Kinetic Release of Alkalinity from Particle-Containing Oil-in-Water Emulsions

Oil-in-water emulsions are typically employed during remediation to promote biotic reduction of contaminants. Emulsions, however, hold promise for encapsulated delivery of many types of active ingredients required for successful site remediation or long-term site stewardship. Our research is currently focused on using alkalinity-containing particles held within oil-in-water emulsions to sustain control of subsurface pH. Here we describe results from laboratory experiments and mathematical modeling conducted to quantify the kinetics associated with the emulsion delivery and alkalinity release process.

Kinetically stable oil-in-water emulsions containing (~60 nmCaCO₃ or ~100 nm MgO particles) were previously developed using soybean oil and Gum Arabic as a stabilizing agent. Batch and column experiments were employed to assess the accessibility and release of the alkalinity from the emulsion. Successive additions of HCl were used in batch systems to produce several pH responses (pH rebounds) that were subsequently modeled to elucidate release mechanisms and rates for varying emulsion compositions and particle types. Initial results suggest that a linear-driving-force model is generally able to capture the release behavior in the batch system when the temporally-constant, lumped mass-transfer coefficient is scaled by the fraction of particle mass remaining within the droplets. This result suggests that the rate limiting step in the release process may be the interphase transfer of reactive species at the oil-water interface.

1-d column experiments were also completed in order to quantify the extent and rate of alkalinity release from emulsion droplets retained in a sandy medium. Alkalinity release from the retained droplets treated a pH 4 influent water for 25-60 pore volumes (the duration depended on particle type and mass loading), and the cessation in treatment corresponded to exhaustion of the particle mass held within the oil. Column experiments were simulated using a transport code containing the linear-driving-force expression evaluated in the batch experiments. In these simulations the lumped mass transfer coefficient was fit and compared with values predicted using existing correlations for liquid-liquid and solid-liquid interfaces in porous media.