Melt-fluid interaction in the formation of peralkaline granite: Evidence from the Baiyinwula intrusion, Inner Mongolia, China

Abstract

The genesis of peralkaline granites (with molar (Na + K)/Al > 1) remains a conundrum, particularly their emplacement accompanied by contemporaneous calc-alkaline rhyolites rather than peralkaline rocks. In this study we present a petrogenetic comparison of calc-alkaline rhyolites (296–293 Ma) and subsequent peralkaline granites (291 Ma) in Baiyinwula, Inner Mongolia of north China. The calc-alkaline rhyolites display a narrow variation of GdN/YbN (1.0–1.4) with increasing Zr/Sm (29–79) and positive correlation of εNd(t) (+ 1.5 to +5.1) versus 1/Nd (0.02–0.08). Accordingly, their formation is interpreted to have involved partial melting of juvenile crust, followed by feldspar-dominant fractionation and crustal contamination. The peralkaline granites exhibit roughly constant variations of MgO (0.06–0.24 wt%), Na2O (4.05–4.89 wt%), K2O (3.77–4.61 wt%) and increasing peralkalinity (0.97–1.12) and Fe-index (0.90–0.97) with increasing Zr (350–1022 μg/g) and Zr/Sm, precluding their genesis from calc- or per-alkaline parental magmas through high degree of fractionation. Interestingly, the mixing hyperbola (R2 = 0.85) of GdN/YbN (0.5–1.2) and Zr/Sm (50–240) for peralkaline granites indicate interaction between magmas and alkali-rich fluids, but is distinctly different from those derived by partial melting of metasomatized sources which exhibit a trend of partial melting. Simulation based on two-component (calc-alkaline rhyolitic melts and alkali-rich fluids) mixing yields highly consistent inference. Both magmatic albite in peralkaline granites and dissolved morphology of zircons from calc-alkaline rhyolites are indicators for onset of melt and fluid interaction. Combining the modelling results with positive correlation between εNd(t) (+ 5.6 to +6.6) and Zr/Sm, we infer that the potential end-member source of alkali-rich fluids is from mantle alkali-magmas or degassing. Considering the petrogenesis of peralkaline granites and their occurrences in the Uliastai Continental Margin (UCM) distributed along the major strike-slip fault, and temporally marking the final stage of the Late Palaeozoic magmatism, we propose a model for the magmatic transition from calc- to peralkaline magmas in an extensional setting during the culmination of the magmatic event. Subsequently, the calcalkaline melts began to wane, corresponding to the gradual weakening of the geothermal gradient in the lowermost crust. In contrast, the activities of mantle degassed fluids became more intense and migrated through the major fault, metasomatizing crustal rocks or injecting into and interacting with the pre-existing calc-alkaline magmas in the roof of the magma chamber.