Nature of the problem

Non-aqueous phase liquids (NAPLs) and other hydrocarbon residuals can represent long-term sources of groundwater contaminants. Excavation and disposal often represent the most rapid and effective means of removing these source materials from the environment. However, such an approach is not always physically possible, hence in-situ source area management strategies are employed.

In-situ source management strategy

In-situ biogeochemical stabilization (ISBS) represents a potential means of removing NAPL mass, and reducing flux of organic and inorganic constituents of interest (COI) into groundwater. ISBS entails the use of a specifically modified (catalyzed, buffered) solution of sodium permanganate (NaMnO₄) or potassium permanganate (KMnO₄) that is introduced into a targeted source zone suspected to contain residual COI.

As relatively small amounts of oxidant migrate horizontally and vertically through the targeted source area, various bio-geochemical reactions destroy COI present in the dissolved phase. This, in turn, increases the release of COI from NAPLs into the aqueous phase. The more water-soluble, lower molecular-weight constituents (e.g. benzene, naphthalene) are treated/removed at a proportionally higher rate, thus leading to a hardening or so-called “chemical weathering” of the NAPL as it steadily loses its more labile components. This increases the viscosity of the NAPL, resulting in a more stable residual mass.

As a result, the flux of COI released into the dissolved phase is significantly reduced and natural attenuation processes are more easily capable of managing associated plumes (e.g. dissolved manganese is an effective electron acceptor supporting bioremediation processes)^[1].

In the presence of an organic compound (R), MnO₄- reactions yield an oxidized intermediate (RO_X) or CO₂ plus MnO₂ (R + MnO₄- ≤ MnO₂ + CO₂ or RO_X). More importantly, when catalyzed via ISBS treatment, MnO₂ precipitates will physically stabilize NAPL residuals by the formation of catalyzed manganese dioxide so called “crusts” or “shells” at the organic interface. This physical encapsulation decreases the permeability of the aquifer adjacent to the NAPL and further reduces the COI flux.

SEM photograph showing uniform coating of ISBS treated soil. Red arrows are the inside lighter region, blue arrows are the outside darker region and yellow arrows show the dried crust from the addition of RemOx-EC.

Technology development, demonstration and validation

In 1997, basic technology development using non-catalyzed permanganate was conducted in collaboration with the University of Waterloo^[2,3]. Multiple field-scale technology validation efforts were conducted between 2001 and 2003, and full scale implementations began in 2004. Details on some of these efforts are given below.

In 2002, pilot scale field studies were initiated at an operating wood treatment facility in Colorado where NAPL creosote (PAHs) and pentachlorophenol were found over an 11,000 cu. meter volume in consolidated shallow alluvium deposit. Approximately 24,050 gallons of a 3-percent aqueous potassium permanganate (KMnO₄) solution were injected into 13 locations within a defined test area (75 x 95 x 10 feet deep).

Performance monitoring was conducted for six months to evaluate the ability of ISBS to stabilize the free-phase NAPL residuals and enhance the natural attenuation processes by: A) mitigating the migration of NAPL; B) reducing the concentration of COI in the dissolved phase; C) decreasing the mass of NAPL residuals (source reduction); and D) reducing the flux of COI from NAPL residuals (especially true with MnO₂ precipitate). Field data showed rapid and complete stabilization of NAPL. In addition, mass was reduced by 10 to 79 percent, and the flux of COI was reduced by 56 to 99 percent.

In December 2003, full-scale application of the technology at the Denver site was approved by the state (Colorado DPHE) and federal (EPA Region 8) regulators, and subsequently completed in May 2004. A total of 82,553 gallons of KMnO₄ at 30 g/L (3 percent) was added to 44 injection points, and 9,789 gallons were applied to three trenches (90 gal/LF trench), yielding a total 10,495 kg KMnO₄ for 4.5 g KMnO₄ per kg of soil.

There was a discernible decrease over time in the thickness of the NAPL layer only for those monitoring piezometers located within the treatment area. Changes in NAPL thickness were not observed outside of the treated area, suggesting that NAPL migration did not occur. Flux reduction should have a significant beneficial affect, thereby accelerating the contraction (i.e., remediation by natural attenuation) of the dissolved phase plume at this site.

Engineering optimization efforts for other projects showed that the generic ISBS technology was not directly applicable to all sites due to variations in soil chemistry and mineralogy. Subsequent development of catalyzed, buffered ISBS reagents resolved these problems, yielding significant reductions in soil permeability (less than 98 percent) after only seven days treatment^[4]. In addition, the amount of high-molecular weight PAHs leached from the soils after seven days of treatment dropped to 60 percent of that emanating from the control; low molecular weight PAHs decreased by 52 percent, and pentachlorophenol decreased by 96 percent.

Notice the telltale purple color of the trench water due to the KMnO4 (acknowledgement to Orin Remediation Technologies – Madison, Wisc., and GeoTrans, Inc. Denver).

Application issues to consider

Proper selection, subsequent regulatory approval and ultimate field implementation of the ISBS technology require understanding and acceptance of the case for partial mass removal and flux reduction^[5]. To date, two other main issues have also been identified and addressed accordingly:

The first is potential for NAPL mobilization. The ISBS reagents are usually applied at 1 to 5 percent of the calculated oxidant demand, which limits the volume of material introduced into the subsurface. However, at some sites, NAPL residuals are present as phase-separated hydrocarbons existing as discrete pools and ganglia. There is always a potential concern associated with NAPL mobilization (this is especially relevant when treating NAPL with surface active agents such as surfactants or other emulsifiers/oils).

The potential for NAPL mobilization during injection can be limited by first installing them around the edges of the targeted source and working inward. They also can be evaluated in advance of injections by numerical modeling of the effects of injection on hydraulic gradients, and analytical modeling of the effects of this change in gradient on NAPL mobility.

For example, the potential effect of ISBS amendment injections on NAPL mobilization at a site in Florida was analyzed using a simple, one-layer Modflow/MT3D model of the amendment injection array. The model calculations assumed the use of nine injection locations throughout the entire saturated thickness of the target aquifer, at a rate of 3.5 GPM for seven hours at each location.

Hydraulic gradient calculations required to mobilize residual and pooled NAPL were made using the equations presented in Dense Chlorinated Solvents^[6]. These calculations showed that:

If the residual NAPL is in connected fingers, the horizontal hydraulic gradient would have to be doubled from 0.007 to 0.014 ft/ft to start incipient mobilization of NAPL through the coarsest pores, and

If the residual NAPL is in smaller scale residual ganglia or blobs, then an additional horizontal hydraulic gradient greater than 3 ft/ft would be required to start incipient NAPL mobilization.

Also, the ISBS injection volume will create a negligible change in local groundwater concentrations after 40 days or more have elapsed, due to displacing in-situ groundwater. And the expected increase in horizontal hydraulic gradient due to amendment injection, during the brief injection period, is incapable of mobilizing NAPL at this site whether the NAPL is in residual or pooled form.

The second main issue is long-term stability. Common questions include “what is the composition of the crust” and “how long will the crust last?” Analyses of so-called ISBS “crusts” generated with MGP contaminated and other soils have shown that these physical coatings are similar in structure and composition to Birnessite, which is an oxide of manganese and magnesium along with sodium, calcium, potassium and other minerals. Scanning electron microscopy has determined how these crusts form in soil. Preliminary calculations of crust longevity suggest an effective lifetime of several hundred years. However, this may represent an overestimate under some in-situ conditions because the calculations assumed Eh -400mV and a pH of 6 (at which Birnessite is sparingly soluble).

The treatment chemicals are being fed into the injection points in the photograph above.

Conclusions

Comparing ISBS to alternate NAPL technologies, the approach has various potential advantages.

Extraction techniques are limited by the rate of flushing and the rate of partitioning between NAPL and mobile phases, whereas ISBS treats NAPL, soil and groundwater contamination simultaneously and in-situ without the requirement of large volumes of injectant.

In-situ thermal treatment is resource intensive, and conventional chemical oxidation can require the introduction of large volumes of fluids, potentially mobilizing NAPL products. ISBS does not require much infrastructure or the injection of large volumes of oxidant to offset natural oxidant requirements resulting in a reduction in capital and manpower expenditures.

Enhanced bioremediation techniques may not be able to handle the complex mixtures of contaminants typically encountered. Some contaminants require aerobic treatment and others require anaerobic treatment, whereas ISBS will treat all of the typical COIs along with certain heavy metals, in particular divalent cations such as lead, mercury, etc.

Total project costs for full-scale ISBS treatment at the sites described in this paper have ranged from $45 to $80 per cu. meter. Soil and groundwater treatment was achieved rapidly (within two to three months) and without rebound to date (stable for at least five years). PE