Over one million gallons of high–level–waste with more than a million curies of 137Cs have leaked from Hanford tank farms into the underlying vadose zone. As part of an LLNL involvement in the Hanford Science and Technology program, both experimental and modeling studies have been carried out to study cesium migration in the Hanford Sediments. Column studies give strong support to the ion exchange model for cesium proposed by Zachara et al (submitted). The competing NaNO3 concentration has a strong effect on cesium retardation—changing the NaNO3 concentration from 1M to 5M reduces the retardation from about 42 to 7.5 (see Cs breakthrough results). Also important is the fact that two distinct exchange sites are involved: 1) a site with a relatively low concentration but with a high affinity for cesium, and 2) a more abundant site with a relatively lower affinity for cesium than Site 1. These sites are believed to correlate with frayed edge sites (FES) on micas and the surfaces of expandable clays respectively (Zachara et al., submitted). The importance of the high affinity FES sites is highlighted by column experiments carried out at lower cesium concentrations (see third panel). Desorption studies indicate that cesium is largely reversible at higher NaNO3 and cesium concentrations, but partly irreversible at lower cesium concentrations where exchange on high affinity FES sites dominates. All modeling shown was carried out with CRUNCH.
While the column studies demonstrate that the ion exchange model works quite well in the experimental system, it is not clear a priori to what extent the experimental results can be applied to the field. To make a direct comparison of experimental and field results, we simulated the SX-115 tank leak using the OS3D option within CRUNCH and a discretization of the domain consisting of 17 and 15 cells in the X and Y direction respectively along with 34 cells in the vertical direction. A transient flow and temperature field were generated by the code NUFT (Nitao, 1998) using the hydrostratigraphy reported below SX-115. The flow and temperature field are then read into CRUNCH as part of a simulation of multicomponent, multi-site ion exchange. Using CECs and selectivity coefficients determined in the column experiments, the 3D simulations show a distinct retardation of sodium relative to the tracers technetium and nitrate (see a 2D slice). A direct comparison of the simulations with the pore water data from a borehole drilled next to SX-115 shows that the simulations capture the leading edge of the ion exchange fronts remarkably well. This 3D calculation, which included 12 component species, 24 aqueous complexes, and 12 ion exchange reactions, took about 1.5 hours to run on a dual processor Pentium III with a 933 MHz Xeon chip.
Results are partly summarized in an abstract published as part of the Goldschmidt 2001 proceedings (Steefel et al., 2001).
Nitao J (1998) Reference Manual for the NUFT Flow and Transport Code, Version 2.0 Lawrence Livermore National Laboratory, UCRL-MA-130651.
Steefel C.I., Carroll S.A., Lichtner P.C., and Yabusaki S.B. (2001) Evaluation of the field exchange capacity of Hanford sediments with implications for 137Cs migration. Presented at the Goldschmidt 2001. UCRL-JC-142754. Livermore, California: Lawrence Livermore National Laboratory.
Zachara J.M., Liu, McKinley, Serne, Gassman (submitted) Sorption of Cs+ to micaceous subsurface sediments from the Hanford Site, USA. Geochim. Cosmochim. Acta.