Sequential Electrolytic Oxidation and Reduction of Aqueous Phase

Nov 3, 2005 - Contamination of soils and groundwater with energetic compounds has been documented at many former ammunition manufacturing plants ...
0 downloads 0 Views 218KB Size
Environ. Sci. Technol. 2005, 39, 9270-9277

Sequential Electrolytic Oxidation and Reduction of Aqueous Phase Energetic Compounds DAVID M. GILBERT* AND TOM C. SALE Department of Civil Engineering, Colorado State University, 1320 Campus Delivery, Fort Collins, Colorado 80523

Contamination of soils and groundwater with energetic compounds has been documented at many former ammunition manufacturing plants and ranges. Recent research at Colorado State University (CSU) has demonstrated the potential utility of electrolytic degradation of organic compounds using an electrolytic permeable reactive barrier (e-barrier). In principle, an electrolytic approach to degrade aqueous energetic compounds such as hexahydro1,3,5-trinitro-1,3,5-triazine (RDX) or 2,4,6-trinitrotoluene (TNT) can overcome limitations of management strategies that involve solely oxidation or reduction, through sequential oxidation-reduction or reduction-oxidation. The objective of this proof-of-concept research was to evaluate transformation of aqueous phase RDX and TNT in flow-through electrolytic reactors. Laboratory experiments were conducted using six identical column reactors containing porous media and expanded titaniummixed-metal-oxide electrodes. Three columns tested TNT transformation and three tested RDX transformation. Electrode sequence was varied between columns and one column for each contaminant acted as a no-voltage control. Over 97% of TNT and 93% of RDX was transformed in the reactors under sequential oxidation-reduction. Significant accumulation of known degradation intermediates was not observed under sequential oxidation-reduction. Removal of approximately 90% of TNT and 40% of RDX was observed under sequential reduction-oxidation. Power requirements on the order of 3 W/m2 were measured during the experiment. This suggests that an in-situ electrolytic approach may be cost-practical for managing groundwater contaminated with explosive compounds.

Introduction Contamination of soils and groundwater with energetic compounds has been documented at many former ammunition manufacturing plants, storage facilities, and ranges (1-3). Concern regarding the toxicity of energetic compounds and their breakdown products has resulted in a variety of remediation efforts. Published research on reductive approaches for degradation of energetic compounds (4-7) has focused primarily on biological processes or zerovalent iron technologies. Unfortunately, many of these approaches result in the accumulation of intermediate products that also pose concerns. Use of oxidative approaches has received increasing attention because the proposed degradation pathways are * Corresponding author e-mail: [email protected]. 9270

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

not as susceptible to formation of toxic intermediate compounds (8-10). However, oxidative pathways are not well understood and it is not clear that complete mineralization always occurs. Given the difficulty in achieving complete mineralization, several authors have suggested coupling of existing technologies utilizing both oxidative and reductive steps (6, 11, 12). Electrolytic degradation of aqueous contaminants, including energetic compounds, has been demonstrated in laboratory studies (6, 13, 14). These experiments generally utilize batch electrolytic reactors or cascades of divided flowthrough reactors with a focus on reductive transformation of the target compound. Advancing this type of technology to an in-situ approach is the focus of our research. Specifically, we have combined principles of a flow-through electrolytic reactor to the concept of a permeable reactive barrier. The aim of this research was to test the efficacy of an electrolytic permeable reactive barrier (e-barrier) that provides an oxidative step followed by a reductive step (or a reductive step followed by an oxidative step) to degrade energetic compounds at a lower cost than conventional technologies. Given the highly oxidized nature of energetic compounds, it is expected that cathodic reduction is the primary degradation mechanism. However, it is not clear which sequence will result in optimal overall transformation. Designing an electrically induced permeable reactive barrier required several alterations to typical electrochemical laboratory apparatus, most notably, selection of electrode materials that are environmentally benign, dimensionally stable, and cost practical at a field scale. To that end, expanded titanium coated with mixed metal oxide (Ti-mmo) electrodes were selected for use in this research. Ti-mmo has low aqueous solubility under the extreme conditions developed at the anode and cathode surface and is relatively inexpensive. In this paper we describe experiments that address the potential efficacy of electrolytic processes applied to a flowthrough reactor for the treatment of energetic compounds in water. Attention is given to removal of both parent and daughter products.

Experimental Section Column reactor studies were conducted to evaluate sequential electrolytic degradation of 2,4,6-trinitrotoluene (TNT) and hexahydro-1,3,5 trinitro-1,3,5 triazine (RDX). The column reactors are one-dimensional analogues to a field scale electrolytic permeable reactive barrier (e-barrier), constructed of Plexiglas, 90 cm long with an interior diameter of 10 cm (Figure 1). Two sets of expanded Ti-mmo electrodes were placed at the midpoint of the columns, 2.5 cm apart. The electrodes covered the entire cross-section of the columns. The Ti-mmo electrodes used in this experiment consisted of expanded titanium with a mixed metal oxide (IrO2/TaO5) coating (ELTECH Systems, Chardon, OH). The use of this material for both the anode and the cathode reflects the necessity for polarity reversal of the electrodes to minimize electrode fouling. Electrode fouling has been reported by others (15) and prevention of electrode fouling may be achieved using a polarity reversal strategy. This requirement of electrode material used in this experiment (stable as both an anode and a cathode over the potentials used in the experiment) probably results in significant inefficiencies with respect to electron transfer and catalytic properties of the electrode surfaces. Electrode polarity reversals, although expected to be necessary during a field application, were not conducted during these experiments. 10.1021/es051452k CCC: $30.25

 2005 American Chemical Society Published on Web 11/03/2005

FIGURE 1. Column reactor schematic and experimental setup. Each column was fitted with glass sample ports along the length of the column, Ag/AgCl reference electrodes (World Precision Instruments, Sarasota, FL), and gas vents. The junctions of the reference electrodes were placed immediately adjacent (approximately 2 mm) to the electrode surfaces to minimize IR drop. The space between and adjacent to the electrodes was packed with glass beads. The remainder of the column was packed with 16-40 quartz feldspar sand (Colorado Silica Sand) to simulate an idealized aquifer material. Columns were packed sub-aqueous to limit air entrapment and to facilitate settlement. Three columns for each contaminant were used to test target compound degradation. Since the reactors can be operated in a positive-negative sequence (sequential oxidation-reduction) with respect to fluid flow or a negativepositive sequence (sequential reduction-oxidation) with respect to fluid flow, reactors were operated to evaluate treatment efficacy of each sequence. The third column in each set consisted of a zero voltage control. The control reactors provided data regarding losses of aqueous phase TNT and RDX due to adsorption to matrix materials and other losses not associated with electrolytic transformation. Comparison of results from the active column reactors to the control columns provided a basis for quantifying electrolytic transformation of the target compounds. Influent RDX and TNT solutions were prepared using a calcium-carbonate water (total dissolved solids ∼70 mg/L, pH 7.2). The water was spiked with analytical-grade RDX and TNT (AccuStandard, Inc., New Haven, CT). Influent TNT and RDX concentrations were 2.25 µM (0.5 mg/L) and 4.40 µM (1.0 mg/L), respectively. The TNT and RDX concentrations used in the experiment were intended to simulate typical contaminated groundwater conditions. Influent solutions to the column reactors were stored in 20-L Tedlar bags with zero headspace and covers to limit photodegradation of the target compound. Solutions were pumped through the columns at a rate of 0.5 mL/min using ISMATEC multichannel peristaltic pumps equipped with Viton tubing. The resulting seepage velocity was approximately 0.3 m/day. Other than minor Viton tubing sections, all plumbing was small-diameter glass tubing. Direct current (at constant voltage) was provided to the electrodes using a GW model GPS-W3030D DC power supply. Aqueous phase influent RDX and TNT concentration, electrode spacing, and inorganic water quality were held constant throughout the experiment. During the experiment, imposed electrical potentials of 5, 10, and 15 V were evaluated for target compound removal. In all cases, the electrode potential was intended to provide overpotential relative to the degradation of energetic materials based on work by others (6). The selection of voltages applied was intended to (1) provide proof-of-concept, and (2) provide information regarding operational voltages for a full-scale system. Measurements taken to evaluate reactor performance included resultant cell current (current density),

electrode potentials relative to Ag/AgCl reference electrodes, oxidation reduction potential (ORP), pH, and contaminant concentration. pH measurements were made with a combination pH/ reference (Ag/AgCl) electrode and Denver Instrument model AP25 meter. ORP measurements were made with a combination platinum/reference (Ag/AgCl) electrode and Denver Instrument AP25 meter. Measurements were corrected to reference the standard hydrogen electrode using the Nernst equation and reported as Eh. ORP and pH measurements were conducted in a low-volume-flow cell that provided isolation from atmospheric conditions. Analysis for energetic compounds was conducted using high-performance liquid chromatography (HPLC) (HewlettPackard 1100, RP C18 column, acetonitrile/water (50:50) isocratic mobile phase, HP1100 UV detector at 254 nm), and gas chromatography/mass spectrometry (GC/MS) (Agilent Series 6890 and Agilent 5973 mass selective detector). Aqueous samples were preconcentrated using Waters Porapak RDX cartridges, prior to analysis by HPLC and GC/MS. Cartridges were conditioned prior to extraction according to manufacturer specifications and column samples were extracted immediately following sampling. Extracted samples were stored at 4 °C and analysis by GC or HPLC was conducted within 5 days of extraction. Calibration of the instruments for TNT, associated transformation intermediates, and RDX was conducted using commercially available standards (AccuStandard, Inc., New Haven, CT). Analytical standards for hexahydro-1-nitroso-3,5-dinitro-1,3,5-triazine (MNX), hexahydro-1,3-dinitroso-5-nitro-1,3,5-triazine (DNX), hexahydro-1,3,5-trinitro-1,3,5-triazine (TNX), and methylenedinitramine (MDNA) (SRI International; Menlo Park, CA) were used to identify reaction intermediates associated with transformation of RDX. The mean practical quantitation limits (PQLs) for the experiment were 10 µg/L for TNT and 40 µg/L for RDX using GC/MS. Estimated practical quantification limits for TNT and RDX using HPLC-UV are 5 and 10 ug/L, respectively.

Results Effluent Comparison. Column effluent concentrations of the target compounds were monitored over time at each applied voltage until steady-state conditions were observed. Under sequential oxidation-reduction (positive-negative sequence), fractional removal of both TNT and RDX was greater than 90% based on GC/MS analysis (Table 1). Under sequential reduction-oxidation (negative-positive sequence), lower fractional removal for both test compounds was observed. In all cases, effluent concentrations from the oxidationreduction sequence were below the PQLs. Variability of the reported results (PQLs) reflects varying sample volumes (due to the solid-phase extraction procedure). VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9271

FIGURE 2. Normalized TNT concentration profile at 5 V. Electrodes are located at position 0 cm.

TABLE 1. Steady State TNT and RDX Transformation Based on GC/MS Analysis Relative to Control at the Three Test Voltages percent transformed relative to control reactor contaminant

applied voltage

positive-negative sequence

negative-positive sequence

TNT TNT TNT RDX RDX RDX

15 10 5 15 10 5

>97 >97 >98 >90 >90 >90

37 45 89 29 43 49

Reactor Concentration Profiles. Concentration profiles were measured under steady-state conditions at 5 V applied potential and analyzed using HPLC-UV (Figures 2 and 3). It can be seen that for both target compounds, degradation relative to the control column occurred upstream and downstream of the electrodes (position 0). Losses of target compound because of adsorption or other processes are also illustrated in Figures 2 and 3 (control columns). Transformation rates in the positive-negative sequence test reactors calculated using HPLC generated concentrations were 97.6% for TNT and 98.1% for RDX relative to the control reactors. This result correlates well with the TNT values reported above, but suggests that the RDX transformation rates reported based on GC/MS analyses may be low. This result also reflects the lower quantification limit relative to the GC/MS method. Observation of concentration profiles suggests transformation of both TNT and RDX upgradient of the electrode set. Approximately 45% and 25% of the influent TNT was transformed within the porous media upgradient of the electrodes under sequential oxidation-reduction and sequential reduction-oxidation, respectively. For RDX approximately 30% and 35% of the influent RDX was transformed within the quartz feldspar sand pack under sequential oxidation-reduction and sequential reduction-oxidation, respectively. Hypotheses as to this upgradient degradation include (a) axial back-mixing within the column resulting in dilution of the upgradient solution, (b) axial back-mixing within the column resulting in homogeneous reaction of the target compound with intermediates formed at the electrode 9272

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

FIGURE 3. Normalized RDX concentration profile at 5 V. Electrodes are located at position 0 cm.

TABLE 2. Mean Electrode Potentials (V vs SHE) Measured in the Reactors positive-negative sequence

negative-positive sequence

contaminant

applied potential (V)

anode potential (V)

cathode potential (V)

anode potential (V)

cathode potential (V)

TNT TNT TNT RDX RDX RDX

15 10 5 15 10 5

1.47 1.47 1.40 1.49 1.44 1.37

-0.80 -0.87 -0.97 -0.66 -0.95 -1.07

1.42 1.41 1.34 1.49 1.47 1.40

-0.87 -0.89 -0.84 -0.83 -0.83 -0.84

surface, or (c) electrolytic reactions occurring as a result of a potential difference between the upgradient electrode and the ground potential of the sand pack. Axial back-mixing was investigated under no applied voltage conditions using a bromide tracer, but was not found to be significant. Far smaller losses of the target compounds were observed in the control reactors (containing electrode sets to which no voltage was applied). These losses are attributed to either adsorption of TNT or RDX to the electrode materials or catalytic transformation on the electrode surfaces. Transformation products were not observed in the control reactors. Electrode Potentials. Table 2 contains measurements of electrode potentials relative to the Ag/AgCl reference electrodes corrected to the standard hydrogen electrode (SHE). The high applied potentials relative to those resulting in the electrolytic degradation of water (1.23 V and -0.83 V) is hypothesized to result in the lack of variation in the measurements as well as the minimal difference in treatment efficacy for both contaminants at all voltages tested. Within the reactors testing the positive-negative electrode sequence (sequential oxidation-reduction), decreasing the applied voltage resulted in decreased upstream anode potential and a more negative downstream cathode potential for both TNT and RDX. This type of behavior was not observed in the reactors testing sequential reduction-oxidation. It is hypothesized that the solubility of oxygen gas relative to

FIGURE 4. pH shifts observed in the reactors testing sequential oxidation-reduction (left) and sequential reduction-oxidation (right) for RDX (top) and TNT (bottom). Electrodes are located at position 0 cm. hydrogen gas generated at the upstream electrode accounts for this difference since the dissolved gases are advected to the downstream electrode surface where they are either reduced or oxidized resulting in differences in the measured electrode potentials. Shifts in pH and Eh. Figure 4 presents pH shifts measured as a function of position in the columns. Under sequential oxidation-reduction pH remains relatively constant at all test voltages. Under sequential reduction-oxidation however, pH shifts were more significant (Figure 4). This result suggests that the high fractional degradation observed in the reactors testing sequential oxidation-reduction was not due to shifts in pH (e.g., base hydrolysis). It also may indicate a lower rate of electrolysis of water compared to the reactors testing sequential reduction-oxidation, since the solubility of oxygen is high compared to hydrogen, advection of dissolved oxygen to the downstream cathode and subsequent reduction to H2O, may provide a pH buffering effect. The pH shift observed in the reactors testing sequential reduction-oxidation highlights a significant advantage of operating the reactor in an oxidation-reduction sequence. In formations containing carbonate geologies, the increase in pH observed at an upgradient negative electrode, while advantageous from a treatment perspective, would result in precipitation of carbonate solids on the cathode surface. The potential for fouling and/or plugging of the cathode in many groundwater systems can be avoided by operating the upgradient electrode as an anode, thus minimizing pH increases at the downgradient cathode as observed (Figure 4, left panels). Operation of the reactor in an oxidationreduction sequence may also provide the acidic conditions necessary for reduction reactions to occur more completely at the downgradient cathode.

Substantial shifts in Eh were observed in the reactors under both electrode sequences (Figure 5), suggesting that Eh shifts may be responsible for the observed degradation. Increased removal of TNT and RDX under sequential reductionoxidation occurred at 5 V relative to 10 and 15 V. This may be the result of lower Eh achieved under 5 V. These results may also indicate that groundwater Eh conditions may influence the efficacy of this approach at a field scale. For example, the behavior of the reactor is likely to be different if the influent water is anoxic. Electrical Performance. Current densities as a function of time and applied (cell) voltage measured during the experiment are illustrated in Figures 6 and 7. The data provide support for observed pH shift differences between electrode sequences. Lower current densities measured under sequential reduction-oxidation at 5 V for TNT and 10 and 5 V cell potential for RDX, relate not only lower rates of target compound transformation, but may also indicate lower rates of H2(aq) oxidation at the downgradient anode, due to lower solubility of cathodically generated H2(g). Intermediate Compound Identification. For purposes of this proof of concept study, only those degradation compounds that were commercially available were evaluated. The following TNT degradation intermediates were explicitly included in the search: the dinitrotoluene isomers (2,6-DNT and 2,4-DNT); 1,3,5-trinitrobenzene (1,3,5-TNB); 1,3-dinitrobenzene (1,3-DNB); the nitrotoluene isomers (2-NT, 3-NT, and 4-NT); nitrobenzene (NB); and the amino-dinitrotoluene isomers (2-ADNT and 4-ADNT). None of the TNT degradation intermediates included in the search was observed in the reactor testing sequential oxidation-reduction, suggesting that any degradation intermediates are present at concentrations below detection. Of the compounds that were included in the explicit search, VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9273

FIGURE 5. Eh shifts observed in the reactors testing sequential oxidation-reduction (left) and sequential reduction-oxidation (right) for RDX (top) and TNT (bottom). Electrodes are located at position 0 cm.

FIGURE 6. Current density measured during the TNT experiment. only 2-ADNT was identified in quantifiable concentrations and was observed only under sequential reduction-oxidation. Maximum 2-ADNT concentration observed was approximately 63 ug/L immediately upstream of the cathode, 9274

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

corresponding to approximately 6% of the influent TNT concentration. The concentration of 2-ADNT decreased downgradient of the electrode set. Observation of 2-ADNT is consistent with reduction of TNT noted during microbial

FIGURE 7. Current density measured during the RDX experiment transformations (16), but was not reported in electrochemical reduction experiments (6). RDX transformation intermediates included in the investigation were MNX, DNX, TNX, and MDNA. None of these compounds were detected under either sequential oxidation-reduction or sequential reduction-oxidation. Power Requirement. Amperage data collected from the reactors during the experiment were used to calculate power requirements for a field-scale application. Calculated power densities ranged from 38 W/m2 at 15 V to 1 W/m2 at 5 V applied potential difference. Overall, the results indicate that power consumption associated with a full-scale application could be quite low.

Discussion On the basis of these experiments and mechanisms identified by others (17), the following proposed mechanisms for the degradation of TNT and RDX have been identified and are discussed below. (i) Direct Electrolytic Oxidation and/or Reduction of the Target Compound at the Electrode Surfaces. In order for direct electrolytic oxidation and/or reduction at the electrode surfaces to be responsible for the observed degradation, the standard potentials for oxidation and/or reduction must be exceeded by the anode and cathode, respectively. Unfortunately, published standard potentials or thermodynamic data from which they could be calculated are not available for the compounds tested (at Ti-mmo electrodes). Furthermore, the reaction pathways are not well understood. However, comparing the removal rates to the measured electrode potentials suggests the standard potentials for the reduction of TNT and RDX were exceeded at all test voltages. Work by others (6, 14) suggests that electrolytic reduction of TNT and RDX occurs between -0.11 and -0.34 V (vs SCE) and -1.32 V (vs Ag/AgCl), respectively. This discrepancy is likely due to the choice of electrode materials (dripping mercury for TNT and glassy carbon for RDX). The results also suggest that a lower applied potential (than 5 V) could result in equally high removal rates. It is not clear from the data if both oxidation and reduction are required or which step is most responsible for the observed

removal. Research by others has indicated that reductive pathways can remove high percentages of TNT under cathodic reduction (6). Conditions sufficient for cathodic reduction appear to exist at the cathode under both sequential oxidation-reduction and reduction-oxidation and reductive conditions propagated downstream in the reactors testing sequential oxidation-reduction as measured by Eh (Figure 5). The difference in removal rates observed between the sequential reduction-oxidation and sequential oxidationreduction could simply be a difference in residence time in the vicinity of a cathode or under reducing conditions in the sand matrix downstream of the cathode. pH conditions developed at the upstream anode may also provide necessary conditions for reduction occurring at the downstream cathode, resulting in greater observed transformation. We are currently unaware of existing research regarding direct electrolytic oxidation of aqueous energetic compounds probably because of the general observation that the oxidized nature of the nitro groups limits the possibility of transformation through oxidative approaches. However, direct electrolytic oxidation of chlorinated phenols and other chlorinated compounds have been reported (15, 18-20). Column profile concentration measurements (Figures 2 and 3) may support minor amounts of anodic oxidation of both TNT and RDX (i.e., apparent transformation of TNT and RDX in the vicinity of the anodes), but these measurements may be artifacts of the sampling volumes required. Another potential hypothesis that would explain the observed transformation in the vicinity of the anode is illustrated by the control column apparent transformation under no voltage conditions. Both control columns indicated minor amounts (30% for TNT and 20% for RDX) of target compound transformation/removal through the electrode sets. It is not clear if this transformation is due to catalytic processes at the Ti-mmo surfaces or adsorption to the electrode surfaces. (ii) Reaction of the Target Compound with Electrolytically Generated Oxygen and/or Hydrogen. At the electrode potentials measured during the experiment, oxygen gas evolution and hydrogen gas evolution were observed at the anode and cathode surfaces (respectively) at all applied cell potentials. The rate of gas generation appeared to be related VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9275

to applied (cell) potential with the highest gas generation rates observed at 15 V applied. Very low gas generation rates were observed at 5 V applied potential. Given the semi-volatile nature of both TNT and RDX, it is unlikely that gas stripping contributed significantly to the observed transformation, although gases were not evaluated specifically for TNT or RDX. The role of electrically generated dissolved oxygen and/ or hydrogen in the degradation of the target compounds is not expected to be significant alone, but the formation of each requires intermediates that may be reactive with the target compounds. Additionally, when applied in-situ to a groundwater system, dissolved oxygen and hydrogen may provide substrate for microbial degradation of the target compounds and any daughter compounds that may require removal. (iii) Reaction of the Target Compounds with Electrolytically Generated Hydroxyl Radicals and/or Peroxide. Several researchers have identified reaction pathways involving electrolytically generated hydroxyl radicals and/or peroxide for degradation of organic compounds (21-23). These approaches are hypothesized to involve a Fenton’s reagent that is generated electrically in the presence of aqueous iron or through catalysis at the electrode surface. The mechanism requires cathodic generation of peroxide by reduction of anodically produced dissolved oxygen. This mechanism may be responsible for the observed removal under sequential oxidation-reduction as the measured potentials would support the formation of dissolved oxygen at the upstream anode. The dissolved oxygen would be advected to the cathode where it would become reduced forming hydroxyl radicals and peroxide in the process. (iv) Alkaline Hydrolysis of Target Compounds under Cathodically Generated pH Conditions. Electrolysis of water under reducing conditions is expected to result in the formation of hydrogen gas and hydroxide ions. The pH conditions generated at the cathode surface can be very high under unbuffered conditions. Alkaline hydrolysis of explosive compounds is well established in the literature and occurs at pH values ranging from 10 to 12, the reaction kinetics dependent on pH (24-29). This mechanism may be important at the cathode surface where sufficiently high pH conditions may exist. The maximum pH values measured in the vicinity of the cathodes was pH 9 (Figure 4), which would not indicate alkaline hydrolysis is significant at a bulk scale. However, the measured pH reflects an integrated volume in the vicinity of the cathode and not at the cathode surface. The results from this research suggest that an electrolytic approach to a permeable reactive barrier may provide a costeffective alternative to traditional reactive barrier technologies. Specifically, high fractional removal of TNT and RDX was observed with minimal generation of deleterious products. Additionally, very low power requirements were demonstrated through laboratory studies.

Acknowledgments Funding for the development of the e-barrier concept was provided by the Solvents in Groundwater Research Consortium. Funding for the work described in this paper was provided by the Strategic Environmental Research and Development Program. We thank the technical reviewers for valuable comments and input.

Literature Cited (1) ) Spalding, R. F.; Fulton, J. W. Groundwater munition residues and nitrate near Grand Island, Nebraska, U.S.A. J. Contam. Hydrol. 1988, 2, 139-153. (2) Jerger, D. E.; Woodhull, P. Applications and costs for biological treatment of explosives-contaminated soils in the U. S. In Biodegradation of Nitroaromatic Compounds and Explosives; Spain, J. C., Hughes, J. B., Knackmuss, H. J., Eds.; Lewis Publishers: Boca Raton, FL, 2000. 9276

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 23, 2005

(3) Card, R. E.; Autenrieth, R. Treatment of HMX and RDX Contamination; Report ANRCP-1998-2; Amarillo National Resource Center for Plutonium: Amarillo, TX, 1998. (4) Agrawal, A.; Tratnyek, P. G. Reduction of nitro aromatic compounds by zerovalent iron metal. Environ. Sci. Technol. 1996, 30, 153-160. (5) Riefler, R. G.; Smets, B. F. Enzymatic reduction of 2,4,6trinitrotoluene and related nitroarenes: kinetics linked to oneelectron redox potentials. Environ. Sci. Technol. 2000, 34, 39003906. (6) Rodgers, J. D.; Bunce, N. J. Electrochemical treatment of 2,4,6trinitrotoluene and related compounds. Environ. Sci. Technol. 2001, 35, 406-410. (7) Schmelling, D. C.; Gray, K. A.; Kamat, P. V. Role of reduction in the photocatalytic degradation of TNT. Environ. Sci. Technol. 1996, 30, 2547-2555. (8) Bose, P.; Glaze, W. H.; Maddox, D. S. Degradation of RDX by various advanced oxidation processes: I. reaction rates. Water Res. 1998, 32, 997-1004. (9) Bose, P.; Glaze, W. H.; Maddox, D. S. Degradation of RDX by various advanced oxidation processes: II. Organic by-products. Water Res. 1998, 32, 1005-1018. (10) Zoh, K.; Stenstrom, M. K. Fenton oxidation of hexahydro-1,3,5trinitro-1,3,5-triazine (RDX) and octahydro-1,3,5,7-tetranitro1,3,5,7-tetrazocine (HMX). Water Res. 2002, 36, 1331-1341. (11) Funk, S. B.; Roberts, D. J.; Crawford, D. L.; Crawford, R. L. Initialphase optimization for bioremediation of munition compoundcontaminated soils. Appl. Environ. Microbiol. 1993, 59, 21712177. (12) Spanggord, R. J.; Yao, C. D.; Mill, T. Oxidation of aminodinitrotoluenes with ozone: products and pathways. Environ. Sci. Technol. 2000, 34, 497-504. (13) Jolas, J. L.; Pehkonen, S. O.; Maloney, S. W. Reduction of 2,4dinitrotoluene with graphite and titanium mesh cathodes. Water Environ. Res. 2000, 72, 179-188. (14) Bonin, P. M.; Bejan, D.; Schutt, L.; Hawari, J.; Bunce, N. J. Electrochemical reduction of hexahydro-1,3,5-trinitro-1,3,5triazine in aqueous solutions. Environ. Sci. Technol. 2004, 38, 1595-1599. (15) Rodgers J. D.; Jedral, W.; Bunce, N. J. Electrochemical oxidation of chlorinated phenols. Environ. Sci. Technol. 1999, 33, 14531457. (16) McCormick, N. G.; Feeherry, F. E.; Levinson, H. S. Microbial transformation of 2,4,6-trinitrotoluene and other nitroaromatic compounds. Appl. Environ. Microbiol. 1976, 31, 949-958. (17) Bunce, N. J.; Merica, S. G.; Lipkowski, J. Prospects for the use of electrochemical methods for the destruction of aromatic organochlorine wastes. Chemosphere 1997, 35, 2719-2726. (18) Oturan, M. A.; Peiroten, J.; Chartrin, P.; Acher, A. J. Complete destruction of p-nitrophenol in aqueous medium by electroFenton method. Environ. Sci. Technol. 2000, 34, 3474-3479. (19) Savall, A. Electrochemical treatment of industrial organic effluents. Chimia 1995, 49, 23-27. (20) Hsiao, Y.-L.; Nobe, K. Hydroxylation of chlorobenzene and phenol in a packed-bed flow reactor with electrogenerated Fenton reagent. J. Appl. Electrochem. 1993, 23, 943-946. (21) Oturan, M. A.; Oturan, N.; Lahitte, C. Trevin, S. Production of hydroxyl radicals by electrochemically assisted Fenton’s reagent. Application to the mineralization of an organic micropollutant, pentachlorophenol. J. Electroanal. Chem. 2001, 507, 96-102. (22) Boye, B.; Dieng, M. M.; Brillas, E. Degradation of herbicide 4-chlorophenoxyacetic acid by advanced electrochemical oxidation methods. Environ. Sci. Technol. 2002, 36, 3030-3035. (23) Oturan, M. A. An ecologically effective water treatment technique using electrochemically generated hydroxyl radicals for in situ destruction of organic pollutants: application to herbicide 2,4D. J. Appl. Electrochem. 2000, 30, 475-482. (24) Jones, W. H. Mechanism of the homogeneous alkaline decomposition of cyclotrimethylenetrinitramine: kinetics of consecutive second- and first-order reactions. A polarographic analysis for cyclotrimethylenetrinitramine. J. Am. Chem. Soc. 1954, 76, 829-835. (25) Emmrich, M. Kinetics of alkaline hydrolysis of 2,4,6-trinitrotoluene in aqueous solution and highly contaminated soils. Environ. Sci. Technol. 1999, 33, 3802-3805. (26) Croce, M.; Okamoto, Y. Cationic micellar catalysis of the aqueous alkaline hydrolyses of 1,3,5-triaza-1,3,5-trinitrocyclohexane and

1,3,5,7-tetraaza-1,3,5,7-tetranitrocyclooctane. J. Org. Chem. 1979, 44, 2100-2103. (27) Heilmann, H. M.; Wiesmann, U.; Stenstrom, M. K. Kinetics of the alkaline hydrolysis of high explosives RDX and HMX in aqueous solution and adsorbed to activated carbon. Environ. Sci. Technol. 1996, 30, 1485-1492. (28) Hoffsommer, J. C.; Kubose, D. A.; Glover, D. J. Kinetic isotope effects and intermediate formation for the aqueous alkaline homogeneous hydrolysis of 1,3,5-triaza-1,3,5-trinitrocyclohexane (RDX). J. Phys. Chem. 1977, 81, 380-385.

(29) Balakrishnan, V. K.; Halasz, A.; Hawari, J. Alkaline hydrolysis of the cyclic nitramine explosives RDX, HMX, and CL-20: new insights into degradation pathways obtained by the observation of novel intermediates. Environ. Sci. Technol. 2003, 37, 18381843.

Received for review July 25, 2005. Revised manuscript received September 26, 2005. Accepted September 28, 2005. ES051452K

VOL. 39, NO. 23, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

9277