The research presented was used to develop the surfactant/foam process for aquifer remediation. The surfactant/foam method was designed to address the problem of removing dense non-aqueous phase liquids (DNAPLs) spread throughout a somewhat heterogeneous aquifer. The developmental research included demonstrating the effectiveness of the surfactant/foam process and determining the effects that various parameters had on the technique. Parameters tested included the effects of different contaminants on foam, the effects of varying surfactant solution formulation, the effects of flow rate and slug size, and the effect of temperature on the process. The final design utilized a 4% surfactant solution at its optimal salinity, 11,500 ppm NaCl with trichloroethylene at room temperature and 10,250 ppm NaCl with field DNAPL at 12\sp∘C. Trichloroethylene was used as a representative DNAPL for the laboratory work. Foam was generated with pressure regulated air injection rather than a rate constrained gas injection. This method allowed a high gas flow rate when little foam was present, and a lower, more maintainable, flow rate as foam was generated. The process was tested in several experiments in layered sandpacks in a two-dimensional model. The permeability ratios of the sand layers ranged from 7:1 to 20:1. Packings included both a high permeability sand overlying a low permeability sand and the opposite configuration. The experiments resulted in complete removal of TCE from the sandpacks after one to two pore volumes (PV) of surfactant injection. In contrast, surfactant floods without foam in the same packs required between 13 PV and 28 PV, depending on the sands' arrangement and permeability ratio. The surfactant/foam process was also tested in a larger scale two-dimensional model. Foam was successfully propagated across the eight foot length of the layered sandpack. A field demonstration was then carried out in a fifteen by twenty foot well pattern. The preliminary field results indicate that the surfactant/foam process recovered more DNAPL from the site than had been present initially due to contaminant migration into the well pattern. The final DNAPL saturation at the field site was approximately 0.0003 in the volume swept.