Permeability Reduction and Emulsified Soybean Oil Distribution in Aquifer Sediments: Experimental and Modeling Results

Abstract

Chlorinated aliphatic hydrocarbons (CAH) are among the most common and difficult to treat contaminants in soil and groundwater. The available information suggests that edible oil permeable reactive barriers (PRBs) can be a very cost effective approach for plume control because of their low capital and O&M costs. However important issues need to be addressed to assure its efficiency.

This work will address permeability losses and oil retention along with oil distribution in the subsurface after injection. PRBs design issues will also be dealt with through the implementation of a model to predict soybean oil spatial distribution after injection.

First, pure soybean oil was injected in laboratory columns packed with different materials varying in grain size distribution and clay content. Water permeability was measured before and after oil injection followed by deaired water flush and sediment oil content assessed at the end of the experiment. This preliminary experiment revealed major difficulties related to the injection of pure NAPL soybean oil. Among others: higher residual saturation (more than 20% of the pore volume), high hydraulic gradient during injection (2 order of magnitude increase over initial gradient with water injection).

Because of these problems emulsified soybean oil was considered as an alternative to the injection of NAPL soybean oil. Using a mixture of surfactants (glycerol monooleate and polysorbate 80), a fine and stable soybean oil in water emulsion was prepared. Injection of the emulsion induced very low to moderate permeability losses and low oil retention.

To investigate emulsion transport in the subsurface, long 1-D columns (80 cm long 1 inch in diameter), packed with a fine clayey sand amended with kaolinite were flushed with emulsion while monitoring effluent concentration. The columns were then cut in 10 sections of 8 cm, with each section analyzed for oil content to characterize the oil spatial distribution. A colloidal transport model based on deep bed filtration was successfully fitted to the long columns data. This model was validated by conducting two 3-D sandbox experiments (1.2 m x 0.98 m x 0.98 m) filled with the same fine clayey sand used previously. The first sandbox experiment consisted of one layer. The second experiment contained three layers varying in clay content. Parameters independently estimated from the long columns experiments were used to predict emulsion transport in the sandbox. Experimental results indicate that emulsified soybean oil can be effectively distributed in the sandbox at least 1 m away from the injection well without any observation of buoyancy effect. The mathematical model calibrated using independently estimated parameters yielded a good fit with measured data. This implies that the colloids transport model can be used to model soybean oil emulsion transport in the subsurface.