Scandium quantification in selected inorganic and organometallic compounds

English: The main objective of this study was to develop an analytical method to accurately quantify scandium in different Sc-containing matrices including inorganic compounds, organometallic complexes and finally low grade Sc mineral ores and residues. The organometallic complexes (Sc(acac)3, Sc(tfac)3, Sc(btfac)3, Sc(dbm)3, Sc(hfac)3 and Sc(sacac)3) were synthesized and characterised using melting point, IR, CHNS-micro element analysis and X-ray crystallography techniques. The study involved the use of different mineral acids such as HNO3, HCl and H2SO4 for the dissolution of the different samples using both open-beaker and microwave-assisted dissolution techniques. Flux fusion using NH4F·HF flux was used for the dissolution of columbite mineral and Ta/Nb residue samples. ICP-OES was used for quantification of Sc in all samples as well as some of the other metals in the mineral samples. The wavelengths were carefully selected to minimise any spectral interferences and the matrix matching was also ensured throughout the study for accurate measurements. ScCl3·H2O was dissolved in water and recoveries ranging from 99.64 - 99.79 % were obtained. Open-beaker dissolution of Sc2O3 with HNO3, HCl or H2SO4 yielded Sc recoveries ranging from 44(7) - 99.2(4) % and these recoveries were improved with the microwave dissolution method to recoveries between 99.8(3) and 101(1) %. Excellent Sc recoveries from the most of the organometallic complexes (Sc(acac)3, Sc(tfac)3, Sc(btfac)3, Sc(dbm)3 and Sc(sacac)3) ranging from 95(3) - 99.4(6) % were obtained while Sc(hfac)3 yielded a lower than expected recovery of 80.4(4) % due to the possibility that the molar mass from the empirical formula known is incorrect or that an impure product was isolated. Excellent results were obtained for the columbite ore and the Ta/Nb residue samples (Sample A and B respectively) after the NH4F·HF fusion dissolution. In the process of the beneficiation of Sc, magnetic separation was attempted for the removal of Fe and Ti from the rest of the sample (both Sample A and Sample B) but was considered unsatisfactory since a large portion of the Fe in the sample did not exhibit ferromagnetic properties and hence remained with the original portion (nonmagnetic) of the samples. Results also indicated the loss of a relatively significant portion of Sc in both samples. Solvent extraction with MIBK was also attempted for the possible extraction of Ta and Nb and the results indicated that this process was only successful to eliminate the Ta in large quantities and that both the Ta and Nb can be extracted with MIBK at high [H2SO4] with relatively insignificant losses of Sc, especially from Sample B. The experimental results for the scandium analysis were validated for a large number of validation parameters, which included accuracy, precision, sensitivity, specificity, linearity, etc. to confirm whether the newly developed analytical procedures were suitable for the scandium determination in terms of internationally required standards (ISO 17025). The limit of detection (LOD) and limit of quantitation (LOQ) for Sc were determined to be 0.000991and 0.00991 ppm respectively, which is sufficient to measure trace amounts of scandium. The linearity of the calibration curves was determined from the regression coefficient (R2) and ranged from 0.996 to 1.00. Statistical tests of the experimental results were calculated using the hypothesis test of the t-statistical at 95 % confidence interval (C.I) to determine whether the results were acceptable as recommended by ISO 17025.