High-precision oxygen isotope microanalysis of ferromagnesian minerals by laser-fluorination

Abstract

A laser-assisted fluorination (LF) technique, modified from the previously published Sharp method, is described for the determination of 18 O 16 O in microgram quantities of ferromagnesian silicate minerals. The system employs a 60-W continuous Nd: YAG laser operating at 1064 nm. Samples are reacted with CIF 3 in a 16-position nickel sample holder within a stainless-steel chamber fitted with a sapphire window. The LF system is on-line to the mass spectrometer microinlet. Procedural blanks are almost negligible, typically <0.02 μmol O 2, permitting analysis of samples weighing <20 μg. Measured oxygen isotope ratios are very sensitive to oxygen yield. Low oxygen yields are associated with the fluorination reaction and, for the analysis of very small samples, the O 2CO 2 conversion step. Fluorination yields are related to laser power, beam diameter, quantity of CIF 3 admitted to the sample chamber and beam rastering speed. Optimisation of these conditions allows 100 ± 2% primary yields to be obtained for minerals with different fusion temperatures and melt properties. δ 18O data for analyses having fluorination yields greater than 95% are reproducible to better than ±0.1‰ (1σ). NBS- 30 biotite (quoted δ 18 O-value + 5.10‰ ) independently calibrated against NBS- 19, gives a mean value of +5.03‰ by LF with an overall reproducibility of ±0.08‰ (1σ, n = 82). Analyses of very small samples weighing less than ∼ 750 μg give δ 18O-values that are systematically lower than expected. Replicate single-grain analyses show systematic relationships between sample weight and apparent oxygen yield, with concomitant isotopic fractionation of the δ 13C and δ 18O of CO 2 generated during conversion over hot graphite. Lower oxygen yields, and low δ 13C- and δ 18O-values are an artifact of the Royal Holloway system caused by enhanced CO production in the graphite reactor at low oxygen pressures. Using pure oxygen gas, a detailed study of the oxygen-carbon dioxide conversion reaction at low oxygen pressures reveals a linear relationship between δ 13C and δ 18O that is largely independent of reactor operating conditions. The measured δ 13C-value of CO 2 converted from oxygen released from small (20–750 μg) silicate samples is used to calculate the true δ 18O-value of the silicate, corrected for CO production, from a calibration obtained using pure oxygen gas. This procedure permits δ 18O analysis of samples weighing as little as 20 μg to better than ±0.2‰.