13C and 18O fractionation effects on open splits and on the ion source in continuous flow isotope ratio mass spectrometry

Measurements of carbon and oxygen isotopes of CO₂ by continuous flow isotope ratio mass spectrometry are widely used in environmental studies and climate change research. Yet, there are remaining problems associated with the reproducibility of measurements, in particular when high precision is required and/or the amount of sample material is limited. Isotopic fractionations in open splits and nonlinear effects occurring in the mass spectrometer due to different sample amounts alter the results. In this study, we discuss the influence and the origin of these two effects and propose procedures for preventing their impact. Fractionation in the open split can be related to diffusion of CO₂ and can lead to shifted δ‐values when measuring a sample gas against a reference gas injected via different open splits. We present a method, where such fractionations can be minimized by adjusting either the position of the capillaries or the flow rates involved or both. The nonlinear peak area dependence of δ¹³C measurements for small sample sizes can be explained by adsorption/desorption processes in the ionization chamber or its vicinity. For constant amplitudes, the magnitude of the nonlinearity only depends on the amount of CO₂ entering the ion source. This nonlinearity can be eliminated by a small additional flux of a conditioning gas fed to the mass spectrometer. The best results were obtained when using carbon monoxide. For the adsorption process in the mass spectrometer we found a fractionation factor of 0.982 ± 0.005 for δ¹³C and 1.002 ± 0.004 for δ¹⁸O. Copyright © 2010 John Wiley & Sons, Ltd.