The effects of experimental heating and alteration on melt inclusions in magmatic zircon

Abstract

Melt inclusions (MIs) in zircon can preserve information on the evolution of magmas. However, MIs in plutonic zircons are typically devitrified, consisting of multiple phases that must be remelted to obtain a homogeneous glass for reconstruction of melt composition and magma processes. We undertake a systematic investigation of melt inclusions in zircons from a ∼ 3300 Ma xenolith of tonalite gneiss from the Barberton Greenstone belt, a well-studied section of cratonic lithosphere with components dating back to 3500 Ma. To better understand the influence of experimental heating on zircon and MI chemistry, multiple aliquots of zircons were heated in an internally heated pressure vessel at 0.4 GPa and temperatures ranging from 900 to 1200 °C (T_step = 100 °C). Homogeneous MIs in domains with low degrees of radiation damage and isolated from cracks in the zircon were found by examination of >5000 zircons by SEM (CL, BSE). Oxygen isotopes (δ¹⁸O), OH/O ratios, U-Pb isotopes, trace and rare earth element (TREE) concentrations in zircon, along with δ¹⁸O, H₂O contents, and major element compositions in glassy MIs were measured by SIMS and EPMA. Investigated MIs have granitic compositions with 67 to 81 wt% SiO₂. Both heated and unheated host zircon possess statistically identical and uniform δ¹⁸O values of 6.02 ± 0.45 ‰ (2SD), while OH/O ratios systematically decrease with increasing temperature of laboratory heating. Inclusion textures (BSE contrast homogeneity) and composition (H₂O, δ¹⁸O) suggest that experimental heating at 1100 °C was the most successful in recovering initial MI compositions. Inclusions in lower temperature experiments either did not homogenize (900 °C) or are rarely homogenized (1000 °C), while those at higher temperature (1200 °C) are systematically dehydrated. Twenty-two hydrous MIs from the 1100 °C experiment have 3.1–11.5 wt% H₂O and an average δ¹⁸O of 7.5 ± 0.9 ‰. Four zircons have a Δ¹⁸O(MI-Zrn) fractionation inconsistent with equilibration at magmatic temperatures. In general, δ¹⁸O values and TREE concentrations measured in the zircons heated at 1100 °C show consistent behavior with unannealed zircons, indicating these systems are not significantly disturbed on the μm-scale during heating experiments, and support the use of anomalous TREE concentrations/patterns as indicators of alteration. These measurements combined with Δ¹⁸O(MI-Zrn) identify the MI-zircon pairs that are unlikely to represent the melt composition at the time of entrapment. Ti-in-zircon temperatures (αSiO₂ = 1, αTiO₂ = 0.3) and rhyoliteMELTS thermometry of the unaltered MI-zircon pairs return similar temperature ranges of 762–833 °C and 750–865 °C respectively, and combined with the granitic major element compositions, suggest that zircons crystallized relatively late during melt fractionation and entrapped residual evolved melt. More generally, these experiments represent the first direct reconstruction of H₂O contents and oxygen isotopes of Archean melts from zircon-hosted MIs, and the approach described here can be used as a model for evaluating potential alteration during experimental heating and the geologic history for MI-zircon pairs from plutonic rocks. As compared with the TREE concentrations of similar-aged detrital zircons in the Barberton terrane, the tonalite xenolith zircons are distinct and attest to the diversity of magma compositions around 3300 Ma that formed the Barberton basement. Given that no comparable rock with zircons of the same age and TREE chemistry is exposed in the Barberton terrane, this tonalite likely represents an unknown component of the Barberton basement.