Accuracy and reproducibility of the triple oxygen isotope measurement of silicate micro-samples by laser-fluorination-IRMS

Recent studies showed that the ¹⁷O-excess of plant leaf biogenic silicates (phytoliths) can be used to quantify the atmospheric relative humidity occurring during leaf water transpiration. The ¹⁷O-excess vs ∂¹⁸O signature of phytoliths can also be used to trace back to the signature of leaf water. In a similar way, the signature of lacustrine diatoms is expected to record the signature of the lake water in which they formed. Therefore, the triple oxygen isotope composition of biogenic silicates extracted from well-dated sedimentary cores may bring new insights for past climate and hydrological reconstructions. However, for high time-resolution reconstructions, we need to be able to measure microsamples (300 to 800 µg) of biogenic silica. In another context, the triple oxygen isotope composition of micro-meteorites constitutes an efficient tool to determine their parent-body. In this case too, micro-samples need to be handled.

Here we report the results of new ∂¹⁸O and ∂¹⁷O measurements of macro- and micro-samples of international and laboratory silicate standards (e.g. NBS28 quartz, San Carlos Olivine, Boulangé quartz, MSG phytoliths and PS diatoms). Molecular O₂ is extracted from silica and purified in a laser-fluorination line, passed through a 114°C slush to condense potential interfering gasses and sent to the dual-inlet Isotope Ratio Mass Spectrometer (IRMS) Thermo-Scientific Delta V. In order to get sufficient 34/32 and 33/32 signals for microsamples the O₂ gas is concentrated within the IRMS in an additional auto-cooled 800 ml microvolume tube filled with silica gel. Accuracy and reproducibility of the ∂¹⁸O, ∂¹⁷O and ¹⁷O excess measurements are assessed. Attention is payed to determine the concentration from which O₂ gas yields offsets in ∂¹⁸O, ∂¹⁷O and ¹⁷O-excess are measured and whether these offsets are reproducible and can be corrected for.