Improved accuracy and precision of water stable isotope measurements using the direct vapour equilibration method

Rapid Commun Mass Spectrom. 2019 Oct 30;33(20):1613-1622. doi: 10.1002/rcm.8494.

Abstract

Rationale: A method to measure the δ2 H and δ18 O composition of pore water in soil samples using direct vapour equilibration and laser spectrometry was first described in 2008, and was rapidly adopted. Here, we describe an improved setup to measure pore water δ2 H and δ18 O values through direct vapour equilibration with a laser spectrometer, combining a liquid and a vapour mode for water isotope analyses, and resulting in improved accuracy.

Methods: We first tested new gas sampling bags as part of the equilibration protocol. Then, to assess measurement accuracy, vapour samples from equilibrated liquid waters of known isotope composition were measured in the liquid mode of the analyser using the new setup as well as the manufacturer's vapour mode. Various modes of preparing liquid water standards, namely equilibration, nebulisation, and vapourisation, were tested to determine the best calibration in terms of accuracy. Finally, the proposed modified liquid setup was validated by analysing water vapour equilibrated from soil pore water of a known composition.

Results: The δ2 H and δ18 O measurements were found to be more accurate by the modified liquid mode than by the factory-setup vapour mode. The strong and non-linear dependence of measured δ2 H and δ18 O values on H2 O concentration in vapour mode, especially at concentrations equal to the vapour pressure saturation typically found in laboratories, is problematic for corrections. Regarding calibration and standards, the use of two equilibrated liquid water standards was found to best calibrate measurements in the modified liquid setup. Finally, the modified liquid mode setup and its calibration, as described here, were shown to be appropriate for soil pore water analysis.

Conclusions: The proposed modified setup results in more precise δ2 H and δ18 O soil pore water values than the usual protocols. An average standard deviation of 0.04‰ for δ18 O values and 0.3‰ for δ2 H values, based on 228 soil sample analyses, was obtained.