Further development of the absorption method for preparing CO2 samples for radiocarbon analysis by liquid scintillation counting

The CO2 absorption method for preparing samples for radiocarbon analysis by liquid scintillation counting has been successfully employed by several laboratories over many years. The main advantage of the method is its simplicity. Although the method allows for only relatively small sample sizes, implying relatively low figures of merit, adequate accuracy can be attained for a conventional dating limit up to 37 000 years - more than adequate for hydrological applications. The method reported by Aravena, Qureshi et al., and further developed by Nair et al., involves bubbling the CO2 sample gas through a liquid cocktail containing the scintillator and the alkaline absorber to the point of saturation. It relies on maintaining constant conditions to achieve reproducibility. This process ensures that a maximum amount of sample material is loaded into the cocktail. The sample is then transferred to the counting vial. We report here on further improvements which have considerably simplified the preparation method and have improved its accuracy. The CO2 sample gas is expanded from the purification line to standard pressure (610 torr) into a 1 litre measuring volume which is then isolated. 10 ml of Carbo Sorbreg E is pipetted into a counting vial which is attached to the system though a flexible connection. The vial is briefly pumped to remove air. The CO2 sample is then quantitatively frozen into a small trap and pumped to high vacuum to remove residual non-condensable gas. The low-volume (∼60 ml) section of the system is isolated and the CO2 allowed to sublime whilst the vial is shaken and cooled in a water bath. Absorption proceeds to a final pressure of some 40 torr. The counting vial is removed from the vacuum system, 10 ml Permafluorreg E+ added, capped and shaken well before counting. The absorption is not taken to saturation, which sacrifices some sensitivity. The advantage is that it allows for the amount of sample gas to be measured accurately. The pressure in the measuring bulb can be determined to some 0.3%. The gas left at about 40 torr in the low-volume section after absorption, represents 0.5% of the total sample volume. This residual pressure varies by less than 10 torr, implying a reproducibility of some 99.8% from sample to sample. As the absorption stays well below saturation, the method allows for the accurate preparation of smaller samples down to some 50% of the standard sample size without dilution, keeping in mind that the residual gas loss becomes more important. A major advantage is that the Carbo Sorbreg which becomes quite viscous after absorbing CO2, does not have to be transferred from the line to the vial, which avoids sample losses and the possibility of memory effects. The entire preparation can typically be completed within 30 minutes. The method has shown good reproducibility, also for samples of different sizes. The improved and simplified technique has been successfully used in hundreds of analyses, mainly for hydrology, and periodic checks on the method have shown consistent performance. CO2 generated from BaCO3 precipitate from recent ground water was used to prepare four samples at the standard pressure of 610 torr, and one each at 300, 400 and 500 torr respectively. The samples were each given four 4 hour counts. Results of the smaller samples were normalised to full sample pressure. All results expressed in pMC are presented. Good reproducibility was obtained for samples at the standard pressure as well as for those at reduced pressure, validating the normalisation procedure. The standard deviation of the overall mean lies well within the 1α counting errors values for the individual samples