Zircon micro-inclusions as an obstacle for in situ garnet U-Pb geochronology: An example from the As Sifah eclogite locality, Oman

Abstract. Garnet is commonly used to calculate pressure (P)-temperature (T) histories of metamorphic rocks, as well as to monitor changes in bulk-rock composition (X) and deformation (d). In situ U-Pb geochronology by laser ablation-inductively coupled mass spectrometry (LA-ICPMS) is a rapid and relatively high spatial resolution technique, which can be used to constrain the timing of the metamorphic P–T–X–d histories preserved in garnet. However, the low U contents (low µg/g to ng/g levels) of most common metamorphic garnet crystals presents unique analytical challenges, including potential contamination of the U-Pb system by high-U inclusions, such as zircon, rutile, and monazite. Here we use LA split-stream (SS)-ICPMS analysis to simultaneously measure the U, Th, and Pb isotopes and trace-element contents of eclogite-facies garnet from metamafic rocks at As Sifah, Oman. We observe abundant zircon micro-inclusions (<2 µm) in all five dated samples. Strong linear correlations in U vs Zr contents in the analysed laser-ablation spots plot along garnet–zircon mixing lines, the slopes of which can only be explained by zircon contamination. Despite clear zircon contamination in the trace-element data, the time-resolved laser-ablation U and Pb signals show some irregularities but lack sharp diagnostic spikes typically indicative of inclusions. Instead, zircon micro-inclusions are sufficiently small, abundant, and dispersed over the scale of the laser spot site (193 μm diameter) such that their contribution to the U, Th, and Pb signals is diluted to produce irregular time-resolved signals that have previously not been identified as inclusions.

Analyses affected by contamination result in well-defined U-Pb regression lines that give concordia intercept dates of 94–89 Ma. After screening, only one sample had sufficient inclusion-free analyses and spread in U-Pb ratios to calculate a statistically meaningful date. The calculated concordia intercept date of 71 ± 7 Ma is consistent within uncertainty of previously published garnet–whole rock Sm-Nd peak metamorphic ages. We suggest that the 94–89 Ma ages represent the growth of micro-zircons produced during low-grade metamorphism or hydrothermal alteration of the mafic tuff protolith during the submergence of and sediment deposition on the Arabian margin at this time. To obviate the effect of micro-inclusions in garnet LA-ICPMS U-Pb geochronology, we recommend a careful examination of garnet grains by electron microscopy prior to analysis and determination of background garnet U, Th, Pb contents and Th/U combined with the rejection of analyses with even slight or moderately irregular signals. We also demonstrate that LASS-ICPMS is a powerful tool to screen for inclusion contamination for in situ U-Pb garnet geochronology, providing confidence in the geologic meaning of the resulting ages.