| dc.description.abstract | Light nonaqueous phase liquids (LNAPLs) petroleum hydrocarbons (PHCs) subsurface contamination is complex and poses considerable risks to human health and the environment. The PHC contaminated sites are often difficult to access and remediate due to subsequent changes in composition and distribution of bulk LNAPL mass following the release. The cleanup process becomes even more challenging as the released product ages and leaves behind several weathered residual LNAPL blobs and ganglia that are heterogeneously distributed in the soil pores within the smear zone. The selection of a suitable remedial program is based on bench-scale treatability studies, and the duration of the program relies heavily on time-consuming, labor-intensive, and, therefore, expensive monitoring activities. In situ chemical oxidation (ISCO) using persulfate has been increasingly recognized as one of the most feasible and reliable tools for remediation of PHC contaminated sites. Despite this recognition, there is no modeling approach available that can capture a persulfate ISCO remediation system behavior, evaluate its efficiency and effectiveness, and assist with design optimization. Existing models that can estimate the efficiency of an ISCO remediation are greatly dependent on the availability of LNAPL mass, composition, architecture, and interphase mass transfer rate data. In reality, however, it is difficult, if not impossible, to determine many of these parameters at a weathered PHC contaminated site. The focus of the research was on developing a combined bench-scale and numerical modeling framework to assist with describing and capturing the persulfate ISCO system during the remediation activities of heavily weathered LNAPL PHC contaminated sites.
A series of bench-scale experiments were designed and implemented to support the development of a numerical model. Aquifer material was collected from a heavily weathered diesel contaminated site (Site). Since a common feature of the gas chromatography analysis of a heavily weathered PHC is the presence of an unresolved complex mixture (UCM) of components, pseudocomponents F2 and F3, based on PHC fractions, were defined to estimate fundamental kinetic data. Aqueous phase treatability studies were performed using a series of well-mixed batch reactors to provide information on the ability of various persulfate systems (unactivated, citric acid chelated-ferrous activated, and alkaline activated persulfate) to degrade pseudocomponents F2 and F3 detected in the Site groundwater. Two concentrations of persulfate (40 and 80 g.L-1) and two different persulfate activators (alkaline and citric acid chelated ferrous) at a persulfate concentration of 40 g.L-1 were evaluated for the weathered diesel fuel contaminated groundwater. The pH of ~12 in alkaline activation of persulfate resulted in the highest reduction of ~99% LNAPL mass in groundwater. Chemical kinetic parameters of the aqueous treatment studies were estimated using a numerical approach in the modeling of soil columns. The alkaline activated and unactivated persulfate systems at a persulfate concentration of 40 g.L-1 were selected for remediation of the contaminated soil using the flow interruption column method with multiple persulfate injections. The LNAPL mass reduction of ~50% and 80% for unactivated and alkaline activated systems, respectively, were achieved. The observed data were used to develop an explanatory and practically useful model considering dissolution, advective–dispersive transport, and complex oxidation reactions occurring in porous media during ISCO remediation. The simulations showed that after two persulfate injections, the LNAPL mass reduction was limited by the dissolution process, and no further mass reduction occurred upon additional exposure to more persulfate.
A Site-specific action plan was developed to scale up the bench-scale research to a pilot-scale for a demonstration study, where research data and findings in the laboratory were tested and evaluated under field conditions. A pilot area was selected at a historical Site with a weathered diesel fuel contaminated aquifer manifested as residual sources in soil and dissolved phase in groundwater. A comprehensive monitoring infra-structure was installed. In this unique pilot-scale experiment, the selected area was divided into three zones of treatment (TTZ), control (CZ), and a buffer area in between the TTZ and CZ. Two injection episodes were conducted at the pilot area. The unactivated persulfate system (or municipal water at CZ) was introduced into the subsurface of the treatment zone at locations/depths of PHCs impacts identified with laser-induced fluorescence (LIF) equipped with an ultra-violet optical screening tool (UVOST™) to maximize the opportunity of persulfate to degrade the dissolved F2 and F3 pseudocomponents. No injection activity was conducted at the buffer zone. The dissolved PHCs mass flux was monitored downstream of the pilot area across a transverse fence-line at 60 multilevel sampling points pre- and post-injection episodes. In general, the pilot test data indicate that the aqueous mass of F2 and F3 decreased following each injection episode; however, both the TTZ and CZ showed a similar decrease. Considering the lack of persulfate presence during 3-week monitoring in the post-injection samples collected from the CZ, oxidation of F2 and F3 in the CZ area with persulfate seems unlikely. One potential explanation can be that the injection of water caused a displacement of the dissolved phase and perhaps any mobile nonaqueous phase liquid. Another likely reason is that by injecting uncontaminated water into this area, oxygen was also supplemented to the subsurface and encouraged microbial activities. Generally, it is believed that oxygen is present within the smear zone, and given the age of this contaminated Site (~70 years), it is expected that aerobic bioremediation had long reached its operational limits. However, the latter explanation is in agreement with the results of groundwater treatment tests in control batch reactors in the absence of persulfate. The soil samples were collected and analyzed pre-injection activities, and after completion of the injection, Episode 2 corresponding to nearby LIF elevated response data. Since the places that persulfate can reach during injection are mainly unknown, the post-injection soil samples were collected from ~0.5 m above to ~0.5 m below the depths that pre-injection core samples were collected. The values in the post-injection column are the average of concentrations F2 and F3 in the soil samples that were collected within the intervals mentioned above. Overall, collected soil samples within the TTZ showed a decrease of ~50% in F2 and F3. Samples collected from CZ showed higher concentrations of F2 and F3 compared to pre-injection samples. This observation is not uncommon in the field because of the heterogeneity in contaminant distributions. However, since the overall 50% decrease in contaminant in soil was observed in TTZ, it can be concluded that the two injection episodes at the pilot-scale area were successful and generally in agreement with model simulation results. | en |
