A large-scale laboratory experiment to determine the mass transfer of CO 2 from a sandy soil to moving groundwater

Abstract

Gaseous 14CO 2 originating from low-level radioactive wastes, buried in the unsaturated zone, can be released to the environment via two major routes: gas-phase diffusion through soils to the atmosphere, and by dissolution in groundwaters followed by aqueous migration. The fate of this gas is important for the safety of waste management areas and for waste disposal, because the exposure to humans depends upon the aqueous versus atmospheric path. One key parameter that influences the relative importance of these paths is the transfer to groundwaters, expressed as the mass transfer coefficient K L. Due to the nature of the capillary fringe, located just above the water table, the K L can not be calculated and it has to be obtained experimentally. In this study, a large sand bin (3.5×2×1 m) that featured a flowing aquifer was used to determine K L. The means of the experimental values of K L ranged from 1.9 to 3.1×10 −4 m/h, depending upon the pH interface condition. These K L-values are ∼15 to 25 times lower than the mass transfer to a quiescent liquid (4.7×10 −3 m/h). This suggests that the capillary fringe offers a significant resistance to mass transfer, and this should be accounted for in the calculations of the fate of 14CO 2 near waste management areas. This mass transfer coefficient was used to determine the fluxes and the fate of both CO 2 and 14CO 2 gases originating from a low-level radioactive waste management area. The fate of these gases was calculated using a simple box model applied to a hypothetical 1 m 2 plot extending from groundwater to the soil surface. Two values of K L were used to calculate the fluxes from groundwater: the experimental K L, and a `theoretical' K L calculated from the mass transfer to a quiescent liquid, corrected for porosity and tortuosity. Box model calculations predicted that most of the CO 2 emitted to the atmosphere originated from the soil column, regardless of the K L used (experimental or calculated). The same box model using site 14CO 2 data, however, predicted that radiocarbon emissions originated from the contaminated aquifer. In this case, the box model was sensitive to K L. If a calculated rather than experimental K L-value was used, the amount of 14CO 2 loss from groundwater was substantially overestimated.