A detailed field pilot test was conducted to evaluate the use of edible oil emulsions for enhanced in situ biodegradation of perchlorate and chlorinated solvents in groundwater. Edible oil substrate (EOS®) was injected into a line of ten direct push injection wells over a 2-day period to form a 15-m-long biologically active permeable reactive barrier (bio-barrier). Field monitoring results over a 2.5-year period indicate the oil injection generated strongly reducing conditions in the oil-treated zone with depletion of dissolved oxygen, nitrate, and sulfate, and increases in dissolved iron, manganese and methane. Perchlorate was degraded from 3100 to 20,000 μg/L to below detection (b4 μg/L) in the injection and nearby monitor wells within 5 days following the injection. Two years after the single emulsion injection, perchlorate was less than 6 μg/L in every downgradient well compared to an average upgradient concentration of 13,100 μg/L. Immediately after emulsion injection, there were large shifts in concentrations of chlorinated solvents and degradation products due to injection of clean water, sorption to the oil and adaptation of the in situ microbial community. Approximately 4 months after emulsion injection, concentrations of 1,1,1-trichloroethane (TCA), perchloroethene (PCE), trichloroethene (TCE) and their degradation products appeared to reach a quasi steady-state condition. During the period from 4 to 18 months, TCA was reduced from 30–70 μM to 0.2–4 μM during passage through the bio-barrier. However, 1–9 μM 1,1-dichloroethane (DCA) and 8–14 μM of chloroethane (CA) remained indicating significant amounts of incompletely degraded TCA were discharging from the oil-treated zone. During this same period, PCE and TCE were reduced with concurrent production of 1,2-cis-dichloroethene (cis-DCE). However, very little VC or ethene was produced indicating reductive dechlorination slowed or stopped at cis-DCE. The incomplete removal of TCA, PCE and TCE is likely associated with the short (5–20 days) hydraulic retention time of contaminants in the oil-treated zone. The permeability of the injection wells declined by 39–91% (average = 68%) presumably due to biomass growth and/or gas production. However, non-reactive tracer tests and detailed monitoring of the perchlorate plume demonstrated that the permeability loss did not result in excessive flow bypassing around the bio-barrier. Contaminant transport and degradation within the bio-barrier was simulated using an advection–dispersion–reaction model where biodegradation rate was assumed to be linearly proportional to the residual oil concentration (Soil) and the contaminant concentration. Using this approach, the calibrated model was able to closely match the observed contaminant distribution. The calibrated model was then used to design a full-scale barrier to treat both ClO4 and chlorinated solvents. © 2007 Elsevier B.V. All rights reserved.