Chlorinated solvents can pose a considerable risk to human health, and their presence in groundwater at many Department of Defense (DoD) sites is a significant problem due to the persistence of these chemicals in the environment. Moreover, plumes of dissolved chlorinated solvents in groundwater may not be efficiently treated in a cost-effective manner due to their size, location, and in some cases low contaminant concentration. Many studies have investigated the possibilities of breaking down chlorinated solvents by chemical oxidation using potassium permanganate (KMnO4). Yet, the full potential of such oxidation-based approaches has not been realized, especially for hard to treat plumes. This project investigated cost-effective ways to utilize the in-situ oxidation approach to remediate relatively dilute plumes of chlorinated solvents (e.g., DCE, PCE, TCE) and contaminants in deeper or lower permeability zones relative to more conventional techniques and approaches. A number of key research questions were addressed:

  1. Can hypersaline KMnO4 fluids effectively invade a heterogeneous medium because of density effects and remain sequestered hydrodynamically?
  2. Is it possible to take advantage of the time-delayed gelling property of KMnO4 doped engineered solutions to develop solids in-situ capable of slow KMnO4 release?
  3. How can we predict the fate of these engineered solutions in groundwater?
  4. Are there alternative cost-effective materials suitable for use as slow-release solids in reactive barriers?

Technical Approach

The flow of hypersaline and/or viscous solutions in aquifers and time-delayed in-situ gelling materials were studied using both flow tank experiments and numerical modeling. In the flow tank experiments, different arrangements of glass beads were used to explore the mixing and flow behavior of concentrated sodium chloride (NaCl) solutions and Na-silicate solutions. A variable density flow code, Mitsu 3D, was used to simulate the flow tank experiments and to increase the predictability of the flow patterns of saline and viscous solutions. Furthermore, engineered Na-silicate and colloidal silica solutions were used in flow tank experiments and column experiments to demonstrate the potential of using the in-situ time-delayed gelling property of such solutions. A numerical model was developed to simulate the variable density flow/gelation behavior of the Na-silicate solutions.

A variety of materials were tested in column experiments to find a cost-effective alternative host material that can be used to support the slow release of KMnO4 in reactive barriers. In particular, the slow-release behavior of geopolymers was studied through 1-D column experiments.


One way of utilizing the in-situ oxidation scheme is to inject concentrated treatment fluids into the zone of chlorinated solvents in groundwater. Initial studies tested if hypersaline solutions mimicking concentrated treatment solutions can invade the heterogeneous medium effectively. Specifically, the project team explored the mixing and flow behavior of concentrated NaCl solutions (10 g/L) in flow tank experiments using different arrangements of glass beads. The first experiment with continuous high and low permeability layers showed that some portion of the saline solution flowed away along high permeability layers. Such fast pathways reduce the effectiveness of the delivery of the treatment solutions into the contaminated zone. The other disadvantage of this arrangement was that once injected, highly concentrated solutions sank rapidly to the bottom of the tank without much downward resistance to flow. In the second flow tank experiment, the presence of discontinuous lenses promoted the formation of instabilities to enhance mixing, and slowed down the migration of the tracer plume, thereby increasing the residence time between the potential treatment chemical and the contaminant plume. Another advantage was that solutes that were in contact with the clay lens diffused into the lens and created a slow release system.

The use of controlled-release reactive barriers is another method in the in-situ oxidation scheme. Creating such reactive barriers in-situ requires time-delayed gelling of potential KMnO4 doped solutions. The project team investigated the possibility of creating a MnO4- gel solution by testing the compatibility and reactivity of gels, such as chitosan, aluminosilicate, silicate, and colloidal silica gels, with MnO4-. The team specifically focused on studies of the gelation and MnO4- release characteristics of two promising gels, silicate and colloidal-silica gels. The approach involved a series of batch and flow-through column experiments. Due to their organic character, chitosan gels were found to be incompatible with MnO4-. The aluminosilicate gel was also unsatisfactory because it was not miscible with KMnO4 solution. Silicate gels derived from silicate solutions are compatible with MnO4-, and the viscosity of the resulting silicate gel could be manipulated by the addition of saline MnO4- solution or the addition of salts. Colloidal silica contains nanoparticles of amorphous silicon dioxide and sodium hydroxide suspended in water solution. Colloidal silica solutions are chemically also non-reactive with MnO4-. The results of gelation batch tests further demonstrated the delayed gelling characteristics of colloidal silica. The gelation lag times of colloidal silica increased as KMnO4 concentrations decreased. In porous media, MnO4- release from the gelated slow-release permanganate gel (SRP-G) yielded a similar pattern to that of open water tests, i.e., a rapid initial peak release followed by exponential decay and an asymptotic release phase. The SRP-G solutions that gelated within columns exhibited characteristic transport and release patterns indicating the effects of diffusion in addition to advection and dispersion.

Dense, viscous, and oxidant resistant Na-silicate solutions may be used to deliver KMnO4 to the deep contaminated zones of the aquifers. It was assumed that the density difference between the groundwater and silicate solutions would cause the sinking of silicate solutions without significant horizontal spread of the silicate solution. However, through time the density difference would decrease due to mixing. Using such dense and viscous solutions, Na-silicate solution (N-Clear) containing KMnO4 could be injected into coarse-grained, high conductivity units, but the delivery of this solution had limitations due to its high viscosity. The silicate solutions sank slowly and the concentrations of the viscous solutions decreased significantly over relatively short distances due to the dilution of the solution by mixing. The project team modeled the experiments involving the viscous solution using Mitsu 3D. Modelling of the dense viscous fluid was successful in reproducing the shape of experimental plumes, especially at later times. The successful results of these experiments suggest that silicate solutions can be useful in delivering oxidants to the deeper zones of contaminated aquifers.

The time-dependent gelling property of silicate solutions can be used to create controlled-release reactive barriers in-situ. Silicate solutions are stable at high pHs (i.e., pH > 11), but once the pH of the solution is lowered, the solubility of the silica is reduced and it polymerizes. The results of flow tank experiment illustrate the potential of this approach. The project team prepared a solution by mixing diluted sodium silicate solution (N-clear) with sodium bicarbonate solution that transforms into a soluble silicate gel through time. The initial viscosity of such a mixture is low and the gelation is not immediate (i.e., several hours). These conditions are useful for initial injection of this solution into the targeted layer in the flow tank experiment. The high conductivity layer was flooded with this solution. After 20 days, less than 60% of the initial tracer concentration remained in the gelled silicate material.

Numerical models required to predict the fate of the solutions that have the time-delayed gelation property should simulate the increase in the viscosity depending on the concentration of the silicate solution. The results of the numerical modelling to simulate the time-delayed gelling of silica solutions showed that the simulated behavior of the solution resembles the general features of the laboratory experiments. The calculated gelation times of the solutions closely match the observed gelation times. The numerical model captures not only the concentration profiles but also the shape of the sinking silicate plumes reasonably well. Experiments showed that gelation takes place only at zones with relatively higher solute concentrations. The numerical model generated a concentration profile similar to experiments.

Finally, novel cost-effective controlled-release geopolymers were investigated to be used as reactive barriers. The MnO4- release experiments from KMnO4 doped geopolymer samples showed that permanganate concentrations were initially high due to the high solubility of the KMnO4, but then gradually decreased to lower concentrations. It took about 19 days for samples with higher KMnO4 concentrations to get exhausted. Based on the fractional release data from the column experiments, the anomalous diffusion is the mechanism controlling the KMnO4 diffusion from the geopolymers.


In this project, novel cost-effective ways to utilize the in-situ oxidation approach were investigated as a potential new strategy for remediating relatively dilute plumes of chlorinated solvents and contaminants in deeper or lower permeability zones. The knowledge gathered through this project improved understanding of different aspects of the in-situ oxidation approach such as flow and mixing behavior of saline solutions in fresh water. Further, Na-silicate or colloidal silica solutions were amenable to engineering providing for time-delayed gelation. This concept will allow MnO4- doped silicate solutions to be injected and to spread over relatively large areas and act as slow-release solids following their gelation. The numerical method developed to predict the fate of these complex gelling solutions can also be used in other engineering applications. Finally, geopolymers were utilized as alternative cost-effective slow-release solids since they are inert for chemical oxidation by KMnO4.