Sorption and fixation of Cs by the upland soils of the US Department of Energy’s Savannah River Site (SRS) have been attributed to micaceous grains consisting mostly of hydroxy-interlayered vermiculite (HIV). Results of experiments to characterize SRS soil samples, to examine aspects of their radiocesium sorption, and to determine how much of their natural Cs is accessible for chemical extraction and isotope dilution are presented in support of mechanistic hypotheses to explain Cs sorption and fixation in HIV grains. The HIV is responsible for most of the soil cation exchange capacity, and concentrations of naturally occurring Cs, Rb, and K in soil samples are closely related to the concentration of HIV. Experiments with 137Cs to examine (1) sorption kinetics, (2) blocking of exchange sites with silver thiourea, and (3) susceptibility of sorbed 137Cs to chemical extraction, support the idea that added Cs is sorbed at different kinds of cation exchange sites in HIV grains. Sites highly selective for Cs but relatively few in number are inferred to exist in interlayer wedge zones within such grains. Little of the naturally occurring Cs in the soil samples was extractable by chemical agents that would remove Cs from ordinary cation-exchange sites and from within non-silicate soil components. Furthermore, most of the natural Cs was inaccessible for isotope dilution under slightly acidic conditions approximating the natural soil environment. These observations support the idea that most of the Cs in these soils has become effectively fixed in the narrower parts of interlayer wedge zones. Control of Cs uptake and fixation by highly Csselective interlayer wedge sites would account for the large distribution coefficients found for 137Cs at the low aqueous Cs concentrations typical of environmental systems and also for the relatively large concentrations of stable Cs in the SRS soils.

K-Ar measurements were used in this study of upland Savannah River Site soils to distinguish between sorbed K and the K remaining in remnants of primary minerals. Study of sorbed K contributes to understanding further the interaction of alkali metals (Cs in particular) with the soils. Primary mineral K and the associated radiogenic Ar were studied to characterize soil mica with respect to its provenance and its relationship to hydroxy-interlayered vermiculite. K-Ar age values of Na-saturated clay fractions from five samples of these soils range in age from 270 to 370 Ma. After a moderate acid treatment (6% HNO3 v/v, ~1 mol dm-3, 3 h, 80°C) of the clay fractions, K-Ar age values (270-325 Ma) were little changed on the whole, but they were more closely grouped near 300 Ma. Earlier work had shown that most of the K in these soils is found in material resistant to moderate acid extraction. The K-Ar age values show that this acid-resistant material is much older than any pedogenic minerals could be, even much older than the sedimentary parent rocks from which the soils were derived. These observations support earlier inferences by others that the K in these well leached soils is largely in remnants of primary muscovite from the parent sediments. Age values near 300 Ma suggest that the muscovite is largely from proximal Piedmont terranes of the Appalachian orogen, where the K-Ar relationship in most micas was set by Alleghanian tectonic processes late in the Paleozoic Era. The structural location of the K within mica, shown by the retention of the associated radiogenic Ar, is in contrast to the sorption-dominated behavior of the Cs and most of the Rb in these soils during pedogenesis. Stronger acid treatment (~6 mol dm-3 HNO3, 3 h, 100°C) extracted substantial fractions of both the K and the radiogenic Ar from bulk-soil portions, indicating destruction of some of the primary mica. K-Ar age values for the sand-rich bulk soils were not useful for this study because the sand contains excess radiogenic Ar, probably in sand-sized vein quartz.