Date of Award

Spring 1-1-2015

Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Harihar Rajaram

Second Advisor

Russell L. Detwiler

Third Advisor

Roseanna Marie Neupauer

Fourth Advisor

John Crimaldi

Fifth Advisor

Angela R. Bielefeldt


Chemical oxidation of dense nonaqueous-phase liquids (DNAPLs) by permanganate has emerged as an effective remediation strategy in fractured rock. Our objectives in this research were to carry out a sequence of experimental, computational and theoretical tasks aimed at improving current understanding of permanganate oxidation in fractured rock systems, and also develop modeling tools that can be used for preliminary design of oxidation schemes at field sites. Our research focused on both free-phase entrapped DNAPL in variable-aperture fractures and dissolved DNAPL in the rock matrix.

In the first section of our research, we present high-resolution experimental investigations in transparent analog variable-aperture fractures to improve understanding of chemical oxidation of residual entrapped trichloroethylene (TCE) in fractures. Four experiments were performed with different permanganate concentrations, flow rates, and initial TCE phase geometry. The initial aperture field and evolving entrapped-phase geometry were measured quantitatively. We present results on the time-evolution of fracture-scale TCE consumption and DNAPL removal rates for all the experiments.

In the next part of this work, we developed theoretical understanding of the reaction front dynamics in the case of chemical oxidation of aqueous-phase DNAPL within fracture-matrix system, backed up by numerical simulations. We also consider the influence of NOD consumption and contaminant sorption to solid aquifer materials in our models. Based on the results from this task we are able to propose simple strategies for remediation design (e.g. the time needed to degrade DNAPL inside the fracture-matrix system and the permanganate injection pattern) for a given set of conditions.

Our numerical simulations of diffusion with bimolecular reaction in the rock matrix demonstrated a transition in the spatially integrated reaction rate - increasing with time initially, and transitioning to a decrease with time. We developed a general non-dimensionalization of the problem and a perturbation analysis to show that there is always an early time regime where the spatially integrated reaction rate scales as the square root of t rather than 1/square root of t. The duration of this early time regime (where the total reaction rate is kinetically rather than diffusion controlled) is shown to depend on the kinetic rate parameters, diffusion coefficients and initial concentrations of the two species.