A procedure has been developed to extract and measure the relative concentrations of O2, and 15N14N, as well as the isotopic abundance of these occluded gases, in ice cores. A wet-extraction procedure is used to extract 99.995% of the gas from the ice. The gas is removed from the extraction vessel by helium cryogenic trapping of all its constituents in a stainless-steel sample tube immersed in liquid helium. The elemental O2/N2 ratio is determined by measuring the ratio of masses 16O16O (O2) and 15N14N (N2) simultaneously, using a Finnigan MAT 251 isotope-ratio mass spectrometer. Mass 29 is used because the instrument cannot be configured to measure masses 28 and 32 simultaneously. The δ18O of O2 and δ15N of N2 are also measured on the same ice sample. The bubble composition is compared to a dry-air standard admitted to the reference side of the inlet system. The notation used to report the O2/N2 ratio is the delta notation:

where

Applying these methods both to shallow and to deep ice cores will help us to understand global carbon cycling during the period covered by the ice cores. Changes in the distribution of carbon within the biosphere, atmosphere, oceans, and sediments will produce changes in the O2/N2 ratio of the recent atmosphere. Thus measurements of the O2/N2 ratio in ice will increase our understanding of the distribution of carbon within various reservoirs.

The δ18O of O2 in the atmosphere has been shown to be coupled with the δl8O of the water used during photosynthesis. The δ18Owater of the whole ocean was 1.3‰ heavier during the last glacial maximum, which has been shown to correspond to a 1.3‰ enrichment in the δ180 of atmospheric O2 (Bender and others 1985). Thus the measurements of the δ18O of O2 in ice will give an insight into the coupling between biological and hydrological cycles. Measurements of the δ15N of N2 serve as a fractionation indicator. Any change in the δ15N must be due to some fractionation process, as δ15N cannot have changed over the time-scales in question.

To date, we have analyzed recent (approximately 200 years B.P.) bubble samples from Crête, D10 and Byrd Station ice cores. For the δ(O2/N2) analysis, the range of averaged values from each core was from –0.051 to –7.5‰ relative to present-day air. We believe that the large range is due to separation of O2 relative to N2, either during bubble close-off or during storage of the collected cores. For δ18O of the O2, the average values ranged from +0.37 to +0.71‰. The δ15N results ranged from +0.18 to +0.33‰. The results indicate that there has been some fractionation of the gases, based on the δ15N of the N2 in the bubbles. There is a striking covariance between δ18O and δ15N. The correlation coefficient r = 0.96. The slope of δ15O versus δ15N is 2.2, which is slightly greater than the slope which would be expected if the two isotopes were fractionated in a mass-dependent fashion. Possible explanations for the isotopic fractionation are: (1) the gases were fractionated during close-off in the firn, or (2) the composition of the bubbles has been changed since coring by diffusion through small cracks in the core. Radial sampling of gases in the core is in progress in order to investigate possible gas diffusion from the core during post-coring relaxation of the ice.