Primary Nature of Spodumene-Hosted Crystal-Rich Inclusions Containing Cookeite, Zabuyelite, and Quartz/Cristobalite in Jiajika Pegmatites, China: Based on Observed Formation of Spodumene from these Minerals During Heating in an Optical Cell

Abstract As a major source of rare-element (e.g., Li, Be, Nb, Ta, and Cs) resources, granitic pegmatites commonly host a featured type of liquid inclusions called crystal-rich inclusions (CIs). These CIs are typically regarded as direct indicators of pegmatite formation and mineralization processes. In the Jiajika pegmatites of Sichuan Province, China, which host the largest spodumene deposit in Asia, spodumene-hosted CIs with a spodumene-negative morphology commonly contain zabuyelite, cookeite, and quartz (or cristobalite), in addition to CO2-rich aqueous solutions. However, spodumene is highly reactive, which raises a contentious issue: whether reactions between the cookeite + zabuyelite + quartz/cristobalite assemblage and spodumene occur along spodumene fractures or cleavages during CI entrapment and within CIs during subsequent microthermometry analyses. This debate has led to uncertainties regarding the genesis of CIs in Jiajika pegmatites and the pegmatite formation. In this study, we conducted experiments using a hydrothermal diamond-anvil cell to investigate these reactions. We observed the following reaction, Reaction 1, in aqueous solutions under conditions of 500–600 °C and 400–550 MPa: 2 Li2CO3(zabuyelite)+7 SiO2(quartz or cristobalite)+ LiAl4(Si3Al)O10(OH)8(cookeite)→5 LiALSi2O6(spodumene)+4 H2O+2 CO2. During the reaction, spodumene crystallization occurred following the decomposition of zabuyelite, which rendered the solution alkaline and thus favored spodumene formation. Conversely, the reverse of Reaction 1 is more likely to proceed under acidic conditions. This implies that spodumene cannot react with an alkali-rich CO2–H2O solution to produce cookeite and zabuyelite within CIs. Even if spodumene were to react with an acidic CO2–H2O solution to produce cookeite and zabuyelite contained in CIs, the spodumene host cannot recrystallize on CI walls to make CIs with spodumene-negative shapes. Therefore, CIs could not have formed in spodumene fractures or cleavages via the reverse of Reaction 1. Additionally, the observed decrease in CI volumes during heating can be attributed to the growth of new spodumene on the CI walls, indicating that Reaction 1 proceeded and produced only spodumene that crystallized on the CI walls. This suggests that the molar ratios of minerals within the CIs must align with those in Reaction 1, as observed. This further excluded the possibility that minerals inside CIs were trapped along spodumene fractures or cleavages during the occurrence of Reaction 1; otherwise, the molar ratios of minerals inside CIs would be random. Consequently, CIs hosted by spodumene that have not been necked or destroyed by late fluids are primary in origin and were trapped during spodumene crystallization. During the heating, these primary CIs, in which Reaction 1 did not occur, yielded their homogenization temperatures of 500–550 °C. These temperatures can be interpreted as the crystallization temperatures of the spodumene host. Furthermore, the CIs exhibit high Li contents (up to ∼13,500 ppm), which could reflect the local melt composition in spodumene-rich zones within the Jiajika pegmatites. These findings provide critical insights into the formation mechanisms of CIs and the crystallization history of spodumene in granitic pegmatites.