Transformation of tetrachloroethylene in a flow-through electrochemical reactor

Abstract

Electrochemical transformation of harmful tetrachloroethylene (PCE) is evaluated as a method for management of groundwater plumes to protect the drinking water resource, its consumers and the environment. In contrast to previous work that reported transformation of trichloroethylene, a byproduct of PCE, this work focuses on transformation of PCE in a saturated porous matrix and the influence of design parameters on the removal performance. Design parameters investigated were electrode configuration, catalyst load, electrode spacing, current intensity, orientation of reactor and flow through a porous matrix. A removal of 86% was reached in the fully liquid-filled, horizontally oriented reactor at a current of 120 mA across a cathode → bipolar electrode → anode arrangement with a Darcy velocity of 0.03 cm/min (150 m/yr). The palladium load on the cathode significantly influenced the removal. Enhanced removal was observed with increased electrode spacing. Presence of an inert porous matrix improved PCE removal by 9%-point compared to a completely liquid-filled reactor. Normalization of the data indicated, that a higher charge transfer per contaminant mass is required for removal of low PCE concentrations. No chlorinated intermediates were formed. The results suggest, that PCE can be electrochemically transformed in reactor designs replicating that of a potential field-implementation. Further work is required to better understand the reduction and oxidation processes established and the parameters influencing such. This knowledge is essential for optimization towards testing in complex conditions and variations of contaminated sites.

1. Introduction

Chlorinated ethenes have been unintentionally introduced to the environment causing extensive and harmful contamination of the subsurface, potentially impacting the groundwater and indoor air quality [1,2]. Effort has been put into remediation of low-permeable zones, where use of electrokinetics has shown to overcome some of the challenges faced [3–7]. However, the dense chlorinated ethenes slowly leak into aquifers, causing an increase in detection above regulatory maximum contaminant levels [8]. Current remedies used for chlorinated ethene plumes include pump-and-treat [9,10], enhanced reductive dechlorination [1,10], in-situ chemical oxidation [11,12], permeable reactive barriers [13,14] and air-sparging [15,16]. These technologies can be challenged by inadequate delivery of reactants [12,17,18], short longevity of reactants [12,19,20] or varying contaminant phase transfer affinities with generation of contaminated waste [16,21]. Challenges, which may be overcome by adaptation of the mechanisms of electrokinetics towards mass flux reduction in these high-permeable zones by establishment of electrochemical zones. This is gaining increasing focus, since reducing and oxidizing zones sufficient for degradation of the chlorinated ethenes can be established [22–25]. The reducing zone is generated from electrolysis at the cathode (Eq. 1) [26,27], which can directly and indirectly reduce the parent components tetrachloroethylene (PCE) and trichloroethylene (TCE) and stimulate microbial degradation [28–31]. In addition, hydrogen peroxide (H₂O₂), which is a strong oxidant that can break down the chlorinated ethenes [32,33], can be effectively formed (Eq. 2) at a sufficiently high dissolved oxygen concentration, or if a catalyst like palladium (Pd) is applied to the cathode. Furthermore, Pd can catalyze formation of •OH radicals (Eq. 3), which also oxidize the chlorinated ethenes [34].

$${\mathrm{H}}_2\mathrm{O}+2{\mathrm{e}}^{-}\to {\mathrm{H}}_2+2{\mathrm{OH}}^{-}$$ (Eq. 1)

$$2{\mathrm{H}}^{+}+{\mathrm{O}}_2+2{\mathrm{e}}^{-}\stackrel{\mathit{Pd}}{\leftrightarrow }{\mathrm{H}}_2{\mathrm{O}}_2$$ (Eq. 2)

$${\mathrm{H}}_2{\mathrm{O}}_2\stackrel{\mathit{Pd}}{\leftrightarrow }2•\mathrm{OH}$$ (Eq. 3)

The anode generates oxidizing conditions (Eq. 4) [26], specifically by Ti/IrO₂-Ta₂O₅ mixed metal oxide (MMO) anodes (Eq. 5) [35], which can oxidize e.g. TCE and the chlorinated intermediates dichloroethylene (DCE) and vinyl chloride (VC). However, the oxygen evolution competes with oxidation of the chlorinated ethenes.

$$2{\mathrm{H}}_2\mathrm{O}\to {\mathrm{O}}_2+4{\mathrm{H}}^{+}+4{\mathrm{e}}^{-}$$ (Eq. 4)

$${\mathrm{IrO}}_3+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{IrO}}_2+{\mathrm{O}}_2+2{\mathrm{H}}^{+}+2{\mathrm{e}}^{-}$$ (Eq. 5)

Research on electrochemical removal of chlorinated ethenes has focused on TCE [22,23,39–44,24,28,32,34–38]. Since TCE is more volatile than PCE, mass balances of TCE can be particularly poor. Also, the reported optimal design parameters are dependent on the specific system studied, and little knowledge is available on the specific mechanisms responsible for degradation of PCE and TCE in flow-through electrochemical reactors. This makes it complex to adapt the current level of knowledge on electrochemical removal of TCE towards complete removal of PCE. The reduction mechanisms have been investigated on iron metal (Fe(0)) surfaces, showing higher dechlorination rates with increased halogenation [45–47], but also the opposite trend [48–50]. Arnold and Roberts [49] found, that 87% of PCE and 97% of TCE follow direct reduction on Fe(0), i.e. contact between the surface of Fe(0) and the chlorinated ethene is essential. During direct reduction on Fe(0) surfaces, the reaction rate constant for PCE is higher than that for TCE [51]. While during indirect reduction via atomic hydrogen formed from electrolysis (Eq. 1) [51], the reaction rate constant for PCE is lower than that for TCE. Thus, a smaller fraction of PCE than TCE follows the direct reduction pathway, but reaction kinetics is faster during direct reduction of PCE than TCE. The rate limiting steps for direct and indirect reduction are chemisorption onto the electrode, electron transfer, bond-breaking or molecular rearrangement [52].

Competing reactions, due to presence of ionic species in the electrolyte, can interfere with the chemical mechanisms responsible for removal of the chlorinated ethenes, which may be non-linearly distributed, and time and space dependent. This may explain observed variations over time, the course of reactors and between batch and column set-ups [51,53]. In addition, electrode compositions and impurities influence the reduction processes [49,54,55]. For development of flexible full-scale remediation designs that can overcome the complexity and variations of contaminated sites, improved understanding of the transformation mechanisms and parameters controlling those is essential.

The objective of this study is to examine the influence of reactor design on the electrochemical removal of PCE in a flow-through reactor using palladized Fe foam cathodes and MMO anodes. The study is a screening of the significance of parameters relevant to a field design on the electrochemical PCE removal performance, e.g. a porous matrix and horizontal orientation. Such knowledge is required for development of a more robust test design for enhanced applicability and feasibility testing at conditions simulating that of potential in-situ applications. Previous work has studied electrochemical removal of mainly TCE in static reactors [37,56], stirred reactors [22,57], recirculated reactors [23,35,40], vertical flow-through reactors [24,34,41–44,56,58], applied constant voltage [39,59,60] or studied electrochemistry in combination with other remediation strategies, e.g. bioremediation [29,61–64]. Assessment of the potential of flow-through electrochemical transformation of PCE is important to cover the full range of chlorinated ethene species detected at contaminated sites. This study differs from previous work by studying the electrochemical removal of PCE in horizontally oriented undivided reactors with flow through a porous matrix and operated at constant current. Studying such a reactor design is a necessary step towards upscaling of electrochemical remediation for treatment of contaminated groundwater. In the field, the primary groundwater flow direction is horizontal, flow is through a porous matrix, and the charge transfer is best controlled with application of constant current.

2. Materials and methods

2.1. Chemicals and materials

Chemicals used were of analytical grade: PCE (99%, Sigma-Aldrich), TCE (99.5%, Sigma-Aldrich), cis-DCE (97%, Sigma-Aldrich), calcium sulfate (JT Baker), hydrochloric acid (HCl) (37%, Carolina) and palladium(II) chloride (PdCl₂, 99.9%, Alfa Aesar). MMO meshes (1.8 mm in width, 3N International), composed of IrO₂ and Ta₂O₅ coating on titanium were used as bipolar electrodes (BPE) and anodes. Cathodes of Fe foam (3 mm in width, 45 pores per inch, 98% Fe, 2% Ni, Aibixi Ltd.) and/or 316 stainless steel discs (SS, 0.7 mm in width, 0.4 mm wire diameter, 20 x 20 mesh size, 0.86 mm opening size, 12% Ni, McMaster-Carr) were used. The diameter of electrodes used was 77 mm. The 316 SS cathode was assembled using four discs to increase the width. Fe foam cathodes were perforated with 4 mm holes for every 10 mm and coated with Pd. Prior to electroless plating with Pd, the Fe foam was rinsed in 1 M HCl for 10 minutes, thoroughly washed in deionized water, submerged in a solution of 0.1 M HCl and PdCl₂, rotated until the orange solution turned colorless and washed with deionized water [41,42]. Solid glass beads (5 mm, 2.23 g/cm³) were purchased from Propper Manufacturing Co. Inc. Nylon meshes (0.1 mm in width, 0.12 mm wire diameter, 61 x 61 mesh size, 0.3 mm opening size, McMaster-Carr) were mounted on each side of the electrodes to prevent glass beads in blocking openings of the electrode meshes. Feedstock solutions were stored in Tedlar® bags (Sigma-Aldrich) and prepared using 18.2 MΩcm high-purity water from a Millipore Milli-Q system. Tygon® tubing (2.06 mm ID, 3.79 mm OD, Cole-Parmer) and a peristaltic pump (0.4-85 ml/min, Cole-Parmer), were used to control the flow of liquid. Constant current was applied using one Agilent E3612A 0-120V, 0-0.25 A power supply.

2.2. Analytical methods

Concentrations of PCE, TCE and cis-DCE were measured in 2 ml samples by an Agilent 1200 Infinity Series HPLC with a 1260 DAD detector and a Thermo ODS Hypersil C18 column (4.6 x 50 mm) as described in Rajic et al. [41]. Release of Pd ions was analyzed using a Varian ICP-EOS 720-ES. Samples were filtrated prior to analysis using 0.2 μm PVDF syringe filters from Jin Teng. Conductivity of the feedstock solutions was measured using a Thermo Orion 5 Star meter with corresponding probe. pH of the feedstock solutions was measured using a Mettler Toledo Seven Compact meter with corresponding probe. pH of samples collected from sampling ports was assessed using EMD Millipore™ MColorpHast™ pH test strips.

2.3. Electrochemical flow-through reactor

Synthetic groundwater simulating the conductivity of a sandy aquifer was prepared by dissolving 1600 mg/l calcium sulfate in deionized water. Calcium sulfate was used instead of calcium carbonate to limit the influence of the carbonate system on the electrochemical mechanisms and pH gradients established [65]. The resulting conductivity was 700-890 μS/cm with a pH of 5.7 ± 0.4 and a constant temperature of 20°C. The synthetic groundwater was spiked with PCE to a concentration of 2.7 mg/l. A saturated stock solution of free-phase PCE in deionized water was used. The feedstock solution was stored in Tedlar® bags with minimal headspace to limit mass loss by volatilization. Flow through the reactor was kept constant at velocities reported in Table 1.

Table 1.

Experimental conditions for the electrochemical treatment of initially 2.7 mg/l PCE.

Electrode configurationa		Electrode material	Electrode spacing [cm]	Column orientation	Current [mA]	Pd catalyst loadingb [mg/cm2]	Darcy velocity [cm/min]	Porosity, water-filledc [%]
0d	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	0	0.76	0.09	31.1 ± 0.3
1	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	60	0.76	0.09	30.8
2	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	120	0.76	0.09	30.8
3d	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	120	1.14	0.08	34.8 ± 0.5
4	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	180	1.14	0.08	35.3
5	C → BPE → A	Fe foam → MMO → MMO	2.5	Horizontal	120	1.14	0.08	34.1
6	C → BPE → A	Fe foam → MMO → MMO	4	Horizontal	120	1.14	0.03	100
7	C → BPE → A	Fe foam → MMO → MMO	4	Vertical	120	1.14	0.03	100
8	C → C → A	SS → Fe foam →MMO	4	Horizontal	120	1.14	0.08	34.2
9	C → C → A	Fe foam → Fe foam → MMO	4	Horizontal	120	1.14	0.08	35.0
10	C → A	Fe foam → MMO	8	Horizontal	120	1.14	0.08	34.3
11	C → A	Fe foam → MMO	2.5	Horizontal	120	1.14	0.25	100
12	C → A	Fe foam → MMO	2.5	Vertical	120	1.14	0.25	100

^aA: Anode, C: Cathode, BPE: Bipolar electrode

^bCatalyst applied on Fe foam only. Load is given as mass per geometric electrode surface area

^cPorosities of 100 % indicate a fully liquid-filled reactor, while tests with porosities < 100 % incorporate the porous matrix

^dDuplicate testing

An acrylic column was used as the flow-through electrochemical reactor (L 16.8 cm and D 7.8 cm in inner dimensions). The influence of reactor design on electrochemical removal of PCE was investigated in a sequence of a) electrode configurations and spacings. For three-electrode configurations, the constant current was split equally between the two cathodes by a rheostat, b) current intensities and Pd catalyst loadings, c) porous matrices and reactor orientations. The inert porous matrix was incorporated to replicate field-conditions and to reduce amount of contaminated waste generated, and d) groundwater flow and test duration (Table 1).

The position of electrodes and sampling ports (P0-P5) in the electrochemical reactor is shown in Figure 1. In horizontally orientated reactors, electrolytically generated gases were extracted manually from sampling ports above the electrodes using syringes with needles. To limit mass loss by volatilization, the gas extraction occurred whenever a gas phase was observed. The frequency of this manual venting depended on current intensity, electrode configuration and electrode material with a range of once every 30-90 seconds. In vertically oriented reactors, released gases moved up through the reactor and out with the effluent.

Figure 1.

The baseline flow-through electrochemical set-up composed of a feedstock solution, peristaltic pump, a cathode, bipolar electrode and an anode connected to a power supply. P0-P5 are sampling ports. Sampling ports above the electrodes served for manual extraction of gases generated. Measures are of inner dimensions.

Prior to the electrochemical application, the reactor was saturated with the feedstock solution at elevated flow to limit volatilization, followed by flow-through for one pore volume (PV) at the flow velocities given in Table 1 to ensure a uniform contaminant distribution. Direct current was applied for 3.5 PV and 1.2 PV for flow through the porous matrix (Table 1, test 1-5, 8-10) and completely liquid-filled reactor (Table 1, test 6-7, 11-12), respectively.

3. Results and discussion

Different designs of the electrochemical reactor were tested. To evaluate the influence of a design parameters on the PCE removal, each experiment conducted is compared to similar ones of which one parameter varied only. Accordingly, the significance of electrode spacing, electrode configuration, current intensity, Pd loading on the cathode, duration of electrochemical treatment, orientation of reactor and porosity on PCE removal, potential difference over the electrodes and pH downstream of the electrochemical zone is assessable (Figure 2). The design parameter category, e.g. electrode spacing, is indicated above comparable experiments, whereas the specific dimension of that design parameter, e.g. 2.5 cm and 4 cm, is given on the horizontal axis. Figures written on the bars refer to the experimental numbers given in Table 1.

Figure 2.

Comparison of the experimental conditions tested on a) PCE removal obtained in sampling port P5, b) potential difference over the reactor measured after app. 1 PV (was rather constant), and c) pH measured in sampling port P5. Error bars are included for duplicate testing. * denotes direct reuse of the experimental set-up. Figures on the bars refer to experiment no. in Table 1.

3.1. Impact of electrode configuration and spacing on PCE removal

Influence of electrode configuration and spacing on electrochemical removal of PCE was investigated in the flow-through reactor with Fe and MMO electrodes at 120 mA and a Darcy velocity of 0.08 cm/min (experiments 3,5,8-10, Figure 2a, Table 1).

A three-electrode configuration of two Pd coated Fe foam cathodes and an MMO anode performed similar to the corresponding two-electrode configuration, reaching removals of 42% and 39%, respectively (exp. 9-10, Figure 2a). Rajic et al. [41] also observed comparable removals of TCE in two- and three-electrode configured reactors using Fe and MMO electrodes, while the reaction rate nearly doubled in the three-electrode arrangement. By replacing the first Pd/Fe foam cathode with 316 SS, the removal increased to 51% (exp. 8, Figure 2a). The high content of Ni in this type of stainless steel can catalyze reduction, since Ni like Pd lowers the cathodic hydrogen overpotential [55,66], enhancing the hydrodechlorination. Use of Ni-containing stainless steels for electrochemical removal of chlorinated ethenes has shown to be cost-efficient [55]. The best performing electrode configuration was C → BPE → A with a removal of 54±3% (exp. 3, Figure 2a). By introducing a BPE, a supplementary oxidizing-reducing zone is established, enhancing the hydrodechlorination and oxidation by increased formation of H₂O₂ (Eq. 2) [34]. Of the electrode configurations and materials tested, the lowest potential difference of 17V was obtained in the 316 SS → Pd/Fe → MMO arrangement, and the highest potential difference of 36V in the Pd/Fe → MMO reactor (exp. 8 and 10, Figure 2b). The BPE lowered the potential difference from 36V to 32V (exp. 3 and 10, Figure 2b); a simple measure to potentially reduce the costs of power consumption in upscaling activities.

Increasing the electrode spacing from 2.5 cm to 4 cm in the C → BPE → A arrangement improved the PCE removal from 41% to 54±3% (exp. 3 and 5, Figure 2a). This is opposite to previous findings in completely liquid-filled reactors, where the observed decrease in removal with an increase in electrode spacing was explained by a resulting increase in potential difference over the reactor [41]. In this study, the potential difference also increased, from 22V to 31.6±0.4V (exp. 3 and 5, Figure 2b), and thereby the resistivity of the system [67], but the increase did not inhibit the removal as previously shown for TCE [41,68]. By increasing the electrode spacing, the redox zone is prolonged, which possibly accounts for the increased PCE removal. Considering upscaling, this can reduce the direct costs for the plant due to a cutback in the network of electrodes.

pH measured in the porewater downgradient the electrochemical zone (sampling port P5) was 4.5-5 in the C → C → A and C → A configurations tested (exp. 8-10, Figure 2c). For C → BPE → A, pH was 8.5±1.5 (exp. 3, Figure 2c). The elevated pH in the latter configuration could be due to enhanced H₂O₂ formation when introducing a BPE: free oxygen and hydrogen ions generated at anodic sites (Eq. 4) are consumed at adjacent cathodic sites (Eq. 2), i.e. the acidic gradient is weakened and incapable of neutralizing the alkaline gradient established at cathodic sites. BPEs have shown to significantly assist in H₂O₂ formation in electrochemical flow-through reactors [34]. For comparison, pH in the control experiment without applied current was 7 (exp. 0, Figure 2c).

3.2. Significance of current intensity and catalyst loading on PCE removal

Currents of 0 mA, 60 mA, 120 mA and 180 mA were tested in the C → BPE → A arrangement for electrochemical removal of PCE with Pd catalyst loadings on the Fe foam cathode of 0.76 mg/cm² and/or 1.14 mg/cm² (exp. 0-4, Figure 2a).

For catalyst loadings of 0.76 mg/cm², PCE removals were 0%, 14% and 18% at currents of 0 mA, 60 mA and 120 mA, respectively (exp. 0-2, Figure 2a). For catalyst loadings of 1.14 mg/cm², PCE removals were 54±3% and 39% at 120 mA and 180 mA, respectively (exp. 3-4, Figure 2a). Thus, the current intensity and PCE removal are parabolic correlated. The decline in removal at 180 mA is explained by increased competition of oxygen evolution at the anode. In addition, the oxygen generated physically covered the anode and thereby occupied active sites for oxidation [41,42].

Catalysts are important for the transformation of PCE, since a 50% increase in Pd load enhanced PCE removal by 36%-point at 120 mA (exp. 2-3, Figure 2a). The significance of Pd on chlorinated ethene removal has been studied previously, where an optimal concentration of 0.76 mg/cm² was found [42]. The inconsistency in findings of optimal load of Pd on Fe foam cathodes possibly is associated with the geometry of the Fe foam cathodes used; the electrode surface areas tend to be estimated based on geometrical shape, since the actual surface area of the foams is unknown. Thus, the electrode surface area ratio between experimental set-ups cannot be applied directly in the dimensioning of Pd load.

In the set-up studied, a further increase in Pd load could optimize the removal performance, since the C → C → A arrangement with 316 SS and Pd/Fe foam cathodes was superior to two Pd/Fe foam cathodes (Section 3.1., exp. 8-9, Figure 2a), with higher catalytic content in the 316 SS than the resulting palladization. However, replacing the Pd/Fe foam cathode with 316 SS in a C → A set-up is insufficient [55]. In the remainder of tests, a Pd load of 1.14 mg/cm² was applied. The voltage stabilized at 32V at 120 mA independent of Pd load (exp. 2-3, Figure 2b). Whereas the potential difference increased from 20V to 32V to 38V at 60 mA, 120 mA and 180 mA, respectively (exp. 1-4, Figure 2b), a linear correlation with R² of 0.93. According to Ohms Law, the electrical resistivity should remain constant when the experimental conditions, besides current intensity, are unaltered. In this study, the resistivity increased with increasing current, which possibly can be ascribed to increased oxygen and hydrogen generation at the anode ((Eq. 4)-(Eq. 5) and cathode (Eq. 1) with increased current. The additional gases formed extended the gas coverage of the electrode surfaces. Evolving gas layers on the electrode surfaces reduce connectivity between electrodes and the conductive liquid phase, wherefore the electrical resistivity increases.

MMO anodes are known to maintain pH of the water treated [41]. However, in the studied C → BPE → A arrangement, pH of the porewater downgradient of the electrochemical zone (sampling port P5) was elevated to 8.5±1.5 at 120 mA and 180 mA (exp. 0 and 3-4, Figure 2c). The elevated pH post electrochemical treatment indicates formation of H₂O₂ (Eq. 2) instead of neutralization of the alkaline front established at cathodic sites.

A concern of using catalysts like Pd and Ni for treatment of water, whether it is for drinking purposes or discharge into the nature, is the release of ions, which may inhibit living organisms [69,70]. If the catalyst is applied on the cathode and polarity is not reversed, release is of little concern. Whether Pd is released when no current is applied, is not reported. Therefore, effluent samples extracted throughout the 360 min of control testing (exp. 0) were analyzed for Pd. Concentration of Pd was below 0.002 mg/l. Tests on Daphnia magna indicate long-term effects of concentrations >1 mg/l [71].

3.3. Influence of a porous matrix and orientation of reactor on PCE removal

Electrochemical performance in porous matrices was assessed for flow through a porous matrix of glass beads with a porosity of 34.3% at a Darcy velocity of 0.08 cm/min in the C → A electrochemical reactor at 120 mA. Presence of the porous matrix enhanced PCE removal to 39% compared to 30% in the completely liquid-filled reactor (exp. 10 and 11, Figure 2a). The experimental parameters electrode spacing and flow are varied in this comparison (Table 1, test 10-11). Sorption of PCE onto glass beads is neglectable [32,43], verified by the insignificant mass loss in control experiments (exp. 0, Figure 2a). The more complex flow pattern through a porous matrix than a completely liquid-filled reactor may improve the contact between generated reactants and the PCE and explain the enhanced removal in porous matrices. In the porewater downgradient of the electrochemical zone (sampling port P5), pH 5 was measured with and without presence of the porous matrix (exp. 10 and 11, Figure 2c). The potential difference increased from 7V to 36V when including the porous matrix (exp. 10 and 11, Figure 2b). Hence, resistivity was 5 times higher in the porous matrix. An increase in potential difference as the alkaline front progresses towards the anode in a porous matrix has been observed [72].

Vertical oriented reactors are applied in most studies on flow-through electrochemical removal of chlorinated ethenes [24,32,58,33,34,40–44,55]. One major advantage is automatic venting of gases generated, since gases released from the electrodes will bubble up through the reactor and leave the reactor with the effluent. For horizontally oriented reactors, the gases will accumulate in a vapor phase above the electrodes, pressurize the system and eventually drain out the fluid, if no means of venting are installed. However, a horizontally oriented reactor may be desirable for feasibility studies of field-implementations of electrochemical contaminant transformation in saturated porous matrices. Whether orientation of the reactor influenced the treatment performance was assessed in the C → BPE → A and C → A arrangement at 120 mA. At the two flow velocities tested, the horizontally oriented reactor was superior to the vertically oriented with 9-16%-point (exp. 6 vs. 7 and 11 vs. 12, Figure 2a). pH in the porewater downgradient the electrochemical zone (sampling port P5) was similar in the horizontal and vertical orientations of 4.5 and 5 for the low and high flow velocity studied, respectively (exp. 6-7, 11-12, Figure 2c). For the potential difference over the electrodes, no clear trend is observed (exp. 6-7, 11-12, Figure 2b).

Based on observations, the differences in performance of horizontal and vertical oriented reactors probably were related to the system’s ability to release gases generated at the electrodes. In the vertically oriented reactors, the gases accumulated on the upgradient side (lower) of the electrodes until a certain gas volume was reached. The bubbles then pushed through the perforated mesh electrodes and escaped the reactor with the effluent. During the buildup of gases, the potential difference increased followed by a sudden drop once the bubbles had migrated through the mesh electrodes. Accumulation of gases was most prevalent during the high flow tested. Significance of stripping of the volatile contaminant into those gas phases is unknown. In the horizontally oriented reactors, gas phases were immediately extracted above the electrodes using syringes. The gases appeared to be more readily and steadily released and extracted in the horizontal reactors, thus disturbing the system less: the potential difference remained steadier, less pressure build up and the vapor phases were less extensive. Mass loss through volatilization is considered less in the horizontally oriented reactors. Several means of automatic venting in the horizontally oriented reactors were tested; automatic vents, gas-liquid separation membranes, gravity and water-locks. Transient changes in the electrochemical system challenged these venting mechanisms, i.e. the manual extraction of gases was most suitable. Venting mechanisms are lab challenges only, because the test set-ups must be closed systems to hinder mass loss through volatilization. In the field, the test area is an open system from which gases can escape through the electrode wells, in which they are generated, or migrate through the soil matrix. If stripping of chlorinated ethenes is an issue, granular activated carbon above the electrodes can capture the volatilized fraction of the contamination.

3.4. Effect of flow and test duration on PCE removal

Flow rate is inverse proportional to residence time of PCE within the electrochemical reactor, i.e. lowering the flow rate prolongs the residence time. Effect of flow was studied in the completely liquid-filled reactor with a high flow in the C → A arrangement (exp. 11-12) and compared to a low flow in the C → BPE → A configuration (exp. 6-7). The experimental parameters electrode spacing, electrode configuration and flow are varied in this comparison (Table 1, test 6-7 vs. 11-12), wherefore changes cannot solely be ascribed to altered flow. However, the effect of flow rate on the PCE removal is significant in comparison to the effect from electrode spacing and configuration alone (exp. 6-7,11-12 vs. 3,5,8-10, Figure 2a).

Reducing the Darcy velocity from 0.25 cm/min (1314 m/yr) to 0.03 cm/min (150 m/yr) increased PCE removal with 56-63%-point reaching a removal of 86%. For electrochemical treatment of TCE at 60 mA, a reduction in Darcy velocity from 2628 m/yr to 1314 m/yr resulted in a 31%-point improvement in TCE removal to 87% [41]. The low flow in the present investigation is representative of common aquifer flows at contaminated sites.

pH of the porewater downgradient the electrochemical zone (sampling port P5) was similar of 4.5-5 (exp. 6-7 vs. 11-12, Figure 2c). The potential difference decreased by 6.5-13V at elevated flow (exp. 6-7 vs. 11-12, Figure 2b), i.e. the resistivity of the high-flow system was lower, indicating a higher mass-transfer of ionic species.

Electrochemical treatment during 120 min and 290 min insignificantly influenced the PCE removal, pH in porewater downgradient the electrochemical zone (sampling port P5) and the potential difference over the electric field in the C → BPE → A configuration at 120 mA and a Darcy velocity of 0.08 cm/min (exp. 3, Figure 2). The electrochemical processes appeared to rapidly establish and stabilize in this system having a simple hydrogeochemistry.

3.5. Charge transfer in the electrochemical set-up

Experimental specifications vary between studies, e.g. cross-sectional areas, flow velocities, current intensities etc. For comparison of data across studies, normalization of the experimental data is necessary. Many procedures exist for normalization. In the present study, normalization for current intensity (I), flow rate (Q) and contaminant concentration (C) was chosen. Flow rate was chosen instead of flow velocity to account for differences in cross-sectional areas of reactors. The output is charge transfer per contaminant mass and the lower the number, the better in terms of power consumption and thereby cost. The normalized data are plotted against PCE removal percentage (Figure 3) and compared to data reported on electrochemical removal of TCE in a similar set-up [41] (Figure 3, squares). The ratio of electrode surface area to liquid volume was different in the two experimental set-ups applied (lower in the present work) and was not optimized for in the experimental series carried out. Therefore, a deviation in levels of normalized data for a specific removal percentage obtained was expected. Figure 3 allows to compare the overall level of the normalized data obtained for electrochemical removal of PCE and TCE and the shape of the resulting curves. Note, if sequential degradation via e.g. hydrogenolysis was in progress, a fraction of the current supplied would be directed towards concurrent transformation of the degradation products formed. In this case, however, TCE nor DCE was detected (VC analysis was not possible), neither in Rajic et al. [41].

Figure 3.

Normalization of experimental data for PCE based on samples extracted in the final sampling port (solid dots) and data for TCE in Rajic et al. [41] (squares) for current intensity (I), flow (Q) and contaminant concentration (C) plotted against the removal of PCE and TCE. Numbers next to markers refer to experiment number (Table 1).

For PCE removals beyond 30%, the charge input required increased, while for TCE a similar increase appeared for removals beyond 77% (Figure 3). Thus, a higher charge transfer was required to remove PCE than TCE. The difference in electrode surface area to liquid volume ratio between the two experimental set-ups possibly contributed to the observed deviation, where the contact between contaminant, Pd catalyst on the cathode and the energy source itself is improved at a higher such ratio, enhancing the conditions for contaminant transformation at a specific charge input. A low ratio of electrode surface area to liquid volume can be compensated for by increasing the current and/or Pd load. The similar levels of normalized data for PCE and TCE at a removal of 30% support this: MMO anodes and an electrode spacing of 2.5 cm were used in both set-ups. However, in the study on TCE, a Pd load of 0.76 mg/cm² and a current of 60 mA was applied [41]. In the study on PCE, having a lower electrode surface area to liquid volume, a Pd load of 1.14 mg/cm² and a current of 120 mA was applied (exp. 11). I.e. the lower electrode surface area to liquid volume was counterbalanced by a 50% increase in Pd load and a 100% increase in current. A further increase in current would possibly not be beneficial (exp. 4 vs. 11, Figure 2a). Instead, a reduction in flow (exp. 6 vs. 11, Figure 2a) and an increase in Pd load is suggested for optimization of electrochemical removal of PCE in the experimental set-up studied.

If comparing the normalized data obtained from all sampling ports, sampling rounds and test designs (data not shown), and neglecting the associated removal levels, a generally best fit with those of Rajic et al. [41] is seen when set-ups are most alike. I.e., for vertically oriented reactors excluding a porous matrix (exp. 7 and 12), which illustrates the sensitivity towards the experimental design. The concept of electrochemical treatment of chlorinated ethenes must be strengthened to overcome the complex conditions of contaminated sites [65,73]. Finally, removals beyond app. 60% and 90% for PCE and TCE, respectively, required a significant additional charge transfer, as indicated by the steep slope of curves in Figure 3. Thus, treating the low levels of contaminant concentrations, which is required to comply with the regulatory contaminant level for PCE and TCE in drinking water, are the most expensive.

3.6. PCE transformation pathways in the electrochemical reactor

Mass loss is a concern when working with volatile contaminants like PCE, because it can give false positive test results and thereby incorrect foundations for decision making. Two control experiments were carried out in a set-up similar to that of experiment 1-4, except that no current was applied (exp. 0, Table 1). No mass loss was observed throughout the 360 min of testing, measured downgradient the electrochemical zone (sampling port P5, exp. 0, Figure 2a). An observed drop in PCE concentration during experiments with applied current therefore is stimulated by the electrochemistry. Stripping of the contaminants due to the gas generating electrode processes (Eq. 1 and 4) rather than transformation is another concern when applying electrochemistry for removal of chlorinated ethenes. A risk especially at stable anodes like MMO, since these produce more oxygen (Eq. 5 vs. 4) [41] and PCE may be more prone to volatilization than oxidation at the anode, as opposite to the higher affinity of reduction near the cathode. The specific electrode’s contribution to removal of PCE was found in the order of cathode > BPE > anode, indicating that electrochemically induced transformation was occurring rather than stripping. Furthermore, increasing the Pd load on the cathode from 0.76 mg/cm² to 1.14 mg/cm² improved the removal at 120 mA from 19% to 54±3% (exp. 2 vs. 3, Figure 2a), suggesting that the cathode induced processes removed PCE.

Transformation mechanisms at the cathode possibly were related to the catalytic properties of Pd through enhanced adsorption of atomic hydrogen, and thereby reductive hydrodechlorination of PCE, and/or formation of the strong oxidative species H₂O₂ and •OH (Eq. 2–3). Formation of H₂O₂ and •OH appeared to contribute to the transformation of PCE, since 12% more PCE was removed in the C → BPE → A arrangement than C → C → A (exp. 3 vs. 9, Figure 2a). Rajic et al. [34] showed that using a BPE in between the cathode and anode strengthened the H₂O₂ formation. This could be due to i) oxygen and dihydrogen being formed concurrently on each side of the BPE, improving the contact necessary for conversion into H₂O₂, and ii) the anodic side of the BPE facing the cathode, i.e. in closer contact and thus, if the oxygen diffused towards the cathode, it could combine with the dihydrogen generated, forming H₂O₂ at the Pd surface [32]. Elevated pH measured downstream the C → BPE → A electrochemical zone (sampling port P5) indicated formation of H₂O₂ (Section 3.1).

The sequential hydrogenolysis products TCE and DCEs were not detected in any sampling point of the experiments conducted. This suggests fast concurrent hydrogenolysis of TCE and DCEs or another transformation pathway for PCE in the electrochemical reactor, e.g. oxidation by H₂O₂ and •OH, and/or β-elimination [73]. In parallel to the Fe cathodes applied in this study, degradation of PCE on Fe(0) has shown to follow β-elimination for 87% of the contaminant mass [49]. The electrochemical mechanisms responsible for PCE transformation along with the parameters controlling those should be studied in detail for design of a robust electrochemical methodology for removal of chlorinated ethenes.

4. Conclusion

Protection of the groundwater resource from harmful chlorinated ethenes is crucial and challenging. Electrochemically induced reduction and oxidation may overcome some of the difficulties experienced. In the development of the method, PCE must be included in the assessment of electrochemical transformation of chlorinated ethenes, but is mostly overlooked. This study screened the significance of parameters influencing electrochemical removal of PCE in a flow-through reactor using a palladized Fe foam cathode and an MMO mesh anode. The major findings are:

• Of the electrode configurations tested, presence of a BPE between the cathode and anode improved removal of PCE to 54±3%. The BPE added an oxidizing-reducing zone and lowered the electrical resistivity.

• Pd played a major role in the transformation of PCE with a 50% increase in Pd load on the cathode improving the removal by 36%-point.

• Increasing the electrode spacing from 2.5 cm to 4 cm enhanced the removal from 41% to 54±3%, due to extension of redox zones, which is a promising trend when considering the installation and cost of field-implementation.

PCE removal in liquid flowing through an inert porous matrix was investigated to simulate in-situ conditions. Presence of the porous matrix enhanced PCE removal by 9%-point compared to a completely liquid-filled reactor due to a more complex flow pattern within the porous matrix. A horizontally orientated reactor improved PCE removal by 9-16%-point compared to vertically oriented reactors. Release of oxygen and hydrogen gas from the electrode surfaces in the horizontal reactor was enhanced, wherefore a higher number of active electrode surface sites were available for PCE transformation. The highest removal of 86% was obtained using a horizontally oriented reactor at a current of 120 mA applied to an electrode configuration of C → BPE → A with a Pd load on the cathode of 1.14 mg/cm² and a Darcy velocity of 0.03 cm/min (150 m/yr) through a fully liquid-filled reactor. No chlorinated ethene degradation products were detected in any of the test designs studied.

Normalization of the data revealed, that the low contaminant concentrations were the most difficult to treat in that a significant charge transfer was required for removal beyond 60%. The removal performance was found sensitive to dimensional changes in experimental reactors. The transformation mechanisms and the parameters controlling those need to be better understood to control and adjust the relevant parameters for a flexible remediation design, that can overcome the variations occurring between lab set-ups and not least at contaminated sites. Thus, this study identified some weaknesses in the current stage of development of electrochemical removal of chlorinated ethenes, but also some promising trends in conceptual designs that better simulate that of a potential field-implementation.

Highlights.

• Harmful chlorinated ethenes threaten the drinking water resource and its consumers

• Transformation of PCE in electrochemical flow-through reactors was studied

• Influence of various reactor design parameters was investigated, reaching 86% removal

• Normalization of data improved understanding of mechanisms across tests and studies

• Electrochemical reduction and oxidation can reduce mass flux of simulated PCE plumes