Unintentional contaminant transfer from groundwater to the vadose zone during source zone remediation of volatile organic compounds

Highlights

• Remediation of TCE via ISCO or biostimulation causes gas exsolution and ebullition.

• TCE is stripped from the treatment zone for both remediation methods.

• ISCO and biostimulation cause unintended spreading of contaminants.

• Both remediation methods lead to TCE transfer across the water table.

• Biostimulation leads to generation and stripping of VC from the treatment zone.

Abstract

Historical heavy use of chlorinated solvents in conjunction with improper disposal practices and accidental releases has resulted in widespread contamination of soils and groundwater in North America and worldwide. As a result, remediation of chlorinated solvents is required at many sites. For source zone treatment, common remediation strategies include in-situ chemical oxidation (ISCO) using potassium or sodium permanganate, and the enhancement of biodegradation by primary substrate addition. It is well known that these remediation methods tend to generate gas (carbon dioxide (CO₂) in the case of ISCO using permanganate, CO₂ and methane (CH₄) in the case of bioremediation). Vigorous gas generation in the presence of chlorinated solvents, which are categorized as volatile organic contaminants (VOCs), may cause gas exsolution, ebullition and stripping of the contaminants from the treatment zone. This process may lead to unintentional ‘compartment transfer’, whereby VOCs are transported away from the contaminated zone into overlying clean sediments and into the vadose zone. To this extent, benchtop column experiments were conducted to quantify the effect of gas generation during remediation of the common chlorinated solvent trichloroethylene (TCE/C₂Cl₃H). Both ISCO and enhanced bioremediation were considered as treatment methods. Results show that gas exsolution and ebullition occurs for both remediation technologies. Facilitated by ebullition, TCE was transported from the source zone into overlying clean groundwater and was subsequently released into the column headspace. For the case of enhanced bioremediation, the intermediate degradation product vinyl chloride (VC) was also stripped from the treatment zone. The concentrations measured in the headspace of the columns (TCE ∼ 300 ppm in the ISCO column, TCE ∼ 500 ppm and VC ∼ 1380 ppm in the bioremediation column) indicate that substantial transfer of VOCs to the vadose zone is possible. These findings provide direct evidence for the unintended spreading of contaminants as a result of remediation efforts, which can, under some circumstances, result in enhanced risks for soil vapour intrusion.

Introduction

Trichloroethylene (TCE/C₂Cl₃H) has historically been one of the most widely used industrial solvents in North America, popular due to its rapid evaporation rate and low reactivity (Doherty, 2000). Historical heavy use in conjunction with improper disposal practices and accidental releases has resulted in widespread contamination of soils and groundwater with TCE. Conventional pump-and-treat systems, the most widely used strategy for groundwater remediation, can be difficult to apply to TCE-contaminated aquifers (Hoffman, 1993, Mackay and Cherry, 1989). As an alternative, in-situ remediation of TCE and other dense non-aqueous phase liquids (DNAPLs) refers to treatment strategies that are applied in the subsurface, such as in-situ chemical oxidation (ISCO) and enhanced bioremediation.

During ISCO remediation, an oxidant solution is released into the zone of NAPL contamination, whereby carbon-based contaminants are oxidized to carbon dioxide (CO₂). For TCE, the most commonly used oxidant is permanganate, which has the advantage of being effective over a wide pH range, and is relatively stable in the subsurface (Yin and Allen, 1999). The oxidation of TCE proceeds as follows:

$${\mathrm{C}}_2{\mathrm{Cl}}_3\mathrm{H}+2{{\mathrm{MnO}}_4}^{-}➔2{\mathrm{CO}}_2+2{\mathrm{MnO}}_{2\left(\mathrm{s}\right)}+3{\mathrm{Cl}}^{-}+{\mathrm{H}}^{+}$$

In experimental studies and applications at TCE-contaminated sites, permanganate oxidation has been found to be successful at effectively treating TCE contamination (Schnarr et al., 1998, West et al., 1997, Yan and Schwartz, 1999). In particular, permanganate oxidation can be very effective at treating source-zone contamination. However, ISCO also tends to generate CO₂ gas as a result of contaminant destruction and the oxidation of soil organic matter.

Bioremediation of TCE has been thoroughly investigated in field studies after first being demonstrated in soil columns by Wilson and Wilson (1985). It has been studied under both aerobic and anaerobic conditions, whereby microbes can achieve remediation of TCE via reductive dechlorination (e.g. Freedman and Gossett, 1989, Hopkins and McCarty, 1995, McCarty et al., 1998). The transformation of TCE to ethene proceeds via the sequential dechlorination to dichloroethylene (DCE), which has three possible isomers, then vinyl chloride (VC), before ethene (Kästner, 1991). The presence of any of these daughter products is likely on sites undergoing bioremediation via co-metabolism of TCE. Incomplete dechlorination of TCE is unhelpful or even harmful to cleanup efforts, since the daughter products dichloroethylene (DCE) and vinyl chloride (VC) are also regulated toxic compounds.

Due to their ubiquitous nature, microbes capable of mediating the reductive dechlorination of TCE are already present within the microbial community at most sites. However, they may require stimulation in order to effectively remediate a contaminated area; this process is known as enhanced bioremediation or biostimulation (Tyagi et al., 2011). During enhanced bioremediation, selected additives can be provided via injection wells to enhance activity of the native microbial population. Additives can include primary substrates, electron acceptors, or nutrients, depending on the site-specific requirements. Primary substrates that have been found to successfully stimulate the microbial community to co-metabolize chlorinated solvents include acetate (Lee et al., 2012, Rittmann and Seagren, 1994), phenol (Fries et al., 1997, Hopkins et al., 1993), methanol (El Mamouni et al., 2002), lactate (Ellis et al., 2000, Song et al., 2002), and vegetable oil (Hunter, 2002), amongst others. However, enhanced biodegradation via co-metabolism enhances the rate of contaminant degradation but also tends to promote the generation of gases such as CO₂ and under anaerobic conditions also methane (CH₄).

Several recent studies on source zone natural attenuation processes in hydrocarbon contaminated aquifers have demonstrated that naturally occurring biodegradation reactions can lead to substantial gas generation, in particular CO₂ and CH₄ (Amos et al., 2005, Amos and Mayer, 2006a, Jones et al., 2014). The amount of released gas that remains dissolved is dependent on the gas solubility. If the gas solubility is exceeded, gas exsolution will occur, whereby dissolved gases nucleate into gas bubbles or partition into pre-existing gas bubbles (Blicher-Mathiesen et al., 1998, Amos et al., 2005, Jones et al., 2014). Following exsolution, gas bubbles undergo vertical transport driven by buoyancy forces, resulting in the process of ebullition (Amos et al., 2005). Using laboratory column experiments, Amos and Mayer (2006a) confirmed ebullition as the primary gas transport mechanism for biogenically produced CH₄.

Other recent studies have shown that in the presence of a pre-existing discontinuous gas phase, NAPL compounds partition into the gas phase and affect vertical mobilization of disconnected gas clusters (Mumford et al., 2008, Roy and Smith, 2007). For example, Mumford et al. (2009) demonstrated that when a seed gas phase of atmospheric gases trapped due to a fluctuating water table exists above a DNAPL pool, the volatilization of the DNAPL results in vertical growth of discontinuous gas fingers, and can lead to upwards mass transfer of volatile organic contaminants (VOCs) derived from the NAPL source.

Considering that remediation methods such as ISCO and enhanced bioremediation tend to lead to vigorous gas production, even more so than under natural attenuation conditions as investigated by Amos and Mayer (2006a), and that gas production takes place in the presence of VOCs at high dissolved concentrations, it can be hypothesized that these remediation techniques will cause gas exsolution, possibly resulting in ebullition and unintended mass transfer of the VOCs out of the treatment region into the overlying vadose zone. A conceptual model of these processes is presented in Fig. 1.

Observations at field sites undergoing groundwater remediation support this hypothesis. For example, at a TCE-contaminated site in Oregon subjected to enhanced bioremediation, indoor air concentrations of TCE in the basement of a building above the treatment zone were found to be significantly higher than those measured prior to remediation. Additionally, DCE and VC in sub-slab soil vapour were only identified after remediation was initiated (Landau Associates, 2008). At a US EPA Superfund Site in Bozeman, Montana with tetrachloroethylene (PCE) contamination, a pilot project was implemented to investigate the suitability of enhanced bioremediation (Montana Department of Environmental Quality, 2011). Soil vapour monitoring during the project revealed substantially increased concentrations of PCE daughter products in the vadose zone after remediation efforts commenced (Cardno, 2015). Based on empirical evidence of gas transfer to the vadose zone, Baciocchi et al. (2014) suggested to include soil vapour extraction (SVE) systems above the treatment zone during ISCO operations.

Although evidence for stripping and ‘compartment transfer’ of VOCs during groundwater remediation exists, this process has not been quantified previously. To address this knowledge gap, laboratory column experiments were conducted to investigate the potential and nature of VOC stripping from a contaminant treatment zone. In these experiments, the mass transfer of VOCs into a discontinuous gas phase generated as a byproduct of in-situ remediation techniques, and the potential for subsequent transport of the volatiles during ebullition, was investigated for permanganate-based ISCO and enhanced bioremediation via co-metabolism under anaerobic conditions. Specifically, our study addresses the following research questions: 1) Is gas production occurring and if so, is it sufficient to drive gas exsolution and ebullition? 2) To what degree is clean groundwater above the treatment zone affected by the remediation efforts? 3) Is mass transfer of TCE and its degradation products into the column headspace occurring? And 4) Are there characteristic differences between ISCO and enhanced bioremediation in the context of unintentional contaminant redistribution? From the overall perspective of groundwater remediation, this work aims to assess how much movement of contaminant may occur from the treatment zone into the vadose zone, and under what conditions is this likely to be exacerbated.