Determination of carbon and oxygen isotopes of geological samples with a complicated matrix: comparison of different analytical methods

Abstract

Several automated on-line carbonate reaction devices coupled with isotope ratio mass spectrometers (IRMS) have been frequently used to determine the isotopic compositions of the carbon and oxygen in carbonates because of their high efficiency and small sample size. However, although the δ¹³C and δ¹⁸O values were measured with good precision with each device, whether the measured data remains consistent between these different analysis methods remains unknown. We report a systematic comparison for the δ¹³C and δ¹⁸O values in geological samples with complicated matrices measured using a Kiel IV and a GasBench II, as well as with dual-inlet isotope ratio mass spectrometry (DI-IRMS). Large variations were observed between the measured δ¹³C and δ¹⁸O values for some loess and lake sediments, with maximum differences of −0.4‰ to +0.3‰ for δ¹³C and −0.5‰ to +0.6‰ for δ¹⁸O relative to the mean values of the three methods. Furthermore, the δ¹³C values obtained using the Kiel IV IRMS method are generally lower than those obtained using the other two methods, whereas the δ¹⁸O values measured by the Kiel IV IRMS method are obviously higher than those obtained using DI-IRMS. These results might have been because of the different effects of the organic matter in the samples analyzed using different methods. However, vacuum roasting did not eliminate this effect and instead increased the variability of the δ¹³C and δ¹⁸O values in some samples. The usefulness of the Kiel IV was limited to the analysis of samples containing small amounts of carbonate because of this pronounced effect. In contrast, the GasBench II might produce data more reliably for these types of samples. Consequently, caution must be exercised when selecting a method for analyzing samples with complicated matrices to minimize any potential error.