Abstract
The diffusion of radionuclides into the pore spaces of a rock matrix and the pore properties in fractured crystalline rocks were studied. The work concentrated on the predominant water-conducting fracture system in the host granodiorite of the Kamaishi In Situ Test Site, which consists of fracture fillings and altered granodiorite. Through-diffusion experiments to obtain effective and apparent diffusion coefficients (De and Da, respectively) for Na+, Cs+, HTO, Cl-, and as a function of ionic charge were conducted through the fracture fillings and altered and intact granodiorite. The total porosity ϕ, density, pore-size distribution, and specific surface area of the pores of the rocks were also determined by a water saturation method and Hg porosimetry. The average ϕ is, in the order from highest to lowest, as follows: fracture fillings (5.6%) greater than altered granodiorite (3.2%) greater than intact granodiorite (2.3%), and gradually it decreases into the matrix. The pore sizes of the intact and altered granodiorite range from 10 nm to 200 μm, and the fracture fillings from 50 nm to 200 μm, but almost all pores are found around 0.1 and 200 μm in the fracture fillings. The De values for all species are in the following order: fracture fillings greater than altered granodiorite greater than intact granodiorite, as with the rock porosity. In addition, no effect of ionic charge on De is found. No significant dependence for Da values on the rock porosity is found. The formation factors FF and geometric factors G of the rocks were evaluated by normalizing the free water diffusion coefficient Do for each species. The FF decreased with decreasing rock porosity, and an empirical equation for the rock porosity was derived to be FF = ϕ1.57±0.02. The G values showed a tendency to slightly decrease with decreasing rock porosity, but they were approximately constant (0.12 to 0.19) in this porosity range. This indicates that accessible pores decrease into the rock matrix from the fractures, and diffusion is also reduced into the rock matrix. Furthermore, it was concluded that a simplified model based on FF and Do values is approximately reasonable to predict De.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have