Redox Control and HydrogenProduction in Sediment Caps UsingCarbon Cloth ElectrodesM E I S U N , † F E I Y A N , § R U I L I N G Z H A N G , §

D A N N Y D . R E I B L E , §

G R E G O R Y V . L O W R Y , * , † , ‡ A N DK E L V I N B . G R E G O R Y * , †

Department of Civil and Environmental Engineering,Department of Chemical Engineering, Carnegie MellonUniversity, Pittsburgh, Pennsylvania 15213-3890, United States,and Environmental and Water Resources Engineering, Universityof Texas, Austin, Texas 78712, United States

Received March 29, 2010. Revised manuscript receivedSeptember 13, 2010. Accepted September 14, 2010.

Sediment caps that degrade contaminants can improve theirability to contain contaminants relative to sand and sorbent-amended caps, but few methods to enhance contaminantdegradation in sediment caps are available. The objective ofthis study was to determine if, carbon electrodes emplaced withina sediment cap at poised potential could create a redoxgradient and provide electron donor for the potential degradationof contaminants. In a simulated sediment cap overlyingsediment from the Anacostia River (Washington, DC),electrochemically induced redox gradients were developedwithin 3 days and maintained over the period of the test (∼100days). Hydrogen and oxygen were produced by waterelectrolysis at the electrode surfaces and may serve aselectron donor and acceptor for contaminant degradation.Electrochemical and geochemical factors that may influencehydrogenproductionwerestudied.Hydrogenproductiondisplayedzero order kinetics with ∼75% Coulombic efficiency. Rateswere proportional to the applied potential between 2.5 and 5V and not greatly affected by pH. Hydrogen productionwas promoted by increasing ionic strength and in the presenceof natural organic matter. Carbon electrode-stimulateddegradation of tetrachlorobenzene in a batch reactor wasdependent on applied voltage and production of hydrogen toa concentration above the threshold for biological dechlorination.These findings suggest that electrochemical reactive cappingcan potentially be used to create “reactive” sedimentscaps capable of promoting chemical or biological transformationsof contaminants within the cap.

IntroductionContaminated sediments often contain complex mixturesof toxic anthropogenic organic and inorganic contaminants,and are ubiquitous and costly to remediate (1).The high cost

and limited effectiveness of dredging (2-4), and slow ratesof monitored natural recovery (5), lead to the use of in situsediment caps for abatement of risk from contaminatedsediments. In brief, a layer of material (usually sand) isoverlain on the sediment to isolate the contaminants fromthe aquatic ecosystem (2, 6). A recent improvement to thebasic sand cap, known as amended capping, employsadsorbents such as activated carbon or apatite in the capmaterial to sequester contaminants and further retard themovement of contamination from the sediments into thebasal trophic levels of the ecosystem (7-9). However, cappingdoes not necessarily provide removal or detoxification, andthe risk to human and environmental health may return oncethe capping material becomes saturated. State-of-the-art capshave also been proposed which include a reactive componentfor biotic or abiotic contaminant degradation, reducing thetime scales for habitat redevelopment (9) and site closure(10).

Reactive caps incorporate engineered components whichact to stimulate biotic or abiotic contaminant degradationto reduce toxicity of or mineralize contaminants. For example,2-chlorobiphenyl is adsorbed and simultaneously dechlo-rinated by granular activated carbon cap material containingFe/Pd nanoparticles (11). Additionally, reactive caps may bedesigned to promote microbial growth and biodegradationof contaminants; Himmelheber et al. demonstrated depth-dependent microbial populations which varied with thestratified geochemical redox zones within a sand sedimentcap (12). The chemical species in each stratified zoneindicated a corresponding biogeochemical process and ledto the suggestion that microbial contaminant degradationmay be engineered within a reactive sediment cap (13).

Some remediation strategies were developed based onelectrochemical technologies. Electrodes may be emplacedin groundwater or sediment for treating contamination withchlorinated solvents and energetic compounds (electrolyticreduction) (14-16), mixed organic/inorganic wastes (Lasagnaprocess) (17), and heavy metals (electrosorption) (18). In thesecases the electrodes are used to directly degrade contami-nants, sequester heavy metals at the electrode, or collectcontaminants by electrokinetic process for further treatment(19). Usually these processes require high voltage or a largenumber of electrodes, increasing costs (19).

In contrast to other electrode-based remediation ap-proaches, we propose to employ geotextiles with polarizedcarbon electrodes in the sediment-cap to impose a desiredelectrical potential gradient to stimulate biodegradation ofcontaminants within the cap, that is, an electrode biobarrier(Figure 1). In brief, the electrodes are placed perpendicularto contaminant transport through the cap, and polarized atlow potentials to accelerate contaminant degradation by, (1)rapidly establishing and maintaining a redox gradient withinthe cap, which may be varied in real time, and (2) supplyingthe cap with electron acceptor and electron donor to stimulatemicrobial growth. As contaminants migrate from the deepersediments, they will be exposed to the electrode biobarriermaintained by microbial growth and respiration, which isscalable in both magnitude and thickness of the respectivereducing and oxidizing zones created by the electrodes. Thecritical advance in this research is the application of lowvoltage and low current to large area electrodes to enhanceand control appropriate microbial activity in a thin horizontallayer within a cap. Microbial growth can be stimulated byhydrolytic reactions at the anode and cathode producingoxygen and hydrogen, respectively, and by direct microbialrespiration of the anode and cathode as electron acceptor

* Address correspondence to either author. Phone: 412-268-9811(K.B.G); 412-268-2948 (G.V.L.). Fax: 412-268-7813 (K.B.G.); 412-268-7813 (G.V.L.). E-mail: kelvin@cmu.edu (K.B.G.); glowry@cmu.edu(G.V.L.).

† Department of Civil and Environmental Engineering, CarnegieMellon University.

‡ Department of Chemical Engineering, Carnegie Mellon University.§ University of Texas.

Environ. Sci. Technol. 2010, 44, 8209–8215

10.1021/es101003j 2010 American Chemical Society VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8209

Published on Web 09/29/2010

and donor, respectively. This approach has enabled thecoupling of microbial oxidation of benzene to the respirationof an anode as the electron acceptor (20), or the microbialreduction and precipitation of U(VI) (21) and nitrate reduction(22) to the respiration of the cathode as the electron donorfor bacteria. The residence time of contaminants within thesezones will, in principle, be controllable in real time bymanipulating the applied potential. In this manner, thereactive electrode-stimulated biobarrier may be specificallyengineered and adjusted for targeted degradation of con-taminants or their mixtures through electrode stimulationof appropriate microbial communities.

An electrode-based biobarrier for reactive sedimentcapping has the potential to address the unique challengesimposed by the sediment environment. First, biodegradationprocesses in sediment caps are commonly limited by theavailability of either electron donor or electron acceptor,depending on the contaminant. The electrode cap will supplyboth spatially and temporally controlled electron donor andacceptor. Second, complete mineralization of some con-taminants (e.g., PCBs and some chlorinated solvents) onlyoccurs through sequential reduction and oxidation, whichmay not develop under natural conditions. Through theelectrode imposed and microbially maintained redox gradi-ent, residence times through reductive and oxidative condi-tions for specific contaminant transformation may beengineered. Third, the common observance of contaminationin mixtures which degrade or detoxify under disparate redoxconditions confounds single-approach remedial or seques-tration designs. As before, an electrode stimulated biobarrier,redox gradient may be engineered to address the specificcontaminants of concern on a site by site basis.

Here, we describe the deployment of carbon electrodesin a simulated sediment cap orientation for engineering redoxgradients to achieve a reactive cap over freshwater sedimentfrom the Anacostia River (Washington, DC). Electrodesrapidly establish oxidizing conditions at the anode andreducing conditions at the cathode as well as a gradientbetween the electrodes, which was maintained for about100 days. The impact of various geochemical porewaterparameters and electrochemical conditions on hydrogenevolution rates from the cathode were examined as well as

the potential for electrodes to stimulate degradation of 1,2,3,5-tetrachlorobenzene.

Materials and MethodsChemicals. Sodium bicarbonate (NaHCO3), sodium chloride(NaCl), calcium chloride (CaCl2 ·2H2O), sodium hydroxide(NaOH), and chloride acid (HCl) were supplied by FisherScientific (Pittsburgh, PA). Magnesium chloride (MgCl2 ·6H2O) was purchased from ICN Biomedicals Inc. (Costa Mesa,CA). Humic acid sodium salt (50-60% as humic acid) waspurchased from Acros Organics (Geel, Belgium). Hydrogenstandard (1%) and nitrogen were purchased from Butler GasProducts (Pittsburgh, PA).

Engineered Redox Gradient in Sediment. Three T-cellreactors were used to evaluate the ability of carbon electrodesto control the redox gradient in a sand cap overlying AnacostiaRiver sediment (Supporting Information (SI) Figure S1a). Thereactor, as previous described (6), was filled with sieved (2mm) sediment from the Anacostia River. A 14 ×7 cm wovencarbon cloth (BASF Fuel Cell, Inc., Somerset, NJ) was placedon top of the sediment as the cathode. A 2.5 cm layer ofsieved (0.425 mm) sand (Riccelli Enterprises, Rush NY) wasplaced over the cathode and a second, identical, carbon clothon the sand layer as the anode, followed by a 0.5 cm sandlayer. The sediment and sand layers were saturated with tapwater. The electrodes of T-cell 1 and T-cell 2 were connectedby copper wire to 4 V Extech 382202 DC power supply (ExtechInstruments Corp., Waltham, MA). Power to T-cell 1 wascontinuously applied for ∼100 days, whereas power to T-cell2 was removed after 30 days of continuous operation in orderto examine the recovery of the sediment cap followingstimulation. A third T-cell served as an unpowered controlreactor. The potential was continuously applied except duringmicroelectrode pH and oxidation-reduction potential (ORP)measurement. pH was measured by MI-405 pH microelec-trode (Microelectrodes Inc., Bedford, New Hampshire) andORP was measured by Pt microelectrode against Ag/AgClreference (fabricated as previously described (23)). pH andORP measurement required 2 and 5 min, respectively, toreach equilibrium. Vertical profiles of pH and ORP from thewater-sand interface to a depth of 44 mm with 2 mm intervals

FIGURE 1. Conceptual model for an electrode-based reactive sediment cap. Through water electrolysis, hydrogen is produced in thecathode to stimulate reductive biodegradation, and oxygen is produced in the anode to stimulate oxidative biodegradation.

8210 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

were acquired and ORP data were converted and reportedas versus standard hydrogen electrode (SHE).

Geochemical Impacts on Electron Donor Supply Rate.Experiments to evaluate the impact of environmentallyrelevant geochemical porewater parameters on H2 productionwere carried out in triplicate H-cell reactors (SI Figure S1b)described previously (22). The two chambers were separatedby a cation-exchange membrane (Nafion 117; Electrosyn-thesis Inc., Lancaster, NY). Electrodes were 12.6 × 6.25 cmcarbon cloth as described above. The electrodes wereconnected to an E3620A DC power supply (Agilent Tech-nologies, Santa Clara CA) via 0.64 cm diameter × 15.2 cmgraphite rods (GraphiteStore.com, Inc., Buffalo Grove, IL) bygraphite epoxy (377H; Epoxy Technology, Inc., Billerica, MA).The reactors were well stirred at room temperature. Reactorscontained 250 mL buffer solution and 60 mL headspace.Unless specified, the buffer solution contained 20 mMNaHCO3 and 20 mM NaCl (supporting electrolyte, conduc-tivity ∼4mS/cm). Buffer solution was prepared anoxicallyand pH adjusted by HCl or NaOH. Identical control experi-ments using the same tap water as used in the T-cellexperiments were also performed to verify comparison andare not presented.

The solution pH in the H-cell was measured using a gel-filled 8 mm pH electrode (Fisher Scientific, Pittsburgh, PA).Current was recorded by a model 2700 digital multimeter(Keithley Instruments, Inc., Cleveland, Ohio). Headspace H2

concentration was determined using GC/TCD as previouslydescribed (24). The H2 production rate was calculatedaccording to Henry’s Law (SI eq S1). Coulombic efficiency(CE), an estimation of the fraction of the current capturedas hydrogen was calculated as the ratio of measuredcumulative hydrogen production (SI eq S2) to the theoreticalhydrogen production calculated by integrating currentequivalent with time (SI eq S3).

Electrode-Stimulated Degradation of 1,2,3,5-Tetra-Chlorobenzene. Degradation of 1,2,3,5-tetrachlorobenzene(TeCB) was examined in borosilicate glass, electrochemicalreactors (SI Figure S1(c)). The effective volume of the reactorwas 600 mL and 2/3 of the reactor was filled with Anacostiariver sediment with approximately 80% water content, andspiked with TeCB (Sigma-Aldrich, St. Louis, MO) to achievethe desired initial concentration. Carbon electrodes (6.15mm diameter × 15.2 cm Alfa Aesar, Ward Hill, MA) wereconnected to an external power supply and agitated on ashaker table. Control experiments without electrodes werealso examined. TeCB concentrations in aqueous phase wereanalyzed by solid-phase microextraction (SPME) as describedelsewhere (25) and in SI.

Results and DiscussionRedox Control and pH Change in Sediment. Poised elec-trodes emplaced in a sediment cap rapidly established adepth-dependent stratification of redox potential over Ana-costia river sediment (Figure 2 shows data in selected days,all the data are shown in SI Figure S2). A 4 V potential waschosen to ensure adequate hydrogen evolution as detailedlater. Prior to the application of potential, the ORP at theanode was +270 mV and increased with depth through thesand cap to around +350 mV at the cathode. The sedimentbelow the cathode became more reduced with depth, a likelyresult of natural microbial processes (13). After 3 days withapplied potential between the electrodes, the ORP in thevicinity of the anode had slightly increased to near +350 mVand the ORP in the vicinity of the cathode had decreasedsharply to approximately -300 mV. The ORP measured inboth powered T-cell reactors changed similarly under appliedpotential, indicating reproducibility; the redox conditionsstratified through the sediment cap and developed a relativesteady state after approximately one month. T-cell 2 was

disconnected from the power supply on day 30 to monitorthe ORP during recovery. The depth-dependent stratificationof ORP through the sand cap in T-cell 1 continued until day98. In contrast, ORP stratification in T-cell 2 graduallydiminished after the power off (Figure 2b): on day 98, theORP in the vicinity of the anode was +240 mV and near thecathode, +210 mV. Over the course of the experiment, thecontrol T-cell 3 (with imbedded, but unpowered electrodes),remained oxidizing in the vicinity of and between electrodes(Figure 2c).

A depth dependent stratification of redox conditions inthe control reactors was only observed within the sediment.Such gradients develop naturally under the sediment-waterinterface and are indicative of the microbial oxidation ofsediment organic matter coupled to the reduction of terminalelectron acceptors in an order which reflects the maximumenergy yield for the microbial community (26). By applyinga voltage through the electrodes, however, the location andsize of the respective oxidation and reduction zones can becontrolled.

The pH profile in the sand cap demonstrated a similar,depth-dependent gradient (Figure 2d and e). On day 0, thepH between the anode and cathode in all T-cells varied fromaround 8 at the anode to 6.5 at the cathode. Soon after externalpower application, the pH gradient through the cap increasedsharply: pH in the vicinity of the anode dropped while nearthe cathode, the pH changed from circumneutral to near pH11 on approximately day 9. The pH in the control reactorremained relatively steady with time and depth (Figure 2f).The rapid increase of pH near the cathode is the result of thereduction of water-derived protons for H2 formation and thelack of sufficient buffer capacity in the sand cap to neutralizeOH- produced from reduction of water. Following removalof power from T-cell 2 on day 30, the pH began to return tocircum neutral, but did not completely recover over thecourse of the experiment (Figure 2e). The pH change wasexpected in this system since no buffer was added to balancethe buildup of H+ and OH- at the electrodes. Under fieldconditions, such large pH increases would not be expected;since electrolyte flow between electrodes may alleviateaccumulation of H+ and OH- (27). Additionally, under fieldconditions, porewater may be buffered by solid and dissolvedorganic and inorganic phases in the cap and porewater seepmay flush excessive H+ to the anode to neutralize excessiveOH-. The addition of modest amounts of buffering solids inthe cap design, for example, iron carbonate (siderite), couldmaintain pH if needed.

Hydrogen Production Rate and Coulombic Efficiency.The ability to control the redox profile and microbialprocesses in sediment will depend on the rate of electrondonor and acceptor evolution. Although electrolysis of waterand proton reduction rates are more efficient when catalyzedby noble metals, for sediment applications, inexpensiveelectrode material such as carbon is more likely to beemployed. H2 evolution from the carbon electrode displayedzero-order kinetics. At an applied potential of 4 V, theelectrode surface area normalized rate constant was 5 mmolH2 m-2 hr-1 (SI Figure S3a), despite an increase in pH from7 to 9.5 The zero-order kinetic relationship agrees with aprevious report of hydrogen production from carbon-basedelectrodes (28). Over 24 h the CE was about 75% (SI FigureS3b), indicating that 25% of the current flow between theelectrodes was not captured as H2. This was not surprisingas O2 produced at the anode may diffuse through the nafionmembrane (29) and consume electrons at the cathode.Additionally, H2 was detected on the anode side of the H-cellwhere it would be consumed at the anode. Non-Faradiaccurrent to maintain charge neutrality at the electrodes mayalso be a source of current which was not captured as electronequivalents in H2 gas.

VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8211

Influence of Voltage on Hydrogen Production. Thetheoretical equilibrium potential necessary for water elec-trolysis is 1.23 V (30). Therefore hydrogen and oxygenaccumulation requires a greater potential difference at theelectrodes. In addition, overpotentials at the anode andcathode, resulting from activation and mass-transfer resistivelosses as well as ohmic loss across the electrolytic cell, mustbe overcome for H2 production to proceed. In order to addressthese concerns, experiments were performed to evaluate H2

production at various cell potentials using carbon clothelectrodes. When cell potential was in the range of 2-5 V,the observed current was between 0.5 and 5 mA. H2

production rate was proportional to the applied cell potentialabove 2.5 V (Figure. 3a). There was no statistically significantdifference between H2 production rate at 2.0 and 2.5 V. Sinceexternal cell potential is the driving force for electrolysis ofwater, the increase in H2 production rate was not surprising.

Similar findings with graphite electrodes are reported (29).The relationship between applied potential and H2 produc-tion rate may enable real-time control of electron donor formicrobial growth within the sediment cap, creating animpenetrable biobarrier if the contaminant degradation rateis sufficiently high relative to advection and diffusion throughthe cap (10).

Lower cell potentials produced lower H2 evolution ratesand resulted in smaller pH changes over the course of theexperiment (Figure 3b). The pH changes are likely to be lessof an issue during field applications, as the required rates ofH2 evolution are will be lower for degradation of contaminantspresent in low concentrations (e.g., PCBs) and, as describedbriefly above, migration of sediment porewater containingnatural organic matter and alkalinity creates an elevatedbuffer capacity (31). Nevertheless, the impact of pH on H2

production rates was examined in the H-cell reactors (SI

FIGURE 2. Vertical (a, b) ORP and (d, e) pH profiles developed in sediment and cap containing carbon cloth electrodes (T-cell 1 andT-cell 2) imposed to a 4 V external voltage. Vertical (c) ORP and (f) pH profiles in control sediment (with electrode unpowered).T-cell 2 showed recovery of ORP and pH toward control after the power is disconnected, while in T-cell 1 ORP and pH profiledeveloped rapidly after ∼25 days and stabilized over ∼100 days. Depth zero was the water-sand interface. The horizontal dashedlines indicate the position of electrodes.

8212 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

Figure S4). H2 production rate was not significantly affectedby initial pH at either 4 V or 2.5 V applied potential, accordingto single factor analysis of variance (ANOVA) (p ) 0.16). Thisresult is consistent with the approximately zero-order kineticsfor H2 production despite the pH increase observed withextended reaction times in Figure 3b. However, the impactof pH on H2 production as well as the change in pH observedwith H2 production should be verified in field samples underconditions which more closely simulate in situ.

Influence of Aqueous Chemical Species on HydrogenProduction. Many chemical species which may affect ratesand efficiencies of the electrolytic cell are present in sedimentporewater. Among these, electrolyte concentration, presenceof divalent cations, and natural organic matter (NOM) wereinvestigated. In reference to Figure 4a, the “diluted electro-lyte” group and “concentrated electrolyte” group simulatedfreshwater (I ) 0.044 N) and seawater (I ) 1.040 N),respectively (32), and addressed the impact of overpotentialchanges on H2 production. The rate of H2 productionincreased over an order of magnitude from freshwater toseawater-like conditions and is the result of increasingconductivity in the solution, therefore reducing the ohmicresistance. The overall result is higher current at a similarelectrode potential, which leads to a higher hydrogenproduction rate. Additionally, depending on the ionic species,they may adsorb to the electrode and alter the kinetics of theH2-producing reaction through occupation of reactive sitesor formation of mineral precipitates. This finding is analogousto those of Call and Logan (33) where H2 production in MECincreased with catholyte concentration. In practice, theelectrode cap operating in high ionic strength porewater

would require less energy to achieve similar delivery rates ofelectron donor and redox gradients through the cap. Theionic strength and species of the sediment porewater will besite-specific and greatly impact remedial design.

Cations may potentially impact remedial design throughtheir precipitation at the electrode. 5 mM of calcium andmagnesium salts were evaluated for their impact on H2

production rates (Figure 4a). H2 production was not signifi-cantly impacted by the addition of Mg2+ and slightly enhancedthrough the addition of Ca2+. Precipitation of solids wasobserved on the bottom of the cathode chamber but not onthe electrode material. These results are different from Franzet al. who observed precipitation and deposition on stainlesssteel plate cathodes and increased resistance across the cell(34). This difference is likely the result of different electrodematerials which may alter the double-layer composition aswell as kinetics of precipitation reactions. Should cathodepH effect precipitation reactions the polarity of the electrodescan be periodically reversed to redissolve the precipitates(35).

Natural organic matter (NOM) is a major component ofsediment porewater and exerts a contribution to conductivity,buffer strength, and reactions at the anode and cathodethrough electron shuttling, and therefore may be expectedto affect electrode reactions. Humic acid was added to arelatively high concentration of 200 mg/L (about 32 mgcarbon/L). The rate of H2 production increased statisticallysignificantly. The potential reasons for its effect are (1)contribution to conductivity (by adding 200 mg/L humicacid sodium salt, sodium concentration increased about 5mM); (2) humic acid may be adsorbed onto the electrodes

FIGURE 3. Influence of applied potential on (a) hydrogen production rate and (b) cathode solution pH. All experiments wereconducted at an initial pH 7 in 20 mM NaHCO3 and 20 mM NaCl solution over 24 h. Dashed vertical line in (a) shows the equilibriumpotential for water electrolysis. When voltage is over 2.5 V, increase of voltage causes linear increase of hydrogen production rateand hence greater pH change.

FIGURE 4. Influence of aqueous chemical species on (a) hydrogen production rate and (b) Coulombic efficiency. All experimentswere conducted at voltage 4 V over 24 h with initial solution pH 7. The chemical composition for each group was: Control: 20 mMNaHCO3 and 20 mM NaCl; Diluted electrolyte: 20 mM NaHCO3 and 2 mM NaCl; Concentrated electrolyte: 20 mM NaHCO3 and 500 mMNaCl; Mg2+: 20 mM NaHCO3, 20 mM NaCl and 5 mM MgCl2; Ca2+: 20 mM NaHCO3, 20 mM NaCl and 5 mM CaCl2; Natural organicmatter: 20 mM NaHCO3, 20 mM NaCl and 200 mg/L humic acid sodium salt. The dashed line showed the average results of controlgroup.

VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8213

and act as electron shuttles (36) or as a donor or acceptor(37). T-test results indicated that except for the “concentratedelectrolyte” group, Coulombic efficiency in all other groupswas not significantly different from the control (Figure 4b).

Electrode-Stimulated Degradation of TeCB in Sedi-ment. Degradation of TeCB in Anacostia river sediments wasbriefly examined to verify the potential for contaminantdegradation using an electrode-based approach. Althoughelectrochemical degradation of a variety of contaminants iswell-known, most studies were conducted in sediment-freesystems (14, 15, 27). SI Figure S5(a) shows that degradationof TeCB was voltage-dependent. An external potential of 4V stimulated removal of TeCB; approximately 90% over 21days. Similar removal was not observed in the absence ofelectrodes or when only 3 V was applied to the system. It isworth noting that TeCB degradation did not occur until theH2 concentration was above the threshold concentration forbiological dechlorination of TeCB (0.4 ppm or 16nM) (38).The H2 concentration did not reach this threshold in the 3V reactor (SI Figure S5a). While not conclusive, the voltageand H2-threshold dependence of TeCB removal suggest thatdegradation may be linked to electrode-stimulated microbialmetabolism.

Power supply to the electrodes decreased ORP from+200mV to +70 mV (vs SHE) (SI Figure S5b). In addition, the pHchange was less than 1 unit and was likely stabilized by thenatural capacity for buffering in sediments. Note that the pHchange in this experiment is not directly comparable withthe T-cell experiments; T-cells were a quiescent incubation.Therefore, neutralizing of excess H+ and OH- was limited bydiffusion and not expected. Additional studies of the elec-trode-based approach in sediments are needed for a betterunderstand the pH control and the mechanisms of con-taminant transformation.

Implications for Electrochemical Sediment Capping.The results presented here show that simple carbon elec-trodes can be used to engineer a redox gradient in a sedimentcap, produce microbial electron donor, and stimulatecontaminant degradation. Successful implementation ofelectrode-based capping of sediments will meet with a varietyof technical and economic challenges which relate to scalingof electrode technology to the field; capital costs for materialsand remote power (such as solar) or recurring costs forconventional electricity are just a few. A recent field demon-stration of electrode-based treatment of groundwater usingan “e-barrier” suggests that electrode technology is cost-comparable with conventional permeable reactive barriertechnology (16). Alternatively, the electrochemical cappingstrategy can be applied in combination with “funnel andgate” treatment (39). In such configuration, the majority ofthe sediment area is covered by a low permeability cap, andthe contaminants can be funneled by hydraulic control toa relatively small area, or “gate”, where electrochemical capis installed for contaminant degradation.

Electrode-based capping with biobarriers may enablecontaminant degradation and in particular those whichrequire or are enhanced by sequential anaerobic/aerobictreatment, such as contaminant mixtures (e.g., aromaticsand chlorinated solvents). Our results show that importantenvironmental variables such as pH, dissolved organic carbonand ionic composition do not greatly affect the rate of H2

production or Coulombic efficiency, suggesting a relativelyrobust technology under a variety of field conditions, butlonger term studies will be necessary and operating param-eters (e.g., applied potential) will likely need to be optimizedfor specific site conditions. The same electrode material wasused for over six months in this study without decrease inH2 production rate or Coulombic efficiency, indicating thepotential for long-term use. Based on this investigation,further study of electrode-based sediment capping is war-

ranted for determination of construction and operationalcosts, the potential for pH control under in situ flowconditions, the effect of geochemical parameters on con-taminant removal rates, and the evolution of microbialcommunities within an engineered redox gradient to confirmthe practicability of this remediation strategy.

AcknowledgmentsWe thank Dr. Guy Sewell (East Central University, Ada, OK)for initial discussions on electrode-based remediation usingBioLance technology. The assistance of Dr. XiaoXia Lu duringproof of concept experiments is also greatly appreciated.The project described was supported by Award No.: 1R01ES016154-01 from the National Institute of EnvironmentalHealth Sciences, National Institute of Health. The content issolely the responsibility of the authors and does not neces-sarily represent the official views of the NIEHS or the NIH.

Supporting Information AvailableCalculation of Coulombic efficiency, method for 1,2,3,5-tetrachlorobenzene (TeCB) measurement, scheme of T-celland H-cell reactors, comprehensive ORP and pH profile inT-cell experiment, figure describing the initial pH impact onhydrogen production rates, and a figure detailing results forthe degradation of TeCB. This material is available free ofcharge via the Internet at http://pubs.acs.org.

Literature Cited(1) The Incidence and Severity of Sediment Contamination in Surface

Waters of the United States, National Sediment Quality Survey;U.S. Environmental Protection Agency: Washington, DC, 2004.

(2) Palermo, M., Maynord, S., Miller, J., Reible, D. Guidance for inSitu Subaqueous Capping of Contaminated Sediments; GreatLakes National Program Office: Chicago, IL, 1998.

(3) A Risk-Management Strategy for PCB-Contaminated Sediments;National Research Council: Washington, DC, 2001.

(4) Reible, D.; Hayes, D.; Lue-Hing, C.; Patterson, J.; Bhowmik, N.;Johnson, M.; Teal, J. Comparison of the long-term risks ofremoval and in situ management of contaminated sedimentsin the Fox River. Soil Sediment Contam. 2003, 12 (3), 325–344.

(5) U.S. Environmental Protection Agency EPA’s monitored naturalattenuation policy. The Army Lawyer 1998, 308 (85).

(6) Wang, X. Q.; Thibodeaux, L. J.; Valsaraj, K. T.; Reible, D. D.Efficiency of capping contaminated bed sediments in situ. I,laboratory-scale experiments on diffusion-adsorption in thecapping layer. Environ. Sci. Technol. 1991, 25 (9), 1578–1584.

(7) Jacobs, P. H.; Forstner, U. Concept of subaqueous capping ofcontaminated sediments with active barrier systems (ABS) usingnatural and modified zeolites. Water Res. 1999, 33 (9), 2083–2087.

(8) Knox, A. S.; Paller, M. H.; Reible, D. D.; Ma, X. M.; Petrisor, I. G.Sequestering agents for active capssRemediation of metalsand organics. Soil Sediment Contam. 2008, 17 (5), 516–532.

(9) McDonough, K. M.; Murphy, P.; Olsta, J.; Zhu, Y. W.; Reible, D.;Lowry, G. V. Development and placement of a sorbent-amendedthin layer sediment cap in the Anacostia River. Soil SedimentContam. 2007, 16 (3), 313–322.

(10) Murphy, P.; Marquette, A.; Reible, D.; Lowry, G. V. Predictingthe performance of activated carbon-, coke-, and soil-amendedthin layer sediment caps. J. Environ. Eng. 2006, 132 (7), 787–794.

(11) Choi, H.; Agarwal, S.; Al-Abed, S. R. Adsorption and simultaneousdechlorination of PCBs on GAC/Fe/Pd: mechanistic aspectsand reactive capping barrier concept. Environ. Sci. Technol.2009, 43 (2), 488–493.

(12) Himmelheber, D. W.; Thomas, S. H.; Loffler, F. E.; Taillefert, M.;Hughes, J. B. Microbial colonization of an in situ sediment capand correlation to stratified redox zones. Environ. Sci. Technol.2009, 43 (1), 66–74.

(13) Himmelheber, D. W.; Taillefert, M.; Pennell, K. D.; Hughes, J. B.Spatial and temporal evolution of biogeochemical processesfollowing in situ capping of contaminated sediments. Environ.Sci. Technol. 2008, 42 (11), 4113–4120.

(14) Gilbert, D. M.; Sale, T. C. Sequential electrolytic oxidation andreduction of aqueous phase energetic compounds. Environ.Sci. Technol. 2005, 39 (23), 9270–9277.

8214 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

(15) Gent, D. B.; Wani, A. H.; Davis, J. L.; Alshawabkeh, A. Electrolyticredox and electrochemical generated alkaline hydrolysis ofhexahydro-1,3,5-trinitro-1,3,5 triazine (RDX) in sand columns.Environ. Sci. Technol. 2009, 43 (16), 6301–6307.

(16) Sale, T.; Petersen, M.; Gilbert, D. Electrically induced redox barriersfor treatment of groundwater. ESTCP Final Report 2005.

(17) Ho, S. V.; Sheridan, P. W.; Athmer, C. J.; Heitkamp, M. A.; Brackin,J. M.; Weber, D.; Brodsky, P. H. Integrated in situ soil remediationtechnology: the Lasagna process. Environ. Sci. Technol. 1995,29 (10), 2528–2534.

(18) Farmer, J. C.; Bahowick, S. M.; Harrar, J. E.; Fix, D. V.; Martinelli,R. E.; Vu, A. K.; Carroll, K. L. Electrosorption of chromium ionson carbon aerogel electrodes as a means of remediating groundwater. Energy Fuels 1997, 11 (2), 337–347.

(19) Acar, Y. B.; Alshawabkeh, A. N. Principles of electrokineticremediation. Environ. Sci. Technol. 1993, 27 (13), 2638–2647.

(20) Zhang, T.; Gannon, S. M.; Nevin, K. P.; Franks, A. E.; Lovley,D. R. Stimulating the anaerobic degradation of aromatichydrocarbons in contaminated sediments by providing anelectrode as the electron acceptor. Environ. Microbiol. 2010, 12(4), 1011–1020.

(21) Gregory, K. B.; Lovley, D. R. Remediation and recovery ofuranium from contaminated subsurface environments withelectrodes. Environ. Sci. Technol. 2005, 39 (22), 8943–8947.

(22) Gregory, K. B.; Bond, D. R.; Lovley, D. R. Graphite electrodes aselectron donors for anaerobic respiration. Environ. Microbiol.2004, 6 (6), 596–604.

(23) Brendel, P. J.; Luther, G. W. I. Development of a gold amalgamvoltammetric microelectrode for the determination of dissolvedFe, Mn, O2, and S(-|) in porewaters of marine and freshwatersediments. Environ. Sci. Technol. 1995, 29 (3), 751–761.

(24) Liu, Y.; Majetich, S. A.; Tilton, R. D.; Sholl, D. S.; Lowry, G. V.TCE dechlorination rates, pathways, and efficiency of nanoscaleiron particles with different properties. Environ. Sci. Technol.2005, 39 (5), 1338–1345.

(25) Mayer, P.; Vaes, W. H. J.; Wijnker, F.; Legierse, K.; Kraaij, R. H.;Tolls, J.; Hermens, J. L. M. Sensing dissolved sediment porewaterconcentrations of persistent and bioaccumulative pollutantsusing disposable solid-phase microextraction fibers. Environ.Sci. Technol. 2000, 34 (24), 5177–5183.

(26) Reimers, C. E.; Tender, L. M.; Fertig, S.; Wang, W. Harvestingenergy from the marine sediment-water interface. Environ.Sci. Technol. 2001, 35 (1), 192–195.

(27) Shimomura, T.; Sanford, R. A. Reductive dechlorination oftetrachloroethene in a sand reactor using a potentiostat. J.Environ. Qual. 2005, 34 (4), 1435–1438.

(28) Sun, M.; Sheng, G. P.; Zhang, L.; Xia, C. R.; Mu, Z. X.; Liu, X. W.;Wang, H. L.; Yu, H. Q.; Qi, R.; Yu, T.; Yang, M. An MEC-MR-coupled system for biohydrogen production from acetate.Environ. Sci. Technol. 2008, 42 (21), 8095–8100.

(29) Logan, B. E.; Call, D.; Cheng, S.; Hamelers, H. V. M.; Sleutels,T. H. J. A.; Jeremiasse, A. W.; Rozendal, R. A. Microbial electrolysiscells for high yield hydrogen gas production from organic matter.Environ. Sci. Technol. 2008, 42 (23), 8630–8640.

(30) Benjamin, M. M. Water Chemistry; McGraw Hill: New York,2002; pp 648.

(31) Khaitan, S.; Dzombak, D. A.; Lowry, G. V. Chemistry of the acidneutralization capacity of bauxite residue. Environ. Eng. Sci.2009, 26 (5), 873–881.

(32) Holland, H. D. The Chemistry of the Atmosphere and Oceans;John Wiley & Sons Inc: New York, 1978; pp 154.

(33) Call, D.; Logan, B. E. Hydrogen production in a single chambermicrobial electrolysis cell lacking a membrane. Environ. Sci.Technol. 2008, 42 (9), 3401–3406.

(34) Franz, J. A.; Williams, R. J.; Flora, J. R. V.; Meadows, M. E.; Irwin,W. G. Electrolytic oxygen generation for subsurface delivery:Effects of precipitation at the cathode and an assessment ofside reactions. Water Res. 2002, 36 (9), 2243–2254.

(35) Petersen, M. A.; Sale, T. C.; Reardon, K. F. Electrolytic trichlo-roethene degradation using mixed metal oxide coated titaniummesh electrodes. Chemosphere 2007, 67 (8), 1573–1581.

(36) Kang, S. H.; Choi, W. Oxidative degradation of organic com-pounds using zero-valent iron in the presence of natural organicmatter serving as an electron shuttle. Environ. Sci. Technol.2009, 43 (3), 878–883.

(37) Motheo, A. J.; Pinhedo, L. Electrochemical degradation of humicacid. Sci. Total Environ. 2000, 256 (1), 67–76.

(38) Loffler, F. E.; Tiedje, J. M.; Sanford, R. A. Fraction of electronsconsumed in electron acceptor reduction and hydrogen thresh-olds as indicators of halorespiratory physiology. Appl. Environ.Microbiol. 1999, 65 (9), 4049–4056.

(39) Reible, D. D., Funnel and Gate Innovations -Stabilization andTreatment of Contaminated Sediments, RO1 2007-2010; Na-tional Institute Of Environmental Health Sciences: ResearchTriangle Park, NC, 2007.

ES101003J

VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8215