Reductive Debromination ofPolybrominated Diphenyl Ethers byZerovalent IronY O U N G - S O O K E U M A N D Q I N G X . L I *

Department of Molecular Biosciences and Bioengineering,University of Hawaii, Honolulu, Hawaii 96822

Polybrominated diphenyl ethers (PBDEs) are a new classof global, persistent, and toxic contaminants, which needproper remediation technologies. PBDE degradation inthe environment is not well understood. In this study,degradation of PBDEs with zerovalent iron was investigatedwith six BDEs, substituted with one to 10 bromines.Within 40 days 92% of BDE congener 209 was transformedinto lower bromo congeners. During the initial reactionperiod of BDE 209 (<5 days), hexa- to heptabromo BDEswere the most abundant products, but tetra- to pentabromocongeners were dominant after 2 weeks. The amount ofmono- to tribromo BDEs was steadily increased during theexperiments. BDEs 28, 47, 66, and 100 also showed astepwise accumulation of lower bromo congeners. Nooxidation products were detected in all experiments. Theresults showed that a stepwise debromination from n-bromo-to (n-1)-bromodiphenyl ethers was the dominant reactionin all congeners. The reaction rate constants of lowerbromo BDEs decreased as the number of bromines decreased.The initial reductive debromination rate constants werepositively correlated with the heats of formation of BDEs.The preferential accumulation of specific congenerswas observed in the experiment with BDEs 28, 47, 66, and100, where the most abundant products were BDEs 15,28, 37, and 47, respectively. Reactions proceeded to formmore stable and less brominated products that havelower heats of formation. Almost all the possible isomersfrom a specific parent BDE were found in all the experiments,which was probably due to the small difference of heatof formation between the products (2-5 kcal/mol). Reactionsof all congeners proceeded fast at the initial phase (<5days) followed by a slow reaction. The rate of reductivedebromination of BDE 209 was slower with environmentallyrelevant sulfide minerals (iron sulfide and sodium sulfide).However, the product congener pattern, produced bysulfide mineral catalysis, was nearly similar with that ofzerovalent iron treatment. This may be a possible source oflower brominated BDEs in the environment. Debrominationof PBDEs by zerovalent iron has high potential valuesfor remediation of PBDEs in the environment.

IntroductionPolybrominated diphenyl ethers (PBDEs) are widely used asflame-retardants in various industrial products (1-3). Unlikepolychlorinated biphenyls (PCBs) that are banned from use,

worldwide demand of PBDEs (e.g., commercial deca-, octa-,and penta-BDE technical products) increased from 40 000ton in 1992 to 67 125 ton in 1999, and approximately half ofthem are consumed in North American regions (1-4). PBDEresidues are found in sediments, freshwater and marineorganisms, birds, plants, and humans (2, 5). Several congenersare found even in Arctic regions (6, 7). Concentrations ofPBDEs in the environment have increased exponentiallyduring the past several decades (6).

The toxicity of PBDEs is not well understood yet. Somecongeners showed weak to moderate dioxin-like activitiesor binding affinities to human estrogen receptors (8, 9).Hydroxy metabolites of PBDEs are found to have strongbinding affinity with the thyroxine transporting protein,transthyretin (10). In addition, polybrominated dibenzo-p-dioxin/dibenzofurans (PBDD/Fs), which exhibit similar tox-icities with their chlorinated analogues (PCDD/Fs), can beformed during the incineration of materials containing flame-retardants (8, 11, 12).

In comparison with chlorinated aromatic contaminants,research on the metabolism and environmental fate of PBDEsis limited. The metabolic pathways of PBDEs in animalsinclude reductive debromination, hydroxylation, and etherbond cleavage (13, 14). PBDEs in the environment undergoanaerobic or aerobic microbial degradation (15-17). Among209 theoretical isomers, decabromodiphenyl ether (BDE 209)makes up 80% of the world consumption of PBDEs. However,tetra- to hexabromo BDEs (e.g., BDEs 47, 99, 100, 154, 155,and 183) are the most abundant congeners in environmentalsamples. Because of its extreme hydrophobicity (log Kow, >10)and limited bioavailability, BDE 209 is generally consideredhighly recalcitrant but safe. However, BDE 209 in an organicsolvent, water, and heterogeneous mixture is rapidly de-composed to lower bromo congeners under UV or simulatedsunlight (18-20). Although the extent is limited, BDE 209was debrominated to penta- to octabromo congeners in carp(14). In addition, metabolized congener patterns differ amongorganisms in various trophic levels (1). BDE-209 or otherhighly brominated congeners can be transformed to lowerbrominated isomers or metabolites and may be a source ofenvironmentally abundant BDEs.

It is important to develop potential remediation tech-nologies for BDEs. Several kinds of elementary metals havebeen studied for remediation of highly oxidized organicpollutants such as polyhalogenated aromatic compoundsand explosives. Zerovalent iron is one such metal and canreduce nitroaromatic pesticides, PCBs, PCDDs, DDTs, andhalogenated phenols (21-25). Much research effort has beenconducted at laboratory and field trial scales due to itseconomic feasibility and low maintenance cost.

The efficiency of zerovalent iron treatment on contami-nants is highly dependent on the properties of the metalsurface (26) because reduction reaction of contaminantsmainly occurs on the surface of iron. Formation of iron oxideor hydroxide over the surface usually decreases the reactivityand is a common problem of field applications (27). Severalapproaches have been suggested to increase the durabilityor reactivity of the surface, including the addition of minerals,application of nanoscale iron powder, and modification ofiron with other metals (22, 25, 28). Among the halogenatedcontaminants, the current research with zerovalent iron ismainly focused on chlorinated alkane/alkene and aromaticcompounds. Studies of contaminants with halogens, otherthan chlorine (e.g., polybrominated flame retardants) arelimited.

* Corresponding author phone: (808) 956-2011; fax: (808) 956-3542; e-mail: qingl@hawaii.edu.

Environ. Sci. Technol. 2005, 39, 2280-2286

2280 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005 10.1021/es048846g CCC: $30.25 2005 American Chemical SocietyPublished on Web 02/10/2005

In this study, we investigated the possible application ofzerovalent iron for the remediation of PBDEs and estimatedthe reactivity of several environmentally important BDEcongeners using molecular modeling. We propose thatdebromination of PBDEs be a possible pathway for occur-rence of lower BDEs from higher BDEs in the environment.

Experimental SectionChemicals and Reagents. Decabromodiphenyl ether (BDE209) and octachloronaphthalene standards were purchasedfrom Accustandard, Inc. (New Haven, CT). The analyticalstandards BDE-CVS-A and EO-4980 were obtained fromWellington Laboratories, Inc. (Ontario, Canada) and Cam-bridge Isotope Laboratories, Inc. (Andover, MA). Bothanalytical standard solutions were used to construct calibra-tion curves for each congener over a range from 5 to 200pg/µL. Congener numbers and substitution patterns are givenin Table 1. Solvents were obtained from Fisher Scientific,Inc. (Pittsburgh, PA). Powdery zerovalent iron (<325 mesh)and iron sulfide were purchased from Aldrich (Milwaukee,WI). Double-distilled water was deionized with a Milli-Q waterpurification system (Millipore) and filtered by a Millipak 40catridge (0.22 µm) before use.

Preparation of BDEs 7, 28, 47, 66, and 100. BDEs 7, 28,47, 66, and 100 for treatment by zerovalent iron weresynthesized according to the methods of Marsh et al. (29).Briefly, brominated diphenyliodonium chlorides preparedfrom 1,2- or 1,3-dibromobenzene and iodine were coupledto phenol, 4-bromophenol, 2,4-dibromophenol, or 2,4,6-tribromophenol in a mixture of aqueous sodium hydroxideand dioxane. Crude products were purified through a silicagel column eluted with a mixture of hexane and dichlo-romethane. Purified BDEs were identified by comparing theirgas chromatographic (GC) retention time and mass spectrawith those of authentic standards.

Treatment of BDEs by Zerovalent Iron. BDEs in ethylacetate (1 mL, 50 mg/L) were applied on powdery iron oriron sulfide (5 g) in a screw-capped amber vial. After solventremoval with a gentle stream of nitrogen, deionized water(10 mL) was added. The vial was covered with a Teflon-linedcap and then incubated at 30 °C while stirring (60 rpm) (C24Rotary shaker, New Brunswick Scientific, NJ). BDE 209 (50mg/L) was suspended in sodium sulfide solution (200 mg in10 mL of water) and incubated at the same conditions.Controls without iron powder were prepared according tothe same procedures. Three replicates were done for all theexperiments.

Extraction and Analysis. After 3 h and 0.5, 1, 2, 3, 5, 7,14, and 40 days of incubation, BDEs were extracted with amixture of ether and dichloromethane (1:1, v/v) for threetimes (3 × 20 mL). The combined organic extracts wereconcentrated to 1 mL. PBDEs in the extracts were analyzedwith a Varian QP-5000 GC (Varian, CA) equipped with electroncapture detector (ECD, for quantitation) and Saturn-2000ion trap mass spectrometer (MS, for structural confirmation).A ZB-1 capillary column (60 m × 0.25 µm film thickness) wasused. The amounts of mono- to heptabromo congeners, forwhich authentic standards were not commercially available,were estimated from the average ECD response factor of theidentified mono- to heptabromo congeners with the samenumber of bromines (19). The operation conditions were asfollows: column temperature, 130 °C for 10 min, raised to300 °C at a rate of 1 °C/min, and held for 40 min; injectionport, 290 °C; ECD, 300 °C; interface for MS, 280 °C. A splitratio between ECD and MS was 1:10. Carrier gas was heliumat a flow rate of 2.5 mL/min. ECD sensitivity was calibratedwith octachloronaphthalene as an external standard in everythree runs of samples.

Molecular Modeling. Energy minimized structures ofPBDEs were calculated with a CAChe Worksystem Pro 6.1(Fujitsu Ltd., Japan). Conformational search was performedprior to refined energy minimization with stepwise changingof dihedral angles between the two phenyl rings. Thegeometries of candidate structures were further optimizedwith AM1 semiempirical force field. Various propertiesincluding heat of formation (Hf, kcal/mol), highest occupiedmolecular orbital (HOMO, eV), and lowest unoccupiedmolecular orbital (LUMO, eV) were calculated with the samemethods.

Results and DiscussionReductive Debromination of BDE 209 and Other LowerBromo-BDEs. BDE 209 was rapidly transformed to lowerbromo congeners by zerovalent iron treatment (Figures 1and 2). Approximately 90% of BDE 209 was converted into

TABLE 1. Congener Number and Substitution of Bromines onBDEs

congener no. full name

7 2,4-dibromodiphenyl ether8 2,4′-dibromodiphenyl ether

11 3,3′-dibromodiphenyl ether12 3,4-dibromodiphenyl ether13 3,4′-dibromodiphenyl ether15 4,4′-dibromodiphenyl ether17 2,2′,4-tribromodiphenyl ether25 2,3′,4-tribromodiphenyl ether28 2,4,4′-tribromodiphenyl ether30 2,4,6-tribromodiphenyl ether32 2,4′,6-tribromodiphenyl ether33 2,3′,4′-tribromodiphenyl ether35 3,3′,4-tribromodiphenyl ether37 3,4,4′-tribromodiphenyl ether47 2,2′,4,4′-tetrabromodiphenyl ether49 2,2′,4,5′-tetrabromodiphenyl ether66 2,3′,4,4′-tetrabromodiphenyl ether71 2,3′,4′,6-tetrabromodiphenyl ether75 2,4,4′,6-tetrabromodiphenyl ether77 3,3′,4,4′-tetrabromodiphenyl ether85 2,2′,3,4,4′-pentabromodiphenyl ether99 2,2′,4,4′,5-pentabromodiphenyl ether

100 2,2′,4,4′,6-pentabromodiphenyl ether116 2,3,4,5,6-pentabromodiphenyl ether118 2,3′,4,4′,5-pentabromodiphenyl ether119 2,3′,4,4′,6-pentabromodiphenyl ether126 3,3′,4,4′,5-pentabromodiphenyl ether138 2,2′,3,4,4′,5′-hexabromodiphenyl ether153 2,2′,4,4′,5,5′-hexabromodiphenyl ether154 2,2′,4,4′,5,6′-hexabromodiphenyl ether155 2,2′,4,4′,6,6′-hexabromodiphenyl ether181 2,2′,3,4,4′,5,6-heptabromodiphenyl ether183 2,2′,3,4,4′,5′,6-heptabromodiphenyl ether190 2,3,3′,4,4′,5,6-heptabromodiphenyl ether209 decabromodiphenyl ether

FIGURE 1. GC-ECD chromatograms of BDE 209 treated withzerovalent iron after (A) 3 h and (B) 3, (C) 7, and (D) 40 days. Numberslabeled at individual peaks are the congener number of the PBDEs.

VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2281

mono- to hexa-BDEs after 40 days. However, BDE 209 wasfairly stable in controls without iron (Figure 3A). Heptabro-minated or higher brominated DEs were the most abundantcongeners at the initial stage of reaction, where the fractionsaccounted for more than 50% of the total BDEs at <5 days.However, the proportion of lower bromo congeners (monoto hexa) gradually increased with concomitant decrease ofhigher brominated DEs (Figure 2, Table 2). Tri- and tetra-BDEs were the most dominant congeners at 40 days ofincubation. Almost all the theoretically possible isomers wereobserved after 7 days in GC-ECD analysiss112 peaks wereassigned as new peaks in comparison with control samples.Forty BDEs, which are usually found in environmental

samples, were found in varying abundance. Additional 36peaks were tentatively identified as di- to hexabrominatedDEs through comparison with mass spectra of the authenticPBDE standards. The GC-ECD peak area of the other 36 peakswas comprised of 2-3% of total peak area. Limited quantitiesexcluded identification of those 36 peaks by GC-MS. Theamount of the identifiable congeners rapidly increased to56% of the total PBDEs and the amount of the environmen-tally relevant congeners (6) increased to 22% after 2 weeksfollowed by a slow decrease (Table 2). Relative abundanceof each identified congener ranged from 0.03 to 6% of totalBDEs at 7 days. Mass balances of identifiable congenersdecreased to 44.9%, and the amount of unidentifiablecongeners was approximately 40-50% of the initial molaramount of BDE 209 at 40 days.

Because the ECD sensitivities vary among congeners, theestimation variation of unidentifiable congeners may be asmuch as 32%. However, Figure 2 and Table 2 clearly showthat the degradation of BDE 209 occurs through stepwisedebromination from n-bromo- to (n-1)-bromo-DE. Stepwisedehalogenations are found in anaerobic biodegradation orchemical degradation of halogenated phenols or PCBs (30,31).

The biodegradation of PBDEs in animals or aerobicmicroorganisms usually occurs through hydroxylation orether bond cleavage to simple phenols or catechols (13-17).Although oxidation was reported from zerovalent irontreatment of pesticides in oxygen-rich conditions (32), noneof such products (e.g., hydroxyl PBDEs, bromophenols) wereobserved in this study.

In the environment, BDE 209 is usually found in itsproduction or treatment site and lower brominated congenersare detected as dominant BDEs at locations far from suchfacilities (33). Although the extreme hydrophobicity andlimited bioavailability restrict BDE 209 biodegradation, theexperimental results show that reductive chemical degrada-tion can occur by strong reducing agent, such as elementarymetals (e.g., Fe in this experiment).

A pseudo-first-order kinetic model (eq 1) was applied forthe description of reductive debromination of parent BDEs:

with C and C0 as concentrations of parent BDEs at anysampling and initial time, respectively; k as a first-order rateconstant (day-1); and t as reaction time. The debrominationof BDEs 7, 28, 47, 66, 100, and 209 fits well to the abovepseudo-first-order kinetics (Figure 3B). The rate constantsdecreased with the decrease of the number of bromines

TABLE 2. Relative Abundance of Identifiable Product PBDEs Generated from Reductive Debromination of BDE 209 in the Presenceof Zerovalent Irona

rel abundance of PBDEs (%)

BDE No. 0 3 h 1 day 7 days 14 days 40 days

17 NDb ND 0.02 ( 0.02 0.37 ( 0.07 0.83 ( 0.09 1.98 ( 0.3428/33 ND ND 0.01 ( 0.00 0.68 ( 0.13 1.75 ( 0.22 3.59 ( 0.51

47 ND 0.03 ( 0.00 0.22 ( 0.10 2.46 ( 0.47 2.96 ( 0.21 4.85 ( 0.5666 ND ND 0.10 ( 0.05 1.59 ( 0.22 1.78 ( 0.56 1.69 ( 0.3799 ND 0.15 ( 0.01 1.00 ( 0.24 2.70 ( 0.63 2.53 ( 0.33 3.05 ( 0.18100 ND 0.11 ( 0.02 1.10 ( 0.17 6.45 ( 0.89 4.74 ( 0.38 2.18 ( 0.51138 ND 1.07 ( 0.35 1.48 ( 0.25 1.43 ( 0.41 1.49 ( 0.47 0.13 ( 0.01153 ND 1.90 ( 0.14 2.21 ( 0.29 2.49 ( 0.27 2.28 ( 0.25 1.83 ( 0.23154 ND 0.44 ( 0.11 1.35 ( 0.33 1.33 ( 0.14 1.63 ( 0.19 0.13 ( 0.01183 ND 3.20 ( 1.23 2.76 ( 0.54 2.12 ( 0.57 1.86 ( 0.45 0.55 ( 0.10190 ND 2.79 ( 0.47 3.10 ( 0.28 0.74 ( 0.17 0.76 ( 0.31 0.11 ( 0.02

othersc ND 8.78 11.67 24.21 33.54 17.81209 100.04 72.01 ( 4.40 48.00 ( 8.90 15.03 ( 1.31 12.00 ( 5.22 7.00 ( 0.80sum 100.04 90.48 73.02 60.86 68.15 44.90

a Percent abundance of BDE products (mono- to hepta-BDEs) ( standard deviations relative to the initial amount of BDE 209. b Not detected.c Sum of the other PBDEs confirmed with authentic standards.

FIGURE 2. Percent distribution of di- to heptabromodiphenyl ethersproduced from BDE 209 after treatment with zerovalent iron at 30°C. All identified di- to heptabromo-BDEs and those of which weretentatively identified by GC-MS were included for the calculation.The total of all these congeners at each given day was calibratedas 100%.

FIGURE 3. Degradation of BDEs 7, 28, 47, 66, 100, and 209 by zerovalentiron in the initial 3 day incubation at 30 °C: (A) no iron control; (B)zerovalent iron treatment. Error bars indicate standard deviationsof three replicates.

C/C0 ) exp(-kt) (1)

2282 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

(Figure 3). BDE 7 (2,4-dibromo DE) and BDE 28 (2,4,4′-tribromo DE) were stable within the first 3 days. Yak et al.(34) observed a similar decrease of reduction rate constantswith the decrease of chlorination of PCBs in subcritical water.In this study, all possible congeners were detected at varyinglevels in the samples. Those were BDEs 7, 8, and 15 fromBDE 28; BDEs 17 and 28 from BDE 47; BDEs 25, 28, 33, and37 from BDE 66; and BDEs 47, 50, 51, and 75 from BDE 100(Figure 4). In the cases of BDEs 47, 66, and 100 (Figure 4B-D), di- to tribromo-DEs were also found after 14 days. In thecontrol samples, more than 95% of BDEs were recovered asparent compounds without any trace of lower BDEs or otherproducts (Figure 3A). None of the products derived fromoxidation, hydroxylation, or heterolytic fission of the etherbond (e.g. bromobenzenes, bromophenols, hydroxyl PBDEs)were detected in the iron-treated samples during the entireexperimental period. In consideration of irreversible adsorp-tion or formation of nonextractable residues (e.g., adsorptionon iron oxide, hydroxide, and complex formation withbromophenols), mass balances were calculated with BDEs28, 47, 66, and 100 and their products. The mass balanceswere approximately quantitative (95-110%) (Figure 5), whichexcluded a possibility of adsorption or formation of non-extractable residues. PBDFs other than those from simpledebromination such as the photolysis product of BDE 209

(20, 35) were not detected in this study. The results showedthat reductive debromination was the dominant reaction forBDEs 7, 28, 47, 66, and 100.

Rapid debromination was observed for all the congenersstudied in the initial 3 days followed by a gradual decreaseduring further incubation (Figure 6). Because the reductionoccurs on the iron particle surface (28), its surface propertychanges govern its reactivity. For example, oxidation of theiron surface decreases the reduction rate for nitro groups ofseveral pesticides (21). Iron oxide or hydroxide, which isusually formed during the zerovalent iron mediated reductionof contaminants, is not as reactive as elementary iron (23).Iron oxidation and concomitant formation of an oxide orhydroxide layer may result in slow reaction after 5-7 daysof incubation (Figure 6).

Relationships between Molecular Properties and Reac-tion Rate Constants. Correlation analysis between thechemical properties and environmental behaviors of con-taminants is of importance in regulatory decision-makingregarding chemical uses. In addition, valuable informationabout abiotic and biotic reactions can be deduced from theanalysis. Because of the global contamination and toxicity,the reductive transformations of halogenated contaminantsare of interest and are a recent research focus for manyresearch groups. In accordance with the accumulation ofexperimental data, various aspects of reductive dehaloge-nation were studied in relation to the molecular properties,derived from theoretical chemistry (34, 36, 37). It is knownthat highly chlorinated compounds are generally moresensitive to reductive dechlorination than their lower chlo-rinated analogues (37, 38). PBDEs also showed a similarpositive correlation between the debromination rate and thedegree of bromination (Figure 3).

Linear free energy relationships (LFER) are usually appliedfor the explanation of differences in reactivity of homologouscompounds with different degrees of halogenation (37).Several molecular properties are used as independentvariables in LFER analysis. For example, Woods and Trobaugh(31) find a good correlation between the rate and regiose-lectivity of reductive dechlorination of PCBs and the Gibbsfree energies of formation. The LUMO was found to be afairly good descriptor in the correlation analysis of reductivedechlorination of chloroalkanes and -alkenes between ze-rovalent iron (37). Reductive debromination of PBDEs showedsimilar positive correlation between the initial reaction rateconstant and LUMO or heat of formation (Figure 7).

FIGURE 4. Occurrence of (n-1)-bromodiphenyl ethers from specificn-bromo-DEs where the parent BDEs were (A) BDE 28, (B) BDE 47,(C) BDE 66, and (D) BDE 100 after zerovalent iron treatment at 30°C. DPE, diphenyl ether. Error bars indicate standard deviations.Unidentified compounds 1 and 2 may be BDEs 50 and 51, respectively.

FIGURE 5. Mass balance of reductive debromination of BDEs 28(A), 47 (B), 66 (C), and 100 (D) by zerovalent iron during 40 days ofincubation at 30 °C. Black and white bars are represented parentand product BDEs, respectively.

FIGURE 6. Degradation of BDEs 100 (-O-) and 209 (-b-) duringthe entire experimental period (40 days). Error bars indicate standarddeviations.

VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2283

Reduction with zerovalent iron may occur through hydro-genolysis or direct electron transfer (37). The reaction ratesof halogenated alkanes correlate well with LUMO (i.e.,electron affinity), which indicates direct electron transfer (37,39). The chemical stability, represented by heat of formation(Hf), can also be a driving force of the reaction. Because ofthe large steric volume of bromine, highly brominated DEshave higher Hf than their lower bromo congeners (Table 3),which leads to a fast reduction reaction of BDEs 209 or 100.Although decreases of the number of bromines in BDEs areaccompanied by increases of bioavailability or the possibilityof biodegradation, the chemical stability against reductivedegradation is also increased. Currently, it is believed thatlower BDEs (di- to tetrabromo BDEs) in environmentalsamples come from the minor ingredients of commercialtechnical penta-BDEs or octa-BDEs products (33). The resultsof this study suggest that lower bromo-DEs derive from higherbromo BDEs through reducing reagents, which may be amechanism of lower BDEs formation in the environment.Complete debromination of BDE 28 to diphenyl ether byhighly reductive iron suggests potential values of zerovalent

iron for remediation of PBDEs in the environment (Figure4).

In the congener-specific analysis of the debrominatedproducts from the reduction of BDEs 28, 47, 66, and 100,almost all the possible congeners were detected. However,the relative abundance of each congener varied largely, wherethe most abundant products were BDEs 15, 28, 37, and 47from BDEs 28, 47, 66, and 100, respectively (Figure 4).m-Bromines or o-bromines were more susceptible to de-bromination than those at the para position. Regioselectiveformation of specific congeners was reported in the reductionof chlorophenols with zerovalent iron and vitamin B12 (25,41). The regioselectivity of dehalogenation can be correlatedwith various molecular properties, including Gibbs freeenergy, LUMO, and redox potential (31, 37, 38, 41).

Although the LUMO correlated well with debrominationrate of parent BDEs, the same relationships could not beapplied to the accumulation of specific congeners with (n-1)-bromines (Table 3). Among the several molecular proper-ties, Hf and the difference between Hf of parent and products(∆Hf) show reasonable relationships with the relative abun-dance of products (Table 3, Figure 8). The reaction favoredto more stable products (i.e., low Hf of products or high ∆Hf

between parent and products), which, in turn, are morepersistent in the environment. Such an accumulation mayexplain the abundance of specific congeners in environmentalsamples (e.g., BDEs 17, 28, and 47). However, the variationsin Hf between (n-1)-BDEs were relatively smallsapproximately 2-5 kcal/mol for mono- to nona-BDEs. Thesmall ∆Hf difference might be the cause of complex BDEproduct profiles from BDE 209 reduction (Figure 1).

Reductive Debromination of BDE 209 by Other Reduc-tants. In the natural environment, there are numerousnumbers of reductants that can catalyze the reductivedehalogenation of contaminants, including iron-bearingminerals and sulfide ions (42, 43). Adriaens et al. (44) reportedthat octa- to pentachlorodibenzodioxins could be reductivelydechlorinated by several environmentally relevant chemicals(e.g., resorcinol, catechol, and benzoic acids), which canshuttle electrons from reductants (e.g., zinc metal) to thecontaminants. High water solubility of reductants mayincrease their distribution in the matrix and contact withsolid bound residues, and consequently may decompose thesolid-bound or hydrophobic contaminants. In this experi-ment, iron sulfide and sodium sulfide were selected asreductants, which can exist in an anaerobic environment,and their reductive behavior on BDE 209 was evaluated. After14 days of incubation, 94, 2, and 33% of BDE 209 wastransformed into lower bromo-BDEs by zerovalent iron, ironsulfide, and sodium sulfide treatment, respectively. Althoughthe degradation rate was considerably slower in sodiumsulfide treatment, the congener profile was similar to that ofzerovalent iron treatment (Figure 9). Various sulfur-contain-ing minerals and sulfide ion can directly act as reductantsin the absence of another redox reagent or they can acceleratethe reductive dehalogenation by zerovalent metal (45, 46).We propose that widespread amounts and abundance ofsuch reductants in the environment may be a driving forceof debromination of higher bromo-DEs such as BDE 209 tolower bromo-DEs.

In general, reductive dehalogenations by water-insolubleminerals or zerovalent metals occur on their surfaces, wherehydrophobic contaminants are adsorbed. However, highlyhydrophobic chemicals tend to be strongly adsorbed on soilor sediments in natural environments, which make theapplication of insoluble reductant for chemical remediationdifficult. Because of this limitation, auxiliary methods toincrease the adsorption or availability of contaminants arerequired for the remediation of highly hydrophobic com-

TABLE 3. Molecular Properties of PBDEs

parent BDEs product BDEs

BDEno Hf

a HOMO/LUMObBDEno rel. Ac Hf ∆Hf

d HOMO/LUMO

7 36.75 -9.256/-0.265

1 31.44 5.31 -9.149/-0.035

3 29.29 7.46 -9.330/-0.17628 41.91 -9.314/

-0.5597 0.0 36.75 5.16 -9.256/-0.265

8 20.9 36.24 5.67 -9.179/-0.32715 79.1 33.67 8.24 -9.178/-0.402

47 50.44 -9.419/-0.737

17 33.9 44.76 5.68 -9.331/-0.483

28 66.1 42.56 7.88 -9.314/-0.55966 47.86 -9.463/

-0.77125 8.8 42.45 5.41 -9.417/-0.450

28 14.7 42.56 5.30 -9.314/-0.55933 32.4 41.86 6.00 -9.336/-0.41937 44.1 39.22 8.64 -9.375/-0.517

100 59.52 -9.471/-0.973

47 42.9 50.44 9.08 -9.419/-0.737

50e 16.9 54.07 5.45 -9.423/-0.84351e 18.2 53.22 6.30 -9.363/-0.62575 22.0 50.82 8.70 -9.341/-0.904

209 98.81 -9.849/-1.843

206e 87.68 11.13 -9.775/-1.423

207e 91.39 7.42 -9.764/-1.260208e 91.58 7.23 -9.841/-1.246

a Heat of formation in kilocalories per mole. b HOMO/LUMO inelectronvolts. c Percent of each BDEs to total product BDEs. d Differenceof Hf between parent and product BDEs. e Not confirmed but possiblyproduced from BDE 100 or 209.

FIGURE 7. Correlations between initial debromination rate constantsand heat of formation (Hf, A) and lowest unoccupied molecularorbital (LUMO, B).

2284 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005

pounds with zerovalent metal or minerals. For example,several kinds of cationic surfactant can enhance the deg-radation of haloalkane contaminant (47). Sulfide ion is water-soluble and can act more easily on solid-adsorbed contami-nants. Therefore, lower bromo-DEs can probably derive fromBDE 209 or other higher brominated congeners in anaerobic

environments where sulfide or related minerals are abundant.The transformation to lower BDEs by sulfide ion suggestedpossible use of sulfides as a BDE remediation reagent.

In summary, BDE 209 was easily transformed to lowerbromo congeners. Rapid debromination of highly brominatedcongeners suggests that zerovalent iron can be used forremediation of BDEs. However, the limited solubility of BDEscan be a problem for the application, which requires furtherstudy. Stepwise reductive debromination to (n-1)-BDEs wasthe dominant reaction in all tested congeners. Hf and LUMOshowed a good correlation with the debromination rate.Chemical stability, represented by Hf and ∆Hf, was relatedto the regioselectivity in the production of (n-1)-bromo-DEs. Reductive debromination of BDE 209 by iron sulfideand sodium sulfide suggests lower bromo-DEs may be formedvia such a mechanism in the environment.

AcknowledgmentsThis work was supported in part by U.S. EPA Grant No.989512-01-1 and USDA/CSREES/TSTAR Grant No. 0034135-9576.

Literature Cited(1) Rahman, F.; Langford, K.; Scrimshaw, M. D.; Lester, J. N.

Polybrominated diphenyl ether (PBDE) flame retardants. Sci.Total Environ. 2001, 275, 1-17.

(2) de Wit, C. A. Brominated flame retardants, report 5065; SwedishEnvironmental Protection Agency, 2000.

(3) de Wit, C. A. An overview of brominated flame retardants in theenvironment. Chemosphere 2002, 46, 583-624.

(4) Renner, R. Increasing levels of flame retardants found in NorthAmerican environment. Environ. Sci. Technol. 2000, 34, 452A-453A.

(5) Akutsu, K.; Kitagawa, M.; Nakazawa, H.; Makino, T.; Iwazaki, K.;Oda, H.; Hori, S. Time-trend (1973-2000) of polybrominated

FIGURE 8. Proposed debromination pathways of BDEs 100, 75, 47, and 28. Numbers in parentheses are heats of formation in kilocaloriesper mole, calculated by AM1 semiempirical force field; thick lines indicate the most favorable pathways according to the difference inheats of formation between parent and product BDEs.

FIGURE 9. GC-ECD chromatograms of BDE 209 treated withzerovalent iron (A), iron sulfide (B, signal enhanced 30-fold), andsodium sulfide (C) at 30 °C for 14 days. Peaks of product BDEs weregrouped into penta-, hexa-, hepta- and octabromo-DEs.

VOL. 39, NO. 7, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2285

diphenyl ethers in Japanese mother’s milk. Chemosphere 2003,53, 645-654.

(6) Ikonomou, M. G.; Rayne, S.; Addison, R. F. Exponential increasesof the brominated flame retardants, polybrominated diphenylethers, in the Canadian Arctic from 1981 to 2000. Environ. Sci.Technol. 2002, 36, 1886-1892.

(7) Betts, K. S. Flame-proofing the Arctic? Environ. Sci. Technol.2002, 36, 188A-192A.

(8) Behnisch, P. A.; Hosoe, K.; Sakai, S. I. Brominated dioxin-likecompounds: In vitro assessment in comparison to classicaldioxin-like compounds and other polyaromatic compounds.Environ. Int. 2003, 29, 861-877.

(9) Meerts, I. A. T. M.; Letcher, R. J.; Hoving, S.; Marsh, G.; Bergman,A.; Lemman, J. G.; van der Burg, B.; Brouwer, A. In vitroestrogenicity of polybrominated diphenyl ethers, hydroxylatedPBDEs, and polybrominated bisphenol A compounds. Environ.Health Perspect. 2001, 109, 399-407.

(10) Meerts, I. A. T. M.; Van Zanden, J. J.; Luijks, E. A. C.; VanLeewenbol, I.; Marsh, G.; Jakobsson, E.; Bergman, A.; Brower,A. Potent competitive interactions of some brominated flameretardants and related compounds with human transthyretinin vitro. Toxicol. Sci. 2000, 56, 95-104.

(11) Dumler, R.; Lenoir, T. D.; Hutzinger, O. Thermal formation ofpolybrominated dibenzodioxins (PBDD) and dibenzofurans(PBDF) from bromine containing flame retardants. Chemosphere1989, 19, 305-308.

(12) Ebert, J.; Bahadir, M. Formation of PBDD/F from flame-retardedplastic materials under thermal stress. Environ. Int. 2003, 29,711-716.

(13) Hakk, H.; Letcher, R. J. Metabolism in the toxicokinetics andfate of brominated flame retardants. Environ. Int. 2003, 29, 801-828.

(14) Stapleton, H. M.; Alaee, M.; Letcher, R. J.; Baker, J. E. Debro-mination of the flame retardant decabromodiphenyl ether byjuvenile carp (Cyprinus carpio) following dietary exposure.Environ. Sci. Technol. 2004, 38, 112-119.

(15) Schmidt, S.; Wittich, R. M.; Erdmann, D.; Wilkes, H.; Francke,W.; Fortnagel, P. Biodegradation of diphenyl ether and itsmonohalogenated derivatives by Sphingomonas sp. strain SS3.Appl. Environ. Microbiol. 1992, 58, 2744-2750.

(16) Schmidt, S.; Fortnagel, P.; Wittich, R. M. Biodegradation andtransformation of 4,4′- and 2,4-dihalodiphenyl ethers bySphingomonas sp. strain SS33. Appl. Environ. Microbiol. 1993,59, 3931-3933.

(17) Rayne, S.; Ikonomou, M. G.; Whale, M. D. Anaerobic microbialand photochemical degradation of 4,4′-dibromodiphenyl ether.Water Res. 2003, 37, 551-560.

(18) Hua, I.; Kang, N.-G.; Jafvert, C. T.; Fabrega-Duque, H. R.Heterogeneous photochemical reactions of decabromodiphenylether. Environ. Toxicol. Chem. 2003, 22, 798-804.

(19) Bezares-Cruz, J.; Jafvert, C. T.; Hua, I. Solar photodecompositionof decabromodiphenyl ether: Products and quantum yield.Environ. Sci. Technol. 2004, 38, 4149-4156.

(20) Eriksson, J.; Green, N.; Marsh, G.; Bergman, A. Photochemicaldecomposition of 15 polybrominated diphenyl ether congenersin methanol/water. Environ. Sci. Technol. 2004, 38, 3119-3125.

(21) Keum, Y.-S.; Li, Q. X. Reduction of nitroaromatic pesticides withzerovalent iron. Chemosphere 2004, 54, 255-263.

(22) Wang, C.-B.; Zhang, W.-X. Synthesizing nanoscale iron particlesfor rapid and complete dechlorination of TCE and PCBs. Environ.Sci. Technol. 1997, 31, 2154-2156.

(23) Kluyev, N.; Cheleptchikov, A.; Brodsky, E.; Soyfer, V.; Zhilnikov,V. Reductive dechlorination of polychlorinated dibenzo-p-dioxins by zerovalent iron in subcritical water. Chemosphere2002, 46, 1293-1296.

(24) Sayles, G. D.; You, G.; Wang, M.; Kuferle, M. J. DDT, DDD, andDDE dechlorination by zerovalent iron. Environ. Sci. Technol.1997, 31, 3448-3454.

(25) Kim, Y.-H.; Carraway, E. R. Dechlorination of pentachlorophenolby zerovalent iron and modified zerovalent irons. Environ. Sci.Technol. 2000, 34, 2014-2017.

(26) Weber, E. J. Iron-mediated reductive transformations: Inves-tigation of reaction mechanism. Environ. Sci. Technol. 1996,30, 716-719.

(27) Philips, D. H.; Gu, B.; Watson, D. B.; Roh, Y.; Liang, L.; Lee, S.Y. Performance evaluation of a zerovalent iron reactive barrier:Mineralogical characteristics. Environ. Sci. Technol. 2000, 34,4169-4176.

(28) Powell, R. M.; Puls, R. W. Proton generation by dissolution ofintrinsic or augmented aluminosilicate minerals for in situcontaminant remediation by zero-valence-state iron. Environ.Sci. Technol. 1997, 31, 2244-2251.

(29) Marsh, G.; Hu, J.; Jakobsson, E.; Rahm, S.; Bergman, A. Synthesisand characterization of 32 polybrominated diphenyl ethers.Environ. Sci. Technol. 1999, 33, 3033-3037.

(30) Magar, V. S.; Stensel, H. D.; Puhakka, J. A.; Ferguson, J. F.Sequential Anaerobic dechlorination of pentachlorophenol:Competitive inhibition effects and a kinetic model. Environ.Sci. Technol. 1999, 33, 1604-1611.

(31) Woods, S. L.; Trobaugh, D. J. Polychlorinated biphenyl reductivedechlorination by vitamin B12s: Thermodynamics and re-giospecificity. Environ. Sci. Technol. 1999, 33, 857-863.

(32) Joo, S. H.; Feitz, A. J.; Waite, T. D. Oxidative degradation of thecarbothioate herbicide, molinate, using nanoscale zerovalentiron. Environ. Sci. Technol. 2004, 38, 2242-2247.

(33) Hale, R. C.; Alaee, M.; Manchester-Neesvig, J. B.; Stapleton, H.M.; Ikonomu, M. G. Polybrominated diphenyl ether flameretardants in the North American environment. Environ. Int.2003, 29, 771-779.

(34) Yak, K.-H.; Lang, Q.-Y.; Wai, C.-M. Relative resistance ofpositional isomers of polychlorinated biphenyls toward reduc-tive dechlorination by zerovalent iron in subcritical water.Environ. Sci. Technol. 2000, 34, 2792-2798.

(35) Soderstrom, G.; Sellstrom, U.; de Wit, C.; Tysklind, M. Photolyticdebromination of decabromodiphenyl ether (BDE 209). Environ.Sci. Technol. 2004, 38, 127-132.

(36) Gibb, C.; Satapanajaru, T.; Comfort, S. D.; Shea, P. J. Remediatingdicamba-contaminated water with zerovalent iron. Chemo-sphere 2004, 54, 841-848.

(37) Scherer, M. M.; Balko, B. A.; Gallagher, D. A.; Trantnyek, P. G.Correlation analysis of rate constants for dechlorination byzerovalent iron. Environ. Sci. Technol. 1998, 32, 3026-3033.

(38) Fu, Q.-S.; Barkovskii, A. L.; Adriaens, P. Reductive Transformationof Dioxins: An assessment of the contribution of dissolvedorganic matter to dechlorination reactions. Environ. Sci. Technol.1999, 33, 3837-3842.

(39) Fleming, I. Frontier orbitals and organic chemical reactions;Wiley: London, 1976.

(40) Smith, M. H.; Woods, S. L. Regiospecificity of chlorophenolreductive dechlorination by vitamin B12s. Appl. Environ. Mi-crobiol. 1994, 60, 4111-4115.

(41) Dolfing, J. Regiospecificity of chlorophenol reductive dechlo-rination by vitamin B12s. Appl. Environ. Microbiol. 1995, 61,2450-2451.

(42) Ulrich, G. A.; Breit, G. N.; Cozzarelli, I. M.; Suflita, J. M. Sourcesof sulfate supporting anaerobic metabolism in a contaminatedaquifer. Environ. Sci. Technol. 2003, 37, 1093-1099.

(43) Lee, W.-J.; Batchelor, B. Reductive capacity of natural reductants.Environ. Sci. Technol. 2003, 37, 535-541.

(44) Adriaens, P.; Chang, P. R.; Barkovskii, A. L. Dechlorination ofPCDD/F by organic and inorganic electron-transfer moleculesin reduced environments. Chemosphere 1996, 32, 433-441.

(45) Lipczynska-Kochany, E.; Harms, S.; Milburn, R.; Sprah, G.;Nadarajah. N. Degradation of carbon tetrachloride in thepresence of iron and sulphur containing compounds. Chemo-sphere 1994, 29, 1477-1489.

(46) Lesage, S.; Brown, S.; Hosler, K. R. Degradation of chlorofluo-rocarbon-113 under anaerobic conditions. Chemosphere 1992,24, 1225-1243.

(47) Alessi, D.; Zhao, H.-Li. Synergistic effect of cationic surfactantson perchloroethylene degradation by zerovalent iron. Environ.Sci. Technol. 2001, 35, 3713-3717.

Received for review July 24, 2004. Revised manuscript re-ceived December 11, 2004. Accepted December 22, 2004.

ES048846G

2286 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 7, 2005