On-stream analysis is a much more rapid means of obtaining true assays of a flowing product when compared to wet chemistry or “grab” samples dried, briquetted and assayed by x-ray. In contrast, on-stream analysis uses real time samples assayed every 30 seconds while each stream is assayed every 5 to 15 minutes.
The main physical and chemical factors which influence the accuracy of the slurry assay by x-ray fluorescence are as follows:
- Variations in percent solids, Cs
- Variations in particle size within the solids
- Variations of the dry specific gravity of the solids or the matrix and the resultant changes in the interelement interferences
Changes that occur in different analytical lines when the percent solids of the slurry, Cs, increases from 0 to 40%. All intensities are shown in relative percent. The concentration of Cu, Mo, and Pb in the tails is between 0.005% and 0.05% which means that the major portion of the x-ray line intensity is a small fraction of the background intensity.
These scatter lines are selected according to the product and its matrix and provide data necessary to calculate both pulp density and particle size variations within the sample. It is very important to correct for all these variations to obtain an adequate assay on S, Si, and Ca.
However using different lines of scattered radiation for determining corrections and calculations of percent solids plus particle size effects is still not enough to obtain first grade assays on-line. A significant part of the assay errors are due to interelement effects. There are two different and mutual elemental relationships in a multicomponent slurry. It is essential to obtain analytical data from more than one wavelength of scattered radiation in addition to the normal six to nine elements as used at present in order to assay to laboratory standards.
On-stream assays equal to that of the wet laboratory are now possible through an x-ray fluorescence on-line slurry analyzer. The justification for using the data for control actions is multiplying rapidly. As the cost of additional data is small, then processes can be controlled to produce improved products at lower costs.
