Degradation ofHexahydro-1,3,5-trinitro-1,3,5-triazine(RDX) Using Zerovalent IronNanoparticlesG H I N W A N A J A , ‡ A N N A M A R I A H A L A S Z , ‡

S O N I A T H I B O U T O T , † G U Y A M P L E M A N , †

A N D J A L A L H A W A R I * , ‡

Defence Research Establishment, Valcartier (Quebec),2459 Blvd Pie IX, Canada G0A 1R0, and BiotechnologyResearch Institute, National Research Council of Canada,Montreal, Quebec, Canada H4P 2R2

Received November 8, 2007. Revised manuscript receivedMarch 18, 2008. Accepted March 27, 2008.

Hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) is a commoncontaminant of soil and water at military facilities. The presentstudy describes degradation of RDX with zerovalent ironnanoparticles (ZVINs) in water in the presence or absence ofa stabilizer additive such as carboxymethyl cellulose (CMC)or poly(acrylic acid) (PAA). The rates of RDX degradation insolution followed this order CMC-ZVINs > PAA-ZVINs > ZVINswith k1 values of 0.816 ( 0.067, 0.082 ( 0.002, and 0.019 (0.002 min-1, respectively. The disappearance of RDX wasaccompanied by the formation of formaldehyde, nitrogen, nitrite,ammonium, nitrous oxide, and hydrazine by the intermediaryformation of methylenedinitramine (MEDINA), MNX (hexahydro-1-nitroso-,3,5-dinitro-1,3,5-triazine),DNX(hexahydro-1,3-dinitroso-5-nitro-1,3,5-triazine), TNX (hexahydro-1,3,5-trinitroso-1,3,5-triazine). When either of the reduced RDX products (MNX orTNX) was treated with ZVINs we observed nitrite (from MNXonly), NO (from TNX only), N2O, NH4

+, NH2NH2 and HCHO. In thecase of TNX we observed a new key product that wetentatively identified as 1,3-dinitroso-5-hydro-1,3,5-triazacyclo-hexane. However, we were unable to detect the equivalentdenitrohydrogenated product of RDX and MNX degradation.Finally, during MNX degradation we detected a new intermediateidentified as N-nitroso-methylenenitramine (ONNHCH2NHNO2),the equivalent of methylenedinitramine formed upon denitrationof RDX. Experimental evidence gathered thus far suggested thatZVINs degraded RDX and MNX via initial denitration andsequential reduction to the corresponding nitroso derivativesprior to completed decomposition but degraded TNX exclusivelyvia initial cleavage of the NsNO bond(s).

IntroductionThe extensive use of explosives such as hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) has led to widespread contami-nation of soil and water (1, 2). RDX is known to be toxic tovarious aquatic and terrestrial organisms (3), thus neces-sitating its removal from polluted environments. Recentlyour group has conducted several studies to elucidate the

degradation pathways of RDX under various abiotic and bioticconditions in an effort to help in the design of in situremediation strategies. For example we found that initialdenitration can lead to decomposition and the formation ofthe two key intermediates 4-nitro-2,4-diazabutanal (NDAB)and methylenedinitramine (MEDINA) whose formationdepends on the stoichiometry of the released nitrite ion. Theloss of 2 NO2

- from RDX normally leads to NDAB as has beenfound during alkaline hydrolysis (4) and aerobic biodegrada-tion with Rhodococcus sp. strain DN22 (5) and XplA (6).Whereas the loss of 1 NO2

- leads to the predominantformation of methylenedinitramine (MEDINA) as has beenobserved during RDX treatment with diaphorase enzyme (7)and XplA (6). RDX transformation to the correspondingnitroso derivatives (8) has also been documented to occurunder various abiotic reducing and biotic anaerobic condi-tions (9, 10). Although more recently Kemper et al. (11) didnot observe any of RDX nitroso products using hydrogensulfide and black carbon.

RDX removal using zerovalent iron has been extensivelyreported (8, 12–14) where several products including MNX,DNX, and TNX have been identified, but little is known onhow these nitroso products cleave. In the present study wechose to examine highly reactive zerovalent iron nanopar-ticles (ZVINs) capable of generating intermediates in suf-ficient concentrations to allow investigation of their decom-position routes. Recently, zerovalent iron nanoparticles(ZVINs) have been developed for several environmentalremediation technologies (15), especially for the treatmentof chlorinated organic compounds (16), metal ions (17),pesticides (18), organic dyes (19) and inorganic anions (20).However, the integration of ZVI nanoparticles in environ-mental processes has been held back by the key technicalbarrier represented by the tendency of iron nanoparticles toagglomerate and, thereby, rapidly lose their chemical reac-tivity and mobility. Extensive studies have been devoted tothe stabilization of the ZVINs. While Schrick et al. (21) usedhydrophilic carbon as delivery vehicles to support ZVInanoparticles, He et al. (22) reported a new strategy forstabilizing palladized iron nanoparticles with sodium car-boxymethyl cellulose.

To our knowledge, no reports are available for thedegradation of RDX with ZVI nanoparticles whose uniqueproperties present a novel technological potential. Thecurrent work examines the reaction between RDX and ZVINsfocusing on the identification of the RDX transformationpathway pinpointing the products, intermediates of thereaction as well as their yields. For this purpose, ZVInanoparticles were used to degrade RDX in water in thepresence and absence of the two stabilizers, carboxymethylcellulose (CMC) and poly(acrylic acid) (PAA). Addition of thepolymeric stabilizers into the system is intended to keep thenanoparticles well dispersed, to facilitate their even distribu-tion and smooth penetration through soil when applied inaccelerated in situ RDX degradation at contaminated sites.Finally, due to the formation of MNX and TNX as potentialRDX intermediates during RDX reduction with ZVI nano-particles, we thus investigated their reaction under the sameconditions to gain further insight into the degradationpathway(s) of RDX.

Experimental SectionChemicals. Hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX)(>99%), hexahydro-1,3,5-trinitroso-1,3,5-triazine (TNX) (99%)and ring-labeled [15N]-RDX (98%) were obtained fromDefence Research and Development Canada (Valcartier, QC).

* Corresponding author phone: +1-514-496 6267; fax +1-514-4966265; e-mail address: jalal.hawari@nrc.ca.

‡ Biotechnology Research Institute.† Defence Research Establishment.

Environ. Sci. Technol. 2008, 42, 4364–4370

4364 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 12, 2008 10.1021/es7028153 CCC: $40.75 2008 American Chemical SocietyPublished on Web 05/14/2008

Hexahydro-1-nitroso-3,5-dinitro-1,3,5-triazine (MNX) (98%),hexahydro-1,3-dinitroso-5-nitro-1,3,5-triazine (DNX) (56%),ring-labeled [15N]-MNX, ring-labeled [15N]-TNX, methylene-dinitramine (MEDINA) and 4-nitro-2,4-diazabutanal (NDAB)were provided by Dr. R. J. Spanggord from SRI International(Menlo Park, CA). Hydrazine sulfate (99%) was obtained fromAldrich, Canada. Sodium borohydride (g98.5%) and poly-(acrylic acid) (M.W. 1800) were purchased from Aldrich, andsodium carboxymethyl cellulose, polymer of cellulose-(OsCH2sCOO-)n, CMC, (M.W. 90 000) and nitrous oxidestandard (1000 ppm) were procured from Sigma-Aldrich.Ferrous sulfate (99%) and sodium hydroxide (99%) wereobtained from Anachemia and EMD, respectively. Nitriteand ammonium standards (1000 ppm) were purchased fromAlltech. All other chemicals were reagent grade, and allsolutions were prepared using Milli-Q-UV Plus Ultrapurewater system (>18 MΩ, Millipore, MA).

Zerovalent Iron Nanoparticles Preparation. Zerovalentiron nanoparticles (ZVINs), poly(acrylic acid) modifiedzerovalent iron nanoparticles (PAA-ZVINs), and carboxym-ethyl cellulose modified zerovalent iron nanoparticles (CMC-ZVINs) were prepared according to the methods publishedby Liu et al. (15), Schrick et al. (21) and He et al. (22),respectively, with modifications detailed in the SupportingInformation.

Physical Characterization. Transmission electron mi-croscope (TEM) micrographs were recorded using a PhilipsCM20 200 kV electron microscope equipped with an OxfordInstruments energy dispersive X-ray spectrometer (Link exlII) and an UltraScan 1000 CCD camera. To obtain the TEMimages, the nanoparticle suspensions were diluted withmethanol and sonicated vigorously. Twenty µL of each samplewas then dropped on a holey carbon film 300 mesh coppergrid and allowed to air-dry.

The N2sBET specific surface area of the nanoparticleswas measured using a TriStar 3000 gas adsorption analyzerand the multipoint method (Micromeritics Analytical Ser-vices, GA) with a resolution of 0.05 mmHg.

Reaction of RDX with ZVINs. Deionized water (10 mL)was mixed with 3 mg of ZVINs (methanol solution) in 15 mLserum bottles and crimp-sealed with Teflon-coated septa.The solution was either made anaerobic by purging theheadspace for 10 min with argon or kept under a blanket ofair for the aerobic experiments. Following 15 min of gentleshaking (rotary shaker, 150 rpm) to equilibrate the pH value(5.9-6.1), 0.82 µmol of RDX in methanol (0.2 mL) was addedto each bottle and the reaction was allowed to take place atroom temperature. In some experiments ring-labeled [15N]-RDX was used under the same conditions to help identifyRDX products. All experiments were made in triplicate.Controls containing ZVINs in water with no RDX were alsoperformed to ensure the ZVINs stability in water. Controlscontaining RDX in water with no ZVINs were not necessarysince RDX hydrolysis could be neglected within the range ofpH values examined (4). The reactions were stopped after 3,6, 10, 20, 40, 60, 120, 240, and 480 min. For each timemeasurement three bottles were sacrificed for analysis.Parallel RDX anoxic batch experiments were performed usingCMC-ZVINs and PAA-ZVINs (0.3 g L-1 of nanometal) tocompare their reactivity with the nonstabilized ZVINs basedon the same initial iron nanoparticle concentration.

Other batch experiments were conducted using ZVINs(0.3 g L-1) and either MNX (80 µM) or TNX (80 µM) todetermine products formed and thus know their role in thedegradation of RDX. In some experiments ring-labeled [15N]-MNX and [15N]-TNX were used under the same conditionsto help identify the degradation products. To determine theeventual fate of nitrogen-containing RDX degradation prod-ucts such as NO2

-, NH4+, N2O, MEDINA, and NH2NH2 we

allowed ZVINs (0.3 g L-1) to react with 20 mg L-1 of eachchemical separately.

Chemical Analysis. The gas phase in the headspace ofthe chemical assays was sampled using a gastight syringe(250 or 100 µL) and then analyzed for nitrogen, nitrous oxide,and hydrogen by a gas chromatograph (Hewlett-Packard 6890GC, Mississauga, ON) connected to either a TCD or an ECDdetector (23). Aliquots of the aqueous phase of the reactionmixtures were filtered through 0.22 µm filters (Millipore) priorto analyses of RDX, intermediates and final products. Thenitroso derivatives MNX, DNX, and TNX and the ring cleavageproducts methylenedinitramine (MEDINA) and 4-nitro-2,4-diazabutanal (NDAB) were analyzed as described by Hawariet al. (24) and by Bushan et al. (25). Formaldehyde, formicacid, ammonium, nitrate, nitrite were analyzed as describedby Monteil-Rivera et al. (26). Denitrosation of TNX wasfollowed by monitoring nitric oxide (NO) using Apollo 4000free radical analyzer (WPI, U.S.) specific for NO analysis.Whereas hydrazine formation was monitored by analyzingaliquots of the reaction mixture after derivatization withsalicylaldehyde (98%, Aldrich) followed by LC/MS analysisas described by Monteil-Rivera et al. (26). The concentrationof iron in CMC-ZVINs and PAA-ZVINs, ferric, and ferrousions were determined by spectroscopic methods as describedby Schrick et al. (21).

Results and DiscussionZVI Nanoparticle Characterizations. The TEM images(Figure 1) indicated that the three types of nanoparticles(ZVINs, CMC-ZVINs, and PAA-ZVINs) were mostly sphericalin shape and formed aggregates. The ZVINs featured a distinctcore and amorphous shell structure as clearly presented inFigure 1b. The TEM images (Figure 1e) also showed that theCMC-ZVINs had the smallest average particle diameter of 15( 4 nm compared to the ZVINs with 32 ( 7 nm and to thePAA-ZVINs with 173(40 nm as an average particle diameter.These measurements were in agreement with those foundin the literature. For instance, Sun et al. (27) used the sameprocedure forming iron nanoparticles with a median diameterof approximately 60.2 nm, whereas He et al. (22) reportedCMC stabilized Fe-Pd nanoparticles with the average particlediameter of 4.3 nm. When using poly(acrylic acid) to prepareiron nanoparticles, Schrick et al. (21) observed iron aggregatesmeasuring approximately 100 nm.

The specific surface areas of ZVINs, CMC-ZVINs, and PAA-ZVINs were 42.6, 11.3, and 5.9 m2 g-1, respectively. Theseexperimental surface areas were compared to the theoreticalvalues calculated using eq 1 (28) (24.0, 51.3, and 4.4 m2 g-1

for ZVINs, CMC-ZVINs, and PAA-ZVINs, respectively).

specific surface area) 6Fd

(1)

where F and d are the particle density (7.8 × 106 g m-3) anddiameter (m), respectively.

The observed difference between the calculated andmeasured areas may be caused by the density difference(27). In fact, the surfaces of those different types of nano-particles differed whereby iron was largely present as ironhydroxides with the ZVINs, and it was surrounded bypolymers in the two other cases coating the nanoparticleswith a thick film and thus decreasing the value of the specificsurface area.

Kinetics of RDX Degradation with ZVINs. Since thephysicochemical properties of RDX indicate that it has animportant potential for leaching, remediation treatments thatrapidly transform RDX and promote its degradation are ofgreat importance. In the present study, ZVINs (3 g L-1)completely degraded 82 µmol L-1 of RDX in five minutesunder both aerobic (98.3%) and anaerobic (100%) conditions.However, when reducing the amount of ZVINs to 0.3 g L-1

VOL. 42, NO. 12, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4365

less than 5% of RDX degraded under aerobic conditions, butmore than 75% of the nitramine degraded under anaerobicconditions in one hour. Hundal et al. (29) reported thecomplete transformation of RDX (144 µmol L-1) with microZVI (10 g L-1) in 96 h, but Wanaratna et al. (13) reported that50 µmol L-1 of RDX can be degraded using micro ZVI in lessthan 10 min at pH 3.5 by applying an excess amount of ZVI(32 g L-1). In the present study, the rapid removal of RDXwith ZVINs was attributed to the high reactivity of the nanometal (30–32) due to the small particle size (32 nm) offeringa large surface area (42.6 m2 g-1) to facilitate the reaction.The slow-down of RDX degradation in the presence of airwas attributed to the possible corrosion of the surface of Fedue to its reaction withy oxygen (33).

The kinetics of RDX (82 µmol L-1) degradation was thenfollowed in anoxic batch experiments using 0.3 g L-1 of thethree types of ZVINs (nonstabilized ZVINs, and stabilizedCMC-ZVINs and PAA-ZVINs). The comparison of the reac-tivity was based on the same amount of nano metal. Withinone hour, more than 75% of RDX was degraded using thethree types of nanoparticles (Figure 2), with CMC-ZVINsshowing the highest degradation percentage. The latter wasvery reactive and 6 min were sufficient to degrade 100% ofRDX.

Assuming that RDX degradation followed a second orderreaction rate with respect to RDX and iron concentrations(13), and by maintaining the change in the iron concentrationinsignificant compared to the change in RDX concentration,

FIGURE 1. TEM micrographs of the (a) nonstabilized ZVINs and (b) its enlargement; (c) CMC-ZVINs; and (d) PAA-ZVINs.; (e)probability density function of the particle size distribution of the iron nanoparticles (2 corresponds to the CMC-ZVINs (n ) 100), 0to the ZVINs (n ) 100), and b to PAA-ZVINs (n ) 45)).

4366 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 12, 2008

a pseudofirst order constant k1 was obtained. The reactionconstants (Supporting Information Table S-1) were deter-mined (min-1) as well as the corresponding BET surface area-normalized rate constants (h-1 m-2 L). The results indicatedthat the CMC-ZVINs application produced the highest kineticconstant (0.816 min-1), almost 40 times higher than the ZVINs(0.019 min-1). The obtained values fall within the range ofthose found in the literature. The k1 constant values obtainedby Wanaratna et al. (13) varied between 0.095 min-1 and0.976 min-1 when ZVI powder was used to remediate RDX-contaminated water. Oh et al. (8) estimated the k1 constantat 0.016 min-1 (value close to the one obtained in the presentstudy) when following the degradation of RDX using ironhaving 25.8 m2 of surface area per liter of solution.

The high reactivity of the CMC-ZVINs has been observedby He et al. (22) who reported that the stabilized nano reagentcan degrade trichloroethylene 17 times faster than thenonstabilized ZVINs. This high reactivity of the CMC-ZVINscould be explained by the higher dispersion of the nano-particles caused by the negatively charged carboxylic groupsthat inhibit aggregation and thus reduce the adhesioncoefficient between the nanoparticles (21). The reactivitydifference could also be partly attributed to the boron contentof these particles since different ratios of iron to boron wereused during the syntheses. Indeed, it has been speculated inseveral studies that boron may be responsible for the uniquereactivity of ZVINs (prepared using borohydride reducingagent) when compared to nano iron synthesized by the gas-phase reduction of iron oxides (34, 35). However, since inthe present case the reactivity was compared for the sameamount of nano metal, the effect of boron on the RDXdegradation could not be definitively determined.

Products and Degradation Pathways. Figure 3 representsRDX degradation with the simultaneous appearance of thering cleavage products (MEDINA, HCHO, NO2

- N2O, NH4+,

NH2NH2, and N2). The degradation of RDX was also ac-companied by the formation of the nitroso derivatives (MNX,DNX, and TNX) (Supporting Information Table S-2A). Mostof the detected products have been observed during the RDXdegradation with micro ZVI (8, 14). However, in the presentstudy several new products were detected giving new insightsinto the initial steps involved in the degradation pathwaysof RDX (discussed below).

At the end of the reaction, which lasted 4 h, most of RDXdetected intermediates transformed further to eventually giveN2O, NH4

+, and N2 as the main N-containing products and

HCHO as the main C-containing product. In all 2.86 HCHOmolecules were produced per one RDX molecule cleaved,accounting for more than 95.6% of the total carbon in RDXafter cleavage (Supporting Information Table S-2B). In thecase of nitrogen-containing degradation products, 1.44molecules of N2O and 1.25 molecules of NO2

-, NH2NH2, NH4+,

and N2 were produced, accounting for more than 79% of thetotal nitrogen of RDX after cleavage (Supporting InformationTable S-2B). The presence of the stabilizer did not seem todrastically affect product distribution. As shown in TablesS-2 and S-3 RDX degradation using the three types of ironnanoparticles led to the same intermediate and final products.But the nitroso intermediates seemed to disappear faster inthe presence of CMC and PAA.

The identity of nitrogen as an RDX degradation productwas confirmed using the ring-labeled [15N]-RDX and GC/MSanalysis. We detected N2 with a molecular mass ion at both28 Da (14N14N) and 29 Da (15N14N), confirming the formationof the gas (29 Da) from the original NsNO2 group in RDXand from further reduction of nitrite following initial deni-tration. When either NO2

- or N2O was allowed to come incontact with ZVINs we detected N2 and NH4

+. Nitrous oxideis a decomposing product of MEDINA in water (36) but itcould also arise from the reduction of NO2

- by ZVINs (datanot shown). Comparatively, NH4

+ was weakly degraded intoN2 and was probably adsorbed on the surface of thenanoparticles. The iron-aided NO2

- reduction into N2 hasalready been reported by Huang et al. (37). The presentexperimental findings mimic the generally known denitri-

FIGURE 2. Time course of RDX (0.82 µmol L-1) degradationexpressed in percentage using the three types of ironnanoparticles. (O corresponds to the CMC-ZVINs, 9 to theZVINs, and 4 to PAA-ZVINs). The concentration of ironnanoparticles was 0.3 g L-1. The standard deviations werewithin 5% of the corresponding values.

FIGURE 3. Time course of RDX degradation (82 µmol L-1) usingZVINs (0.3 g L-1) showing the formation of (a) formaldehyde,nitrous oxide, and MEDINA; (b) ammonium, nitrogen, and nitrite.The standard deviations were within 7% of the correspondingvalues, except for nitrogen where the standard deviation waswithin 15% of the value.

VOL. 42, NO. 12, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4367

fication process in an anoxic environment or in the presenceof specific kind of anaerobic bacteria (38).

Hydrazine (NH2NH2) was also detected as has been thecase when HMX was treated with ZVI (26). As it is toxic, weconducted experiments to determine its origin in thedegradation process and its eventual fate. We found thatNH2NH2 was a transient species which was transformedfurther to give NH4

+. After nine days, almost 90% of the initialamount of hydrazine disappeared (Figure S-1). When TNXwas treated with ZVINs under the same conditions used forRDX, hydrazine was also detected, suggesting that NH2NH2

might have originated from the NsNO functional group ofTNX formed during RDX reductive transformation (Sup-porting Information Table S-4). The hydrazine derivative wasdetected at a retention time of 15 min with a m/z [M+H]+

of 241 Da, but when the 15N-labeled RDX (or [15N]-TNX) wasused the product showed a m/z [M+H]+ of 242 Da,representing an increase of one Da due to the introductionof one 15N atom (the aza nitrogen) (Supporting InformationFigure S-1).

Furthermore we detected traces of nitric oxide (NO) duringtreatment of RDX with ZVINs (Figure 4a). To confirm theorigin of its formation we treated TNX with the nano metalunder the same conditions and found appreciable amounts

of NO being formed that did not persist indefinitely (Figure4a). The formation of NO from TNX was consistent with theobservation of the novel intermediate II with [M - H]- at144 Da, representing an empirical formula of C3H7N5O2

(Figure 4b). When ring labeled [15N-TNX] was used, the [M- H]- was detected at 147 Da, representing an increase of3 Da corresponding to the three 15N aza labeled N in theoriginal nitramine (Figure 4c). We tentatively identified theinitial TNX degradation intermediate as 1,3-dinitroso-5-hydro-1,3,5-triazacyclohexane (Figures 4b and c). The ex-pected denitrohydrogenated product of RDX (I) (Figure 5)was not detected. Presumably the latter was less stable thanII under these reducing conditions. For example, Bonner etal. (39) reported that intermediate I was unstable anddecomposed into MEDINA.

Likewise when MNX was treated with the nano metal weobserved nitrite, indicating initial denitration, and traces ofNO possibly from the denitrosation of its reduced TNXproduct (Figure 4a). Denitration of MNX was supported bythe detection of another new intermediate with [M - H]- atm/z 119 Da matching an empirical formula of CH4N4O3. Whenthe ring-labeled [15N]-MNX was used the [M - H]- wasdetected at m/z 121 Da representing an increase of 2 Da,

FIGURE 4. (a) Nitric oxide (NO) production during TNX, MNX, and RDX (80 µmol L-1) degradation with ZVINs (0.3 g L-1); (b) ES(-)ion mass spectrum of denitrosed hydrogenated compound (II) from TNX; (c) ES(-) ion mass spectrum of denitrosed hydrogenatedcompound (II) from [15N]-TNX.

4368 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 12, 2008

indicating the involvement of the two aza nitrogens in theformation of CH4N4O3. We tentatively identified the inter-mediate as N-nitroso-methylenenitramine (ONNHCH2-NHNO2) (III). Compound III is the equivalent of MEDINAformed following denitration of RDX (Figure 5). We did notobserve MEDINA when MNX was treated with ZVINs,suggesting that NO originated from its reduced product TNX.Supporting Information Table S-4 summarizes the productdistribution observed from MNX and TNX treatment withZVINs (Figure 5).

Experimental evidence gathered thus far on productsdistribution, stoichiometry, and time courses indicate thatRDX degraded via two initial routes. The first route involvedinitial denitration of RDX giving the suspected unstabledenitrohydrogenated intermediate I which would decomposein water to produce MEDINA (Figure 5, path a). The secondroute involved the stepwise reduction of the NsNO2 func-tional groups to give MNX, DNX, and TNX. Likewise MNXwould undergo either denitration prior to ring cleavage orreduction to eventually give TNX which underwent denit-rosation (cleavage of NsNO bond) followed by ring cleavage(Figure 5, path b).

Environmental Significance. The use of CMC as stabiliz-ers for ZVINs, which kept the metal well dispersed in water,degraded RDX forty times faster than the nonstabilized ZVINs.The use of surfactants in many industrial remediationtechnologies often enhances the remediation process byincreasing mobility and solubility in water of insoluble orsparingly soluble contaminants that, in turn, improves theirmass removal and the overall process performance. Also thepresent reaction system degraded RDX to obnoxious productssuch as formaldehyde (biodegradable), nitrous oxide, am-monia, and nitrogen. Although the three nitroso productsMNX, DNX, and TNX were also detected as RDX productsnone of these hazardous chemicals persisted indefinitelyrather they all degraded further to produce HCHO andhydrazine, the latter degraded to ammonia. These experi-mental findings can constitute the basis for the developmentof in situ remediation technologies for contaminated sites.Understanding the dynamics and pathways of RDX degra-dation would help optimizing the in situ remediation of water

contaminated with explosives including groundwater andmarine sediment.

AcknowledgmentsWe thank Mr. Dashan Wang from the ICPET for thetransmission electron microscope images. We also thank Dr.Fanny Monteil-Rivera for helpful discussions and LouisePaquet and Stephane Deschamps for conducting the analy-ses. Financial support was provided by DRDC, Valcartier,Canada, and Office of Naval Research (ONR) U.S. Navy (AwardN000140610251).

Supporting Information AvailableExperimental section (ZVI nanoparticles preparation andmodification), Figure S-1 and Tables S-1, S-2, S-3 and S-4 asmentioned in the text. This material is available free of chargevia the Internet at http://pubs.acs.org.

Literature Cited(1) Brannon, J. M.; Price, C. B.; Yost, S. L.; Hayes, C.; Porter, B.

Comparison of environmental fate and transport processdescriptors of explosives in saline and freshwater systems. Mar.Pollut. Bull. 2005, 50, 247–251.

(2) Eisentraeger, A.; Reifferscheid, G.; Dardenne, F.; Blust, R.;Schofer, A. Hazard characterization and identification of aformer ammunition site using microarrays, bioassays, andchemical analysis. Environ. Toxicol. Chem. 2007, 26, 634–646.

(3) Robidoux, P. Y.; Dubois, C.; Hawari, J.; Sunahara, G. I. Assessmentof soil toxicity from an antitank firing range using Lumbricusterrestris and Eisenia andrei in mesocosms and laboratorystudies. Ecotoxicology 2004, 13, 603–614.

(4) Balakrishnan, V. K.; Halasz, A.; Hawari, J. Alkaline hydrolysis ofcyclic nitramine explosives RDX, HMX, and CL-20: New insightsinto degradation pathways obtained by the observation of novelintermediates. Environ. Sci. Tecnol. 2003, 37, 1838–1843.

(5) Fournier, D.; Halasz, A.; Spain, J.; Fiurasek, P.; Hawari, J.Determination of key metabolites during biodegradation ofhexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) with Rhodococcussp. strain DN22. Appl. Environ. Microbiol. 2002, 68, 166–172.

(6) Jackson, R. G.; Rylott, E. L.; Fournier, D.; Hawari, J.; Bruce, N. C.Exploring the biochemical properties and remediation applica-tion of the unusual explosive-degrading P450 system XplA/B.Proc. Natl. Acad. Sci. USA 2007, 104, 16822–16827.

FIGURE 5. Proposed pathway for degradation of RDX in the presence of ZVINs: (a) Denitration route; (b) Nitro reduction to nitrosofollowed by denitrosation of TNX formed.

VOL. 42, NO. 12, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4369

(7) Bhushan, B.; Halasz, A.; Spain, J. C.; Thiboutot, S.; Ampleman,G.; Hawari, J. Diaphorase catalyzed biotransformation ofHexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) via N-denitrationmechanism. Biochem. Biophys. Res. Commun. 2002, 296, 779–784.

(8) Oh, S. Y.; Cha, D. K.; Kim, B. J.; Chiu, P. C. Reductivetransformation of hexahydro-1,3,5-trinitro-1,3,5-triazine, oc-tahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine, and methylene-dinitramine with elemental iron. Environ. Toxicol. Chem. 2005,24, 2812–2819.

(9) McCormick, N. G.; Cornell, J. H.; Kaplan, A. M. Biodegradationof hexahydro-1,3,5-trinitro-1,3,5-triazine. Appl. Environ. Mi-crobiol. 1981, 42, 817–823.

(10) Zhao, J.-Z.; Paquet, L.; Halasz, A.; Hawari, J. Metabolism ofhexahydro-1,3,5-trinitro-1,3,5-triazine through initial reductionto hexahydro-1-nitroso-3,5-dinitro-1,3,5-triazine followed bydenitration in Clostridium bifermentans HAW-1. Appl. Microbiol.Biotechnol. 2003, 63, 187–193.

(11) Kemper, J. M.; Ammar, E.; Mitch, W. A. Abiotic degradation ofhexahydro-1,3,5-trinitro-1,3,5-triazine in the presence of hy-drogen sulfide and black carbon. Environ. Sci. Technol. 2008,ASAP Article.

(12) Comfort, S. D.; Shea, P. J.; Machacek, T. A.; Satapanajaru, T.Pilot-scale treatment of RDX-contaminated soil with zerovalentiron. J. Environ. Qual. 2003, 32, 1717–1725.

(13) Wanaratna, P.; Christodoulatos, C.; Sidhoum, M. Kinetics ofRDX degradation by zero-valent iron (ZVI). J. Hazard. Mater.2006, 136, 68–74.

(14) Oh, B. T.; Just, C. L.; Alvarez, P. J. J. Hexahydro-1,3,5-trinitro-1,3,5-triazine mineralization by zerovalent iron and mixedanaerobic cultures. Environ. Sci. Technol. 2001, 35, 4341–4346.

(15) Liu, Y. Q.; 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, 1338–1345.

(16) Doong, R. A.; Chen, K. T.; Tsai, H. C. Reductive dechlorinationof carbon tetrachloride and tetrachloroethylene by zerovalentsilicon-iron reductants. Environ. Sci. Technol. 2003, 37, 2575–2581.

(17) Leupin, O. X.; Hug, S. J. Oxidation and removal of arsenic (III)from aerated groundwater by filtration through sand and zero-valent iron. Water Res. 2005, 39, 1729–1740.

(18) Ghauch, A.; Rima, J.; Amine, C.; Martin-Bouyer, M. Rapidtreatment of water contaminated with atrazine and parathionwith zero-valent iron. Chemosphere 1999, 39, 1309–1315.

(19) Pereira, W. S.; Freire, R. S. Azo dye degradation by recycledwaste zero-valent iron powder. J. Braz. Chem. Soc. 2006, 17,832–838.

(20) Li, Z. H.; Jones, H. R.; Bowman, R. S.; Helferich, R. Enhancedreduction of chromate and PCE by palletized surfactant-modified zeolite/zerovalent iron. Environ. Sci. Technol. 1999,33, 4326–4330.

(21) Schrick, B.; Hydutsky, B. W.; Blough, J. L.; Mallouk, T. E. Deliveryvehicles for zerovalent metal nanoparticles in soil and ground-water. Chem. Mater. 2004, 16, 2187–2193.

(22) He, F.; Zhao, D. Y.; Liu, J. C.; Roberts, C. B. Stabilization of Fe-Pdnanoparticles with sodium carboxymethyl cellulose for en-hanced transport and dechlorination of trichloroethylene insoil and groundwater. Ind. Eng. Chem. Res. 2007, 46, 29–34.

(23) Sheremata, T. W.; Hawari, J. Mineralization of RDX by the whiterot fungus Phanerochaete chrysosporium to carbon dioxide andnitrous oxide. Environ. Sci. Technol. 2000, 34, 3384–3388.

(24) Hawari, J.; Halasz, A.; Groom, C.; Deschamps, S.; Paquet, L.;Beaulieu, C.; Corriveau, A. Photodegradation of RDX in aqueoussolution: A mechanistic probe for biodegradation with Rhodo-coccus sp. Environ. Sci. Technol. 2002, 36, 5117–5123.

(25) Bhushan, B.; Paquet, L.; Halasz, A.; Spain, J. C.; Hawari, J.Mechanism of xanthine oxidase catalyzed biotransformationof HMX under anaerobic conditions. Biochem. Biophys. Res.Commun. 2003, 306, 509–515.

(26) Monteil-Rivera, F.; Paquet, L.; Halasz, A.; Montgomery, M. T.;Hawari, J. Reduction of octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine by zerovalent iron: Product distribution. Environ.Sci. Technol. 2005, 39, 9725–9731.

(27) Sun, Y. P.; Li, X. Q.; Cao, J. S.; Zhang, W. X.; Wang, H. P.Characterization of zero-valent iron nanoparticles. Adv. ColloidInterface Sci. 2006, 120, 47–56.

(28) Cao, J. S.; Clasen, P.; Zhang, W. X. Nanoporous zero-valent iron.J. Mater. Res. 2005, 20, 3238–3243.

(29) Hundal, L. S.; Singh, J.; Bier, E. L.; Shea, P. J.; Comfort, S. D.;Powers, W. L. Removal of TNT and RDX from water and soilusing iron metal. Environ. Pollut. 1997, 97, 55–64.

(30) Li, X. Q.; Elliott, D. W.; Zhang, W. X. Zero-valent iron nano-particles for abatement of environmental pollutants: Materialsand engineering aspects. Crit. Rev. Solid State Mat. Sci. 2006,31, 111–122.

(31) Zhang, W. X. In situ remediation with nanoscale iron particles.Abstr. Pap. Am. Chem. Soc. 2003, 225, U961–U962.

(32) Zhang, W. X. Nanoscale iron particles for environmentalremediation: An overview. J. Nanopart. Res. 2003, 5, 323–332.

(33) Keenan, C. R.; Sedlak, D. L. Factors affecting the yield of oxidantsfrom the reaction of nanoparticulate zero-valent iron andoxygen. Environ. Sci. Technol. 2008, 42, 1262–1267.

(34) Liu, Y. Q.; Choi, H.; Dionysiou, D.; Lowry, G. V. Trichloroethenehydrodechlorination in water by highly disordered monome-tallic nanoiron. Chem. Mater. 2005, 17, 5315–5322.

(35) Ponder, S. M.; Darab, J. G.; Bucher, J.; Caulder, D.; Craig, I.;Davis, L.; Edelstein, N.; Lukens, W.; Nitsche, H.; Rao, L. F.; Shuh,D. K.; Mallouk, T. E. Surface chemistry and electrochemistry ofsupported zerovalent iron nanoparticles in the remediation ofaqueous metal contaminants. Chem. Mater. 2001, 13, 479–486.

(36) Halasz, A.; Spain, J.; Paquet, L.; Beaulieu, C.; Hawari, J. Insightsinto the formation and degradation mechanisms of methyl-enedinitramine during the incubation of RDX with anaerobicsludge. Environ. Sci. Technol. 2002, 36, 633–638.

(37) Huang, Y. H.; Zhang, T. C. Nitrite reduction and formation ofcorrosion coatings in zerovalent iron systems. Chemosphere2006, 64, 937–943.

(38) Garber, E. A. E.; Hollocher, T. C. Positional isotopic equivalenceof nitrogen in N2O produced by the denitrifying bacteriumPseudomonas-Stutzeri. Indirect evidence for a nitroxyl pathway.J. Biol. Chem. 1982, 257, 4705–4708.

(39) Bonner, T. G.; Hancock, R. A.; Roberts, J. C. The N-nitroxymethylderivatives of 1,3-dinitroperhydro-1,3,5-triazine, piperidine, andsuccinimide. J. Chem. Soc., Perkin Trans. 1. 1972, 1902–1907.

ES7028153

4370 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 12, 2008