Analysis of Niobium-Tantalum Ore by X-Ray Fluorescence Spectrometry Combined with Low Dilution Fusion

Low dilution fusion, i.e. fusion of sample with low mass ratio of flux to sample, was used as sample preparation for use in XRFS analysis of crude ore and concentrate containing niobium and tantalu. Optimum conditions for sample fusion were studied thoroughly and reported in detail. A mixture of lithium tetraborate and lithium metaborate in the mass ratio of 12 to 22 was used as flux. Mass ratio of the flux mixture to sample of 4 to 1 was taken for the fusion, and the optimum conditions were chosen as follows:2.000 g of the ore sample were mixed thoroughly with 8.000 g of the mixed flux in a Pt-Au (w_Pt:w_Au=95:5) crucible. After addition of 0.15 mL of saturated lithium bromide solution, the fusion was carried out with TNRY-01C automatic sample fusion melter at 1 200℃ for 180 s, the crucible was swung for 240 s and then stayed for 20 s. When the melt was cooled down naturally, the melt disc of the sample was turned over and the fusion was repeated once under the same conditions described above, with a supplementary addition of 0.05 mL of saturated lithium bromide solution. By the above fusion process either the crude ore samples containing ≥ 0.01% of niobium tantalum oxide or the concentrate sample containing ≥ 60% of niobium tantalum oxide can all be well melted. In case only analysis of ores containing Nb and Ta is considered, effects of time of fusion can be ignored, but in case of some ore samples hard to melt, the fusion time can be prolonged, with attention to supplementary addition of LiBr solution. In case of analysis of ore samples containing alkali metals or halogens, definite and consistent time of fusion should be kept. The purpose of the re-fusion process mentioned above were ① to expel gas bubbles from the melt in order to keep a plain and smooth surface of the melt; and ② to ensure a complete melting of the sample. Precision of determination of the 11 components in the ore samples was tested, values of RSDs (n=6) <1.0% were found for the major components and ≤ 4.0% for those micro-components.