Environ. Sci. Technol. 2008, 42, 2118–2123
Abiotic Degradation of Hexahydro-1,3,5-trinitro-1,3,5-triazine in the Presence of Hydrogen Sulfide and Black Carbon JEROME M. KEMPER, EMAAN AMMAR, AND WILLIAM A. MITCH* Department of Chemical Engineering, Yale University, Mason Lab 313b, 9 Hillhouse Avenue New Haven, Connecticut 06520
Received September 24, 2007. Revised manuscript received November 30, 2007. Accepted December 3, 2007.
We report that hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) was rapidly destroyed by sulfides in the presence of black carbon, forming nitrite and formaldehyde, rather than toxic nitrosated reduction products. Although traditionally viewed as inactive sorbents, black carbons have been noted to participate in the destruction of certain contaminants, such as azo dyes, via quinonoid groups. However, in our experiments sulfide modification of quinones did not seem to be involved. Although at least 1.2 mM sulfides were needed for the reaction to proceed, abiotic natural attenuation of RDX in marine sediments may occur, because these concentrations are found in certain marine sediments, together with black carbon. In the absence of natural black carbons, synthetic black carbons, such as activated carbon, may be added to sediments. As compared with other in situ techniques, such as bioremediation and zerovalent iron cutoff trenches, which often generate nitrosated byproducts, this in situ, abiotic technique may be an attractive alternative.
Introduction The explosive compound, hexahydro-1,3,5-trinitro-1,3,5triazine (RDX), has replaced trinitrotoluene (TNT) as the primary explosive used by the United States military (1), but it is considered a possible human carcinogen by the U.S. Environmental Protection Agency (EPA) (2). Because RDX migrates more rapidly in aquifers than TNT (1), RDX contamination of soils and aquifers underlying munitions manufacturing sites and military bombing ranges is of current concern. The U.S. EPA has placed RDX on the Candidate Contaminant List (3), and the California Department of Health Services has assigned RDX a notification level of 300 ng/L (2). More effective in situ RDX remediation techniques are needed. Via several techniques, including bioremediation (4, 5), reduction by zero-valent iron (6), and Fe(II) bound to magnetite (7), RDX destruction can generate undesirable nitrosated byproducts such as hexahydro-1-nitroso-3,5dinitro-1,3,5-triazine (MNX), hexahydro-1,3-dinitroso-5nitro-1,3,5-triazine (DNX), and hexahydro-1,3,5-trinitroso1,3,5-triazine (TNX). Because of the microbial reduction of sulfate, hydrogen sulfide occurs naturally in many anaerobic soil and ground* Corresponding author telephone: (203) 432-4386; fax: (203) 4324387; e-mail:
[email protected]. 2118
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 6, 2008
water systems, such as marine sediments. Sulfides can be present in marine sediment pore waters from tens of micromolar to several millimolar concentrations (8). Black carbons, including wood chars and soots, constitute 15–30% of total organic carbon in marine sediments and soils (9, 10). Black carbons are important because, although RDX is only moderately hydrophobic (i.e., log Kow ) 0.9 (11)), it will concentrate on black carbons. Although traditionally viewed as inactive sorbents, black carbons were of interest because they often contain quinone groups that can mediate chemical reactions. Previous research has examined the sulfide reduction of nitroaromatics (12, 13) and azo dyes (14) in the presence of redox-mediating compounds, such as quinones, and the reduction of azo dyes by hydrogen sulfide via redoxmediating quinone groups on granular activated carbon (GAC), an anthropogenic form of black carbon (15). Alternatively, previous research noted that hydrogen sulfide can add to quinones to form mercaptoquinones (16) that may serve as potent nucleophiles. We evaluated the potential for hydrogen sulfide to mediate the destruction of RDX in the presence of black carbon. Specific goals included the characterization of reaction timescales, reagent requirements, and the evaluation of whether toxic nitrosated byproducts formed. Although the heterogeneous nature of black carbon inhibits our ability to draw firm conclusions regarding reaction pathways, we also performed experiments to initiate the characterization of the reaction pathway. Our findings indicate that RDX is rapidly degraded in the presence of sulfides and black carbons to low molecular weight ring-cleavage products. Because sulfides and black carbons often occur in marine sediments, our results indicate that abiotic natural attenuation of RDX is likely to be rapid when RDX is released to marine sediments.
Materials and Methods Materials. Chem Service RDX (1000 µg/mL) and SRI International MNX (98.5%), TNX (99.9%), DNX (57%), methylenedinitramine (98%), and 4-nitro-2,4-diazabutanal (99%) were used as received. Dimethylnitramine was synthesized according to the procedure of Mezyk et al. (17). The purity was estimated to be 94% by 1H NMR (Figure SI-1 of the Supporting Information). Nitrous oxide gas was purchased from Airgas (99% purity) and was diluted with ultra high purity nitrogen by a factor of 100 before use. Hydrogen sulfide stock solutions were made fresh daily by dissolving Acros sodium sulfide nonahydrate in deionized water and were standardized iodometrically (18). All other chemicals were reagent grade. Fisher activated carbon (GAC; 6–14 mesh) and graphite were used as received. Ponderosa Pine and Red Oak wood chars were prepared by placing wood shavings in a covered ceramic crucible and heating the contents at 400 °C for 2 h (19). Surface areas were characterized by N2 sorption (Table SI-1 of the Supporting Information). A seawater sample was collected from Jones Beach (New York). A sandy sediment was collected from an Atlantic coastal beach near Wellfleet, Cape Cod (Massachusetts). Experimental Procedures. Unless otherwise stated, all experiments were buffered with 20 mM phosphate buffer and were performed in duplicate. Solutions were purged with nitrogen gas prior to adding hydrogen sulfide and organic precursors. Solutions were decanted into 25 mL vials containing a weighed amount of carbon (GAC, wood shavings, or wood char), capped without headspace, and 10.1021/es702402a CCC: $40.75
2008 American Chemical Society
Published on Web 02/13/2008
FIGURE 1. Effect of mixing on RDX decay. Deionized water buffered at pH 7 with 20 mM phosphate buffer and containing 4.5 µM RDX, 3 mM sulfides, and 4 g/L GAC was either mixed (0) or not mixed (4). Error bars represent one standard deviation of experimental duplicates. Controls containing RDX with either sulfides without GAC or GAC without sulfides exhibited no significant RDX loss (not shown). Note the break in the x-axis. were placed on a rotating bed until analyzed unless otherwise stated. Periodically, the aqueous and solid phases of samples were analyzed for RDX, MNX, DNX, TNX, methylene dinitramine, and 4-nitro-2,4-diazabutanal by HPLC-UV (20, 21). Dimethylnitramine was analyzed by GC/MS. Formate, thiosulfate, and sulfate were measured by ion chromatography with conductivity detection. Aqueous nitrite was analyzed either by ion chromatography or colorimetrically (18). Aqueous formaldehyde was measured by HPLC-UV following derivatization with 2,4-dinitrophenylhydrazine (22). Nitrous oxide was measured by GC with thermal conductivity detection (23). Derivatization of GAC was carried out by the procedure of Langley and Fairbrother (24). Briefly, 0.5 g of GAC was placed in a 50 mL beaker within a vacuum chamber containing a flourinated derivatizing agent. The GAC was derivatized sequentially using three derivatizing agents: 2,2,2-trifluoroethylhydrazine for carbonyl groups, trifluoroacetic anhydride for hydroxyl groups, and 2,2,2-trifluoroethanol for carboxylic acid groups. The combined apparatus was cooled with acetone/dry ice slush and was opened to vacuum at approximately 50 mTorr for 30 min. The vacuum was sealed and left overnight. Although we could not calculate the percent conversion of carbonyls, Dr. Howard Fairbrother (Johns Hopkins University) verified that derivatization of carbonyls had occurred using X-ray photoelectron spectroscopy (22).
Results When 4.5 µM RDX (0.1125 µmoles) was mixed with 3 mM sulfides at pH 7 without black carbon, RDX was recovered quantitatively even after 21 h. Similarly, when a 4.5 µM RDX solution was mixed with 4 g/L GAC without sulfides, RDX sorbed nearly completely to the GAC surface within 0.5 h, but was recovered quantitatively from the GAC even after 32 h. However, in the presence of 3 mM sulfides and 4 g/L GAC, total destruction of RDX (i.e., aqueous and sorbed RDX) followed first order kinetics and was nearly complete within 2 h (Figure 1). When this same reaction was conducted without mixing, total RDX decayed slowly over the course of 1 week, indicating that mass transfer of RDX to the GAC surface was rate-limiting. In all further experiments, the solutions were mixed. A plot of ln(RDX/RDX0) versus time was linear over 2 log orders, indicating that the rate of RDX decay was first order with respect to RDX (Figure 2A). Plotting the pseudo-first order observed rate constants for RDX decay (kobs) versus
FIGURE 2. Dependence of RDX decay rates on the concentrations of RDX, GAC, and sulfides and on pH. All experiments were mixed and initially contained 4.5 µM RDX in deionized water buffered with 20 mM phosphate buffer. (A) RDX decay with 4 g/L GAC and in the presence (0) or absence (4) of 3 mM sulfides at pH 7. (B) Observed RDX decay rate at pH 7 in the presence (0) or absence (4) of 3 mM sulfides as a function of GAC concentrations. (C) Observed RDX decay rate as a function of sulfide concentrations in the presence of 4 g/L GAC at pH 7. Controls lacking sulfides showed no RDX decay. (D) Observed RDX decay rate as a function of pH with 2 g/L GAC in the presence (0) or absence (4) of 3 mM sulfides. Error bars represent one standard deviation of experimental duplicates. GAC concentration indicated that RDX removal was first order with respect to GAC concentrations up to 4 g/L (Figure 2B). Between 1.25 and 3 mM sulfides, kobs was first order in sulfide concentrations in the presence of 4 g/L GAC (Figure 2C). However, no significant decay was observed at sulfide concentrations less than 1.25 mM. To evaluate whether RDX removal occurs at combinations of lower GAC and sulfide concentrations, the decay of 4.5 µM RDX was examined over 7 d in the presence of 2 g/L GAC and 400 µM sulfides; no RDX decay was observed. The rate of RDX decay in the presence of sulfides and GAC increased with increasing pH over the range of 5–9 (Figure 2D). Controls lacking sulfides showed MNX > TNX > dimethylnitramine). No residual RDX or MNX was observed after 1 week (Figure 4). However, over two weeks, TNX degraded slowly, and no degradation of dimethylnitramine was observed. We evaluated the capability of other forms of black carbon to mediate the removal of 4.5 µM RDX in the presence of 3 mM sulfides buffered at pH 7. With 2 g/L Ponderosa Pine and Red Oak chars, controls lacking sulfides showed that RDX sorbed significantly to the char after 16 h, but no decay was observed (data not shown). In the presence of sulfides, 63 and 50% decreases in total RDX (sum of adsorbed and dissolved) concentrations were observed for Ponderosa and Red Oak chars, respectively, over 16 h (Figure 5). On the other hand, 2 g/L of the uncharred woods did not significantly sorb RDX or mediate RDX removal over 16 h. Like GAC, 4 g/L
Discussion
FIGURE 4. Decay of 4.5 µM RDX, MNX, TNX, or dimethylnitramine (DMN) mixed with 3 mM sulfides and 4 g/L GAC in deionized water buffered at pH 7 with 20 mM phosphate buffer. Left, center, and right bars represent concentration measurements (sum of adsorbed and dissolved) taken at 0, 7, and 14 d, respectively. After t ) 0 days no RDX, MNX, TNX, or DMN were detected in solution. Error bars represent one standard deviation of experimental duplicates.
FIGURE 5. Decay of 4.5 µM RDX in the presence of 3 mM sulfides in deionized water buffered at pH 7 with 20 mM phosphate buffer. Solutions were mixed with 2 g/L GAC, 4 g/L graphite, 2 g/L Pine, 2 g/L charred Pine, 2 g/L Oak, or 2 g/L charred Oak. White bars ) dissolved RDX, and grey bars ) sorbed RDX. RDX loss in controls lacking sulfides was
