Environ. Sci. Technol. 2000, 34, 3143-3147
Dechlorination-Controlled Polychlorinated Dibenzofuran Isomer Patterns from Municipal Waste Incinerators FUKUYA IINO,† TAKASHI IMAGAWA,‡ AND B R I A N K . G U L L E T T * ,§ Oak Ridge Institute for Science and Education, Oak Ridge, Tennessee 37831, National Institute for Resources and Environment, 16-3 Onogawa, Tsukuba, Ibaraki, 305-8569 Japan, and U.S. Environmental Protection Agency, National Risk Management Research Laboratory, Air Pollution Prevention and Control Division, Air Pollution Technology Branch, MD-65, Research Triangle Park, North Carolina 27711
The ability to predict polychlorinated dibenzofuran (PCDF) isomer patterns from municipal waste incinerators (MWIs) enables an understanding of PCDF formation that may lead to preventive measures. This work develops a model for the pattern prediction, assuming that the chromatograph peak ratios relative to the total four heptachloro-dibenzofuran (H7CDF) isomers are determined by position-specific dechlorination kinetics from an octachloro-dibenzofuran (O8CDF) parent to H7CDFs and that these probabilities can subsequently predict the relative concentrations of lower chlorinated isomers. Agreement of PCDF isomer patterns between the model and sampled data from eight MWIs is consistent with formation of tetrachloro-dibenzofurans to H7CDFs by dechlorination from an O8CDF parent. The application of the analogous theory to predict isomer patterns of polychlorinated dibenzo-p-dioxins (PCDDs) did not provide any significant results, which implies that the formation mechanism of PCDDs is controlled by other factors such as condensation of precursors and further chlorination. This method can be a fundamental basis to develop a prediction model for total PCDFs emission and toxic equivalent values.
Introduction The formation mechanisms of polychlorinated dibenzofurans (PCDFs) and polychlorinated dibenzo-p-dioxins (PCDDs) from municipal waste incinerators (MWIs) have been extensively studied due to their strong toxicity. The emission of PCDDs and PCDFs (PCDD/Fs) was first reported by Olie et al. in 1977 (1). Since then, many efforts (2-8) have been made to understand what mechanisms control the homologue and isomer patterns of PCDD/Fs. There are eight homologues each for PCDDs and PCDFs, determined by the number of chlorine (Cl) atoms on the molecule, and this results in 75 PCDD and 135 PCDF species. * Corresponding author phone: (919)541-1534;
[email protected]. † Oak Ridge Institute for Science and Education. ‡ National Institute for Resources and Environment. § U.S. Environmental Protection Agency. 10.1021/es9913131 CCC: $19.00 Published on Web 06/28/2000
2000 American Chemical Society
e-mail:
Despite a variety of incineration conditions due to different incinerator types, feed compositions, temperature distributions, and air supply methods, the similarity of the homologue-specific isomer patterns from these incinerators has been well-known (9) with some exceptions (10). Addink et al. (11) investigated the similarity of the isomer patterns by using equilibrium concentrations relative to each total homologue that were predicted from calculated Gibbs free energy of formation (∆G°f,T) values (12-18). However, it was not confirmed that the isomer patterns were thermodynamically controlled, and the question that still remains to be answered is what mechanism causes the similarity of the PCDD/F isomer patterns in these disparate incinerators? There are two main formation mechanisms considered to play an important role in MWIs. One is the condensation pathway of precursors such as chlorophenols, and the other is de novo synthesis from fly ash-bound C, H, O, and Cl (19). The de novo mechanism is not rigidly defined, but a definition by Iino et al. (20) considers this process to be direct PCDD/F formation from a carbon matrix that has been chlorinated and oxidized. By identifying products formed from a 12C and 13C mixture, Hell et al. (21) suggested that 99% of PCDFs and 75-90% of PCDDs are directly released from preformed structures in the amorphous carbon matrix, e.g., biphenyls. In addition, Iino et al. (22, 23) showed that the main product formed from graphite and copper chloride was octachlorodibenzofuran (O8CDF). The reason for O8CDF as a main product from graphite is that full substitution, or termination, by chlorine atoms is necessary in order for the biphenyl structures to be released from the graphitic carbon plane under a hydrogen-scarcity condition. These experimental results on PCDFs from de novo synthesis are consistent with a theory suggesting that O8CDF is the initial PCDF product from the carbon matrix and can be subsequently dechlorinated into lower chlorinated PCDFs. On the other hand, we should point out that a recent study by Hell et al. (24) proved that O8CDF can also be formed from 2,4,6trichlorophenol on fly ash. The possibility that isomer patterns can be explained, in part, by dechlorination kinetics from O8CDF was explored by Hagenmaier et al. (25) and Wiesmu ¨ ller (26) to explain PCDD/F isomer patterns in field data. In Wiesmu¨ller’s model, however, all simple possibilities of chlorination and dechlorination of each PCDD/F isomer were considered in order to derive and solve, via the Runge-Kutta method, differential equations for the isomer patterns. The lack of conceptual theory in this model limits its usefulness and applicability.
Calculation Methods In this paper, we attempted to predict the field sampled isomer patterns of PCDFs by assuming that the measured chromatograph peak areas of the four heptachlorodibenzofuran (H7CDF) isomers, relative to the total H7CDFs, are determined by position-specific dechlorination kinetics from the O8CDF parent and that these probabilities can subsequently predict the relative concentrations of lower chlorinated isomers such as hexachlorodibenzofurans (H6CDFs) and further still to the pentachlorodibenzofurans (P5CDFs), etc. The relative peak ratios of 1,2,3,4,6,7,8-, 1,2,3,4,6,7,9-, 1,2,3,4,6,8,9-, and 1,2,3,4,7,8,9H7CDFs are termed a-d, respectively. The values of a-d are the assumed constant dechlorination probabilities from the 9- or 1-, 8- or 2-, 7- or 3-, and 6- or 4-substitution positions, respectively, on a PCDF molecule. For example, VOL. 34, NO. 15, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3143
TABLE 1. Predicted Peak Ratios of H6CDFs, P5CDFs, and T4CDFs When the Relative Peak Ratios of the Four H7CDFs Are a-d, Respectively, in Order isomers
equations
1,2,3,4,6,7,8-H7CDF 1,2,3,4,6,7,91,2,3,4,6,8,91,2,3,4,7,8,9-
a b c d
1,2,3,4,6,8-H6CDF 1,3,4,6,7,9-/1,3,4,6,7,81,2,4,6,7,81,2,4,6,7,91,2,3,4,7,9-/1,2,3,4,7,81,2,3,6,7,81,2,4,6,8,91,2,3,4,6,71,2,3,6,7,91,2,3,6,8,9-/1,2,3,4,6,91,2,3,7,8,91,2,3,4,8,92,3,4,6,7,8-
2ac b2 + 2ab 2ac 2bc 2bd + 2ad 2ad c2 2ab 2bd 2cd + 2bc d2 2cd a2
isomers
equations
isomers
equations
1,3,4,6,8-P5CDF 1,2,4,6,81,3,6,7,81,3,4,7,91,2,3,6,8-/1,3,4,7,81,2,4,7,81,3,4,6,7-/1,2,4,7,91,2,4,6,71,4,6,7,8-/1,2,3,4,71,3,4,6,91,2,3,7,8-/1,2,3,4,81,2,3,4,61,2,3,7,91,2,3,6,71,2,4,6,9-/1,2,6,7,81,2,6,7,91,2,3,6,92,3,4,6,81,2,3,4,91,2,4,8,92,3,4,7,81,2,3,8,92,3,4,6,7-
6abc 6ac2 6abd 6 b 2d 6acd + 6abd 6acd 6ab2 + 6bcd 6abc 6abc + 6abd 6 b 2c 6ad2 + 6acd 6abc 6bd2 6abd 6bc2 + 6acd 6bcd 6bcd 6 a 2c 6bcd 6 c 2d 6 a 2d 6cd2 6 a 2b
1368-T4CDF 1378-/1379134714681247-/136713481346-/12481246-/12681237-/1369-/14781678-/12342468-/1467-/1238-/1236134912781267-/12791469-/1249236824671239-/23471269237823482346236734671289-
24abcd 24abd2 + 12b2d2 24ab2d 24abc2 24abcd + 24ab2d 24abcd 24ab2c + 24ac2d 24abc2 + 24ac2d 24abd2 + 24b2cd + 24abcd 24abcd + 24abcd 12a2c2 + 24ab2c + 24acd2 + 24abcd 24b2cd 24acd2 24abcd + 24bcd2 12b2c2 + 24bc2d 24a2cd 24a2bc 24bcd2 + 24a2bd 24bc2d 12a2d2 24a2cd 24a2bc 24a2bd 12a2b2 12c2d2
FIGURE 1. (A) Predicted H6CDFs isomer patterns and (B) the eight MWI sample data (1-4 from left are fluidized-bed incinerators, and 5-8 are stoker-type incinerators). loss of a chlorine atom from the 9-substitution position followed by loss of another chlorine atom from the 8-position on an O8CDF molecule has a probability of ab. The reverse 3144
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 15, 2000
dechlorination pathway has a probability of ba, resulting in a joint probability of forming 1,2,3,4,6,7-H6CDF of 2ab. The predicted concentration of H6CDFs, P5CDFs, and tetra-
FIGURE 2. (A) Predicted P5CDFs isomer patterns and (B) the eight MWI sample data (the same as in Figure 1).
FIGURE 3. (A) Predicted T4CDFs isomer patterns and (B) the eight MWI sample data (the same as in Figure 1). chloro-dibenzofurans (T4CDFs) are shown in Table 1. Those values show only the relative isomer concentrations within each homologue; they do not represent the quantitative relation between homologues. Each coefficient of the equations in Table 1 indicates the number of possible dechlorination pathways from O8CDF to a PCDF isomer. The more symmetric isomer in a homologue has a smaller
coefficient, such as 12 of 12a2d2 for 2,3,7,8-T4CDF because it has a smaller number of potential dechlorination pathways from its O8CDF parent in the T4CDF homologue. All of the P5CDF isomers have the same coefficient, that is 6, due to the same symmetries in the P5CDF homologue. The predicted results were obtained by using the average H7CDF peak area ratios, normalized to a total H7CDF homologue, from two VOL. 34, NO. 15, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3145
FIGURE 4. PCDD/Fs homologue profiles of the field samples (1, 4, 6, and 7). T4CDD to O8CDD stand for tetrachloro- to octachloro-dibenzo-pdioxins, respectively.
TABLE 2. Calculated Values of Similarity, Sa S1 S2 S3 S4 S5 S6 S7 S8 Se
H6CDF
P5CDF
T4CDF
0.86 0.95 0.93 0.94 0.85 0.91 0.86 0.98 0.70
0.94 0.93 0.95 0.96 0.79 0.82 0.81 0.93 0.45
0.88 0.80 0.89 0.92 0.79 0.70 0.80 0.89 0.48
a S shows similarity between the predicted patterns and eight n samples from eight MWIs in Japan, and Se shows similarity between the predicted and the equilibrium concentrations calculated from ∆G.
field data (out of eight) to derive values of a ) 0.630, b ) 0.166, c ) 0.129, and d ) 0.0755 in Table 1. This reported research applies this method to data derived from eight different MWI fly ash samples analyzed according to the method described elsewhere (20). Four of these incinerators are fluidized bed incinerators (numbers 1-4 in Figures 1-3 and in Table 2) and the other four are stoker type incinerators (numbered 5-8). The equilibrium data were obtained from ref 15 as selected by Addink et al. (11).
Results and Discussion The predicted results (blank bars) of H6CDFs, P5CDFs, and T4CDFs are compared with the field sampled MWI data in Figures 1-3, respectively. The field samples (filled bars) are 1-8 from left to right for each isomer. The isomer contents (y-axis) were normalized to each homologue total. The predicted isomer patterns of H6CDFs (Figure 1) and P5CDFs (Figure 2) showed almost identical isomer patterns with the MWIs. The T4CDF isomer pattern was only fairly predicted (Figure 3). Agreement between the model and sampled data is consistent with formation of T4CDFs to H7CDFs by dechlorination from an O8CDF parent. It is known that the isomer patterns are independent of combustion conditions even under conditions that provide different homologue profiles (9). Homologue profiles of the four field data (the sample numbers are 1, 4, 6, and 7) are shown in Figure 4. The field samples of 1 and 7 showed H7CDFs as the highest PCDF, and T4CDFs were the highest in 4 and 6. The abundance of O8CDF and H7CDFs in homologue patterns does not necessarily mean that highly 3146
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 15, 2000
chlorinated species are first formed and dechlorination follows. The homologue profiles are more easily changed by combustion conditions, such as a chlorine/hydrogen ratio and temperature (23), than isomer patterns; the homologue patterns provide us information on chlorine abundance and do not exactly tell us which formation pathway is predominant, chlorination or dechlorination. However, our prediction model on isomer patterns leads to a possible answer that the formation of isomer-specific PCDFs is mainly controlled by dechlorination. As an evaluating method of isomer pattern matching, Morita et al. (27) introduced similarity, S, as follows:
S)
∑AiBi
x∑(Ai)2∑(Bi)2
where Ai and Bi are the homologue-normalized isomer contents of the isomer i in samples A and B, respectively. The Sn (number of samples, n ) 1-8) values between the predicted isomer contents in this paper and the MWI samples n are shown in Table 2. Likewise, Se shows the similarity between the predicted and the equilibrium isomer concentrations. The Sn of H6CDFs, P5CDFs, and T4CDFs between the predicted and MWI samples are within 0.98-0.85, 0.96-0.79, and 0.92-0.70, respectively. These values indicate that the data are well-predicted by the dechlorination-controlled isomer patterns. The decreasing of Sn with the lower degree of chlorination can be attributed to two possible reasons. One reason is that the error that derives from the assumption of constant position-specific dechlorination probabilities propagates every time in multiplying the probability to get lower chlorinated PCDFs. The dechlorination probability from a position on a PCDF molecule may be different before and after dechlorination occurs at other positions. The other possible reason is that different formation mechanisms are controlling the lower chlorinated PCDFs such as chlorination from further lower chlorinated PCDFs and de novo synthesis from polycyclic aromatic hydrocarbons (28). Similarity calculations with the dechlorination model and the equilibrium models result in Se values of P5CDFs and T4CDFs that are 0.45 and 0.48, respectively, from which it can be concluded that the predicted and the equilibrium concentrations do not have similar distributions. A theoretical model for the buildup of PCDF isomer patterns was also attempted, on the assumption that four
monochloro-dibenzofurans are formed by chlorination from dibenzofuran. No significant results were found, suggesting that dechlorination mechanisms are better able to explain these field results than mechanisms of chlorination from lower chlorinated isomers. We also tried to predict PCDD isomer patterns according to the same procedures, but no reasonable correlations were obtained. The chemical structure of PCDFs may be more amenable to the de novo synthesis mechanism, which tends to form higher chlorinated PCDFs from particulate carbon, than PCDDs; hence, PCDFs are more applicable for this dechlorination mechanism. The mechanism of dechlorination of octachloro-dibenzo-p-dioxin (O8CDD) does not seem to play a crucial role for PCDDs formation in MWIs, and other formation pathways (e.g., condensation of the precursors and further chlorination) or other factors (29) are probably predominant. Wehrmeier et al. (29) mentioned these factors by concluding that “a model which was derived from a superposition of thermodynamic stability and reactivity of PCDD isomers gave a qualitative description of the typical PCDD combustion patterns”. The excellent agreement between the predicted T4CDF through H7CDF isomer patterns and the MWI field data indicates that we have to recognize that the formation mechanism of PCDFs is completely different from that of PCDDs. Therefore, approaches to emission-control technology or pollution prevention may also differ between the two toxic byproducts from MWIs.
Acknowledgments This research was supported in part by the appointment of F.I. to the Postdoctoral Research Program at the National Risk Management Research Laboratory, administered by the Oak Ridge Institute for Science and Education through Interagency Agreement DW89938167 between the U.S. Department of Energy and the U.S. Environmental Protection Agency. This work was performed at the U.S. EPA/NRMRL through participation in the ORISE Postdoctoral Research Program.
Literature Cited (1) Olie, K.; Vermeulen, P. L.; Hutzinger, O. Chemosphere 1977, 6, 455.
(2) Eklund, G.; Pedersen, J. R.; Stromberg, B. Nature 1986, 320, 155. (3) Karasek, F. W.; Dickson, L. C. Science 1987, 237, 754. (4) Stieglitz, L.; Zwick, G.; Beck, J.; Roth, W.; Vogg, H. Chemosphere 1989, 18, 1219. (5) Stieglitz, L.; Zwick, G.; Beck, J.; Bautz, H.; Roth, W. Chemosphere 1990, 20, 1953. (6) Jay, K.; Stieglitz, L. Chemosphere 1991, 22, 987. (7) Gullett, B. K.; Bruce, K. R.; Beach, L. O. Chemosphere 1990, 20, 1945. (8) Gullett, B. K.; Bruce, K. R.; Beach, L. O.; Drago, A. M. Chemosphere 1992, 25, 1387. (9) Schramm, K. W.; Wehrmeier, A.; Lenoir, D.; Henkelmann, B.; Hahn, K.; Zimmermann, R.; Kettrup, A. Organohalogen Compd. 1996, 27, 196. (10) Weber, R.; Hagenmaier, H. Chemosphere 1999, 38, 2643. (11) Addink, R.; Govers, H. A. J.; Olie, K. Environ. Sci. Technol. 1998, 32, 1888. (12) Shaub, W. M. Thermochim. Acta 1982, 55, 59. (13) Shaub, W. M. Thermochim. Acta 1982, 58, 11. (14) Shaub, W. M. Thermochim. Acta 1983, 62, 315. (15) Unsworth, J. F.; Dorans, H. Chemosphere 1993, 27, 351. (16) Thompson, D. Chemosphere 1994, 29, 2545. (17) Thompson, D. Chemosphere 1994, 29, 2583. (18) Zimmermann, R.; Wehrmeier, A.; Lenoir, D.; Schramm, K. W.; Kettrup, A. Organohalogen Compd. 1996, 27, 237. (19) Addink, R.; Olie, K. Environ. Sci. Technol. 1995, 29, 1425. (20) Iino, F.; Imagawa, T.; Takeuchi, M.; Sadakata, M. Environ. Sci. Technol. 1999, 33, 1038. (21) Hell, K.; Stieglitz, L.; Zwick, G.; Will, R. Organohalogen Compd. 1997, 31, 492. (22) Iino, F.; Imagawa, T.; Takeuchi, M.; Sadakata, M. Organohalogen Compd. 1998, 36, 89. (23) Iino, F. Ph.D. Dissertation, University of Tokyo, Japan, 1999. (24) Hell, K.; Altwicker, E. R.; Stieglitz, L.; Addink, R. Chemosphere 2000, 40, 995. (25) Hagenmaier, H.; Kraft, M.; Brunner, H.; Haag, R. Environ. Sci. Technol. 1987, 21, 1080. (26) Wiesmu ¨ ller, T. Ph.D. Dissertation, University of Tu ¨ bingen, Federal Republic of Germany, 1990. (27) Morita, M.; Yasuhara, A.; Ito, H. Chemosphere 1987, 16, 1959. (28) Weber, R.; Iino, F.; Imagawa, T.; Takeuchi, M.; Sakurai, T.; Sadakata, M. Organohalogen Compd. 1999, 41, 301. (29) Wehrmeier, A.; Lenoir, D.; Schramm, K. W.; Zimmermann, R.; Hahn, K.; Henkelmann, B.; Kettrup, A. Chemosphere 1998, 36, 2775.
Received for review November 24, 1999. Revised manuscript received April 18, 2000. Accepted April 25, 2000. ES9913131
VOL. 34, NO. 15, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3147
