Analysis of sources of bulk conductivity change in saturated silica sand after unbuffered TCE oxidation by permanganate

Highlights

• TCE oxidation increases fluid and bulk electrical conductivity in saturated sand.

• PHREEQC was used to model changes in pore fluid conductivity from TCE oxidation.

• Silica sand protonation removed at least 94% of H⁺-related pore fluid conductivity.

• PHREEQC underpredicted pore fluid conductivity by 2.1 to 5.5% in saturated sand.

• Formation factor changes from MnO₂ production were not statistically significant.

Abstract

Time lapse resistivity surveys could potentially improve monitoring of permanganate-based in situ chemical oxidation (ISCO) of organic contaminants such as trichloroethene (TCE) by tracking changes in subsurface conductivity that result from injection of permanganate and oxidation of the contaminant. Bulk conductivity and pore fluid conductivity changes during unbuffered TCE oxidation using permanganate are examined through laboratory measurements and conductivity modeling using PHREEQC in fluid samples and porous media samples containing silica sand. In fluid samples, oxidation of one TCE molecule produces three chloride ions and one proton, resulting in an increase in fluid electrical conductivity despite the loss of two permanganate ions in the reaction. However, in saturated sand samples in which up to 8 mM TCE was oxidized, at least 94% of the fluid conductivity associated with the presence of protons was removed within 3 h of sand contact, most likely through protonation of silanol groups found on the surface of the sand grains. Minor conductivity effects most likely associated with pH-dependent reductive dissolution of manganese dioxide were also observed but not accounted for in pore-fluid conductivity modeling. Unaccounted conductivity effects resulted in an under-calculation of post-reaction pore fluid conductivity of 2.1% to 5.5%. Although small increases in the porous media formation factor resulting from precipitation of manganese dioxide were detected (about 3%), these increases could not be confirmed to be statistically significant. Both injection of permanganate and oxidation of TCE cause increases in bulk conductivity that would be detectable through time-lapse resistivity surveys in field conditions.

Introduction

In situ chemical oxidation (ISCO) using permanganate ion (MnO₄⁻) has become a common method for degrading organic contaminants such as trichloroethene (TCE) in aquifers (Petri et al., 2011, Schnarr et al., 1998, Thomson et al., 2007). As with any chemical remediation technique, ISCO requires physical contact between the injected media and the contaminant. Aquifer heterogeneity can often influence the subsurface movement of MnO₄⁻ and other oxidants, making delivery unpredictable and leading to uneven contaminant removal (Smith et al., 2008). Furthermore, monitoring wells on site may be insufficient in number to adequately track oxidant movement (Halihan et al., 2012).

Electrical resistivity tomography (ERT) offers a possible non-intrusive solution for monitoring MnO₄⁻ fate and transport with improved resolution over monitoring wells alone. This method uses an array of electrodes emplaced either at the ground surface or in boreholes to collect resistivity measurements at multiple locations and depths over a period of time (see for instance Kowalsky et al., 2011, Revil et al., 2013). Most commonly, four electrodes are used at a given time, with two injecting a current and the other two measuring the difference in electrical potential. An inversion algorithm is then used to produce a model of the bulk resistivity (or conductivity) distribution below the electrodes or between the wells in the case of cross-hole resistivity tomography. Reviews of the methodologies used to perform ERT measurements and to invert these measurements can be found in Daily et al. (2004) and Revil et al. (2012).

Time-lapse ERT involves the use of different strategies to measure changes in electrical resistivity over time. One common technique is to repeat resistivity surveys multiple times and separately invert the different snapshots and map the conductivity changes among each snapshot. Alternately, using permanent electrode installations, resistivity can be monitored continuously over time with more measurements. Then, rather than inverting individual snapshots, newer inversion algorithms process the data as a whole, regularizing both over space and time (Karaoulis et al., 2011, Karaoulis et al., 2013). This type of inversion can reduce noise and improve recovery of conductivity changes that are correlated with time. Time-lapse ERT has been used to monitor the movement of salt plumes and also to provide flow and transport information in heterogeneous aquifers (Jardani et al., 2013, Kemna et al., 2002, Kemna et al., 2006, Vanderborght et al., 2005). Because permanganate-based ISCO involves injecting ions into the aquifer, time-lapse ERT could potentially be used to track the fate and transport of the injected oxidizing solution.

Although investigators have yet to use improved time-lapse inversion methods to monitor ISCO, a few researchers have used multiple ERT snapshots with independent inversions to locate changes in subsurface resistivity that could be attributed to MnO₄⁻ injection (Halihan et al., 2012, Harte et al., 2012, Nyquist et al., 1999). While each study was successful in demonstrating changes in aquifer resistivity after MnO₄⁻ injection, certain anomalies appeared that were not explained by the investigators. For instance, after injecting MnO₄⁻ to oxidize TCE, Nyquist et al. (1999) were unable to explain why increased conductivity values were observed in certain locations where MnO₄⁻ could not be detected in groundwater samples. Harte et al. (2012) also observed increased bulk conductivity in areas where MnO₄⁻ could not be detected through groundwater sampling during tetrachloroethene (PCE) oxidation. Additionally, they noted the opposite response in certain locations where they expected MnO₄⁻ to be located but where conductivity did not increase.

While these initial studies of ERT use during ISCO with MnO₄⁻ demonstrated that significant conductivity changes can be observed, additional work is needed before such measurements can be used for more quantitative observations of MnO₄⁻ and contaminant fate and transport. While increased electrode coverage, improved electrode placement, or newer inversion methods would improve the resolution of these measurements, this paper focuses on improving the understanding of subsurface conductivity changes that occur as a result of ISCO when MnO₄⁻ is the oxidant. Specifically, this study examines the changes in bulk conductivity that occur in saturated sand from the addition of MnO₄⁻, the reaction between MnO₄⁻ and TCE, and protonation of the silica sand surface resulting from post-reaction acidity production.