Abstract The growing demand for Lithium – due to its use in modern economies and green energy – is prompting closer examination of all geological processes that control Li mobilisation. For example, while studies on the genesis of enriched peraluminous granitoids and pegmatites have traditionally focused on the final magmatic-hydrothermal stages of enrichment, recent studies have addressed how the preceding metapelite melting stage controls Li mobilisation. However, these studies employed different partition coefficients to estimate Li fluxes and thus predict contrasting Li behaviour. In this study, we investigate Li mobilisation during metapelite melting by comparing experimental and natural examples (starting compositions, partition coefficients and in-situ mineral concentrations) with equilibrium melting models. Melting models based on experimental partition coefficients predict that biotite-dehydration melting producing either peritectic garnet or cordierite generates more enriched melts; however, the model predicted concentrations in cordierite are lower, and concentrations in muscovite are higher than those observed in nature. In contrast, equilibrium melting models based on natural partition coefficients suggest that initial muscovite melts will be more enriched, and biotite-dehydration melts more enriched, if formed with peritectic garnet (as cordierite sequesters Li). In addition, the lack of temperature-dependent partition coefficients fails to optimally reproduce natural mineral concentration trends. Based on the comparison of experiments and natural examples, we suggest that Li behaviour during equilibrium melting is affected by protolith and mineral major element compositions, particularly water and fluorine contents.
Costa et al. (Tue,) studied this question.