The sorption of 85Sr2+ in the form of 10−6 M Sr(NO3)2 in synthetic granitic water (SGW), and its desorption using the same radiotracer-free solution, were investigated under dynamic conditions in columns loaded with crushed granitic materials. The goal of study was to evaluate the influence of grain size on the retardation (R) and distribution (Kd) coefficients of the soluble 85Sr2+, as well as on the other transport parameters type of Peclet number (Pe) and hydrodynamic dispersion coefficient (Dd). Pure granitic sample and granitic fissure infill samples were used, crushed and sieved into 4 different grain size from 0.063 to1.25 mm were used. In order to determine migration parameters, the model based on erfc-function was used, assuming reversible equilibrium linear or non-linear (Freundlich) sorption/desorption isotherms. By means of both model approaches, the experimental breakthrough curves were fitted using non-linear regression procedure according to Newton–Raphson method. The obtained results also validated the applicability of the linear equilibrium isotherms of the 85Sr2+ sorption/desorption in the studied systems. It was found that in the case of linear isotherm approach, both retardation and distribution coefficients increased with decreasing grain size. Moreover, their values for fracture infill samples are higher than comparing to granite. Depending on the grain size, the retardation coefficient R varied between 11 and 25 for pure granite and 33–58 between for fissure infill material. These values correspond to distribution coefficients Kd of 2–7 and 9–24 cm3/g, respectively. Consequently, the sorption capacity of crushed rocks also increases with decreasing grain size and are about 2.5-times higher for fracture infill than in pure granite. The values of Dd increase with increasing grain size. Due to inverse proportion, values of Pe number are decreasing.
Crystalline rocks Groundwater Sorption Desorption 85Sr Dynamic conditions Grain size Breakthrough curve Linear and non-linear isotherm Erfc-function model