WM’01 Conference, February 25-March 1, 2001, Tucson, AZ TOWARD AN ACCURATE MODEL OF METAL SORPTION IN SOILS Randall T. Cygan, Howard. L. Anderson, Sara E. Arthur, Patrick V. Brady, Carlos F. Jove- Colon, Jian-Jie Liang, Eric R. Lindgren, Malcolm D. Siegel, David M. Teter, Henry R. Westrich, and Pengchu Zhang Sandia National Laboratories, Albuquerque, NM 87185-0750 ABSTRACT Radionuclide transport in soils and groundwaters is routinely evaluated in performance assessment (PA) using simplified conceptual models ( e.g., K D method) to describe radionuclide sorption. However, the K D approach with linear and reversible sorption of metal cations is rarely observed in the field. Inaccuracies of this model are typically addressed by conservativeness in the use of the chemical partitioning parameters, and often result in failed transport predictions or in increased costs for the cleanup of a site. Realistic assessments of radionuclide transport over a wide range of environmental conditions can proceed only from accurate and mechanistic models of the metal sorption process. Our research has recently examined the sorption mechanisms and partition coefficients for Ba 2+ (analog for 226 Ra 2+ ) onto soil minerals (iron oxides and clay phases) using a combination of isothermal sorption/desorption measurements, synchrotron spectroscopic analyses of metal sorbed substrates, and computer molecular modeling simulations. Research goals include 1) evaluation and quantification of the critical mechanisms and geochemical parameters that control the retardation of radionuclides on the sorbing phases in near-field soils, 2) use of atomistic computer simulations to predict radionuclide K D values based on the partitioning of the metal cations between the solution and mineral surface, and 3) identification of the general trends in metal plume length associated with field sites. Results should improve our ability to estimate radionuclide migration at contaminated sites. INTRODUCTION The sorption of metals onto mineral surfaces has significant influence on the estimates of radionuclide transport in the environment. Unfortunately, most current performance assessment (PA) tools use a simplified conceptual models for radionuclide sorption. Linear and reversible sorption ( i.e., K D approach) is rarely observed in soils and groundwaters because of complex geochemical factors that can significantly affect radionuclide transport mechanisms and kinetics (e.g., pH, fluid composition, ionic strength, mineral substrate structure and composition, organic complexation). Conceptual sorption model inaccuracies are typically addressed by added layers of conservativeness such as the use of low K D values taken from default compilations of radionuclide partition coefficients. This approach often results in failed transport predictions, including such recent examples as the vadose zone transport of radioactive cesium at the DOE Hanford site and fracture flow transport of plutonium at the Nevada Test Site. Each failure decreases public confidence in the risk assessment process. Realistic assessments of