Triple Oxygen Isotope Compositions of Isotopic Reference Waters: Implications for VSMOW-Scale and VSMOW–SLAP-Scale Δ′17O Calibration

Triple oxygen isotope ratios of waters are commonly reported on the VSMOW–SLAP scale, whereas comparison with theoretical calculations and assessment of instrumental scale distortion require accurate constraints on the measured isotopic compositions of the primary isotopic reference materials (iRMs) themselves. In particular, the measured δ17OVSMOW and Δ′17OVSMOW values for SLAP and its substitute, SLAP2, remain insufficiently constrained and reported values differ among laboratories. Here, we measured triple oxygen isotope ratios for two primary iRMs (VSMOW and SLAP) and twelve secondary iRMs (VSMOW2, SLAP2, GISP, GRESP, USGS45, USGS46, USGS46a, USGS47, USGS48, USGS49, USGS50 and USGS53) using BrF5 fluorination and dual-inlet isotope-ratio mass spectrometry. For SLAP, we obtained δ18OVSMOW = -55.50 ± 0.15‰ and δ17OVSMOW = -29.66 ± 0.08‰, giving Δ′17OVSMOW = 34 ± 6 per meg (all expanded uncertainties, k = 2). The corresponding values for SLAP2 agree within uncertainty. When expressed on the conventional VSMOW–SLAP scale, our Δ′17OVSMOW–SLAP values for all secondary iRMs with available literature values agree with previously published values within uncertainty, confirming interlaboratory comparability on that scale. In contrast, waters with low δ18O values, Δ′17OVSMOW values are systematically higher than the corresponding Δ′17OVSMOW–SLAP values. These results further suggest that accurate determination of the VSMOW-scale isotopic composition of SLAP and/or SLAP2 across laboratories is essential, particularly for low-δ18O samples, for which the difference between VSMOW-scale and VSMOW–SLAP-scale Δ′17O values becomes large, and for meaningful comparison between measured data and theoretical calculations.