matrix effects in x-ray fluorescence analysis, there are few as efficient in terms of storage requirement or ease of programming for digital computers as the multiple regression method. In the case of ore slurry, it has the additional advantage of allowing a unified approach to the problems of slurry density correction and matrix effect corrections. The data used in the present work were obtained with an on-stream x-ray analyzer capable of analyzing 15 slurry streams for as many as nine elements simultaneously. One of its fixed monochromators was set to accept scattered radiation at a discretely selected wavelength as a means of monitoring the slurry's pulp density. The data were reduced to obtain the correlation equations and a computer program combining slurry density with matrix corrections was written in FORTRAN IV to calculate concentrations.