Mendeleev Communications Electronic Version, Issue 4, 1999 (pp. 129–170) Role of electronic factors in the formation of a standard quasi-stable mixture of toxic polychlorinated dibenzo-para-dioxins and polychlorinated dibenzofurans Sergei S. Yufit N. D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, 117913 Moscow, Russian Federation. Fax: +7 095 135 5328; e-mail: yufit@ioc.ac.ru Only electronic factors are responsible for the kinetics and mechanism of consecutive chlorination of 2,3,7,8-tetrachlorodibenzopara- dioxin and 2,3,7,8-tetrachlorodibenzofuran; this results in the formation of a standard quasi-stable mixture of chlorination products (congeners), polychlorinated dibenzo-p-dioxins and dibenzofurans, whose composition was independent of the sample type and the place of sampling.Polychlorinated dibenzo-para-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) containing chlorine atoms in the 2,3,7,8-positions are extremely toxic substances (the entire group of these compounds is usually referred to as dioxins). The major quantity of these substances is formed as a result of incineration of municipal and hospital wastes and other hightemperature processes in which chlorine-containing substances are produced, destroyed or treated.1 Because dioxins are a serious hazard to the environment and human health, the mechanism of their formation is studied intensively (see the recent data in ref. 2). Unfortunately, the data acquired hitherto are inadequate to develop the commonly accepted mechanism of the formation and reactions of dioxins because of great experimental difficulties and very expensive analysis for dioxins.Here a new approach to the elucidation of some details in the mechanism of dioxin formation is suggested. It is based on the analysis of mixtures formed in different sources of contamination. It is believed that, in all cases, the chlorination of 2,3,7,8- tetrachlorodibenzo-para-dioxin (TCDD) or 2,3,7,8-tetrachlorodibenzofuran (TCDF) proceeds consecutively with the participation of the same chlorinating agent at all stages of the production of completely chlorinated products.The chlorination path depends on the ortho-para-directing effects of oxygen and chlorine in the molecule. The quasi-steady-state concentrations of products (congeners) in the mixture formed depend on only the activity of a particular site in the reacting dioxin molecule and the rate of ‘discharge’ (consumption) of the congener formed.In the course of a thorough examination of the contamination of human milk with PCDDs and PCDFs in Russia, which was performed within the framework of the WHO programme ‘Second Round of Exposure Studies on Levels of PCBs, PCDDs and PCDFs in Human Milk’, we found that the composition of PCDD/PCDF contaminants in human milk sampled in different Russian cities was very similar.3–5 A comparison with the data obtained by other researchers in different countries6 supported this observation.In this case, the toxicity of milk expressed in the toxic equivalents I-TEQ (i.e., referred to the toxicity of TCDD, which was taken to be unity on the international scale of the equivalent toxicity factors I-TEF) could differ widely (Figure 1).As an example, Table 1 summarises the data for milk from two cities in Russia and for bottom sediment samples from the river of Severnaya Dvina (these data were obtained in different analytical laboratories). To explain the similar distribution of congeners (not the different TEQ level), we used a mechanism of the consecutive chlorination of aromatic rings in TCDD and TCDF, the precursors of toxic PCDD and PCDF groups, and compared these qualitative data to the experimental data published in terms of ‘more–less’. We considered the slopes between two neighbouring points in graphs of the kind shown in Figure 1.The expression of analytical data in terms of molar concentrations did not affect the results obtained.The qualitative composition of the produced mixture can be determined on the basis of simple chemical notions using the model of the formation of a mixture of 17 toxic PCDDs and PCDFs. This approach seems to be very useful taking into account that the analytical data are extremely difficult to obtain.In this mechanism, TCDD 1 and TCDF 8 (see Table 1 for the compound numbers) are the starting substances. In the consideration of the sequence (probability) of replacement by chlorine in aromatic rings, only a single assumption will be used: the reaction proceeds as usual electrophilic substitution in the aromatic series; therefore, ortho-substitution with respect to aThe data for Salavat (Russia) were obtained in the laboratory of A.K. D. Liem (RIVM, the Netherlands) and for Volgograd (Russia), in the laboratory of W. A. Traag (RIKILT-DLO, the Netherlands). bThe data (averaged over eight samples) were obtained in the laboratory of N. A. Klyuev (Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences, Russia). cn. d.= not detected (the limit of detection was ~0.1 pg per gramm of fat). Table 1 Dioxin concentrations in human milka (pg per gramm of fat) and bottom sediments from the river Severnaya Dvinab (pg kg–1) (data in terms of 10– 4 pmol g–1 for milk and pmol kg–1 for bottom sediments are given in parentheses). Compound Code Congener Mr Salavat Volgograd Severnaya Dvina 1 D4 2,3,7,8-tetraCDD 322 5.3 (164.6) 2.73 22 (0.068) 2 D5 1,2,3,7,8-pentaCDD 365.5 3.9 (106.7) 1.47 20 (0.056) 3 D6(1) 1,2,3,4,7,8-hexaCDD 391 1.0 (25.58) 1.14 39 (0.1) 4 D6(2) 1,2,3,6,7,8-hexaCDD 391 2.7 (69.05) 3.26 48 (0.123) 5 D6(3) 1,2,3,7,8,9-hexaCDD 391 0.7 (17.9) n.d.c 40 (0.102) 6 D7 1,2,3,4,6,7,8-heptaCDD 425.5 3.4 (79.9) 3.98 4420 (0.987) 7 D8 octaCDD 460 16.4 (356.5) 18.32 3900 (8.48) 8 F4 2,3,7,8-tetraCDF 306 0.8 (26.14) 1.44 54 (0.18) 9 F5(1) 1,2,3,7,8-pentaCDF 340.5 0.4 (11.7) 0.62 20 (0.06) 10 F5(2) 2,3,4,7,8-pentaCDF 340.5 6.9 (202.6) 8.37 17 (0.05) 11 F6(1) 1,2,3,4,7,8-hexaCDF 375 3.7 (98.7) 4.64 48 (0.13) 12 F6(2) 1,2,3,6,7,8-hexaCDF 375 2.0 (53.4) 2.32 28 (0.08) 13 F6(3) 1,2,3,7,8,9-hexaCDF 375 0.7 (18.7) n.d. 7 (0.02) 14 F6(4) 2,3,4,6,7,8-hexaCDF 375 0.1 (2.7) 0.82 68 (0.18) 15 F7(1) 1,2,3,4,6,7,8-heptaCDF 409.5 1.5 (36.6) 1.34 370 (0.9) 16 F7(2) 1,2,3,4,7,8,9-heptaCDF 409.5 0.1 (2.4) n.d. 41 (0.1) 17 F8 octaCDF 444 0.3 (6.76) n. d. 1100 (2.48)Mendeleev Communications Electronic Version, Issue 4, 1999 (pp. 129–170) the oxygen atom is predominant. Correspondingly, isomers that are formed by ortho-substitution will exhibit higher concentrations in the mixture.The consumption of a formed isomer at the following step of chlorination will be considered as a correction factor. The resulting mixture of congeners is considered to be standard (independent of the sources of contamination) and quasi-stable. As an example, we consider the chlorination of TCDF 8 (Scheme 1). At the first step of the reaction, two pentachloro isomers 9 and 10 can be formed because, in contrast with dioxin 1, not all positions accessible to chlorination in initial furan 8 are equivalent.Using the accepted model, it is easy to predict that the amount of isomer 10 formed at the first step will be significantly greater than that of isomer 9. The ratio between these isomers can serve as a measure of the reactivity of the ortho-position with respect to oxygen in this system (the 4- and 6-positions) and of the ortho-position with respect to chlorine atoms (1- and 9-positions).To determine the excess of isomer 10 over isomer 9, the rates of consumption (‘discharge’) of these isomers at the following step of chlorination should be considered. Furan 9 has two ortho-positions accessible to chlorination, whereas furan 10 has only one position of this kind.Hence it follows that 9 will be consumed at the second step of the reaction more rapidly than 10. Thus, furan 9 will be formed more slowly and consumed more rapidly than furan 10. As a consequence, the concentration of furan 9 in the formed mixture of congeners will be lower than that of furan 10. This conclusion ([10] > [9]) is supported by data obtained by different authors.At the third step of chlorination, four hexachloro derivatives of furan (11, 12, 13 and 14) are formed. Among these compounds, furan 13 can be mentioned. To form 13, the substitution in the ortho-position with respect to chlorine in 9 is required. At the same time, the consumption of 13 is maximally facilitated by two active sites in the 4- and 6-positions.Consequently, among the hexachloro isomers, the rates of formation and consumption of isomer 13 are lowest and highest, respectively, and isomer 13 will exhibit the lowest concentration among all hexachloro isomers. This fact is in complete agreement with analytical data. The formation of isomer 11 also requires unfavourable replacement in the 1-position of isomer 10.However, isomer 11 will be formed at a higher rate than 13 because, in this case, 11 can also be formed from isomer 9. It is evident from a comparison between the rates of formation of isomers 14 and 12 that the formation of 12 is more probable than the formation of 14: there are two active sites (the 4- and 6-positions) for the formation of 12 from 9 and only a single active site (the 6-position) for the formation of 14 from 10.Thus, at the stage of production, isomer 13 will be formed in the smallest amount. However, the distribution of the other three isomers is not so obvious, and it is more reasonable to compare the rates of ‘discharge’ (consumption). Isomer 13 is the most readily consumed, and its concentration in the mixture will be low (see above).This is a good reference compound for testing the model. Isomer 14 is poorly consumed (the two unfavourable 1- and 9-positions); this fact would result in its accumulation. However, this isomer cannot be formed from 9; consequently, it has only a single channel of formation in contrast with 11 and 12. Isomers 11 and 12 are not dissimilar in the ortho-positions; however, 11 is produced from both of pentachloro isomers 9 and 10, and 12 cannot be produced from 10.It is believed that the concentrations of 11 and 12 will be O Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl O Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl 8 9 10 11 12 13 14 15 16 17 [10] > [9] [11] � [12] � [14] > [13] [15] > [16] 1 2 3 4 6 7 8 9 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 2 3 7 8 4 6 9 1 6 9 1 9 4 6 4 9 6 9 9 6 Scheme 1 Chlorination of furans (see Table 1 for the numeration of compounds).Honolulu, USA South Vietnam West Germany Suzdal, Russia Salavat, Russia Congeners Concentration (pg per gramm of fat) 102 101 100 10–1 10–2 10–3 10–4 10–5 D4 D5 D6 D6 D6 D7 D8 F4 F5 F5 F6 F6 F6 F6 F7 F7 F8 Figure 1 Comparison between the contamination of human milk in different countries, expressed as the toxic equivalents TEQ of the scale I-TEF.The data for Honolulu, Suzdal (Russia) and Salavat (Russia) were taken from refs. 4 and 5, for West Germany and South Vietnam, from refs. 6 and 7. The points at a level of 10–4 are below the limit of detection. The codes of congeners correspond to those in Table 1.Mendeleev Communications Electronic Version, Issue 4, 1999 (pp. 129–170) not strongly different; however, 11 will be somewhat predominant. Thus, the concentration distribution in the mixture of hexachloro isomers will be as follows: [11] � [12] > [14] > [13].At the fourth step of the reaction, isomer 15 can be produced from three hexachloro isomers 11, 12 and 14, and isomer 16, from isomers 11, 12 and 13 (the amount of 13 in the mixture is always small). The comparison between the rates of ‘discharge’ of these isomers shows that isomer 16 will be consumed more rapidly than 15, which has no free ortho 6-position favourable for chlorination.Hence it follows that isomer 15 will be formed more rapidly and consumed more slowly than 16. Therefore, the concentration of isomer 15 will be higher than that of 16, which is the case. As for the concentration of 17, it will not be used for evaluating the reliability of the mechanism suggested, as is the case with octachloro dioxin 7. The reason is that concentrations of these compounds, as well as concentrations of tetrachloro isomers 1 and 8, can vary significantly for unknown reasons.The same mechanism can also be applied to the chlorination of TCDD. Thus, the following concentration ratio in the standard quasistable mixture of toxic PCDD and PCDF congeners was found: For statistical checking, published data on the PCDD/PCDF contamination of various matrices of both biological (blood, milk and foods) and anthropogenic origin (sludge, slimes, bottom sediments, soils, wastewater etc.)8,9 were considered.The total number of analyses was 70; the total number of experimental points was 1190; 910 points were used for comparison with the predicted composition of the mixture. The slopes for the first and last members of the series were not compared because of the absence of data on the initial concentrations.Deviations from the predicted slops were observed for 40 points (4.3%). References 1 The Inventory of Sources of Dioxin in the United States, US EPA, External review draft, April 1998. Report #EPA/600/P-98/002Aa. 2 Organohalogen Compounds. Annual Reports of the International Symposium on Chlorinated Dioxins and Related Compounds. 3 S. S.Yufit, Organohalogen Compounds, 1997, 33, 165. 4 W. A. Traag and S. S. Yufit, Organohalogen Compounds, 1997, 33, 524. 5 A. K. D. Liem and R. M. C. Theelen, Dioxins: Chemical Analysis, Exposure and Risk Assessment, Tauw Milieu, The Netherlands, 1997, p. 186. 6 A. Schecter, J. R. Startin, V. Rose, C. Wright, I. Parker, D. Woods and H. Hansen, Chemosphere, 1990, 20, 919. 7 A. Schecter, P. Fürst, C. Fürst and O. Päpke, Chemosphere, 1991, 23, 1903. 8 Z. Amirova and E. Kruglov, Dioksiny v okruzhayushchei srede, nagruzka na cheloveka i immunologicheskie aspekty vliyaniya iskhodnogo urovnya zagryazneniya dioksinami v kogortnykh gruppakh (Dioxins environment, load on human being and immunological aspects of dioxin influence on background level in cohort groups), Reactiv, Ufa, 1998, p. 115 (in Russian). 9 S. S. Yufit, N. A. Klyuev and E. S. Brodsky, in Dioksiny — supertoksikanty XXI veka (Dioxins as the supertoxicatns of XXI century), ed. Yu. A. Arsky, VINITI RAS, Moscow, 1998 (in Russian). Dioxins [1]/[2] not found [4] > [3], [5] [6]/[7] not found Furans [8]/([9] + [10]) not found [10] > [9] [13] < [11], [12], [14] [11] � [12] � [14] > [13] [15] > [16] ([15] + [16])/[17] not found Received: 4th December 1998; Co