Mechanisms of Dioxin Formation from the High-Temperature

Technol. , 2005, 39 (1), pp 122–127. DOI: 10.1021/es049355z .... Environmental Science & Technology 0 (proofing), ..... Procedia Engineering 2013 62...
0 downloads 0 Views 151KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 122-127

Mechanisms of Dioxin Formation from the High-Temperature Oxidation of 2-Chlorophenol CATHERINE S. EVANS AND BARRY DELLINGER* Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803

The homogeneous, gas-phase oxidative thermal degradation of 2-chlorophenol was studied in a 1 cm i.d., fused silica flow reactor at a concentration of 88 ppm, reaction time of 2.0 s, over a temperature range of 300 to 1000 °C. Observed products in order of yield were as follows: 4,6dichlorodibenzofuran (4,6-DCDF) > dibenzo-p-dioxin (DD) > 1-monochlorodibenzo-p-dioxin (1-MCDD), 4-chlorodibenzofuran (4-MCDF), dibenzofuran (DF), naphthalene, chloronaphthalene, 2,4-dichlorophenol, 2,6-dichlorophenol, phenol, chlorobenzene, and benzene. In contrast to pyrolysis, 4,6-DCDF is the major product rather than DD, and 1-MCDD and naphthalene are formed at temperatures as low as 400 °C. Under oxidative conditions, •OH and Cl• are the major carriers, which favors 4,6-DCDF formation over DD or 1-MCDD through abstraction of H• through diketo- and ether- intermediates. It is proposed that below 500 °C, unimolecular tautomerization/HCl elimination and CO elimination/isomerization reactions result in the formation of 1-MCDD and naphthalene, respectively.

products (13-16). In our recently published study of the high-temperature pyrolysis of 2-chlorophenol, we reported that formation of PCDDs (dibenzo-p-dioxin (DD) and 1-monochlorodibenzo-p-dioxin (1-MCDD)) was favored over the formation of PCDFs (4,6-dichlorodibenzofuran (4,6DCDF)) (17). These studies suggest that molecular oxygen and oxygen containing radicals, i.e., hydroxyl radical, are playing a large role in the product distribution and PCDD/F ratio. In this article, we report the thermal degradation of 2-chlorophenol under oxidative conditions for a reaction time of 2.0 s over the temperature range of 300 to 1000 °C and compare these results to the results of our previous study of the pyrolysis of 2-chlorophenol under otherwise similar conditions (17). Reaction pathways to PCDD/F products are proposed that are consistent with the experimental data. Pathways to non-PCDD/F products are also analyzed, and the intermediates in these pathways are discussed in their relation to formation of PCDD/F and other pollutants. The effect of molecular oxygen on the PCDD to PCDF ratio is explained based on these reaction pathways.

Experimental Section

It has recently been recognized that the contribution of gasphase pathways to the formation of polychlorinated dibenzop-dioxins and dibenzofurans (PCDD/F) have probably been underestimated (1, 2). Chlorinated phenols are known to form PCDD/F via gas-phase pathways through reactions of phenoxyl radicals (3-7). In fact, it was the overestimation of the rate of consumption of phenoxyl radical through reaction with molecular oxygen that led to the presumption that gasphase pathways were too slow to contribute significantly to PCDD/F emissions from full-scale combustors (1, 2, 8). In the original mechanistic and reaction kinetic models, it was presumed that chlorinated phenols form only PCDDs through radical-molecule reactions involving phenoxyl radicals and phenols (8-10). Subsequently, it has been suggested that chlorinated phenols also form PCDFs as well as PCDDs through radical-radical reactions of phenoxyls (8, 11). The first experimental studies of the high-temperature, gas-phase reactions of 2,4,6-trichlorophenol (T ) 300-800 °C, reaction time ) 2.0 s) reported the formation of PCDDs without PCDF (12). Subsequent studies of the thermal oxidation of 2,4,6-trichlorophenol in the presence of hexane under similar reaction conditions resulted in the formation of PCDFs in addition to PCDDs. Other studies of the oxidation of chlorophenols (i.e. 2-chlorophenol, 3-chlorophenol, 4-chlorophenol, 2,4-dichlorophenol and pentachlorophenol) under “slow combustion conditions” (i.e. T ) 300-600 °C, reaction times between 10 s and 10 min) reported PCDFs as the major

All experiments were performed using a high-temperature flow reactor system referred to in the archival literature as the System for Thermal Diagnostic Studies (STDS). The detailed design has been published elsewhere (18). In short, the STDS consists of a high-temperature, 1 cm i.d. fused silica flow reactor equipped with an in-line Varian Saturn 2000 GC/MS. Previous work found that an inside diameter of 1 cm was acceptable in reducing the wall effects (19). The flow reactor is housed inside a furnace located inside a Varian GC where the temperatures surrounding the reactor are controlled. Pressure inside the reactor is also maintained at 1.00 ( 0.15 atm. Gas-phase products are cryogenically trapped at the head of the GC column in preparation for chemical analysis. To maintain a constant concentration of 88 ppm, 2-monochlorophenol (2-MCP) (Aldrich) was injected into a 20% O2 in helium gas stream by a syringe pump through a vaporizer maintained at 280 °C. Gas-phase samples of 2-MCP then were swept by the 20% O2 in helium flow through heated transfer lines (300 °C) into a 35 cm long, 1.0 cm i.d. fused silica tubular flow reactor where the temperature was parametrically varied between 300 and 1000 °C. The 20% O2 in helium flow rate was varied with temperature so that the residence time inside the reactor was held at 2.0 s. The unreacted 2-MCP and the thermal degradation products were then swept through a heated transfer line to another Varian GC where they were cryogenically trapped at the head of a CP-Sil 8 phase capillary column (30 m, 0.25 mm i.d., 0.25 µm film thickness). To separate the individual reaction products, the column was temperature programmed from -60 to 300 °C at 15 °C/min. Detection and quantification of the products were obtained using a Varian Saturn Mass Spectrometer, which was operated in the full-scan mode (40 to 650 amu) for the duration of the GC run. The length of each experimental run was approximately 45 min. Product concentrations were calculated based on the calibrations with standards of the products (Aldrich and Cambridge Isotope Lab) and the peak area counts from the chromatogram. The yields of the products were calculated using the expression

* Corresponding author phone: (225)578-6759; fax: (225)578-3458; e-mail: [email protected].

Y ) {([Product])/[2-MCP]0}*100

Introduction

122

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 1, 2005

10.1021/es049355z CCC: $30.25

 2005 American Chemical Society Published on Web 11/24/2004

FIGURE 1. “Dioxin” products from the gas-phase oxidation of 2-MCP. [2-MCP]o ) 88 ppm in helium. Gas-phase reaction time of 2.0 s. where [Product] is the concentration of the particular product formed (in moles) and [2-MCP]0 is the initial concentration of 2-MCP (in moles) injected into the reactor. (Please note that in our previous publication on the pyrolysis of 2-MCP (17) the concentration of the chlorinated products that contained two phenyl rings was multiplied by two to account for the fact that the maximum yield for total PCDD/F from chlorinated phenols can be no greater than 50%.) These observed oxidation products were dibenzo-p-dioxin, 4,6dichlorodibenzofuran, 1-chlorodibenzo-p-dioxin, dibenzofuran, naphthalene, chloronaphthalene, and diphenylethyne. Multiple runs were performed for each temperature to ensure the repeatability of the experiments. Once the experimental procedure was fully developed, the repeatability of the experiments was within 10%. The heats of reaction, ∆Hrxn, for key steps in product formation pathways were calculated using AM1, semiempirical molecular orbital formalism. The calculations were performed using the MOPAC computation program that is contained within the ChemBats3D Pro computer application. Without experimental benchmarks, the calculated ∆Hrxn cannot be considered to be completely accurate. They are shown to assess the likelihood of potential parallel pathways. Pseudoequilibrium calculations were performed to estimate the concentrations of reactive species such as •OH, O•, H•, and Cl•. The Chemkin Equil code was used to calculate the concentrations of these species over a range of reaction temperature from 570 to 1270 K. The initial inputs for the calculations were the same as the experimental runs in that the initial 2-MCP concentration was held at 88 ppm and the initial O2 concentration was held at 20%. Over this temperature range, the major products species resulting in the calculation were CO2, H2O, HCl, and Cl2. Other species included were •OH, O•, H•, H2, HO2, H2O2, CO, and Cl•.

Results The temperature dependence of the oxidative thermal degradation of 2-MCP and the yield of “dioxin” products are presented in Figure 1 and Table 1 for a reaction time of 2.0 s. The nondioxin products are presented in Table 1 and Figure 2. Figures 1 and 2 are presented as the percent yields of products (or percent of unconverted 2-MCP) on a logarithmic scale versus temperature. The thermal degradation of 2-MCP initially increased gradually from 400 to 650 °C, where the rate drastically accelerated, and achieving 99% destruction at 850 °C. The anticipated PCDD/F products, dibenzo-p-dioxin (DD), 1-chlorodibenzo-p-dioxin (1-MCDD), and 4,6-dichlorodibenzofuran (4,6-DCDF), were all observed for the oxidation of 2-MCP (cf. Figure 1 and Table 1). 4-Chlorodibenzofuran (4-MCDF) and dibenzofuran (DF) were additionally

detected PCDD/F products. Initially at 400 °C, 1-MCDD was the only PCDD/F product observed, achieving a maximum yield of 0.92% at 600 °C. DD and 4,6-DCDF were detected between 500 and 900 °C with maximum yields of 1.32 and 9.4%, respectively. The other PCDD/F products, 4-MCDF and DF, were not observed until 650 °C where they achieved maximum yields of 0.69 and 0.20% respectively. Above 900 °C no chlorinated aromatic products were detected. Nondioxin products were also detected from the oxidation of 2-MCP. (cf. Figure 2 and Table 1). Chlorinated products included chlorobenzene (maximum yield of 0.03% at 650 °C), 1-chloronaphthalene (maximum yield of 0.10% at 650 °C), 2,4-dichlorophenol, and 2,6-dichlorophenol (maximum yields of 0.28 and 0.59, respectively at 600 °C). Nonchlorinated products included naphthalene, benzene, and phenol. Naphthalene is the only non-PCDD/F product observed at 400 °C and achieves a maximum yield of 0.22% at 700 °C. However the percent yield of naphthalene remains reasonably constant between 400 °C and 900 °C. Benzene and phenol were observed between 600 and 700 °C and achieved maximum yields at 650 °C.

Discussion The formation of dioxin products (DD, 1-MCDD, and 4,6DCDF) indicates that stable phenoxyl radicals are formed in significant yields through loss of the hydroxyl-hydrogen. The formation of aromatics (phenol, chlorobenzene, and benzene) indicates that simple substitution reactions are occurring. The formation of 2,4-dichlorophenol and 2,6dichlorophenol indicate evidence of chlorination of 2-MCP. The formation of larger aromatic molecules at low temperatures (naphthalene and chloronaphthalene) is the result of reactions involving the release of CO from the phenoxyl radicals that will then recombine to form naphthalenes. 2-MCP Decomposition. The addition of oxidative destruction pathways with the addition of molecular oxygen results in the rapid decomposition of 2-MCP beginning at 650 °C rather than 750 °C, the temperature observed under pyrolytic conditions (17). The decomposition of 2-MCP can in principle be initiated by loss of the phenoxyl hydrogen by unimolecular, bimolecular, or possibly other low energy pathways (including heterogeneous reactions). Unimolecular decomposition of the oxygen-hydrogen bond (reaction 1) is rapid with a reported rate coefficient for phenol of k1(430-930 °C) ) 3.2 × 1015exp(-86 500/RT) s-1 (8, 20). The bimolecular reaction with O2 via reaction 1a is endothermic by 28 kcal/mol and is viable only as minor initiation reaction.

C6H4ClOH f C6H4ClO• + H• ∆Hrxn ) 78 kcal/mol (reaction 1) C6H4ClOH + O2 f C6H4ClO• + HO2• ∆Hrxn ) 28 kcal/mol (reaction 1a) Bimolecular propagation reactions under oxidative conditions include attack by H•, Cl•, •OH, and O•. In our previous paper on the thermal degradation of 2-MCP under pyrolytic conditions, the ∆Hrxn for reactions with H• and Cl• were discussed (17). It was determined that the most favorable reactions for generating chlorophenoxyl radicals were the abstraction of hydrogen by H• or Cl•. With the addition of O2, one can easily generate •OH and O• that can also abstract hydrogen via highly exothermic reactions (reactions 2 and 3).

C6H4ClOH + •OH f C6H4ClO• + H2O ∆Hrxn ) -42 kcal/mol (reaction 2) C6H4ClOH + O• f C6H4ClO• + •OH ∆Hrxn ) -25 kcal/mol (reaction 3) VOL. 39, NO. 1, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

123

TABLE 1. Percent Yield of Products of Gas-Phase Oxidation of 2-Chlorophenol temperature (°C) product

400

2-chlorophenol dibenzo-p-dioxin 1-chlorodibenzo-p-dioxin dibenzofuran 4-chlorodibenzofuran 4,6-dichlorodibenzofuran naphthalene 1-chloronaphthalene phenol 2,4-dichlorophenol 2,6-dichlorophenol benzene chlorobenzene a

98 0.12

450

500

550

600

650

700

750

800

850

900

82.1

62.7

51.6 1.32 0.92

9.34 0.09 0.16

5.74 0.05 0.09

3.52

0.36

0.05

0.13

0.34 0.16

9.43 0.16

31.8 0.49 0.44 0.20 0.69 3.62 0.18 0.10 0.81 0.21 0.39 0.04 0.03

3.99

0.16

57.2 0.23 0.29

0.27 1.23 0.22

0.13 0.69 0.11

0.24 0.08

0.18 0.08

0.07 0.07

0.07 0.21 0.01 0.02

0.05 0.02

0.14

0.14

0.21 0.28 0.59 0.004

0.005

0.006

Percent yield ) {([Product])/[2-MCP]0}*100.

SCHEME 1. Reaction Mechanism for the Formation of 2,4-Dichlorophenol and 2,6-Dichlorophenol from the 2-Chlorophenoxyl Radical

FIGURE 2. “Nondioxin” products from the gas-phase oxidation of 2-MCP. [2-MCP]o ) 88 ppm in helium. Gas-phase reaction time of 2.0 s. Rate coefficients based on analogous reactions with phenol for reactions 2 and 3 are k2 (1000-1150 K) ) 6.0 × 1012 cm3/mol/s and k3 (340-870 K) ) 1.28 × 1013exp(-2900/ RT) cm3/mol/s (21). Based on pseudoequilibrium calculations of the concentrations of reactive radicals and their rate coefficients for reaction with phenol, the rate of reaction 2 is 190× faster than reaction 3 and a factor of 30 faster than reaction 1. Thus, reaction 2 is one of the dominant sources of chlorophenoxyl radical under oxidative conditions. Our pseudoequilibrium calculations also indicate that with the addition of oxygen, there is an increase in the concentration of Cl•. This can be explained by reaction 4 in which •OH abstracts hydrogen from HCl to form Cl•.

•OH + HCl f H2O + Cl•

(reaction 4)

Reaction 4 has a rate coefficient of k4 (300-700 K) ) 1.77 × 1012exp(-0.89 kcal/RT) cm3/mol/s (21). With the increase in the concentration of Cl•, the abstraction of hydrogen by Cl• with a rate coefficient for phenol of k (25 °C) ) 1.43 × 1014 cm3/mol/s continues to be the dominant source for the formation of chlorophenoxyl radicals (21). In fact based on our pseudoequilibrium calculations at 700 °C, the rate for the abstraction of hydrogen by Cl• is 5000 times faster than the rate for reaction 2. We have previously shown that above 725 °C, the purely gas-phase reactions of 2,4,6-trichlorophenol (2,4,6-TCP) can account for its decomposition behavior as well as observed PCDD/F products (1). However the initiation of the decomposition of 2,4,6-TCP and 2-MCP at temperatures as low as 400 °C requires us to consider other possible low energy 124

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 1, 2005

initiation pathways. Heterogeneous wall reactions may result in a wall collision assisted reaction analogous to reaction 1. Trace impurities may also result in low temperature decomposition. The potential contribution of these reactions have been previously analyzed for the oxidation of 2,4,6trichlorophenol and been shown to not significantly impact subsequent propagation reactions (2). Formation of Phenol, Chlorobenzene, Benzene, 2,4Dichlorophenol, and 2,6-Dichlorophenol. The formation of phenol is likely due to the exothermic displacement of Cl• by H• (∆Hrxn ) -17 kcal/mol). Chlorobenzene and benzene are formed with much lower yields and at slightly higher temperatures than phenol. These lower yields are due to the more endothermic displacement of •OH from 2-MCP by hydrogen to form chlorobenzene (∆Hrxn ) 0 kcal/mol) and the displacement of •OH from phenol by hydrogen to form benzene (∆Hrxn ) 4 kcal/mol). The temperature ranges of formation of all these products are lower than for previous results of 2-MCP under pyrolytic conditions (17). This is due to the early onset of reaction of 2-MCP under oxidative conditions and the oxidation of the products at higher temperatures. 2,4-Dichlorophenol and 2,6-dichlorophenol are produced from the chlorination of the 2-MCP. Since displacement of H• by Cl• is endothermic, the direct reaction of Cl• with 2-MCP is unlikely. The formation of dichlorophenol is instead due to recombination of 2-chlorophenoxyl radicals and Cl•. Scheme 1 depicts the formation of 2,4-dichlorophenol and 2,6-dichlorophenol by Cl• attack at the resonance-stabilized, ortho- or para- carbon sites of the chlorophenoxyl radicals (∆Hrxn ) -43 kcal/mol). Subsequent tautomerization results in the formation of the respective dichlorophenols (∆Hrxn ) -20 kcal/mol) (22). Under pyrolytic conditions, HCl is a largely unreactive sink for chlorine. Under oxidative condi-

SCHEME 2. Postulated Pathways for the Formation of DD, 1-MCDD, and 4,6-DCDF

tions, •OH and HO2• can react exothermically with HCl to produce a relative abundance of Cl• (23, 24). Consequently, chlorination reactions are enhanced under oxidative conditions (25-27). Formation of Naphthalene and Chloronaphthalene. Formation of polycyclics such as naphthalene and chloronaphthalene has been traditionally ascribed to molecular growth pathways involving largely C2 fragments (28-30). However, the low-temperature onset of formation of naphthalene (400 °C) suggests a pathway that does not require the complete fragmentation of 2-MCP. The recombination of two cyclopentadienyl radicals has been previously shown to be a favorable pathway for formation of naphthalene where the reported ∆Hrxn for this overall reaction is 9.23 kcal/mol (31, 32). In our previous work on the pyrolysis of 2-MCP, we presented a reasonable pathway for the formation of naphthalene from the 2-chlorophenoxyl radical through elimination of CO to form a cyclopentadienyl radical with a calculated ∆Hrxn of 20 kcal/mol and an assigned rate coefficient of k6 (730-1300 °C) ) 1 × 1011.4exp(-22100 ( 450/T) s-1 (17, 20). Another possible low-temperature route to the formation of naphthalene is described in Scheme 2 as a competitive pathway to the formation of 4,6-DCDF (33).

The yields of naphthalene and chloronaphthalene are significantly lower under oxidative than pyrolytic conditions, most probably due to the more rapid rate of oxidation of the cyclopentadienyl radical (17). The oxidation rate of naphthalene is also increased. In addition, the yield of PCDD/F is expected to increase quadratically with the increase in chlorinated phenoxyl radical concentration. PCDD/F formation competes with the formation of naphthalene by elimination of CO to cyclopentadienyl radical (vide infra), and the concentration of naphthalene never becomes a major product as it did under pyrolytic conditions. Formation of Dibenzo-p-dioxin, 1-Chlorodibenzo-pdioxin, 4,6-Dichlorodibenzofuran, 4-Chlorodibenzofuran, and Dibenzofuran. Scheme 2 summarizes previously identified reaction pathways to DD, 1-MCDD, and 4,6-DCDF from the reaction of the different mesomers of 2-chlorophenoxyl radical. Pathway IA in Scheme 2 depicts the mechanism for DD formation; the oxygen-centered radical mesomer recombines with the carbon (chlorine substituted) centered radical mesomer to form a keto-ether. Following the abstraction of chlorine by H•, Cl•, or •OH, DD is formed by intra-annular elimination of Cl•. Another possible pathway for the formation VOL. 39, NO. 1, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

125

SCHEME 3. Proposed Pathways for the Formation of 4-MCDF

of DD is through a radical-molecule reaction, IB, shown in parentheses below the radical-radical pathway in Scheme 2. This reaction depicts the oxygen centered radical mesomer reacting with 2-MCP via Cl• displacement to form a chlorohydroxy diphenyl ether (HDE) followed by the abstraction of H• by •OH. Finally, DD is formed by intra-annular displacement of Cl•. It has been previously suggested that this radical-molecule reaction is too slow to account for the observed yields of the DD (1, 2, 11). Formation of 1-MCDD, shown as pathway IIA in Scheme 2, is initiated by the recombination of the oxygen-centered radical mesomer and the carbon-hydrogen- centered radical mesomer to form a keto-ether. Following loss of hydrogen to form the phenoxyl diphenyl ether (PDE), ring closure to form 1-MCDD occurs through intra-annular displacement of Cl•. Pathway IIB depicts an alternate pathway in which ring closure is by unimolecular elimination of HCl. Pathways IIIA and IIIB in Scheme 2 depicts the proposed pathway of 4,6-DCDF formation and the competing pathway of naphthalene formation. Initially, for both pathways, two carbon-hydrogen centered radical mesomers react to form a common diketo-dimer intermediate. The dimer can then follow pathway IIIA by the abstraction of hydrogen by •OH or Cl• and undergo tautomerization followed by the displacement of hydroxyl to form 4,6-DCDF. In pathway IIIB, the initially formed diketo-dimer possibly forms naphthalene by elimination of two CO molecules followed by a unimolecular isomerization and elimination of two Cl• (33). We believe that DF is simply a recombination product of unchlorinated phenoxyl radical formed from decomposition of phenol (34). The reaction proceeds by mechanisms analogous to those shown for formation of 4,6-DCDF in pathways IIIA in Scheme 2. Scheme 3 depicts reasonable pathways for the formation of 4-MCDF, which was observed under oxidative but not pyrolytic conditions. Two pathways are depicted: (1) the recombination of the carbon (hydrogen)-centered radical mesomer with the carbon (chlorine)-centered radical mesomer and (2) the recombination of the carbon (hydrogen)centered radical mesomer with an unchlorinated carboncentered mesomer to form a diketo-dimer. For the first pathway, H• abstracts chlorine and, in the second, •OH abstracts another hydrogen. Both pathways then undergo tautomerization followed by displacement of hydroxyl to form 4-MCDF. Based on comparisons to simple alkane reactions for the abstraction of chlorine by H• and the abstraction of hydrogen by •OH, the lower pathway is a much more favorable route for 4-MCDF formation under oxidative conditions, while the upper pathway of Cl• abstraction by H• is more favorable under pyrolytic conditions (35-37). Oxidation versus Pyrolysis of 2-MCP. Four major differences were observed in the results of oxidation and pyrolysis of 2-MCP: 1. A decrease in non-PCDD/F products under oxidative conditions. 2. An increase in total PCDD/F product yield. 3. 4,6-DCDF is the major product under oxidation, while DD and 1-MCDD are the major products 126

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 1, 2005

for pyrolysis. 4. 1-MCDD is observed at temperatures as low as 400 °C under oxidative conditions. The first two observations are readily explained. The addition of oxygen to the systems results in oxidation of the decomposition products of 2-MCP before they can form the observed aromatic molecular growth products. The addition of oxygen also results in lower temperature formation of chlorophenoxyl radical where it can react to form PCDD/F products rather than decompose. The explanations of observations 3 and 4 are slightly more subtle. The maximum yields of DD under oxidative and pyrolytic conditions are 1.32 and 0.28%, respectively. The maximum yields of 1-MCDD are 0.92 and 0.05%. In contrast the maximum yield of 4,6-DCDF, which was not detected under pyrolytic conditions, was 9.43% under oxidative conditions. Thus DD only increased by a factor of 4.7, 1-MCDD increased by a factor of 18.4, while 4,6-DCDF increased at least 100× with the addition of oxygen to the system. Examination of the major pathways to DD, 1-MCDD, and 4,6-DCDF in Scheme 2 reveals that chlorine is abstracted to form DD, one hydrogen is abstracted to form 1-MCDD, and one of two hydrogens is abstracted to form 4,6-DCDF. H• is the major chain carrier in the 2-MCP system under pyrolytic conditions, while •OH and Cl• are the major carriers under oxidative conditions. The ∆Hrxn for abstraction of Cl• by •OH (pathway IA in Scheme 2) is endothermic by 45 kcal/mol. Thus, chlorine abstraction by •OH is not expected to be competitive with the abstraction of chlorine by H• (∆Hrxn ) -64 kcal/mol) and Cl• (∆Hrxn ) -19 kcal/mol). The addition of oxygen consequently has little effect on the yield of DD other than the increase in Cl• present that increases the concentration of chlorophenoxyl radicals. However in pathways IIA and IIIA, ∆Hrxn for the abstraction of hydrogen by •OH is exothermic by 16 kcal/mol. Thus, the addition of oxygen to the system increases the •OH and Cl• concentrations, which increases the rate of abstraction of hydrogen to form 1-MCDD and 4,6-DCDF. The increase in yield of 4,6DCDF due to an increase in •OH and Cl• concentrations is exponentially larger than for 1-MCDD because the hydrogen abstraction rate coefficient is doubled due to the presence of two sites for hydrogen abstraction in the diketo intermediate involved in 4,6-DCDF formation versus the “single” keto intermediate in formation of 1-MCDD. Formation of 1-MCDD and Naphthalene from 400 to 600 °C. Previous studies of the oxidation of 2-MCP over a wide temperature range and longer reaction times have also reported 4,6-DCDF as the major product (13-16). Only the study of Weber and Hagenmaier who studied lower temperature reactions (330-600 °C) reports 1-MCDD as a major product at a low temperature of 375 °C (15, 16) in addition to DD, 4,6-DCDF, and 4-MCDF. The lower formation temperature is likely due to the longer residence times of 10 min of their experiments. Our results also show a low temperature formation of 1-MCDD as well as naphthalene. The observation of naphthalene and 1-MCDD as the only organic products below 500 °C suggests (1) that a somewhat

different mechanism of formation may be operative at these lower temperatures for which the radical pool is not fully developed, and (2) their formation mechanisms have a common intermediate or are at least closely related. After the recombination of chlorinated phenoxyl radicals to form the diketo intermediate, the formation of naphthalene shown in pathway IIIB in Scheme 2 is unimolecular (33). This can explain the high yields at low temperatures before the radical pool has developed. The observation of significant yields of 1-MCDD at low temperature suggests that it may also be formed by a unimolecular pathway following the formation of the keto-ether intermediate via radical-radical recombination. Alternatively to the abstraction of hydrogen by •OH in pathway IIA in Scheme 2, a simple, intra-ring one proton tautomerization results in the formation of a hydroxydiphenyl ether intermediate that can then form 1-MCDD by inter-ring elimination of HCl. Above 500 °C, the radical pool increases rapidly, and bimolecular pathways involving hydrogen and chlorine abstraction begin to dominate the formation of 1-MCDD and other PCDD/F products. In summary, we have proposed reasonable mechanisms for the formation of each observed product of the oxidation of 2-MCP. We have also identified mechanistic rationales for the differences in product distribution and PCDD to PCDF branching ratios for oxidative versus pyrolytic conditions. Comparison of oxidation and pyrolysis results have also identified a possible lower temperature, primarily unimolecular routes to formation of naphthalene and 1-MCDD that can occur before the radical pool increases significantly at 600 °C. These results have implications in the mechanism of molecular growth and particle formation from resonance stabilized radicals such as cyclopentadienyl and phenoxyl radicals and their derivatives.

Acknowledgments We gratefully acknowledge the assistance of our colleagues, Dr. Lavrent Khachatryan and Alexander Burcat, in evaluation of the thermochemistry presented in this manuscript as well as helpful discussions concerning the mechanisms of dioxin formation. We acknowledge the partial support of this work under EPA contract 9C-R369-NAEX, EPA grant R828166, and the Patrick F. Taylor Chair foundation.

Literature Cited (1) Khachtrayan, L.; Burcat, A.; Dellinger, B. Combust. Flame 2003, 132, 406. (2) Khachtrayan, L.; Asatrayan, R.; Dellinger, B. Chemosphere 2003, 52, 695. (3) Karasek, F. W.; Dickson, L. C. Science 1987, 237, 754. (4) Dickson, L. C.; Lenoir, D.; Hutzinger, O. Environ. Sci. Technol. 1992, 26, 1822.

(5) Altwicker, E. R.; Milligan, M. S. Chemosphere 1993, 27, 301. (6) Addink, R.; Olie, K. Environ. Sci. Technol. 1995, 29, 1425. (7) Milligan, M. S.; Altwicker, E. R. Environ. Sci. Technol. 1996, 30, 225. (8) Shaub, W. M.; Tsang, W. Environ. Sci. Technol. 1983, 17, 721. (9) Bruce, K.; Beach, L. O.; Gullett, B. K. Waste Manage. 1983, 11, 97. (10) Altwicker, E. R. Sci. Total Environ. 1991, 104, 47. (11) Louw, R.; Ahonkhai, S. T. Chemosphere 2002, 46, 1273. (12) Sidhu, S. S.; Masqud, L.; Dellinger, B.; Mascolo, G. Combust. Sci. Technol. 1995, 100, 11. (13) Waiter-Protas, I.; Louw, R. Eur. J. Org. Chem. 2001, 3945. (14) Born, J. G. P.; Louw, R.; Mulder, P. Chemosphere 1989, 19, 401. (15) Weber, R.; Hagenmaier, H. Chemosphere 1999, 38, 529. (16) Weber, R.; Hagenmaier, H. Organohalogen Compd. 1997, 31, 480. (17) Evans, C. S.; Dellinger, B. Environ. Sci. Technol. 2003, 37, 1325. (18) Rubey, W. A.; Grant, R. A. Rev. Sci. Instrum. 1988, 59, 265. (19) Taylor, P. H.; Tirey, D. A.; Dellinger, B. Combust. Flame 1996, 106, 1. (20) Colussi, A.; Zabel, F.; Benson, S. W. Int. J. Chem. Kinet. 1977, 9, 161. (21) NIST Chemical Kinetics Database 17, Gaithersburg, MD, 1998. (22) Capponi, M.; Gut, I.; Hellrum, B.; Persy, G.; Wirz, J. Can. J. Chem. 1999, 77, 605. (23) Soderstrom, G.; Marklund, S. Environ. Sci. Technol. 2002, 36, 1959. (24) Wilkstrom, E.; Ryan, S.; Touati, A.; Telfer, M.; Gullett, B. K. Environ. Sci. Technol. 2003, 37, 1108. (25) Procaccini, C.; Bozzelli, J. W.; Longwell, J. P.; Sarofim, A. F.; Smith, K. A. Environ. Sci. Technol. 2003, 37, 1684. (26) Huang, J.; Onal, I.; Senkan, S. M. Environ. Sci. Technol. 1997, 31, 1372. (27) Satyapal, S.; Werner, J. H.; Cool, T. A. Combust. Sci. Technol. 1995, 106, 229. (28) Appel, J.; Bockhorn, H.; Frenklach, M. Combust. Flame 2000, 121, 122. (29) Richter, H.; Howard, J. B. Prog. Energy Combust. Sci. 2000, 26, 4. (30) Richter, H.; Mazyer, O. A.; Sumathi, R.; Green, W. H.; Howard, J. A.; Bozzelli, J. W. J. Phys. Chem. A 2001, 105(9), 1561. (31) Melius, C. F.; Calvin, M.; Marinov, N. M.; Ritz, W. J.; Senkan, S. M. 26th Symp. (Intern.) on Combustion, The Combustion Inst. 1996, 685. (32) Lu, M.; Muholland, J. A. Chemosphere 2001, 42, 623. (33) Kim, D. H.; Mulholland, J. A.; Ryu, J.-Y. 26th Symp. (Intern.) on Combustion, The Combustion Inst. 2004, in press. (34) Waiter, I.; Born, J. G. P.; Louw, R. Eur. J. Org. Chem. 2000, 921. (35) Green, J. H. S.; Maccoll, A. J. Chem. Soc. 1955, 2449. (36) Price, S.; Shaw, R.; Trotman-Dicksenson, A. J. Chem. Soc. 1956, 3855. (37) Kale, M.; Maccoll, A. J. Chem. Soc. 1957, 5020.

Received for review April 28, 2004. Revised manuscript received August 7, 2004. Accepted October 8, 2004. ES049355Z

VOL. 39, NO. 1, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

127