Comparison of laboratory methodologies for evaluating radiostrontium diffusion in soils: planar-source versus half-cell methods.

A planar-source method, initially designed to obtain diffusion coefficients in compacted clay, is adapted here to determine the apparent diffusion coefficient (D(a)) of radiostrontium in soils representative of the Spanish territory. Experiments were carried out by varying the moisture content (F(moist)), and bulk dry density (ρ(bulk)) of the soil samples, in order to study the influence of these soil packing parameters on D(a) values. The moisture in the soil samples was established as the percentage of occupancy of each soil's field capacity (OFC). For a similar OFC, D(a) values in the examined soils ranged by approximately one order of magnitude (e.g. from 6.2 × 10(-)(11) to 6.5 × 10(-)(12)m(2)s(-)(1), at 100% of OFC; from 3.0 × 10(-)(11) to 3.8 × 10(-)(12)m(2)s(-)(1), at 60% of OFC). For a given soil, D(a) values increased when water content was increased. F(moist), and tortuosity (τ) explained D(a) variability, with R(2) values usually over 0.9. However, no good simple or multiple regressions between the soil packing parameters and D(a) were obtained with the whole dataset of all soils, which indicated that soil sorption capacity affects the diffusion of reactive radionuclides in soils. The inclusion of calculated K(d) values in the multiple regressions improved the correlations in all cases. Finally, D(a) values were compared with those obtained by the application of a half-cell method. The values of D(a) obtained by the planar-source methods were systematically lower than the half-cell ones, with a good correlation between the D(a) derived from both methods (R(2)=0.98).

Diffusion experiments for estimating radiocesium and radiostrontium sorption in unsaturated soils from Spain: comparison with batch sorption data.

As sorption data obtained from batch tests are often used to estimate pollutant transport in unsaturated soils, comparison between sorption data obtained in the two conditions is required to ensure a correct risk assessment. With this aim, radiostrontium and radiocesium apparent diffusion coefficients (D(a)) were quantified in nine unsaturated soils, and the derived distribution coefficients (K(d)) were compared with K(d) data from batch experiments. The D(a)(Sr) and the D(a)(Cs) ranged from 1.8x10(-11) to 1.5x10(-10) m(2) s(-1), and from 1.0x10(-13) to 5.9x10(-11) m(2) s(-1), respectively. The D(a)(Sr) varied according to both soil packing parameters and properties governing Sr interaction. For Cs, the soil sorption properties explained the variation on D(a)(Cs). The K(d) values derived from D(a) (from 0.014 to 1.8 L kg(-1) for Sr; from 0.55 to 942 L kg(-1) for Cs) were lower than from batch tests (from 1 to 97 L kg(-1) for Sr; from 10 to 14,600 L kg(-1) for Cs), thus indicating that batch data may not accurately describe radionuclide transport in unsaturated soils. However, the two sets of data correlated well, thus suggesting that radionuclide transport can be estimated from batch tests, which are faster than diffusion experiments.

Laboratory experiments to characterize radiochloride diffusion in unsaturated soils.

Diffusion transport of (36)Cl was examined in seven soils under unsaturated conditions in tubes packed with two portions of each soil having different (36)Cl activity concentrations. Apparent diffusion coefficients (D(a)) derived from diffusion profiles varied within a narrow range (from 3x10(-10) to 7x10(-10) m(2) s(-1)) confirming the minor effect of soil properties on the diffusion of a non-reactive radionuclide like (36)Cl. Instead, packing conditions had a major effect. Solid-liquid distribution coefficients (K(d)) derived from D(a) (0.02-0.2 L kg(-1)) were systematically lower than those obtained from batch experiments (0.6-1.0 L kg(-1)), but with a similar variation pattern among soils. The low values of K(d) (Cl) confirmed an almost negligible radiochloride-soil interaction.