The Development and Application of a New High Precision GC-IRMS Technique for N2O-Free Isotopic Analysis of Atmospheric CO2

Abstract

A new GC-IRMS technique has been developed for isotopic and mixing ratio analysis of atmospheric CO2. The technique offers for the first time, N2O-free, high precision (<0.05 [per mil]) analysis of d13C and d18O from small whole-air samples. On-line GC separation of CO2 and N2O from these small samples is combined with IRMS under elevated ion source pressures. A specialised open split interface is an integral part of the inlet system and ensures a continuous flow of either sample gas or pure helium to the IRMS. The analysis, including all flushing, uses a total of 45 ml of an air sample collected at ambient pressure. Of this, three 0.5 ml aliquots are injected onto the GC column, each providing [approximately] 0.8 nmol CO2 in the IRMS source. At this sample size, d13C precision obtained is at the theoretical shot-noise limit. Demonstrated precisions for d13C, d18O, and CO2 mixing ratio (all measured simultaneously)are 0.02 [per mil], 0.04 [per mil] and 0.4 ppm respectively. The initial results from an inter calibration exercise with Atmospheric Research at the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia achieved the International Atomic Energy Agency (IAEA) target precision for d13C. During this exercise, agreement for d18O and CO2 mixing ratio was outside the IAEA and World Meteorological Organization (WMO) target precisions for these species, however, when the measurement uncertainties of the two laboratories were considered, the differences were not significant. An inter comparison program using air samples collected at Baring Head, New Zealand and Cape Grim, Australia was also established with CSIRO and d13C, d18O and CO2 mixing ratio showed excellent agreement when combined measurement uncertainties were considered. Further inter comparisons with the Carbon Cycle Group at the National Oceanic and Atmospheric Administration Climate Monitoring and Diagnostics Laboratory (NOAA CMDL), the Institute of Arctic and Alpine Research (INSTAAR), and Scripps Institution of Oceanography (SIO) were also established. No significant differences for d13C were observed during these inter comparison programs. Therefore, these preliminary measurements suggest that the current situation between these laboratories for d13C comparisons from whole-air in glass flasks may be improved compared to the 1995 IAEA inter comparison from whole-air in high-pressure cylinders. Following these inter calibration and inter comparison exercises, temporal and spatial variations in the mixing ratio and isotopic composition of atmospheric CO2 were determined over a large region of the Pacific Ocean to demonstrate the successful use of the GC-IRMS technique. Temporal variations were observed at long-term monitoring sites in the Southern Hemisphere (Baring Head, Cape Grim, and Arrival Heights, Ross Island, Antarctica). Seasonal cycles of CO2 mixing ratio and d13C, with amplitudes of [approximately] 1 ppm and [approximately] 0.05 [per mil] respectively, were measured at Baring Head. A decline in d13C of [approximately] -0.1 [per mil]/year was observed at Arrival Heights between 1997 and 1999. Spatial variations in the Pacific Ocean were investigated by shipboard sampling programs between [approximately] 62 degrees S and [approximately] 32 degrees N. These data were consistent with a Southern Ocean sink between [approximately] 43 degrees S and [approximately] 57 degrees S. In addition, inter hemispheric gradients of d13C and CO2 mixing ratio in March and September 1998 were determined and the position and intensity of the SPCZ and ITCZ were important for the strength of these inter hemispheric gradients. Measurements performed during an upper tropospheric flight from New Zealand, to Antarctica show elevated CO2 levels and depleted d13C compared to samples obtained in the marine boundary layer over this region. A small-scale application of the technique measured soil-respired CO2 in a New Zealand Mountain Beech forest from 150 ml sample flasks that were filled to ambient pressure. These measurements determined a difference between the d13C source signature from the young and old trees of [approximately] 0.3 [per mil], which was in the correct direction but of smaller magnitude than that expected. The small sample requirements of the GC-IRMS technique ease sample collection logistics for varied research. Since initial results from an inter calibration exercise with CSIRO obtain the IAEA target precision for d13C and the technique has demonstrated its ability to successfully monitor atmospheric CO2 species from small whole-air samples, without contamination by atmospheric N2O or the use of cryogen, the technique will be a powerful tool in global carbon cycle research.