Environ. Sci. Technol. 2007, 41, 6425-6432
Formation of Polychlorinated Dibenzo-p-Dioxins and Polychlorinated Dibenzofurans (PCDD/F) in Fires of Arsenic-Free Treated Wood: Role of Organic Preservatives NIGEL W. TAME, BOGDAN Z. DLUGOGORSKI,* AND ERIC M. KENNEDY Process Safety and Environment Protection Research Group, School of Engineering, The University of Newcastle, Callaghan, NSW 2308, Australia
This article demonstrates that biocidal organochlorines such as tebuconazole and permethrin, employed in formulations of wood preservatives, produce significant quantities of polychlorinated dibenzo-p-dioxin and polychlorinated dibenzofuran (PCDD/F) when subjected to thermal decomposition under oxidative conditions. Both tebuconazole and permethrin form PCDD/F during gas-phase oxidation, but much greater yields occurred in the presence of surrogate ash corresponding to wood treated with copper-based fungicides. The significant yields have implications for the increased toxicity of PCDD/F emissions during fires of wood impregnated by combination of organic and copper-based preservatives. The oxidative pyrolysis of tebuconazole and permethrin over simulated wood ash generated amounts of PCDD/F exceeding those of untreated wood by 3 orders of magnitude. We obtained yields of 1000 ng WHO97-TEQ/g tebuconazole and 5500 ng WHO97-TEQ/g permethrin when reacting the organochlorines in an oxidizing atmosphere over surrogate wood ash. Gasphase oxidation also produce measurable quantities of PCDD/ F, corresponding to 1 ng WHO97-TEQ/g tebuconazole and 36 ng WHO97-TEQ/g permethrin. In the case of tebuconazole, the present measurements correlate well with those obtained from oxidative pyrolysis of CBA-treated wood in the cone calorimeter. It appears that permethrin and tebuconazole provide phenyl and diphenyl precursors to formation of PCDD/F and both constitute a source of chlorine upon fragmentation.
Introduction Wood preservatives containing organic components are currently gaining popularity as the use of heavy metals such as chromium and arsenic is being restricted. Future methods of controlled and uncontrolled disposal of the waste wood will ultimately center on thermal treatment, along with the constant possibility of accidental combustion. As a consequence, the impact of wood treatment on the combustion emissions are being investigated for these additives (1, 2). * Corresponding author; phone: +61 2 4921 6176; fax: +61 2 4921 6920; e-mail:
[email protected]. 10.1021/es0703980 CCC: $37.00 Published on Web 08/21/2007
2007 American Chemical Society
Alternative preservatives that are emerging contain constituents that exhibit minimal mammalian toxicity while remaining effective biocides. Copper is a notable inclusion in the main replacement solutions, alkaline copper quaternary (ACQ) and copper boron azole (CBA). CBA contains Cu(II) as copper carbonate, boron as boric acid, and tebuconazole (Figure 1). The latter is an organic compound that has historically been used in the agricultural industry as a fungicide. Our earlier research has established that the presence of CBA significantly increases the levels of polychlorinated dibenzo-p-dioxin and polychlorinated dibenzofuran (PCDD, PCDF, PCDD/F) emitted during oxidative pyrolysis of the impregnated wood, whereas no difference was observed over untreated wood upon ignition and flaming combustion (2). Recently, the formation and emission of PCDD/F have attracted significant research scrutiny owing to their classification as highly carcinogenic and persistent organic pollutants, as reflected by hundreds of research papers and extensive governmental regulations. Domestic or accidental fires are recognized as potentially significant sources of PCDD/F as a consequence of favorable temperature and oxygen environments (3). Such fires also include domestic heating and burn barrels, however, all biomass fires, regardless of combustor design, will ultimately involve oxidative pyrolysis of the fuel, ignition of the volatiles and flaming combustion, and burnout of the char residue. A recent literature survey illustrates that additional sources of chlorine and structural precursors present in wood as preservatives have potential for enhancing the formation of PCDD/F (4). There has only been limited research published on the formation of PCDD/F from the combustion of organochlorines other than chlorophenols, benzenes, and PVC. Vikelsøe and Johansen (5) burnt pure compounds, such as pentachlorophenol (PCP), to represent fires using standard fire simulation apparatus. Compounds containing halogenated benzene substituents yielded greater amounts of PCDD/F than those without, although PCP does contain levels of PCDD/F contamination. While this observation is useful as an initial estimate of the potential of organic wood preservatives to form PCDD/F, the relevant yields cannot be extrapolated to real fire scenarios because of significant differences in the combustion conditions (such as much greater fuel to oxygen ratios) and very high concentrations of precursors and chlorine in the product gases. While the thermal decomposition of organochlorines has been studied (6-8), only a few publications have examined compounds relevant to wood preservatives. One relevant study has reported the results from pyrolysis coupled with mass spectrometry for common preservative components including permethrin and tebuconazole (9). Others have examined the pyrolysis of wood impregnated with similar compounds with the addition to pentachlorophenol (PCP) (10) and the decomposition of cis-permethin and cypermethrin at low temperatures (11). None of these studies have established detailed degradation mechanisms. This article constitutes a first-order assessment of the oxidative decomposition of biocidal organochlorines via the gas phase or surface interaction as valid pathways for the formation of PCDD/F. The aim here is to investigate the potential of combustion of treated wood to become a new source of dioxins in the environment. The methodologies adopted in this study, including the selection of organochlorines and experimental conditions, reflect this intention. In particular, the homo- and heterogeneous oxidation of tebuconazole and permethrin are investigated under condiVOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6425
FIGURE 1. Structures of (i) tebuconazole, the organic preservative in CBA; and (ii) permethrin. tions designed to simulate the oxidative pyrolysis of these preservatives in wood fires (as illustrated in Figure 1 of Supporting Information).
Experimental Section To simulate the oxidative pyrolysis of impregnated timber, the experiments involved generating dilute streams of tebuconazole and permethrin and oxidizing them in a 12 mm i.d. quartz tube reactor, which was heated by a horizontal tube furnace (tube o.d 40 mm, length 600 mm), as illustrated in Figure 2. The inlet stream was generated upstream of the reactor by passing N2 through a heated vaporizer unit containing the solid tebuconazole or permethrin substrate, which volatilized slowly upon heating. The vaporizer was heated to 150 °C to slowly evaporate the solid material under the flow of nitrogen. A heated quartz transfer line fed the dilute reactant/N2 mixture to the reactor. Oxygen, introduced as the mix, entered the reactor, resulting in a N2 to O2 ratio of 9:1. A plug of glass wool was placed in the tube entrance to ensure the reactants were uniformly mixed as they were brought to the reaction temperature. Dilute fluxes of reactants ensured a low fuel to oxygen equivalence ratio and avoided sooting. The average reactant evaporation rate was measured to be 0.6 ((10%) mg/min, which corresponded to a concentration of tebuconazole and permethrin of 0.26 ((0.03)% (w/w) for all experiments, except experiment 4. In experiment 4, the gas flow was reduced to 100 cm3/min, which elevated the concentration of tebuconazole to 0.52% (w/w), probing the sensitivity of PCDD/F formation to reactant loading of the inlet gas. Reaction temperatures were achieved by manipulating the three zones of the tube furnace. Measurements with a 1.5 mm type K chromel alumel thermocouple, by insertion into the surrogate ash plug, confirmed the desired temperature (350 °C for heterogeneous experiments) by insertion into the surrogate ash plug. The solid was used to simulate the residue arising from burning of wood treated with a copper preservative and composed of copper (1%), as CuO, with SiO2 as balance. Typically, 0.6 g of this surrogate ash was loaded into the reactor and held in place by glass wool. The reactor length occupied by this mass of solid combined with the gas flow of 200 cm3/min (as measured at 25 °C and 1 atm) permitted a contact time of 0.15 s. Note that in experiment 4 (Table 1), the flowrate was reduced to 100 cm3/min and the resultant contact time increased to 0.3 s. Two different residence times (2 and 5 s) were utilized for gas-phase oxidation of tebuconazole, and the interaction with CBA ash was represented by performing the experiments with a surrogate ash made up from copper oxide/silica (1% Cu) at contact times of 0.15 and 0.3 s. A similar program was undertaken with permethrin. Copper oxide was selected as 6426
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
the active component of the surrogate ash because it is the thermal decomposition product of copper carbonate (present in CBA). Reaction temperatures (350 °C for heterogeneous and 500 °C for homogeneous experiments) were selected to simulate possible temperatures of oxidative pyrolysis in fires, which also corresponded to temperatures known to enhance the formation of PCDD/F (12). Nitrogen to oxygen ratio was kept constant at 9:1 for all experiments. Product gases were bubbled through a chilled impinger containing dichloromethane and acetone and a XAD2 resin cartridge to collect PCDD/F. The experimental program is provided in Table 1. Upon completion of each experiment, the XAD2 and surrogate ash plugs were Soxhlet extracted and the solvent combined with that from the impinger. Hot Soxhlet extraction was performed in an automated Soxhlet unit (Buchi, Switzerland) for 6 h. The sample cleanup and analysis for PCDD and PCDF were as discussed elsewhere (13). The reactor tube was filled with solvent between experiments and extracted in an ultrasonic bath to remove residual contaminants. All glass ware was rinsed with additional solvent, washed in detergent, and annealed at 450 °C for 1 h. Additional details about the material and the apparatus can be found in Supporting Information accompanying the article.
Results and Discussion We added no external chlorine to the reactor feeds to assess whether the chlorine present in the permethrin and tebuconazole molecules themselves was sufficient for significant PCDD/F formation to occur. This approach is consistent with simulating wood decomposition where the timber contains only trace levels of chlorine. The presence of chlorinated precursors is important to consider in assessing the propensity of these precursors to produce PCDD/F, but formation of PCDD/F may also be driven by the generation of secondary HCl and Cl2 to facilitate further chlorination of intermediates or dibenzo-p-dioxin (DD) and dibenzofuran (DF). Table 1 summarizes the pertinent experimental measurements. Homologue distributions from the experiments are presented in Figure 3a, while example chromatographic congener patterns (db-5 ms) can be found in Supporting Information accompanying this article. Tetra-octa-PCDD/F yields are presented in addition to TEQ normalized to 1 g of feed reactant (i.e., tebuconazole or permethrin). Significant amounts of PCDD/F were generated by tebuconazole and permethrin without additional chlorine sources, up to 340 000 ng/g permethrin. The reaction of permethrin with the surrogate ash (experiment 7) was responsible for the most PCDD/F generated and represented a chlorine utilization of 0.089%. Tebuconazole decomposed with 0.029% chlorine incorporation into PCDD/F under the same conditions (experiment 3). Notably, increasing the reactant contact time from 0.15 to 0.3 s (experiment 4) diminished the conversion of tebuconazole to PCDD/F from 78 000 to 34 000 ng/g, however, the homologue (Figure 3a) and congener patterns remained consistent. Shorter residence times favored gas phase formation of PCDD/F from tebuconazole. The longer reaction time of 5 s at 500 °C did not render quantifiable levels of PCDD/F, but at 2 s, 46 ng/g was detected. The trend reversed for the homogeneous oxidation of permethrin, as 1400 and 1800 ng/g resulted for 2 and 5 s at 500 °C, respectively. Chlorination was more efficient at the longer residence time, and a shift toward a greater level of tetrachlorodibenzofuran (TCDF) in the homologue contributions accompanied the difference in yields. Incorporation of chlorine into PCDD and PCDF was more efficient during the gas-phase oxidation of permethrin than tebuconazole (experiment 6 versus experiment 2) with a 16-
FIGURE 2. Schematic diagram of the experimental apparatus, including examples of the temperature profiles along the reactor.
TABLE 1. Summary of the Experimental Program and Measurements expt
reactant
conditions: temp., residence/contact time, expt duration
tebuconazolea gas phase, 500 °C, 5 sb, 90 min tebuconazole gas phase, 500 °C, 2 sb, 135 min tebuconazole CuO/SiO2, 350 °C, 0.15 sc, 200 min tebuconazole CuO/SiO2, 350 °C, 0.3 sc, 45 min permethrin gas phase, 500 °C, 5 sb, 90 min permethrin gas phase, 500 °C, 2 sb, 135 min permethrin CuO/SiO2, 350 °C, 0.15 sc, 45 min
1 2 3 4 5 6 7 a
PCDD/F ng TEQ/gd
1.1
PCDD ng/g
2.5
PCDF ng/g
PCDD/F ng/g
44
46
PCDD/F yield mol % × 103
Cl conversion to PCDD/F, %
PCDD/ ∑2,3,7,8s/ PCDF ∑PCDD/F
0.0037
0.000024
0.058
0.71
1 000
23 000
46 000
78 000
6.40
0.029
0.50
0.062
340
9 000
25 000
34 000
3.2
0.014
0.36
0.045
36
560
1 200
1 800
0.21
0.00051
0.45
0.17
14
160
1 300
1 400
0.18
0.00037
0.13
0.050
5 500
160 000
0.089
0.85
0.080
Concentration of PCDD/F below detection limit.
b
180 000 340 000
Residence time. c Contact time.
fold increase in PCDD/F detected per mole of Cl. The difference receded when surrogate ash was introduced, and permethrin produced 3 times more PCDD/F (experiment 7 compared with experiment 3). These experiments were also responsible for the greatest yields, which corresponded to 0.089 and 0.029% conversion of chlorine, respectively. PCDF were preferred products over PCDD for all experiments. PCDD:PCDF ratios below unity in Table 1 indicate this trend, which correlates with typical combustion and oxidative formation of PCDD/F (12). The combined effect of surrogate ash and lower temperature led to higher relative production of PCDD for both reactants, particularly in the oxidation of permethrin with a PCDD:PCDF of 0.85. Gasphase oxidation of permethrin at 500 °C yielded a greater fraction of PCDD than the corresponding experiments with tebuconazole. The presence of surrogate ash raised PCDD/F production above that for gas-phase oxidation at higher temperature,
d
40
WHO97-TEQ.
emphasizing the importance of an active solid catalyst. Permethrin reacted to form greater quantities of PCDD/F than tebuconazole under identical conditions, perhaps as a consequence of its elevated chlorine content (18% for permethrin, 11% for tebuconazole), higher reactivity, and the existence of the diphenyl ether fragment in its structure. The tetra- to octachloro-PCDD/F homologue profiles for all experiments are collated in Figure 3a. Most experiments produced distributions characterized by prominent content of tetra-DD and -DF homologues, with descending contribution for the higher chlorinated homologue groups. Notable variations occurred in experiment 7 (permethrin and surrogate ash) with greater fraction of PCDD, and the gas-phase oxidation of tebuconazole (experiment 2), which was characterized by a greater proportion of higher chlorinated PCDF. Table 1 also summarizes the concentrations of PCDD/F expressed as TEQ. High PCDD/F yields confirm that the VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6427
FIGURE 3. Homologue profiles derived from the experimental measurements: (a) summary of the current results; (b) a comparison of homologues from oxidative pyrolysis of CBA-impregnated wood and from oxidation of tebuconazole over surrogate ash. combination of tebuconazole and copper oxide increases the toxicity of emissions in fires of treated wood. Tebuconazole generated 1 ng TEQ/g during gas-phase oxidation at 500 °C (experiment 3), but this increased to 1000 ng TEQ/g upon introducing CuO at 350 °C (experiment 4). Permethrin formed 36 and 14 ng TEQ/g homogeneously, and 5500 ng TEQ/g with CuO. These yields are considerably greater than those typically observed in untreated wood fires, which range from 7.5 × 10-6 to 1.4 × 10-3 ng TEQ/g wood (13). PCDD/F emissions from CBA-treated wood may potentially increase 3 orders of magnitude over those usually observed in wood fires, assuming a nominal retention of 0.025% (w/w) for impregnation of tebuconazole into wood (14). This increase is similar to that observed for the oxidative pyrolysis of CBA-treated wood compared to that of untreated pine (2) when the oxidative pyrolysis of CBA reached a maximum of 2200 ng TEQ/kg wood, compared to 2.4 ng TEQ/kg for untreated pine. The thermal decomposition of wood preserved with permethrin, in the absence of copper-based fungicides, would not appreciably increase emissions of PCDD/F. However, under more favorable temperatures, a higher residual chlorine content or the 6428
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
introduction of copper oxide, the formation becomes significant (experiment 7). Pentachlorophenol (PCP) is another organochlorine compound capable of protecting wood. Its use has been in decline due to an ability to form PCDD/F during synthesis and upon combustion of impregnated wood (15). Several researchers have studied this behavior, and comparisons with the results in the present study are useful. Unfortunately variations in experimental techniques mean that direct comparisons with previous research are difficult. Vikelsøe and Johansen (5) burnt several grams of pure pesticides at 500 °C, which emitted PCDD/F yields ranging from 740 and 57 ng TEQ/g for PCP and down to 0.10 ng TEQ/g for dichlorprop (2,4dichlorophenoxypropionic acid). Sooting would have been unavoidable during flaming combustion of pure chemicals, which would complicate PCDD/F formation pathways and make mechanistic conclusions difficult. Annealing small quantities of organic reactants, e.g., PCP (16) and hexachlorocyclohexane (17) (the pesticide lindane), was a further experimental technique for assessing potency for the formation of PCDD/F. The experiments generated extremely high levels of PCDD/F because very long residence
times were used (up to 2 h). PCP (67% Cl) mixed with CuO and annealed at 280 °C for 2 h generated 6.2 mg PCDD/F/g, with OCDD as the predominant congener (>90%). Similarly, lindane, with a chlorine content of 73.2%, produced 4.5 mg PCDD/F/g in the presence of a Fe3O4 catalyst at 350 °C. PCDD/F Formation from Tebuconazole. Experiments 1-4 were designed to reconcile the oxidation of tebuconazole with increased PCDD/F emissions from CBA-treated wood in the cone calorimeter, as depicted schematically in Figure 1 of Supporting Information. Figure 3b compares the distribution of the PCDD/F from the oxidation of tebuconazole in the presence of copper in surrogate ash (present results, mean of experiments 3 and 4) with the average solid and gaseous emissions from the combustion of CBA-treated wood in the cone calorimeter (2). The similar nature of the profiles, PCDF > PCDD, and strong contributions from the TCDF through HxCDF, suggests that a similar “bulk” formation pathway is present and that tebuconazole contributed to the emissions from CBA-impregnated timber. An additional examination was conducted of the congener profiles for the dominant TCDF, PeCDF, and HxCDF homologues to establish further correlation. When the congener profiles were analyzed, the patterns were found to be quite different. The TCDF, PeCDF, and HxCDF chromatograms are provided in Figure 2 of Supporting Information to heighlight these differences. Similar congeners eluted from the high-resolution GC/MS analysis of the samples from tebuconazole oxidation and decomposition of CBA-treated wood, although not at the same relative abundance. It would be unrealistic to anticipate that the complexity of the chemistry involved (i.e., decomposing wood impregnated with preservatives, in comparison with oxidation of pure tebuconazole) allows chlorine concentration, residence times, and other experimental conditions to be replicated, leading to the same relative abundance of the congeners. Hence, the oxidative decomposition of tebuconazole exhibited the same PCDD/F homologues as those observed from the decomposition of CBA-treated wood under the cone calorimeter, although both profiles differed in detailed distribution of PCDD/F congeners produced. There have been no dominant congeners in the tebuconazole experiments, indicating no favoring of any specific precursor species. This is because the formation of PCDD and PCDF involves condensation of various chlorinated precursors and competing chlorination and dechlorination of the product PCDD and PCDF. An interesting observation was noted in the PCDD/F congener pattern resulting from the gas-phase oxidation of tebuconazole (experiment 2). Figure 3a shows that higher chlorinated PCDF, dominating the homologue distribution, are unlike those of the other experiments. In addition, the 2,3,7,8-substituted congeners are the most prominent in each homologue, which leads to the considerably higher value of the 2,3,7,8- to total PCDD/F ratio (Table 1; 0.71 for experiment 2 compared to 0.045-0.170 for other experiments). Ryu and Mulholland reported a congener pattern identical to that of experiment 2 for the sequential chlorination of dibenzofuran by CuCl2 (18). PCDD/F formation via decomposition of tebuconazole can be attributed to fragmentation of the chlorobenzene substituent. The preference for PCDF observed in experiments 3 and 4, and to a greater extent in experiment 2 (gasphase oxidation), is consistent with formation of PCDF from benzene derivatives (19). Chlorobenzene is an established precursor for the gas-phase formation of PCDF via oxidation to a chlorophenoxy radical prior to dimerization and further chlorination. Sommeling et al. (19) reported a mechanism to this effect to account for the gas-phase oxidation of chlorobenzene in the presence of additional chlorine. Conversion of the chlorinated benzene is much lower than
other commonly investigated precursors such as phenol and chlorophenols. Without excess chlorine, chlorobenzene at 500 °C is less likely to yield polychlorinated species (chlorobenzenes, CBz, and chlorophenols, CP) (20), as the homogeneous generation of secondary HCl and Cl2 only becomes significant above 650 °C. Therefore significant quantities of PCDD and PCDF from the gas-phase oxidation of tebuconazole may occur at temperatures higher than those investigated here. The presence of a catalytic surface, however, allows efficient chlorination at much lower temperatures. Chlorobenzene also forms PCDD/F or precursors when a solid catalyst is present. Further chlorination of the ring occurs at temperatures as low as 250 °C (21), and PCB are produced between 200 and 400 °C (22). van den Brink et al. (23) suggested that sequential gas-phase chlorination of chlorobenzene is unlikely when the concentration of oxygen far exceeds that of Cl2, rather, additional chlorination occurs through the reactions of adsorbed chlorine with catalystbound chlorobenzene. The lack of favored congeners in the heterogeneous oxidation of tebuconazole (experiments 3 and 4) then reflects chlorination via interactions with copper in the surrogate CBA ash without generating specific precursors. PCDD/F Formation from Permethrin. Pyrolytic decomposition pathways for permethrin have been empirically described by Meier et al. (10) and Audino et al. (11) and have been studied by means of quantum chemistry calculations by Altarawneh et al. (24). The investigations of Meier et al. and Audino et al. (10, 11), the latter conducted at a relatively low temperature of 210 °C, have identified the major degradation products based upon fragmentation of permethrin at the ester oxygen (i.e., O-CH2 bond), therefore incorporating the diphenyl ether (DPE) substituent (3phenoxybenzaldyhyde, 3-chloromethylphenoxybenzene, 3-phenoxybenzyl methanol, 3-methylphenoxybenzene, and 3-(2,2-dichlorovinyl)-2,2-dimethyl-cyclopropane-carbonic acid) (10). Altarawneh et al. (24) found the energy of the O-CH2 bond as 65.3 kcal/mol (at 0 K), less than the energy of the two O-C linkages in the H2CC6H4-O-C6H5 moiety, of around 71 kcal/mol (at 0 K). This means that benzyl and phenoxy products from the rupture of O-C (ether) linkages in H2CC6H4-O-C6H5 will coexist with those forming as a consequence of the fission of the O-CH2 bond. Figure 4 outlines these products within the context of permethrin decomposition. Formation of PCDD and PCDF from permethrin occurs via the production of DPE, utilizing Cl provided by degradation of the dichlorovinyl substituent. The ability of DPE to behave as a PCDD/F precursor has been assessed for gas phase formation by Wiater and Louw (25) and, to test precursor reactivity on flyash, by Wilhelm et al. (26). The direct conversion of polychlorinated diphenyl ethers to PCDF and PCCD has also been examined (27, 28). Altarawneh et al. (24) identified a direct corridor for formation of furans by bridging between two ortho carbons in R1 and subsequent formation of R2 and ejection of H.
A direct path from R1 to dioxins in gas-phase reactions is absent under pyrolytic conditions, resulting in substantially higher yields of PCDF (24). Under these conditions, PCDD and PCDF also form via secondary pathways involving coupling of phenoxy radicals (as indicated schematically in the bottom right corner of Figure 4) (29), which are however less abundant than R1 in the present system. Under oxidative conditions, a direct channel from DPE to PCDD unlocks (24), VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6429
FIGURE 4. Mechanism proposed for the formation of PCDD and PCDF from the decomposition of permethrin.
increasing in importance as the oxygen content enlarges. In other words, the less oxygen subsists in the system, the smaller the PCDD/PCDF ratio. Under truly nonoxidative conditions, the ratio remains extremely small, although PCDD still form owing to phenoxy-phenoxy coupling (30). This constitutes the reason for higher formation rates of PCDF than PCDD observed in the present, low-oxygen content, gas-phase experiments. Direct gas-phase oxychlorination of DPE occurs at meta and para positions to the bridging O (24, 25). Abstraction of hydrogen will also occur under these conditions to yield either dibenzofuran or, via cleavage at the ether oxygen, chlorobenzene. Wiater and Louw (25) observed product ratios of 3-chloro-DPE:4-chloro-DPE of between 4:1 and 5:1, and DF approximately equal to the sum of 3- and 4-chloro-DPE at 750 K. The annealing of DPE in the presence of model flyash results in similar products (26). PCDF, PCDD, polychlorinated benzenes, and polychlorinated phenols were detected upon annealing DPE with synthetic fly ash and CuCl2 for 1 h in 5% O2 at 250 °C. DPE clearly favored production of PCDF over PCDD, similarly to the gas-phase reactions covered in the two preceding paragraphs. These results demonstrate that, while PCDF are the preferred products from DPE, the parent molecule is capable of decomposing to give phenoxy radicals, which can then condense as PCDD (31). This pathway competes with consumption of PCDD/F by molecular oxygen. The yield of PCDD increases in the presence of surrogate ash, suggesting higher selectivity of phenoxy coupling to PCDD on ash particles (32) in preference to consumption of phenoxy radicals by oxygen (32). The possible mechanism for the oxidative decomposition of permethrin to yield PCDD and PCDF is summarized in Figure 4. PCDD/F generated from the oxidation of permethrin can also result from substitution of Cl in DPE fragments prior to closure of the furan or dioxin ring. Ross et al. (27) and Lindahl et al. (28) investigated this reaction step using polychlorinated diphenyl ether isomers and their PCDD and PCDF products. PCDPE with a Cl ortho to the bridging oxygen demonstrated much greater conversions to PCDF, for example, 3,3′,4,4′,66430
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
CDPE generated predominantly TCDF. Homogeneous PCDD/F reached a maximum between 500 and 700 °C. This temperature range is consistent with that observed for the maximum gas-phase formation of PCDF from the oxidation of the pesticide chloronitrofen (1,3,5-trichloro-4′-nitrodiphenyl ether) (8). Several major PCDD/F congeners were identified for the heterogeneous conversion of permethrin, with 1,2,8,9-TCDF (37% of total TCDF detected, 15% of total PCDD/F) being preponderant, as demonstrated in Figure 3 and Table 1 included in Supporting Information. Other notable TCDF included the 2,3,4,7-, 2,3,4,6-, and 2,3,4,8-TCDF. The gasphase oxidation of permethrin did not yield a consistent PCDF congener profile to that formed in the presence of surrogate ash. TCDF were the major products for oxidation at 500 °C for 2 s, notably the 1,3,6,8- and 1,4,6,8- congeners, while dominant congeners were not apparent when the reaction time increased to 5 s. While the PCDF resulting from gas-phase oxidation differed from that involving surrogate ash, the congener patterns of the PCDD homologues were similar, indicating a common formation route. Experiment 7 produced the largest fraction of PCDD due to the major contributions of several congeners; 1,2,8,9-, 1,2,7,9-, and 1,2,3,9-TCDD together accounted for 14% of the total native PCDD/F. The gas-phase oxidative decomposition of permethrin (experiments 5 and 6) also favored these congeners. GC outputs for the experiments involving permethrin are provided in Figure 4 of Supporting Information to highlight the similarities. The major difference between the three TCDD profiles are the relatively greater quantities of TCDD substituted in the 1 and 3 positions for oxidation at 2 s (experiment 6). These congeners are widely attributed to the condensation of trichlorophenols (33-35). The homologue profiles presented in Figure 3a are similar to those reported for the solid and gas-phase oxidation of a CP mixture (36); The table included in Supporting Information lists all major congeners identified in the oxidative decomposition of permethrin over surrogate ash. Of the major products of experiment 7, 1,2,8,9-TCDF has been noted for
the heterogeneous oxidation of 3,4-dichlorophenol, 1,2,7,9TCDD in the coupling of 3,4-dichlorophenol and 2,4,6trichlorophenol (37). Both 1,2,7,9- and 1,2,8,9-TCDD resulted from the pyrolysis of a trichlorophenol mixture (38). Lindahl et al. (28) produced 1,2,8,9-TCDD upon oxidation of the 2,3,4,2′,3′,4′-hexachlorodiphenyl ether. Sequential chlorination from lower to higher PCDD/F homologues was also observed in the heterogeneous oxidation of permethrin and is described in Table 1 of Supporting Information. For example, 1,2,8,9-TCDF produces 1,2,3,8,9PeCDF and 1,2,3,7,8,9-HxCDF upon further substitution with Cl. Similarly, 1,2,8,9-TCDD was accompanied by appreciable contributions of 1,2,3,8,9-PeCDD and 1,2,3,6,8,9-HxCDD. The PCDD congener patterns for the gas-phase oxidations of permethrin were also consistent with these observations. The measurements presented in this article demonstrate that preservation of wood with treatment formulations that include both an organochlorine biocide and a copper fungicide promotes the formation of toxic PCDD/F in fires. Especially, we have established that two organochlorines, tebuconazole and permethrin, commonly adopted in compositions of wood preservatives that replace CCA formulations, form significant amounts of PCDF and PCDD during their oxidative decomposition.
Acknowledgments We thank the Australian Research Council for funding this study and the staff of the Dioxin Analysis Unit at the National Measurement Institute, Sydney, Australia, for their support and use of analytical facilities. We express our gratitude to Professor John Mackie for his insightful comments and discussions.
Supporting Information Available (i) Additional experimental details regarding materials and apparatus; (ii) a schematic diagram for the decomposition of tebuconazole during oxidative pyrolysis of CBA impregnated wood; (iii) a comparison of chromatographs of toxic PCDF from oxidation of pure tebuconazole over simulated ash with those obtained from emissions from oxidative pyrolysis of CBA-treated wood in cone calorimeter; (iv) identification of major PCDD/F congeners and chromatograph traces for all TCDD congeners observed for oxidative decomposition of permethrin. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Tame, N. W.; Dlugogorski, B. Z.; Kennedy, E. M. Increased PCDD/F formation in the bottom ash from fires of CCA-treated wood. Chemosphere 2003, 50, 1261-1263. (2) Tame, N. W.; Dlugogorski, B. Z.; Kennedy, E. M. PCDD/F formation in flaming combustion, smoldering, and oxidative pyrolysis of “eco-friendly” treated wood. Proc. Combust. Inst. 2005, 30, 1237-1243. (3) Lemieux, P. M.; Lutes, C. C.; Santoianni, D. A. Emissions of organic air toxics from open burning: A comprehensive review. Prog. Energy Combust. Sci. 2004, 30, 1-32. (4) Tame, N. W.; Dlugogorski, B. Z.; Kennedy, E. M. Formation of dioxins and furans during combustion of treated wood. Prog. Energy Combust. Sci. 2007, 33, 384-408. (5) Vikelsøe, J.; Johansen, E. Estimation of dioxin emission from fires in chemicals. Chemosphere 2000, 40, 165-175. (6) Klusmeier, W.; Ohrbach, K. H.; Kuehn, P.; Kettrup, A. Investigations into the thermal decomposition of selected pesticides. J. Anal. Appl. Pyrolysis 1989, 16, 205-211. (7) Klusmeier, W.; Ohrbach, K. H.; Kuehn, P.; Kettrup, A. Investigations into the thermal decomposition of the pesticides metoxuron and chloridazon. J. Anal. Appl. Pyrolysis 1989, 16, 213219. (8) Onodera, S.; Saitoh, K. Formation of chlorodibenzofurans upon thermochemical reactions of diphenyl ether herbicide (CNP). Jpn. J. Toxicol. 1997, 43, 293-299.
(9) Horn, W.; Marutzky, R. A rapid pyrolytical method for the determination of wood preservatives in treated wood. Fresenius’ J. Anal. Chem. 1994, 348, 832-835. (10) Meier, D.; Ollesch, T.; Faix, O. In Progress in Thermochemical Biomass Conversion, 5th Conference, Tyrol, Austria, Sept 1722, 2000; Vol. 2, pp 1405-1416. (11) Audino, P. G.; Licastro, S. A.; Zerba, E. Thermal decomposition and isomerization of cis-permethrin and β-cypermethrin in the solid phase. Pest Manage. Sci. 2001, 58, 183-189. (12) Stanmore, B. R. The formation of dioxins in combustion systems. Combust. Flame 2004, 136, 398-427. (13) Lavric, E. D.; Konnov, A. A.; Ruyck, J. D. Dioxin levels in wood combustionsa review. Biomass Bioenergy 2004, 26, 115-145. (14) Specification for preservative treatmentsssawn and round timber. Australian Standard, AS 1604.1-2005, 2005. (15) Rappe, C.; Marklund, S. Formation of polychlorinated dibenzop-dioxins (PCDDs) and dibenzofurans (PCDFs) by burning or heating chlorophenates. Chemosphere 1978, 7, 269-281. (16) Qian, Y.; Zheng, M.; Liu, W.; Ma, X.; Zhang, B. Influence of metal oxides on PCDD/fs formation from pentachlorophenol. Chemosphere 2005, 60, 951-958. (17) Zheng, M.; Bao, Z.; Xu, X.; Wang, K. Formation of polychlorinated dibenzo-p-dioxins and dibenzofurans from the pyrolysis of hexachlorocyclohexane in the presence of iron(III) oxide. Chemosphere 1996, 32, 595. (18) Ryu, J.-Y.; Mulholland, J. A.; Chu, B. Chlorination of dibenzofuran and dibenzo-p-dioxin vapor by copper(II) chloride. Chemosphere 2003, 51, 1031-1039. (19) Sommeling, P. M.; Mulder, P.; Louw, R. Formation of PCDFs during chlorination and oxidation of chlorobenzene in chlorine/ oxygen mixtures around 340 °C. Proceedings of the 13th International Symposium on Chlorinated Dioxins and Related Compounds, Sept 20-24, 1993; Chemosphere 1994, 29, 2015. (20) Born, J. G. P.; Louw, R.; Mulder, P. Formation of dibenzodioxins and dibenzofurans in homogenous gas-phase reactions of phenols. Chemosphere 1989, 19, 401-406. (21) De Jong, V.; Cieplik, M. K.; Louw, R. Formation of dioxins in the catalytic combustion of chlorobenzene and a micropollutantlike mixture on Pt/γ-Al2O3. Environ. Sci. Technol. 2004, 38, 52175223. (22) Lago, R. M.; Green, M. L. H.; Tsang, S. C.; Odlyha, M. Catalytic decomposition of chlorinated organics in air by copper chloride based catalysts. Appl. Catal., B 1996, 8, 107-121. (23) van den Brink, R. W.; Louw, R.; Mulder, P. Formation of polychlorinated benzenes during the catalytic combustion of chlorobenzene using a Pt/γ-Al2O3 catalyst. Appl. Catal., 1998, 16, 219-226. (24) Altarawneh, M.; Ziolkowski, A.; Kennedy, E. M.; Dlugogorski, B. Z.; Mackie, J. C. Theoretical and experimental study of pyrolysis of organochlorine pesticide permethrin: Precursor of polychlorinated dibenzo-p-dioxin and dibenzofurans. Chemosphere, submitted for publication. (25) Wiater, I.; Louw, R. Reactions of diphenyl ether with chlorine and bromine atoms around 750 Ksrelevance for gas-phase “dioxin” formation. Eur. J. Org. Chem. 1999, 261-265. (26) Wilhelm, J.; Stieglitz, L.; Dinjus, E.; Will, R. Mechanistic studies on the role of PAHs and related compounds in PCDD/F formation on model fly ashes. Chemosphere 2001, 42, 797-802. (27) Ross, B. J.; Naikwadi, K. P.; Karasek, F. W. Effect of temperature, carrier gas, and precursor structure on PCDD and PCDF formed from precursors by catalytic activity of MSW incinerator fly ash. Chemosphere 1989, 19, 291-298. (28) Lindahl, R.; Rappe, C.; Buser, H. R. Formation of polychlorinated dibenzofurans (PCDFs) and polychlorinated dibenzo-p-dioxins (PCDDs) from the pyrolysis of polychlorinated diphenyl ethers. Chemosphere 1980, 9, 351-361. (29) Altarawneh, M.; A.; Dlugogorski, B. Z.; Kennedy, E. M.; Mackie, J. C. Quantum chemical investigation of formation of polychlorodibenzo-p-dioxins and dibenzofurans from oxidation and pyrolysis of 2-chlorophenol. J. Phys. Chem. A 2007, 111, 25632573. (30) Altarawneh M., A.; Dlugogorski, B. Z.; Kennedy, E. M.; Mackie, J. C. Theoretical study of reaction pathways of dibenzofuran and dibenzo-p-dioxin under reducing conditions. J. Phys. Chem. A 2007, 111, 7133-7140. (31) Lomnicki, S.; Dellinger, B. Formation of PCDD/F from the pyrolysis of 2-chlorophenol on the surface of dispersed copper oxide particles. Proc. Combust. Inst. 2002, 29, 2463-2468. (32) Lomnicki, S.; Dellinger, B. J. A detailed mechanism of the surfacemediated formation of PCDD/F from the oxidation of 2-chlorophenol on CuO/silica surface. J. Phys. Chem. A 2003, 107, 4387-4395. VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6431
(33) Khachatryan, L.; Burcat, A.; Dellinger, B. An elementary reactionkinetic model for the gas-phase formation of 1,3,6,8- and 1,3,7,9tetrachlorinated dibenzo-p-dioxins from 2,4,6-trichlorophenol. Combust. Flame 2003, 132, 406-421. (34) Yamamoto, T.; Inoue, S.; Sawachi, M. Post furnace formation and progressive chlorination of PCDD and PCDF in municipal waste incinerator. Chemosphere 1989, 19, 271-276. (35) Sidhu, S. S.; Maqsud, L.; Dellinger, B.; Mascolo, G. The homogeneous, gas-phase formation of chlorinated and brominated dibenzo-p-dioxin from 2,4,6-trichloro- and 2,4,6tribromophenols. Combust. Flame 1995, 100, 11-20. (36) Ryu, J.-Y.; Mulholland, J. A.; Kim, D.; Takeuchi, M. Characteristic homologue and isomer patterns of polychlorinated dibenzofurans from phenol precursors. Organohalogen Compd. 2004, 66, 1050-1055.
6432
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007
(37) Ryu, J.-Y.; Mulholland, J. A. Metal-mediated chlorinated dibenzop-dioxin (CDD) and dibenzofuran (CDF) formation from phenols. Chemosphere 2005, 58, 977-988. (38) Zoller, W.; Ballschmiter, K. Formation of polychlorinated dibenzodioxins and dibenzofurans by heating chlorophenols and chlorophenates at various temperatures. Fresenius’ J. Anal. Chem. 1986, 323, 19-23.
Received for review February 16, 2007. Revised manuscript received July 13, 2007. Accepted July 17, 2007. ES0703980
