Li hyper-enrichment in granitic pegmatites driven by preferential partitioning into transitional melt-fluids: Example from the Dangba Li deposit, NW China

This is the authors' abstract. We don't add key points for this paper.

The formation of Li-rich pegmatites remains a petrological challenge, as conventional magmatic processes alone cannot account for the extreme Li enrichment (>5000 ppm) required for spodumene saturation. Although multi-stage enrichment models have been proposed, the coupled role of boundary-layer accumulation, melt/fluid separation, and subsequent immiscibility during post-emplacement magmatic-hydrothermal evolution remain insufficiently evaluated. By integrating apatite-spodumene geochemistry, melt/fluid inclusion analysis, and quantitative Li partitioning data from the Dangba pegmatite (NW China), this study identifies a two-stage immiscibility-assisted Li enrichment process along the melt-fluid continuum. Here, the term transitional melt-fluids is used descriptively for hydrous silicate liquids with intermediate H 2 O contents (approximately 20–80 wt%) along the melt-fluid continuum, rather than necessarily representing thermodynamically demonstrated critical fluids. In this study, fluid-rich melts refer to melt-dominated transitional liquids, whereas solute-rich fluids refer to later internally derived, fluid-dominated liquids represented by carbonic inclusions containing abundant dissolved components and daughter minerals. The data reveal that initial immiscibility segregated Li-richer, fluid-rich melts (Li approximately 92 ppm in coexisting apatite) from Li-poorer, fluid-poor melts (Li approximately 43 ppm in coexisting apatite), followed by further Li concentration into solute-rich carbonic fluids (Li approximately 364 ppm in coexisting apatite) relative to solute-poor aqueous fluids (Li approximately 4.98 ppm in coexisting apatite). Using apatite-melt/fluid partition coefficients, we show that both stages can generate melt/fluid Li contents approaching or exceeding spodumene saturation. This model is corroborated by microtextural observations that distinguish two spodumene generations: (1) prismatic Spd1 crystallized from fluid-rich transitional melt-fluids and (2) fine-grained Spd2-quartz intergrowths crystallized from internally derived, solute-rich, fluid-dominated residual liquids that retained residual melt components during late magmatic-hydrothermal evolution. Tof-SIMS analyses indicate 6Li enrichment in late-stage spodumene, reflected in systematically lower 7 Li/ 6 Li ratios. These results support a genetic model in which local boundary-layer enrichment may have promoted volatile and solute accumulation and thereby triggered immiscibility, whereas the coexistence of compositionally distinct fluid-rich and fluid-poor melt inclusions remains petrographic evidence for phase separation. Overall, phase separation, late fluid-dominated residual liquids, and boundary-layer-assisted local enrichment jointly promoted Li hyper-enrichment, spodumene crystallization, and Li isotopic modification in the Dangba pegmatite.