Oxygen isotope analysis of olivine by ion microprobe: Matrix effects and applications to a serpentinised dunite

Abstract

In order to resolve inter- and intracrystalline oxygen isotopic heterogeneities in olivine crystals encountered in mantle peridotites, basaltic lavas, chondritic meteorites and metamorphic rocks, in situ techniques such as ion microprobes are needed. Accurate ion microprobe analysis requires not only well-characterised reference materials, but also calibration of the matrix bias for compositional variations within a given mineral.

We investigated matrix bias effects related to Mg/Fe variations in olivine during in situ analysis of oxygen isotopes with sensitive high-resolution ion microprobe (SHRIMP) by analysing chemically homogenous olivine samples with forsterite contents in the range Fo74–Fo100. The isotopic measurements were calibrated against San Carlos olivine (SCO; Fo91). The repeatability achieved for all samples was ±0.21–0.50‰ (standard deviation, SD, at 95% confidence level, c.l.) comparable to that of San Carlos olivine (±0.31–0.48‰, SD at 95% c.l.). A matrix bias up to ~−2.0‰ was observed in olivine with forsterite content above 92 mol%, conversely to what has been reported for Cameca instruments. The relationship between the magnitude of matrix bias and fayalite content (mol%) is described by the quadratic function:

The correction scheme for the matrix bias was applied to chemically zoned olivine crystals from a partly serpentinised dunite from the Archean Nuasahi massif (eastern India). Olivine cores (Fo92) preserve their typical mantle-like signature with a δ18O value of 5.16 ± 0.30‰ (σ at 95% c.l.). During a low temperature stage of serpentinisation, olivine transformed to lizardite1 + brucite + magnetite. Olivine rims (Fo98; δ18O = 1.92 ± 0.60‰, σ at 95% c.l.) and the surrounding lizardite2 (4.87 ± 0.53‰, σ at 95% c.l.), formed during a later stage of rock-fluid interaction, are in isotopic equilibrium at ~405–430 °C, with a fluid having a δ18O of ~5.3–6.9‰. Evolved seawater enriched in 18O by isotopic exchange during infiltration could have been responsible for this later serpentinisation stage observed in the Nuasahi massif. The concomitant analysis of oxygen isotopes at the microscale in both olivine and serpentine represents a powerful tool to constrain the nature and source(s) of serpentinising fluid(s) as well as the temperature of serpentinisation.