Using septum-capped vials with continuous-flow isotope ratio mass spectrometric analysis of atmospheric CO2 for Keeling plot applications

Recent advances in continuous-flow isotope ratio mass spectrometry have made possible automated analysis of δ13C and δ18O in CO2 in air from 10-mL septum-capped vials. With the potential to collect, transport, and analyze a large number of air samples using these small inexpensive vials, the opportunity exists for routine and statistically robust sampling of CO2 for Keeling plot applications. With the above-mentioned objectives in mind, we measured the precision of carbon and oxygen stable isotope ratios in the CO2 of air within septum-capped vials using a Finnigan MAT Deltaplus XL continuous-flow isotope ratio mass spectrometer (CF-IRMS) interfaced to a GasBench II. The standard deviation of 45 δ13CPDB and δ18OVSMOW measurements from five vials filled with identical tank air was 0.08‰ and 0.14‰, respectively. This translated into a measurement precision (standard error; SE) of 0.01‰ δ13CPDB and 0.02‰ δ18OVSMOW. This precision was similar to that obtained with the tank air introduced directly into the GasBench II by bypassing the vials/autosampler (also with a SE = 0.01‰ for both δ13CPDB and δ18OVSMOW), suggesting minimal vial-related effects with no storage time at ambient pressure and CO2 concentration. Further, puncturing the butyl rubber septa during filling just prior to analysis did not affect the precision (SE = 0.01‰ for both δ13CPDB and δ18OVSMOW, n = 45). After nine days of storage in vials with unpunctured septa, neither δ13CPDB nor δ18OVSMOW changed significantly, although the precision decreased to ∼0.04‰ for both δ13CPDB and δ18OVSMOW. A greater decrease in precision was observed for vials with punctured septa (from 0.01‰ to 0.10‰ SE for both δ13CPDB and δ18OVSMOW). Further, whereas vials with punctured septa showed no change in δ18OVSMOW, δ13CPDB decreased by 0.07‰/day (r2 = 0.96). In practice, such storage-related effects could be accounted for based on concurrent changes in reference standards. These results suggest that air may be collected and stored in 10-mL septum-capped vials and analyzed in an automated fashion with the Gas Bench II/Deltaplus with sufficient precision for Keeling plot applications. Further, a relatively high sample throughput of about 4 vials per hour or about 100 vials or 2000 injections per day offers the potential for high precision with large sample sizes. Higher throughput should permit experimental advances that were not possible before. Copyright © 2001 John Wiley & Sons, Ltd.