Photodechlorination of octachlorodibenzothiophene and octachlorodibenzofuran: Comparison of experimental degradation pathways with degradation pathways predicted by DFT

Chemosphere 73 (2008) 1005–1010

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Photodechlorination of octachlorodibenzothiophene and octachlorodibenzofuran: Comparison of experimental degradation pathways with degradation pathways predicted by DFT Shingo Yamada a, Saeko Kishita a, Satoshi Nakai b, Makoto Takada a, Masaaki Hosomi a,* a b

Department of Chemical Engineering, Graduate School of Engineering, Tokyo University of Agriculture and Technology, 2-24-16, Naka-cho, Koganei, Tokyo 184-8588, Japan Cluster III, Faculty of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8527, Japan

a r t i c l e

i n f o

Article history: Received 5 February 2008 Received in revised form 19 April 2008 Accepted 15 May 2008 Available online 26 July 2008 Keywords: Polychlorodibenzothiophenes Polychlorodibenzofurans Density functional theoretical method Gaussian Prediction of degradation pathways

a b s t r a c t Polychlorodibenzothiophenes (PCDTs) are sulfur analogues of polychlorodibenzofurans (PCDFs) and have been detected in environmental samples. We used density functional theory calculations (Gaussian 98W) to predict the photodechlorination pathways of octachlorodibenzothiophene (OCDT) and octachlorodibenzofuran (OCDF) in hexane, and we compared the predicted pathways with those observed during UV irradiation experiments. OCDT and OCDF were observed to degrade through first-order dechlorination processes, and the rate constant for OCDT was less than one-third that for OCDF. The main experimental photodechlorination pathways of OCDT and OCDF led to hexachlorinated and tetrachlorinated congeners, respectively; that is, the photodechlorination pathway of OCDT differed from that of OCDF. On the assumption that the dechlorination mechanisms involved radical reactions, we used DFT calculations to estimate bond-dissociation energies and single-point energies of OCDT and OCDF and their dechlorinated congeners, and we used the resulting information, along with hypotheses regarding the rate-controlling step of the degradations, to predict theoretical degradation pathways. We propose that reaction of dechlorinated radicals with a hydrogen donor was the rate-controlling step for OCDT and that C–Cl bond dissociation by UV light was the rate-controlling step for OCDF. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Polychlorodibenzothiophenes (PCDTs) are sulfur analogues of polychlorodibenzofurans (PCDFs). The structural similarity of these two groups of molecules suggests that, like PCDFs, PCDTs may disrupt the endocrine system, and in fact they have been shown to bind to the aryl hydrocarbon receptor (Nakai et al., 2004). Like PCDFs, PCDTs are generated through the burning of waste, and PCDTs have been detected in soil, dust, and river sediments (Sinkkonen, 1997; Wiedmann et al., 1997; Claus et al., 1998; Nakai et al., 2004; Nakai et al., 2007). Therefore, the long-term adverse environmental effects of PCDTs are of concern, and information about the persistence and degradation processes of PCDTs in the environment is required. Photodegradation is the most important degradation process for dioxins in the environment (Neidhard et al., 1987), and this process is likely to be similarly important for PCDTs. The photodegradation of PCDFs and PCDTs involves dechlorination (Nakai et al., 2004; Nakai et al., 2007), although the fate of the degradation products has not been clarified. PCDF congeners with chlorines at * Corresponding author. Tel.: +81 42 388 7070; fax: +81 42 388 7693. E-mail address: [email protected] (M. Hosomi). 0045-6535/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2008.05.057

the 2, 3, 7, and 8 positions are highly toxic. If the same is true of PCDTs, it is important to determine whether any of these congeners are produced during the dechlorination of PCDTs and to determine their fate in the environment. Researchers have made rapid progress in using computational methods such as the density functional theory (DFT) method to calculate bond energies and other parameters necessary for the prediction of reaction pathways under particular conditions. The DFT method has been used to analyze the dechlorination of chlorinated aromatic compounds such as dioxins (Berkout et al., 1999; Fueno et al., 2002; Arulmozhiraja and Morita, 2004), but it has not been applied to PCDTs. We have used the method to analyze the photodegradation of pentachlorophenol (Suegara et al., 2005), hexachlorobenzene (Yamada et al., 2007a), and chlordane (Yamada et al., 2007b). In this work, we determined the experimental photodechlorination pathways of octachlorodibenzothiophene (OCDT) and octachlorodibenzofuran (OCDF), and we compared the experimental pathways with theoretical pathways calculated by means of the DFT method. n-Hexane solutions of OCDT and OCDF were irradiated with UV light, and the degradation rates and pathways of the two compounds were compared. The theoretical dechlorination pathways were estimated by DFT calculation of the C–Cl bond

1006

S. Yamada et al. / Chemosphere 73 (2008) 1005–1010

dissociation energies of the two compounds and calculation of the single-point energies of their radical intermediates.

3. Results and discussion 3.1. Photolysis

2. Materials and methods 2.1. Materials OCDT were purchased from EnBioTec Laboratories (Tokyo, Japan) and Cambridge Isotope Laboratories (Andover, MA, USA), respectively. For identification of the degradation products, standard samples of 1,2,3,4,7,8,9-HpCDT and 1,2,3,7,8,9-HxCDT were purchased from the same companies. n-Hexane was purchased from Kanto Chemical (Tokyo, Japan). 2.2. Photolysis n-Hexane solutions of OCDT (100 ng ll1) were injected into a 4-cm3 quartz cell (S11-UV-10, GL Sciences, Tokyo, Japan) and irradiated with 254-nm UV light from a 1.7-W low-pressure mercury lamp (Germipak UV Cell GCL212, LightSources Inc., Orange, CT, USA). The distance between the cell and the UV light source was 5 mm. The intensity of the UV lamp was 5.0 mW cm2. After irradiation, the hexane was replaced with toluene and samples were adjusted to toluene for analysis by gas chromatography/mass spectrometry. Each sample (1 ll) was analyzed by lowresolution gas chromatography/mass spectrometry (LR-GC/MS; HP 5973 MS detector, Agilent, US) or high-resolution gas chromatography/mass spectrometry (HR-GC/MS; JMS-700, JEOL Ltd., Tokyo, Japan). Quantitative analysis of the residual OCDT was carried out in the selected-ion-monitoring (SIM) mode, and qualitative analysis of the degradation products was carried out in the scan mode. The monitoring ions used for the SIM mode analysis of the PCDTs had previously been determined on the basis of the scan-mode analysis of PCDT standards. PCDTs were identified by comparing retention times of the components of the irradiated samples with the retention times of PCDT standards. 2.3. Quantum chemical calculations Gaussian 98W (Frisch et al., 2001), a non-empirical molecular orbital calculation package that uses density functional theory (B3LYP functional, 6-311G basis set), was used for the quantum chemical calculations (Suegara et al., 2005; Yamada et al., 2007a; Yamada et al., 2007b). Solvent effects were considered by inclusion of a dielectric constant as a calculation parameter. Photodechlorination of dioxins proceeds through a radical mechanism (Fueno et al., 2002; Arulmozhiraja and Morita, 2004), so we assumed that photodechlorination of OCDT and OCDF proceeded through a process involving C–Cl bond dissociation and production of an organic radical and a chlorine radical. We optimized the geometry of the organic radical, calculated its single-point energy, and estimated the C–Cl bond dissociation energy, DE. The bond dissociation energy at each position was calculated with Eq. (1) (shown for OCDT):

ME ¼ EOCDT  Ec1  ECl

OCDF and OCDT could not be detected after 5 and 30 min of irradiation, respectively. Both OCDF and OCDT degraded through first-order dechlorination processes, and the first-order rate constant for OCDT (9.7 h1) was less than one-third that for OCDF (31 h1). This result suggests that PCDTs are more environmentally persistent than PCDFs. Next, we evaluated the rate constants in terms of the lowest C–Cl bond dissociation energies for each molecule calculated with Gaussian 98W. Table 1 shows the C–Cl bond dissociation energies for OCDT and OCDF and the single-point energy differences of the dechlorinated radicals produced by bond dissociation at the various positions. If the rate-controlling step in the photodegradation of OCDT and OCDF is production of a monodechlorinated radical by means of C–Cl bond dissociation, the compound with the lower bond-dissociation energy should dissociate more readily and be more reactive. However, OCDF was more reactive than OCDT, even though OCDT had the lower bond-dissociation energy (Table 1). This result indicates that the above hypothesis – if the rate-controlling step is production of a monodechlorinated radical, then the compound with the lower bond-dissociation energy should be more reactive – does not apply simply and uniformly to these compounds and that the photodegradation processes of OCDT might differ from those of OCDF. 3.2. Experimental degradation pathways UV irradiation resulted in the dechlorination of both OCDT and OCDF in n-hexane solution. The dechlorinated products of OCDF or OCDT produced through the photodegradation process have been reported to include the dominant isomers (Wagenaar et al., 1995; Nakai et al., 2007). We assumed that the isomer with the largest peak area observed upon analysis of the irradiated samples was the main degradation product, and we determined the main degradation pathways on the basis of the main products. Although some additional, minor peaks were detected, we discuss only the main dechlorination pathways. The degradation pathways for OCDT are shown in Fig. 1. Note that standard samples of PCDTs with fewer than six chlorines could not be purchased, and therefore complete quantitative analysis of all the PCDT congeners could not be accomplished. The main dechlorination pathway for OCDT was OCDT ? 1,2,3,4,7,8,9-HpCDT ? 1,2,3,7,8,9-HxCDT, indicating that OCDT was preferentially dechlorinated at the 4- and 6-positions. In contrast, OCDF has been reported to be preferentially dechlorinated at the 1-position (Wagenaar et al., 1995; Nakai et al., 2004). Thus, the main dechlorination pathway of OCDT differed from that of OCDF, which suggests that the mechanism for dechlorination of OCDT differed from that of OCDF. Table 1 Comparison of the bond dissociation energies and single-point energy differences of the dechlorinated radicals of OCDT and OCDF calculated by Gaussian 98W Bond

Bond dissociation energy (kJ mol1)

Energy difference of dechlorinated radicals (kJ mol1)

OCDT

C1–Cl C2–Cl C3–Cl C4–Cl

301.9 318.8 323.7 327.6

25.7 8.7 3.8 0.0

OCDF

C1–Cl C2–Cl C3–Cl C4–Cl

305.4 314.8 321.1 330.0

24.6 15.2 8.9 0.0

ð1Þ

where EOCDT is the single-point energy of OCDT, Ec1 is the energy of the organic radical generated by chlorine dissociation at the 1-position of OCDT, and ECl is the energy of the chlorine radical. We performed the same procedures for the bonds at the other positions, and all the other compounds were treated in the same fashion. We then predicted the degradation pathways on the basis of these results and hypotheses regarding the ratelimiting step of the degradation processes (the hypotheses are described below).

For both OCDT and OCDF, the dechlorinated radical with the lowest single-point energy was defined as having an energy of 0.

1007

S. Yamada et al. / Chemosphere 73 (2008) 1005–1010

thetical mechanism for photodechlorination of OCDT is shown as an example in Fig. 2. UV irradiation results in the dissociation of a C–Cl bond and the production of a monodechlorinated radical, which then reacts with a hydrogen donor, presumably a solvent molecule (Dung and O’Keefe, 1994), to generate a dechlorinated product. The rate-controlling step is either production of the monodechlorinated radical (bond dissociation rate-controlling hypothesis) or reaction of the radical with the hydrogen donor (radical reaction rate-controlling hypothesis). If the bond dissociation is the rate-controlling step, the relative reactivity of the bonds is the main determinant of the molecular structure of the product. That is, the C–Cl bond that dissociates most easily is expected to be the most reactive, and the compound generated by hydrogen substitution at the position with the lowest bond-dissociation energy is expected to be the main product. In the case of OCDF, mono-dechlorination can be expected to produce 1,2,3,4,6,7,8HpCDF because the 1-position has the lowest C–Cl bond-dissociation energy (Table 1). In contrast, if the radical reaction is the rate-controlling step, the reactivity of the dechlorinated radical is the key determinant of the molecular structure of the product. That is, the stability of the produced radical can be expected to determine the reactivity. The radical with the lowest single-point energy should be the most stable radical generated by C–Cl bond dissociation, and the compound produced by hydrogen donation to the most stable radical is expected to be the main product. In the case of OCDT, monodechlorination can be expected to produce 1,2,3,4,7,8,9-HpCDT because the radical dechlorinated at the 4-position has the lowest single-point energy (Table 1). The same logic applies to subsequent dechlorination reactions of the initial dechlorination products. With these hypotheses in mind, we used DFT calculations to estimate bond-dissociation energies and single-point energies of OCDT and OCDF and their dechlorinated congeners, and we used the resulting information to predict the dechlorination pathways. The dechlorination pathway of OCDT predicted by DFT calculations and the bond dissociation rate-controlling hypothesis is shown in Fig. 3, and the pathway predicted by the radical reaction rate-controlling hypothesis is shown in Fig. 4. Fig. 5 shows the pathway for OCDF predicted by the bond dissociation rate-controlling hypothesis. We compared these theoretical degradation pathways with the main experimental degradation pathways. For OCDF, the monodechlorinated product predicted by the radical reaction rate-controlling hypothesis (not shown) was 1,2,3,4,7,8,9-HpCDF. In contrast, under the bond dissociation rate-controlling hypothesis (Fig. 5), the predicted degradation pathway from the octa- to the tetrachlorinated congener was OCDF ? 1,2,3,4,6,7,8-HpCDF ? 1,3,4,6,7,8-HxCDF ? 1,3,4,6,8-PeCDF ? 1,4,6,8-TeCDF.The predominant mono-photodechlorinated product of OCDF is 1,2,3,4,6,7,8-HpCDF (Wagenaar et al., 1995; Nakai et al., 2004), and thus the product predicted by the bond dissociation rate-controlling hypothesis was the observed dominant

Cl

Cl

Cl

Cl

Cl

Cl S Cl

Cl

OCDT

Cl

Cl

Cl

Cl

Cl

Cl S

Cl

1,2,3,4,7,8,9-HpCDT

Cl

Cl

Cl

Cl

Cl

Cl S

1,2,3,7,8,9-HxCDT

Further dechlorination Fig. 1. Main experimental photodechlorination pathway of OCDT.

For PCDFs, toxic equivalency factors, were established only for 2,3,7,8-chloro-substituted congeners. 1,2,3,4,7,8,9-HpCDT and 12,3, 7,8,9-HxCDT have chlorine substitutes at the 2, 3, 7 and 8positions, and therefore these are concerns about their toxicity. 3.3. Rate-controlling hypotheses and theoretical degradation pathways The degradation pathways of OCDT and OCDF were predicted by using DFT calculations and rate-controlling hypotheses. A hypo-

H Hydrogen donor Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

UV Cl

Cl S Cl

Cl

Reactants

Bond dissociation

Cl

Cl Cl

S

Radical reaction

Dechlorinated radical Fig. 2. A hypothetical mechanism for photodechlorination of OCDT.

Cl

Cl S Cl

Dechlorinated products

1008

S. Yamada et al. / Chemosphere 73 (2008) 1005–1010

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl S Cl

Cl

Cl Cl

Cl

Cl

Cl

Cl Cl

Cl S

S

Cl

Cl Cl

Cl S

S

S Cl

Cl

Cl

Cl

Cl

Cl

S

S

S

Cl

Cl

Cl

Cl

Fig. 3. Photodechlorination pathway of OCDT predicted with the bond dissociation rate-controlling hypothesis.

Cl

Cl

Cl

Cl

Cl

Cl S

Cl Cl

Cl

Cl

S

Cl

S

Cl

Cl

Cl

S

S Cl

Cl

Cl

Cl Cl

Cl

Cl S

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl

S

Cl Cl

Cl

S

S

Fig. 4. Photodechlorination pathway of OCDT predicted with the radical reaction rate-controlling hypothesis.

product. Therefore, the bond dissociation rate-controlling hypothesis can be expected to adequately predict OCDF degradation pathways as is also the case for hexachlorobenzene (Yamada et al., 2007a). For OCDT, the degradation pathway predicted under the bond dissociation rate-controlling hypothesis (Fig. 3) was not observed in the irradiation experiment (Fig. 1). However, the predicted pathway from the octa- to the hexachlorinated congener, OCDT ? 1,2,3,4,7,8,9-HpCDT ? 1,2,3,7,8,9-HxCDT, was observed in the irradiation experiment (Fig. 1). Although not all of the degradation products of OCDT could be unequivocally identified (owing to the lack of standards), our results suggest that the predictive method based on the radical reaction rate-controlling hypothesis is better suited for the photodechlorination of OCDT than for the photodechlorination of OCDF. The only difference between OCDT and OCDF is the cross-linking heteroatom, S or O. The heteroatom can be expected to affect the molecular structure and charge distribution in the molecule, and we assume that the differences between the two heteroatoms are responsible for the differences in the degradation mechanisms of OCDT and OCDF. Fig. 6 shows the optimized molecular structures and the Mulliken charges of OCDT and OCDF estimated

with the Gaussian 98W program. The Mulliken charges indicate the charge distribution in the molecules. The heteroatom seems to have a greater effect on the charge distribution than on the molecular structure; for example, the charge on the S atom in OCDT is positive, whereas the charge on the O atom in OCDF is negative. The charges on the carbon atoms that make up the benzene ring of OCDT differ from the charges on the corresponding carbon atoms in OCDF. We suggest that the differences in the degradation mechanisms of OCDT and OCDF were affected by these differences in the charge distributions. However, the reasons for the difference between the charge distributions and the effect of this difference on the degradation mechanisms are topics beyond the scope of the present study, and further experimental studies are needed to determine the importance of this effect. 4. Conclusion We determined the dechlorination degradation pathways of OCDT to hexachlorinated congeners and of OCDF to tetrachlorinated congeners in UV irradiation experiments, and we compared the experimental pathways with the pathways predicted by DFT calcu-

1009

S. Yamada et al. / Chemosphere 73 (2008) 1005–1010

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl O

O

O Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl

O

O Cl

O

O

O

Cl

Cl

Cl

Cl

O

Fig. 5. Photodechlorination pathway of OCDF predicted with the bond dissociation rate-controlling hypothesis.

Fig. 6. Bond length in Å and bond angles in degrees (left) and Mulliken charges (right) for OCDT (a) and OCDF (b).

lations. We proposed two possibilities for the dechlorination ratecontrolling step–bond dissociation or reaction of the resulting radical—and we discussed the degradation pathways expected for each possibility. We found that the radical reaction rate-controlling hypothesis was suitable for OCDT, whereas the bond dissociation rate-controlling hypothesis was better suited for OCDF. These results suggest that PCDTs in the environment should not be treated or managed in the same way as PCDFs. Further study of the dechlorination mechanisms of these analogues is warranted. References Arulmozhiraja, S., Morita, M., 2004. Electron affinities and reductive dechlorination of toxic polychlorinated dibenzofurans. J. Phys. Chem. A 108, 3499–3508.

Berkout, V.D., Mazurkiewcz, P., Deinzer, M.L., 1999. Electron energy dependence of regioselective chloride ion loss from polychlorodibenzo-p-dioxins. J. Am. Chem. Soc. 121, 2561–2570. Claus, E., Heininger, P., Bade, M., Juering, H., Raab, M., 1998. Mass spectrometric identification of polychlorinated dibenzothiophenes (PCDBTs) in surface sediments of the River Elbe. Fresenius J. Anal. Chem. 361, 54–58. Dung, M.H., O’Keefe, P.W., 1994. Comparative rates of photolysis of polychlorinated dibenzofurans in organic solvents and in aqueous solutions. Environ. Sci. Technol. 28, 549–554. Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A., Cheeseman, J.R., Zakrzewski, V.G., Montgomery Jr., J.A., Stratmann, R.E., Burant, J.C., Dapprich, S., Millam, J.M., Daniels, A.D., Kudin, K.N., Strain, M.C., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, G.A., Ayala, P.Y., Cui, Q., Morokuma, K., Rega, N., Salvador, P., Dannenberg, J.J., Malick, D.K., Rabuck, A.D., Raghavachari, K., Foresman, J.B., Cioslowski, J., Ortiz, J.V., Banoul, A.G., Stefanov, B.B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R.L., Fox, D.J., Keith, T., Al-Laham, M.A., Peng, C.Y., Nanayakkara, A., Chakkacombe, M., Gill,

1010

S. Yamada et al. / Chemosphere 73 (2008) 1005–1010

P.M.W., Johnson, B., Chen, W., Wong, M.W., Andres, J.L., Gonzalez, C., HeadGordon, M., Replogle, E.S., Pople, J.A., 2001. Gaussian 98. Gaussian Inc., Pittsburgh, PA. Fueno, H., Tanaka, K., Sugawa, S., 2002. Theoretical study of the dechlorination reaction pathways of octachlorodibenzo-p-dioxin. Chemosphere 48, 771–778. Nakai, S., Nomura, Y., Hosomi, M., Espino, M.P., 2004. Detection of polychlorinated dibenzothiophenes (PCDTs) in environmental samples and investigation of their photodegradability and dioxin-like endocrine disruption potency. J. Environ. Chem. Jpn. 14, 835–844. Nakai, S., Kishita, S., Nomura, Y., Hosomi, M., 2007. Polychlorinated dibenzothiophenes in Japanese environmental samples and their photodegradability and dioxin-like endocrine-disruption potential. Chemosphere 67, 1852–1857. Neidhard, H., Herrmann, M., Abbau, P., 1987. Transport polychlorierter dibenzodioxine und dibenzofurane in der Umwelt. VDI Ber Germany 634, 303–316. Sinkkonen, S., 1997. PCDTs in the environment. Chemosphere 34, 2585–2594.

Suegara, J., Lee, B.D., Espino, M.P., Nakai, S., Hosomi, M., 2005. Photodegradation of pentachlorophenol and its degradation pathways predicted using density functional theory. Chemosphere 61, 341–346. Wagenaar, W.J., Boelhouwers, E.J., de Kok, H.A.M., Groen, C.P., Govers, H.A.J., Olie, K., De Gerlache, J., De Rooij, C.G., 1995. A comparative study of the photolytic degradation of octachlorodibenzofuran(OCDF) and octachlorodibenzo-pdioxin(OCDD). Chemosphere 31, 2983–2992. Wiedmann, T., Riehle, U., Kurz, J., Ballschmiter, K., 1997. HRGC-MS of polychlorinated phenanthrenes (PCPhen), dibenzothiophenes (PCDT), dibenzothianthrenes (PCTA), and phenoxathiins (PCPT). Fresenius J. Anal. Chem. 359, 176–188. Yamada, S., Naito, Y., Takada, M., Nakai, S., Hosomi, M., 2007a. Photodegradation of hexachlorobenzene and theoretical predictions of its degradation pathways using quantum chemical calculation. Chemosphere 70, 731–736. Yamada, S., Naito, Y., Funakawa, M., Nakai, S., Hosomi, M., 2007b. Photodegradation fates of cis-chlordane, trans-chlordane, and heptachlor in ethanol. Chemosphere 70, 1669–1675.