Lithium extraction from lepidolite using molten ammonium bifluoride: applying response surface methodology to model and optimize the thermal process

Abstract

This work presents the modeling and optimization of a new lithium extraction method from lepidolite with molten NH₄HF₂. The interaction between the mineral and the fluorinating agent was studied by DSC, indicating that the reaction occurs between 75 and 160 °C, approximately. Thus, the temperature ($T$), time ($t$), and NH₄HF₂ amount ($m$) of the thermal process were studied by design of experiments (DOE) and response surface methodology (RSM). The DOE experimental results were modeled using least squares to obtain a predictive model of the system (R² = 0.9730). The ANOVA of the reduced cubic model indicated the influence of the three operating parameters investigated, and the interactions $Tm$, $T^2$, $m^2$, and $T^{2}m$ were significant. Finally, the process was optimized, obtaining Li extractions of 95 ± 4 % at 144 °C, 60 min of reaction, and using 13.5 g of NH₄HF₂ per gram of lepidolite. In this way, the new process minimizes operating conditions and costs and maximizes lithium extraction compared to traditional methods. Furthermore, this process allowed us to show the benefits and details of the correct application of DOE and RSM in extractive metallurgy processes.