THE USE OF SAND-TANK EXPERIMENTS TO SIMULATE FIELD CONDITIONS IN SURFACTANT-ENHANCED REMEDIATION OF DNAPL

Since their introduction by Schwille in the early 1980s, sand-tank experiments have become widely used to examine dense, non-aqueous phase liquid (DNAPL) behavior in heterogeneous aquifer materials. Sand-tank experiments should be consistent with conditions measured in the field that provide the motivation for the experiments. We review typical conditions of chlorinated solvent DNAPLs in alluvial geosystems at five sites across the USA, each representing a significantly different hydrogeological environment that we have characterized in considerable detail, and make recommendations on how to conduct relevant sand-tank experiments based upon these field data, particularly with respect to surfactant-enhanced aquifer remediation (SEAR). Among the phenomena to consider are: [1] the wettability of aquifer materials associated with chlorinated solvent wastes and their potential to be trapped in less permeable aquifer sediments beneath aquifer materials of much higher permeability, [2] typical DNAPL saturations measured in-situ in granular aquifer materials, and [3] the permeability fields typically measured at DNAPL sites and the structure of their heterogeneity. Our experience indicates that achievement of high DNAPL recoveries (i.e., > 90%) requires [a] high lateral viscous forces and [b] a surfactant chosen on the basis of rapid microemulsion coalescence as well as high DNAPL solubilization capacity. Numerical simulations of a sand-tank experiment conducted at Sandia National Laboratory are used to show how these two critical issues are incorporated into the SEAR design process to ensure that low-permeability as well as the high-permeability aquifer sediments are swept by the surfactant solution, and how surfactant-inducted mobilization of DNAPL is anticipated and controlled by good engineering design. We demonstrate that, even in the particularly difficult geosystem represented by the Sandia sand tank, it is possible to limit vertical DNAPL migration. More importantly, we show how surfactant flooding can be accomplished without loss of control of DNAPL under typical field conditions.