Lithium distribution and isotopic composition in zircon megacrysts as constraints for their mantle origin and volcanic transfer timescales

Lithium in zircon has attracted considerable attention for identifying magma sources and as a diffusion chronometer, despite the exact diffusion mechanisms remaining controversial. Zircon megacrysts sourced from Quaternary intraplate basaltic volcanoes in the Ratanakiri Volcanic Province of northeastern Cambodia were analyzed here for Li isotopic and elemental concentration profiles to constrain diffusion behavior of Li in zircon megacrysts entrained in high temperature melts. This is complemented by high-spatial resolution U-Pb geochronology, O-Hf isotope compositions and trace element abundances to refine the long-term pre-eruptive history of these zircon megacrysts. Ratanakiri zircon megacrysts yield a Tera-Wasserburg regression U-Pb age of 0.98 +/- 0.02 Ma (n = 615 spot analyses) that is indistinguishable from eruption ages reported for the host basalts, although zircon was unstable in the enclosing magma as evidenced by marginal corrosion and abundant tubular channels within the crystals. Extreme Li-isotopic variations (within -30 to +96% on the delta Li-7(LSVEC) scale) of the zircon megacrysts require kinetic Li fractionation due to diffusive mobilization, which is further supported by offsets in intracrystalline domain boundaries with different Li and Y abundances detected in scanning ion images. Complex abundance and isotopic ratio patterns for Li in zircon megacrysts point to multi-mode Li diffusion that is influenced by variable trace element inventories within individual crystals, which correlate with minor differences in the O-Hf isotopic compositions. The overall tendency of decreasing Li abundances along with increasing delta Li-7 values toward crystal rims indicates diffusion-driven equilibration between zircon and Li-depleted basaltic melt. Diffusion modeling of Li and delta Li-7 on domain-boundaries within crystals show pre-eruptive heating timescales of about 18 days, which is consistent with the magma transfer duration estimated from zircon megacryst settling rate calculations, implying the zircon megacrysts should be transported from their source region to the surface by ascending alkali basaltic magma within few weeks. Nearly one to two orders of magnitude longer Li diffusion model timescales were computed for crystal rim domains. This apparent faster diffusion of Li is attributed to H+ loss into a volatile phase during eruption, providing an additional driving force for Li+ diffusion, which might become more pronounced at the crystal boundary where zircon is exposed to strong chemical gradients in volatile elements due to decompression and concomitant degassing. Lithium isotope compositions for the innermost domains of zircon megacrysts are relatively heavy with delta Li-7 values of +12 to +22%. These values are considered to be less influenced by Li-loss after entrainment by the host basalts than rim compositions. These values in combination with O (delta O-18 = +3.5 to +5.3%) and Hf (epsilon(Hf(0.98 Ma)) = +6.5 to +7.9) isotopic compositions as well as inclusions and internal structures suggest that zircon megacrysts were originally derived from partial melting of a metasomatized lithospheric mantle source altered by fluid released from dissolved carbonates during subduction. Lithium isotopes in Ratanakiri zircon megacrysts provide novel insights into using Li isotopes as a tool to constrain the durations of magma transfer from the mantle to the surface, although Li diffusion mechanisms in zircon need to be further scrutinized. (C) 2021 Elsevier Ltd. All rights reserved.