The cumulative grade–recovery curve shows a maximum Li₂O grade of approximately 25% at the S.G 2.70, corresponding to 76% lithium recovery. Furthermore, reducing S.G. to 2.60 increases lithium recovery but results in significant mass dilution, indicating that the optimal separation interval is between 2.80 and 2.70 g/cm³. To better represent these results, the selectivity index was calculated (Table 4). Table 4 – Selectivity index (Li2O recovery/sink mass recovery) pecific Gravity Interval Selectivity Index >2.89 0.42 2.89-2.80 1.47 2.80-2.70 1.74 2.70-2.60 1.01 As shown in Table 4, the selectivity index reached a maximum value of 1.74 at the S.G 2.80-2.70 interval, confirming that this interval provides the best compromise between lithium recovery and mass weight. At lower S.G, the selectivity decreases due to increased gangue dilution. CONCLUSIONS DMS is the dominant pre-processing method for recovering lithium from pegmatites in Minas Gerais projects. This is justified by the coarse spodumene crystal size and the pronounced difference in specific gravity of spodumene and gangue minerals. The pegmatite samples used as a case study have mainly quartz, spodumene, fluorite, lithium mica (lepidolite), and microcline. Furthermore, it was identified as amphibole and as middling particles with intergrowths of Ti, Zr, Cr, and Fe. The float–sink tests demonstrated that lithium is preferentially concentrated within a narrow specific density interval, specifically between 2.80 and 2.70. Sink-based grade–recovery curves revealed a maximum cumulative Li₂O grade of approximately 25% at 76% recovery, indicating a favorable trade-off between upgrading and mass yield. These results confirm that dense media separation is technically feasible for lithium upgrading in pegmatites originally from Minas Gerais, Brazil, and are compatible with reports found in the literature. Overall, the study demonstrates that sink-based separability analysis provides a robust framework for evaluating lithium beneficiation performance and defining operational cut densities. 151
RkJQdWJsaXNoZXIy MTM0Mzk2