Environ. Sci. Technol. 2005, 39, 9683-9688
Increasing Fe0-Mediated HMX Destruction in Highly Contaminated Soil with Didecyldimethylammonium Bromide Surfactant JEONG PARK, STEVE D. COMFORT, PATRICK J. SHEA,* AND JONG SUNG KIM School of Natural Resources, University of Nebraska-Lincoln, Lincoln, Nebraska 68583-0915
Mixtures of energetic compounds pose a remediation problem for munitions-contaminated soil. Although treatment with zerovalent iron (Fe0) can be effective, RDX and TNT are more readily destroyed than HMX. Adding didecyldimethylammonium bromide (didecyl) at 2% w/v with 3% (w/v) Fe0 to a 20% slurry of Los Alamos National Laboratory soil containing solid-phase HMX (45 000 mg/kg) resulted in >80% destruction within 6 days. Because the HMX concentration did not increase in solution and the didecyl equilibrium concentration was well below the critical micelle concentration, we conclude that the solution primarily contained didecyl monomers. The adsorption isotherm for didecyl on iron is consistent with electrostatic adsorption of monomers and some hydrophobic partitioning at low equilibrium concentrations. Fe0 pretreated with didecyl was superior to Fe0 alone or mixed with didecyl in removing HMX from solution, but it was less effective than Fe0 + didecyl when solid-phase HMX was present. Reseeding HMX to mimic dissolution indicated an initial high reactivity of didecyl-pretreated Fe0, but the reaction slowed with each HMX addition. In contrast, reaction rates were lower but reactivity was maintained when Fe0 and didecyl were added together and didecyl was included in fresh HMX solutions. Destruction of solid-phase HMX requires low didecyl concentrations in solution so that hydrophobic patches are maintained on the iron surface.
Introduction Explosives are broadly categorized as primary or secondary on the basis of their susceptibility to initiation. Because secondary explosives are formulated to detonate only under specific conditions and are the main charge or boostering explosive used in munitions, they are much more prevalent at military sites than primary explosives (1). Examples of secondary explosives include TNT (2,4,6-trinitrotoluene), RDX (hexahydro-1,3,5-trinitro-1,3,5-triazine), and HMX (octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine). In a survey aimed at demonstrating the prevalence of secondary explosives in contaminated soils, the U.S. Army Cold Regions Research and Engineering Laboratory (CRREL) found that TNT was most frequently detected (85%), followed by RDX (44%), and HMX (22%) (1). Successful remediation of these soils requires removal of all three contaminants. We previously showed that zerovalent iron (Fe0) is an effective reductant for RDX and TNT (2, 3), but HMX is more * Corresponding author telephone: +1 402 472 1533; fax: +1 402 472 7904; e-mail:
[email protected]. 10.1021/es050948z CCC: $30.25 Published on Web 11/10/2005
2005 American Chemical Society
difficult to remove in highly contaminated soils (4). We also confirmed that Fe0 preferentially reduces RDX over HMX in a binary mixture (4). One difficulty in treating HMX is its low water solubility (less than 5 mg L-1) (5). Increasing the contaminant concentration in solution is usually desirable for remediation purposes. Several studies have shown that surfactants increase the aqueous solution concentrations of hydrophobic organic compounds by partitioning the solute into the hydrophobic interior of micelles when surfactant concentrations are above the critical micelle concentration (CMC) (6-8). Although increasing the solution concentration is desirable, contaminant reduction by Fe0 is a surfacemediated reaction that requires close contact between the contaminant and the iron surface. For sparingly soluble compounds, destruction rates are expected to increase as the contaminant concentration at the iron surface increases. Surfactant sorption at the iron-water interface might facilitate this process by increasing the hydrophobic contaminant concentration at the reactive Fe0 surface (4, 9, 10). Facilitated destruction is supported by the enhanced tetrachloroethylene (PCE) destruction rates of iron pretreated with hexadecyltrimethylammonium bromide (HDTMA cationic surfactant) (9). However, this process is complicated by the heterogeneous composition and temporal nature of the oxides forming on the iron surface. During Fe0 corrosion, magnetite, hematite, lepidocrocite, and goethite are produced, depending on the oxygen supply (11). These oxides have zero points of charge (zpc) ranging from 6.5 to 8.5 (12), and cationic surfactants will be electrostatically (ionically) adsorbed at pH > zpc (13). Hydrophobic adsorption will also occur on neutral oxide surfaces and carbon inclusions via surfactant hydrocarbon tails. We previously demonstrated that cationic surfactants can increase HMX solubility and facilitate HMX destruction with Fe0 (4). Didecyldimethylammonium bromide (didecyl) was more effective than hexadecyltrimethylammonium bromide (HDTMA). Our present objective was to use didecyl to increase HMX destruction by Fe0 in highly contaminated soil containing solid-phase HMX.
Experimental Section Reagents, Iron Source, and Soil. Technical-grade RDX and TNT were obtained from the U.S. Biomedical Research and Development Laboratory (Fredrick, MD). Technical-grade HMX (91% purity) was a gift from Sandia National Laboratories (Albuquerque, NM). Analytical standards of the explosives were obtained from the Indian Head Division, Naval Warfare Center (Indian Head, MD); additional analytical RDX was obtained from Accustandard (New Haven, CT). Unannealed iron (Peerless Metal Powders, Detroit, MI) was used in all experiments and had a specific surface area of 2.55 m2 g-1 (Micromeritics, Norcross, GA) and contained approximately 2% carbon (Peerless Metal Powders). Didecyl cationic surfactant {[CH3(CH2)9]2N(CH3)2Br, molecular weight 406.54} was obtained from Aldrich (Milwaukee, WI). All experiments were conducted using 50-mL Teflon tubes, and treatments were replicated at least three times. Soil was obtained from the southwestern sector of the Los Alamos National Laboratory (LANL) property (Los Alamos, NM) known as Technical Area 16. Activities in this area included high-explosives (HE) manufacturing, research, development, and testing. The soil was collected from the middle and sides of a discharge pond previously used for munitions wastewater disposal. The composited soil contained TNT (360 mg kg-1), RDX (15 000 mg kg-1), and HMX (45 000 mg kg-1), as determined by HPLC. VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9683
Destruction of Solid-Phase HMX. A batch experiment was conducted to determine Fe0-mediated destruction of solid-phase HMX in aqueous solution containing didecyl. Solid-phase HMX (1.12 g of technical-grade HMX in 25 mL of H2O) was treated with 4% (w/v) Fe0 (1 g in 25 mL) with and without 2% (w/v) didecyl (49 mM). Because RDX was present in the technical HMX (approximately 0.36%), both the HMX and RDX concentrations were monitored during agitation of the mixture on a rotary shaker for 170 days at 25 °C. To determine didecyl- and Fe0-mediated destruction of HMX in highly contaminated soil, slurries were prepared by mixing 5 g of LANL soil with 25 mL of H2O and adding didecyl at 49, 74, and 98 mM (2, 3, and 4% w/v). The mixtures were agitated on an oscillating shaker (100 rpm) for 48 h at 25 °C. HE concentrations in the solution phase were measured by removing 1.5 mL, centrifuging (12 000g), and analyzing by HPLC. Zerovalent iron (3 and 4% w/v) was then added to produce the following treatments: control (soil slurry only); soil slurry + Fe0, and soil slurry + didecyl (2, 3, and 4%) + Fe0. After 6 days of additional shaking, 100 mL of CH3CN was added, and the mixture was shaken for 24 h to determine total extractable HE concentrations. To determine didecyl sorption onto the soil, slurries were prepared by mixing 2 g of LANL soil with 10 mL of H2O and adding didecyl at 25, 49, 78, and 98 mM (1, 2, 3, and 4% w/v). The mixtures were agitated for 48 h before determination of didecyl concentrations. To determine equilibrium concentrations of didecyl following iron addition, Fe0 was added to each tube at 4% (w/v). After 6 days of additional shaking, tubes were centrifuged (5000g), and 1.5 mL of supernatant was removed and centrifuged at 12 000g, and 1 mL was used to determine didecyl concentrations. Sorption of didecyl to the iron was determined by equilibrating 2 g of Fe0 with 25 mL of didecyl solution (initial concentration 0.5-20 mM). After being agitated for 24 h, the tubes were centrifuged, and the supernatant was analyzed as in the previous experiments. The amount adsorbed was determined from the change in solution concentration, and isotherms were constructed. Dynamics of Enhanced Destruction. To determine the dynamics of the didecyl treatment, we pretreated Fe0 with didecyl before using it in batch experiments. The pretreated iron was prepared by equilibrating 1 g of Fe0 with 25 mL of 2% didecyl (w/v) solution in an anaerobic chamber. After being agitated for 24 h, the tubes were centrifuged (5000g), and the pretreated Fe0 was vacuum filtered through Whatman nuclepore 0.45-µm filter paper and air-dried before use. The morphology and mineralogical characteristics of the pretreated iron surface were characterized by scanning electron microscopy (SEM) and X-ray diffraction (XRD). Didecylpretreated Fe0 was compared to Fe0 alone and Fe0 + 2% didecyl by agitating 1 g with 25 mL of HMX (3 mg L-1) solution an orbital shaker. The HMX concentration was measured at 1, 2, 4, 8 and 12 h. The capacity of pretreated iron to continuously remove HMX was evaluated. This was accomplished by mixing 1 g of pretreated Fe0 with 25 mL of HMX solution and measuring changes in the HMX concentration at 0, 10, 40, and 60 min. After the last sampling, the solutions were centrifuged at 5000g for 10 min, and the supernatants removed. Fresh HMX solution (3 mg L-1) was reseeded back into the same pretreated Fe0 at 1, 2, 3, and 4 h (five cycles). The capacity of untreated Fe0 to remove HMX in the presence of didecyl was also studied. Solutions were sampled (1.0 mL) at 0, 1, 2, 4, and 8 h and then reseeded with HMX solution at 8, 16, 24, and 32 h. Changes in the HMX concentration were monitored between the reseeding cycles. Physical and Chemical Analyses. Subsamples (n ) 3) of LANL soil were characterized for particle size (57% sand, 9684
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 24, 2005
FIGURE 1. Destruction of solid-phase HMX and RDX (1.12 g of technical-grade HMX containing approximately 0.36% RDX) from aqueous solution (25 mL) by (A) Fe0 alone and (B) Fe0 + 2% (w/v) didecyl. Error bars indicate standard deviations of the means. 29.3% silt, 13% clay), organic matter (2.8%), pH (7.1), and cation-exchange capacity (6.7 cmolc kg-1) (Midwest Laboratories, Inc., Omaha, NE). HE concentrations were determined by HPLC by injecting 10 µL of sample into a 4.6 × 250 mm Keystone Betasil NA column (ThermoHypersil-Keystone, Bellefonte, PA) connected to a Shimadzu (Kyoto, Japan) photodiode array detector with quantification at 220 nm. The mobile phase was 30:70 methanol/water at a flow rate of 1 mL min-1. Didecyl analysis was performed by HPLC by injecting 5 µL of sample into a 4.6 × 250 mm Keystone nucleosil CN column (ThermoHypersil-Keystone, Bellefonte, PA) connected to a Shimadzu (Kyoto, Japan) ultraviolet detector with quantification at 254 nm. The mobile phase was 55:45 methanol/ water with 5 mM p-toluenesulfonic acid monohydrate at a flow rate of 0.8 mL min-1 (14). The surface morphology of the iron following various HMX treatments [with and without didecyl and didecyl-coated (pretreated) Fe0 before use] was determined by mounting with carbon tabs, sputter-coating with gold/palladium, and observing with a Hitachi S-3000N SEM operated at 15 kV. Mineralogical characteristics were determined by XRD with a Scintag XDS 2000 diffractometer (Scintag, Sunnyvale, CA) operating at 2° 2θ min-1. Cobalt K radiation (45 kV, 40 mA) was used to minimize fluorescence of Fe-rich minerals.
Results and Discussion Enhanced HMX Destruction. For Fe0 to effectively remediate highly contaminated soils, reactivity must be sustained as the solid phase dissolves. Our previous research indicated that HDTMA and didecyl were equally effective in bringing HMX into solution, but Fe0 + didecyl removed all of the solid-phase HMX (0.075 g of technical-grade HMX in 25 mL) (4). Destruction of solid-phase HMX by Fe0 with and without 2% (w/w) didecyl was monitored in aqueous solution for 170 days (Figure 1). Because RDX is an impurity (approximately 0.36%) in technical-grade HMX, changes in the RDX concentration were also monitored. Without didecyl in the matrix, the concentration of RDX in solution was approximately 40 mg L-1, and that of HMX was
