A modified procedure for gas-source isotope ratio mass spectrometry: the long-integration dual-inlet (LIDI) methodology and implications for clumped isotope measurements

摘要

RATIONALE: High-precision stable isotope measurements in gas-source isotope ratio mass spectrometry are generally carried out by repeated comparison of the composition of an unknown sample with that of a working gas (WG) through a dual-inlet (DI). Due to the established DI protocols, however, most of the sample gas is wasted rather than measured, which is a major problem when sample size is limited. Here we propose a new methodology allowing the measurement of a much larger portion of the available sample. METHODS: We tested a new measurement protocol, the long-integration dual-inlet (LIDI) method, which consists of a single measurement of the sample for 200 to 600 seconds followed by a single measurement of the WG. The isotope ratios of the sample are calculated by comparison of the beam ratios of the WG and sample at equivalent intensities of the major ion beam. RESULTS: Three isotopically very different CO_2 samples were analyzed. The LIDI measurements of large samples (50 to 100 μmol of CO_2) measured at quasi-constant beam sizes, and of small samples (1.5 to 2 μmol of CO_2) measured in microvolume mode, generated results that are indistinguishable from the standard DI measurements for carbon, oxygen and clumped isotope compositions. The external precision of Δ_(47) using the LIDI protocol (~±0.007‰) is similar to that of the state of the art DI measurements. CONCLUSIONS: For traditional and clumped isotope measurements of CO_2, the LIDI protocol allows the measurement of a much larger portion of the sample gas rather than only ~20 % of it. In addition, the sample can be measured at higher signal intensity and for longer time, allowing the measurement of smaller samples while preserving precision. We suggest that other gases commonly used for stable isotope measurements with gas-source mass spectrometry would also benefit from this new protocol.