Date of Award


Document Type


Degree Name

Master of Science (MS)

Legacy Department

EE&S-Environmental Health Physics


Fjeld, Robert

Committee Member

Kaplan , Daniel

Committee Member

DeVol , Timothy

Committee Member

Coates , John


Distribution coefficients, Kd, were measured for a suite of gamma-emitting radionuclides in uncontaminated soils from the E-Area burial grounds at the Savannah River Site. The objectives of the study were to (1) characterize the spatial variability of Kd, (2) evaluate the effect of spatial variability on vadose zone transport, and (3) develop statistical models for predicting Kd from physical or chemical characteristics.
Equilibrium batch sorption tests were performed on 27 soil samples collected from depths ranging from 11 ft to 100 ft from the BGO-3A core with 241Am, 109Cd, 139Ce, 137Cs, 57Co, 60Co, 203Hg, 85Sr, and 88Y. Distribution coefficients were calculated from solid and aqueous phase concentrations measured by gamma spectroscopy. Graphical and statistical analyses were performed on Kd distributions for the entire data set and for three stratified samples: the upper vadose zone, the lower vadose zone, and the aquifer zone. The effect of Kd variability on vadose zone transport was assessed by comparing calculated transport times for three deterministic transport models to a stochastic calculation of vadose zone transport time. The impact of Kd variability on stochastic vadose zone transport was based on a Monte Carlo analysis utilizing the distributions of Kd for each isotope. Finally, SAS regression was utilized to develop statistical correlation models for predicting Kd from soil characterization parameters.
The Kd values were generally ranked as follows: 88Y >> 57, 60Co > 109Cd > 203Hg > 137Cs >> 85Sr. Insufficient data were generated for 241Am and 139Ce to be included in this trend analysis. Also, Kd values were generally greatest in the aquifer zone followed by the upper vadose zone with the lower vadose zone generally exhibiting the lowest Kd values. Variability in the Kd values generally followed this same trend.
In terms of the distributions of Kd values, when the BGO-3A core was taken as a whole all of the examined isotopes were most closely log-normally distributed. When the core was stratified into the three zones described above, differences in the distributions were noted. In the upper vadose zone, the isotopes with higher Kd values (109Cd, 57,60Co, and 88Y) were most nearly normally distributed while those with lower Kd values were either log-normally distributed or could not be characterized. In the lower vadose zone all of the isotopes were most closely normally distributed. Finally, in the aquifer zone, cadmium and strontium were log-normally distributed while the distribution for yttrium was characterized as normal. The remaining isotopes could not be characterized in the aquifer zone.
Based on the Monte Carlo analysis, it was noted that all of the travel time distributions were log-normal with truncated tails. Transport times for the relative concentration of C/C0 = 0.5 were about three times as long as those measured for the C/C0 = 0.01 level and about one-third the length of the C/C0 = 0.99 travel times for each of the isotopes. It was also observed that isotopes with higher Kd values (specifically americium, cerium, and yttrium) displayed much greater differences between the transportation times required to reach the C/C0 = 0.5 level and the times required to reach the 0.01 and 0.99 levels. These differences decreased in magnitude with decreasing Kd values.
Viable prediction models were developed for 137Cs, 57Co, 60Co, and 85Sr that were capable of accounting for about 70% or more of the variability observed in Kd for those isotopes. Models which could explain only about 55% of Kd variability were developed for 109Cd and 203Hg. These models were more suited as a first approximation for Kd estimation. All of the developed models generally utilized the expected geochemical variables based on the literature. Aluminum, iron, and titanium content were significant factors in the final statistical model for 109Cd Kd values. For 137Cs Kd values, the primary factors affecting sorption were CEC and clay content. For cobalt sorption, aluminum, iron, and titanium content confirmed the dependence on complexing ions. Mercury sorption correlated with CEC, aluminum, and iron content. Finally, CEC was shown to be the most significant factor for modeling strontium sorption.