Document oDj2Lz6JjOkqYv1yZdN8O9j2o

Ct-pmo-inhore Ho. J, op - ?fil, 1973. Perpan-on FYo-s. rrin*ed in Great Hritain FORMATION OF POLYCHLORINATED DIBENZO-p-DIOXINS (PCDDs) AND DIBENZOFURANS (PCDFs) BY BURNING OR HEATING CHLOROPHENATES Chnstoffer Rappe and Stellan Marklund Department of Organic Chemistry, University of UmeA B"901 87 Umei, Sweden and Hans Rudolf Buscr and Hans-Paul Bosshardt Swiss Federal Research Station CH-B620 WSdenswll, Switzerland (Rereivad in UK 23 February 1978; accepted for piblication 2 Karoh 1978) Introduction Thousands of tons of chlorophenols are annally produced for use as wood preservatives (fungicides), as slimacides in paper mills, bactericides In cutting fluids and industrial oils, as herbicides and a = starting materials for a scries of other products like the chlorinated phenoxy acids 2,4-D and 2,4,5-T. Chlorophenols contain a variety of contaminants and byproducts. In ad dition to oilier chlorophenols, they may contain tip to several percent of polychloi limit'd phenoxyphenols (predloxins). Polychlorinated dibenzo-p-dioxins (PCDDs) , dibenzof urans (PCDFs) and diphenyl ethers often aiti present in the range of tens to hundreds of ppm.^ PCDDs arid PCDFs are two groups of compound*; whose presence in chlorophcmls as well as in other ploducts give rise to nuch concern. In case of the HONS 213778 ?69 270 T'n. 1 pCDDs, 75 different Isomers exist ranging from the mono- to the octachloro compound; the corresponding number for the PCDFs is 135. Some of these iso mers have extreme toxicological properties, the most toxic and biologically active Isomers are 2,3,7,8-tetra-CDD, 1,2,3,7,8-penta-CDD, 1,2,3,6,7,8- and 1,2,3,7,8,9-hexa-C'DDs . Moreover, the acute toxicity, biological activity and enzym induction has been found to vary greatly for closely related isomers.^"4 PCDDS can be formed by dimerization of chlorophenates during pyrolysis. Higginbotham et al.5 first reported on the formation of various PCDDs in 1968, but no quantifications were made and the authors did not discuss the possibili ty of positional isomers. The pyrolysis of 2,3,4,6-tetrachlorophenate was stu died by Aniline.6 The product contained two isomeric hoxa-CDDs but no attempts were made to isolate the individual Isomers or to perform a structural deter mination. Buser synthesized a large number of tetra-, penta- and hexa-CDDs 7-9 using micropyrolysis of different chlorophenates. The compounds were cha racterized by high-resolution gas chromatography and mass spectrometry without actual isolation. Large quantities of chlorophenols are used as fungicides in the wood in dustry. Prior to the final use, the top layers with the applied chlorophenabe are removed and the wood shavings burned for heating purpose or used in the plywood production. During this burning, PCDDs could conceivably be formed. In the present Investigation we have identified and quantified PCDDs formed during the uncontrolled open burning of leaves and wood wool Impreg nated with commercial and purified chlorophenates. The analytical technique used has allowed the Identification of individual isomers of PCDDs. So attempts wore made to find the optimal condItIons in burning for the formation of PCDDs and PCDFs. For comparison we have also studied the mioropyrolysls of a few chlo rophenates at 280C. Experimental Compounds The chlorophcnate formulations used without further purification in MQNS 213779 Ho. 3 ?n the burning experiments were Servarex teknisk (GullvikS Fabrics AB, Malmtt, Sweden-, lot Npnr 524 ) and Kyiraiene KY-5 (Kymmene OY, Kuusankoski , Finland; lot 132/16). These formulations contain approximately 5 1 of 2,4,6-trl-, 50 % of 1,3,4,6-tetra- and 10 % of pentachlorophenol os their sodium salts. The le vels Of polychlorinated phenoxyphenols (predioxins) are about 0.2 to 0.5 t of the penta- to octachloro compounds.* The amounts of PCDDs and PCDFs were de termined according to known methods** and found to be 20 and ISO ppm, respec tively (see Table 1 and Results and Discussion). 2,4,6-Trl- and pentachlorophenol of an Isomeric purity of > 99 l were prepared by multiple recrystallisations. no predioxins were detected (< 1 ppm). The levels of PCDDs ate reported in Table 2. Both compounds were used In the burning experiments as their sodium salts. Burning and sampling procedures The chlcrophenates (1.0 g) were dissolved in 20 ml of water and sprayed over 30 g birch leaves or wood wool. After spraying, the materials were dried overnight at room temperature. For burning, they were placed on a metal screen in a glass bowl, this allowed an air stream from the sides. The materials were ignited and burned spontaneously in open fires. The smoke gases were drawn through a glass funnel into a glass filter using a vacuum pump (capacity: 100 1/min). Pumping was started before ignition and was terminated 15 min after burning ceased. The filter contained 15 g of charcoal (Norit PK, 0.5-1 mm) or Ambetlite XAD-2 (20-50 mesh), both purified by Soxhlet extraction (methylene chloride, 24 hours). After burning, the glass equipment was washed with 150 mi of methylene chloride; the adsorbant was Soxhlet extracted for 10 hours using this washing solution. Sample purification Aliquots (25 %J of the burning extracts were concentrated and purified on silica gel micro-columns (o.5 g silica gel, 70-230 mesh, Merck, in 150 x 5 mm disposable Pasteur pipette) using 6 ml of n-iiexane as eluant. Further puri fication was carried out on alumina micro-columns (1,0 g basic alumina, Woelm, in Pasteur pipette) by elution with 10 ml each o 2 i and 50 % methylene chlo ride in n-hexane.** The 50 ^-fractions containing the PCDDs and rcor* were HONS 213780 Z1Z No. 3 concentrated and the residues redissolvod In 500 pi of n-tctradecane. Aliquots of 2 pi were used for analysis. Mlcropyrolysls Samples of 10 mg of chlorophenates were subjected to micropyrolyses at 280 C for 30 minutes. The exact experimental conditions and the clean-up procedure used have been previously described.^ GC-HS Analysis A Finnigan 4000 quadrupole GC-MS instrument equipped with El source (70 eV, 250 C) and a Finnigan Sill data system was used. An OV-17 glass ca pillary column (50 m, 0.36 mm ID) was coupled via a platinum capillary inter face leading directly into the ion-source. Sample aliquots of 2 pi were Injec ted splltlessly (vaporizer temperature 270 C) and the column programmed from 200 to 240 C with 2 C/min. The helium carrier gas pressure was 1.40 at. The samples were analyzed for the presence of PCDDs and PCDFs using mass specific detection (mass fragmentography) by continuously monitoring molecular ions at m/e 320, 354, 388, 422 and 456 for the tetra- to octa-CDDs and at m/e 304, 338, 372, 406 and 440 for the tetra- to octa-CDFs. Two runs for a sample were required. The detection limits for the tetra- to octachloro compounds in the present case ranged from Q.Q1 to Q.2Q pg/g chlorophenate. Some of the samples were reanalyzed by recording complete mass spectra (m/e 35-460, 1.4 sec/scan} using the data system. Results and Discussion In our burning experiments we selected two of the most commonly used chlorophenol formulations on the Scandinavian market. They both consist of a mixture of 2,4,6-tri-, 2,3,4,6-tetra- and pentachlorophenale. The levels of PCDDs found In the two formulations are similar (see Table 1) and this Is also the case for the PCDFs (see below). Wood wool and birch leaves were impregnated with these formulations, ignited and burned. The burning of the leaves proceed slower and at a lower temperature with formation of more smoko and steam than the burning of the HONS 213761 Ho. 1 273 wood wool. Several individual burnings were carried out and the smoke gases were analyzed for PCDDs and PCDFs after adsorption on a charcoal or Amberllte XAD-2 filter. These two adsorbants were selected to study the efficiency for the adsorption of these compounds. Amounts of PCDDs formed In the burnings In Table 1 we report the amounts of PCDDs found after the burning of impregnated wood wool and birch leaves. It can be seen that the amounts of PCDDs increased and that PCDDs of all degrees of chlorination are formed. Table 1. Amounts of PCDDs found in burning experiments of commercial chlorophenates (pg PCDDs / g chlorophenate] SE RVARE X KY -5 original birch leaves wood wool original birch leaves sample charcoal ash XAD-2 charcoal XAD-2 sample charcoal Tetra-CDDs Penta-CDDa Hexa-C'DDs Hepta-CDDs Octa-CDD 0.7 5 .2 9.S 5.6 0.7 35 17 26 96 210 0.4 90 58 59 120 357 3 .5 80 74 57 110 347 5.3 e 18 S 65 29 2.1 0.3 6.4 0.2 1.2 1.2 0.3 30 84 82 8.2 0.4 The burning of Impregnated birch leaves did yield very similar amounts of PCDDs for both formulations when trapping on charcoal was used. Somewhat lower values were found when using Amberllte XAD-2, but this is not considered to be significant. The burning of birch leaves did yield considerable amounts of ash. Ana lysis of ash from a Servarex burning indicates a similar amount of PCDDs as in the smoke gases with some preference of the higher chlorinated and somewhat smaller amounts of the lower chlorinated dioxins. MONS 213782 m Ho. i The burning of Servarex was also carried out using wood wool and the smoke gases were trapped on charcoal or XAD-2. The amounts of PCDDs formed in these experiments were higher with both adsorbents than in the burning of birch leaves. The higher amount of PCDDs found In the XAD-2 filter as compared to the charcoal filter could be due to some variations in the burning condi tions. In addition to the PCDDs reported in Table 1, we have also found some tri-CDDs, which however were not identified or quantified. The results from the burning of purified 2,4,6-tri- and pentachlorophenate are given in Table 2. In these experiments we used only birch leaves and trapping of the smoke gases on charcoal filters. The levels of PCDDs dramati cally increased during these burnings. In the case of the 2,4,6-trlchlorophenate, the dominating PCDDs are tetra-CDDs. In the case of the pentachlorophenate the highest level was found for the octa-CDD, but the amounts of hexaand hepta-CDDs were found to exceed 50 pg/g pentachlorophenate. Table 2. Amounts of PCDDs found In burning experiments of purified chlorophenates (pg PCDDs / g chlorophenate) 2,4, 6-Trichloro- Pentachloro- phenate phenate original sample birch leaves charcoal Orlalnal sample birch loaves charcoal Tetra-CDDs Penta-CDDs liexa-CDDs llepta-CDDS Octa-CDD < 0.02 < 0.03 < 0.03 < 0.1 < 0.1 2100 5.0 1 .0 3.0 6.0 < 0.02 < 0.03 < 0.03 0.3 0.9 5.2 14 56 172 710 HONS 213783 No. J 275 Table 3. Identification of PCDCsln burning extracts (see Figure 1) Peak PCDD isomer Peak PCDD Isomer 1 2 3 4 5 6 7 s' 9' 10 . 11 12 1,3,6,7-tetra-CDD 1,3,7,9- 2,3,7,8- 1,2,3,0- a) 1,2,6,9- 1,2,6,7- 1,2,0,9- penta-CDD* from pyro- lysis of 2,4,6- trl- and 2,3,4,6-tetrachlorophenate 1,2,4,7,8-penta-ODD 1,2,3,4,7- 13 14 15 16 17 18 19 20 21 22 23 24 1,2,3,7,8-penta-CDD 1,2,4,6,7,9-hexa-CDD*^ 1,2,3,4,6,81,2,3,6,8,9- cl 1,2,3,4,6,9- 1,2,3,4,7,81,2,3,6,7,8- 1,2,3,7,8,91,2,3,4,6,7- 1,2,3,4,6,7,9-hepta-COO 1,2,3,4,6,7,8octa-CDD a) 'or the Smiles product 1,2,3,7-tetra-CDD b) or the Smiles product 1,2,4,6,8,9-hexa-CDD c) or the Smiles product 1,2,3,6,7,9-hexa-CDD Identification of PCDDs formed in the burnings The analytical technique used allowed the identification of individual isomers of PCDDs. A typical run from the burning of Servarex on birch leaves is given in Figure 1 a. In the case of the major Isomers, more information concerning discrete PCDDs and PCDFs could be obtained by running complete El mass spectra. These showed intense molecular ions with the expected ion clustering due to the chlo rine Isotopes and the known chaiacteiistic fragmentation.^ HONS 213784 Figure 1 Mass fragmentograms (50 m OV-17 glass capillary column, 200-240 C) showing elution of PCDDs from the burrung3 on birch leaves of a) Servarex, b) purified pentachlorophenate and c) purified 2,4,6-trlchlorophenate; m/e 320, 354 etc. for the tetra- to octa-CDDs. Peak identification sec Table 3. Ho. HONS 21375 No. i 277 ft comparison of retention times of PCDDs in the burning extracts with synthetic PCDD standards made it possible to identify the major constituents. In addition, the substitution patterns were checked by studying ions in the g low mass range as recently described. This technique gave additional sup port for the assignment of the various PCDDs given in Fig.1 and Table 3, The substitution pattern was defined as the number of chlorine atoms in each car bon ring of a dioxin molecule. Among the tetra-CDDs in Fig.l a, we can eee up to 14 of the theoretical ly possible 22 isomers. The major components were found to have a 2:2 substi tution pattern (two chlorine atoms in each carbon ring). The main isomers are the 1,3,6,8- and 1, 3,7,9-tetra-CDD, both of which are found comparably little biologically active. 2-4 These two Isomers are the normal dimerization7-9 and the Smiles rearrangement 14 ' 15 products f^wn 2,4,6-trichlorophenate. A minor peak has the same retention time as the highly toxic 2,3,7,8-tetra-CDDj in addition this peak also showed the 2:2 chlorine substitution pattern (peak 3). Among the penta-CDDs, we can see up to 9 of the theoretically possible 14 isomers. With one exception,al1 of them had a 3:2 chlorine substitution pattern. However, there are few penta-CDD standards available; one peak, has the same retention time as the highly toxic 1,2,3,7,8-p*nta-CDO (peak 13). Among the hrxa-CODs (10 theoretically possible isomers), we see 5 iso mers in medium to higher concentrations (3 additional isomers are present in minute quantities as well). The major hexa-CDD is the 1,2,3,6,8,9-subst1tuted isomet; the others are 1,2,4,6,7,9-, 1,2,3,4,6,8-, 1,2,3,6,7,8- and 1,2,3,7,8,9-hexa-CDD. 1,2,3,4,6,8-Hexa"CDD is a condensation product of 2,4,6-trl- and a pcntachlorophar.te with a 4.2 chlorine substitution pattern, the other Isomers have 3:3 substitutions and are dimerization products of 2,3,4,6-tetrachlorophenate. The most toxic of these Isomers are considered to be 1,2,3,6,7,8(pcak 19) and 1,2,3, 7,8, 9-hexa-CDD (peak 20).2 Both the hepta-CDDs and octa-CDL) weLe also present. Typical runs from the burning of the purified chlorophonates are given in Fig. 1 b and 1 c. In the case of the 2,4,6-trlclilorophenate (rig, 1 c) the only letra-CDDs were the 1,3,6,8- and 1,3,7,9-isomers. In the case of the MQNS 213786 278 No. 3 pentachlorophenate (Fig. 1 b) , the octa-CDD was the main product In addition to the octa-CDD we found both hepta-CDDs, eight hexa-CDDs and several pentaand tetra-CDDs. In the formation of the lower chlorinated PCDDs a nonspecific dechlorination process must be Involved. A comparison of the chromatograms In Fig. 1 reveals that the distribu tion of the individual iBOmers Is different in the burning of technical pro ducts ( Fig.i a) s compared to the purified samples (Fig. 1 b and c) . The ob vious dimerization products (peaks 1.2 and 24) dominate to a higher degree in the case of the purified samples. It is worth noticing that very similar chromatograms were obtained in burning experiments of Servarex and in micropyrolyses of the same formulation, Most of these pcdd Isomers are also found In fly ash samples. 8 ' 9 PCDF8 In starting materials and burning extracts Using the same technique as described for the PCDDs above, we have quan tified the PCDFs found in the starting materials and burning extracts. Due to the lack of suitable PCDF standards, no identification of the discrete Isomers was possible. In both formulations 40-S0 Individual Isomers were piesent, while the burning extracts contained less than 10, but In part some of these were newly formed. The main PCDFS In the starting materials are the hexa- and hepta-CDFs present at levels of 60-70 pg/g each, and the amounts found for the tetra-, penta- and octa-CDF were about 10 pg/g. The same levels and the same Isomers were found in the two formulations studied. After burning, the levels of most PCDFs were generally much decreased (as opposed to PCDDs) but the level of a few Individual l'CPFs nevertheless In creased, *.3. the major tetra-CDF (unknown isomer) Increased during burning more than 100-fold (from 0.04 to S pg/g). Two of these isomers were not at all found In the starting materials. 2,3,7,8-Tetra-CDF, which is considered to be the most toxic of all PCDFs,was only a minor component in all these samples, in the starting material less than 0.1 4 of the total tetra-CDFs. The same PCDFs were formed in the micropy1olysis of Servarex as in the burning experiments, but from the burning and the micropyrolyses of the HONS 213787 No. 3 279 purified chlorophenates no PCDFs could be found. Consequently, the newly formed PCDFs are formed from Impurities In the technical products, probably in reactions with higher absolute yields than those leading to PCDDs. Conclusions From a mechanistic point of view the PCDDs found in the burning extracts can be formed in three different ways: A. by dimerization of chlorophenates, B. by dechlorination of higher chlorinated PCDDs and C. by cyclization of predioxins. The fact that the purified trl- and pentachlorophenates also yield large amounts of PCDDS strongly suggests a dimerization reaction. The fact that hep- ta-CDDs and lower chlorinated PCDDs are formed from pure pentachlorophenate sug gests that a dechlorination reaction is also involved. The thermal cyclization of predloxins has been reported previously. 17 It seems quite plausible that the three reactions A, B and C yield various and different PCDD isomers. An examination of the kinetic situation reveals that reaction A, which is a dimerization should be a bimolecular reaction, while dechlorination (B) and cyclization (C) are monomolecular. Consequently, the ratio of the iso mers formed in burning reactions should be dependent on the chlorophenate con centration. The dimerization reaction (A) gives lower yields In more diluted systems. The concentrations used In the burnings here are quite high, and further work Is needed to verify that hypothesis. However, it Is worth noti cing that within an evaporating droplet the concentration of nonvolatile com ponents gradually increases resulting in spotwise high concentrations. Conse quently, the local concentration is not as low as indicated by the mean concen tration found. Other variables in the burnings are temperature and flow rate of air. Both arc difficult to control during open burning. However, as mentioned pre viously, no attempts were made to find the optimal conditions for the formation of I'CDDs and PCDFs. HONS 213788 Of special interest is the observation of PCDD isomers with the same re tention time and chlorination pattern as the highly toxic 2,3,7,8-tetra-CbO and 1,2,3,7,8-pcnta-CDD. Although they are found to be only minor constitu ents, in individual burnings both have been found at levels exceeding 10 *q/q chlorophenate. It can also be pointed out that these two isomers cannot bo formed In the expected dimerization reactions (A>, consequently they are for med In dechlorination reactions (B) or cyclizations of prcdioxins (C> , present as impurities in the commercial formulation .No 2,4,5-trichlorophenol could be detected in the commercial formulations. It Is also of interest to compare the formation of PCDOs reported on here with the potential formation of 2,3,7,8-totra-CDO from various 2,4,5-T precur sors, a question associated with some controvereory in the past, Previously, we reported that no 2,3,7,8-tetra-CDJ could be found by burning varlou1. samples of vegetation, sprayed or spiked with 2,4,5-T ester or salts. 18 The combustion gases, Boot particles and ashes were analyzed by the same technique as used here, and the detection limit was 4 ug 2,3,7,8-tetra-CDD/g 2,4,5-T burned. Our results are supported by the recent publications of Stehl et al. on burning vegetation sprayed with purified 2,4,5-T esters and salts, and by Ahling e_t a_l. on burning a 2,4,5-T ester formulation. 5 A The burning of chlorophenates or material like wood shavings, plywood or waste oil containing chlorophenates seems to be a more Important source to environmental pollution by PCDDs Including 2,3,7,8-totra-CDD than the acciden tal burning of 2,4,5-T derivatives or vegetation treated with 2,4,5-t derivati ves . References 1. C.-A. Nilsson, A. Norstrbm, k. Andersson and C. Rappe in Pcntachjoro- phenoli Chemistry, Pharmacology and Environmental Toxicolog'/, R. Rangn Rao (Ed.), R. Eiiv. Set, Res. Series Vol. 11 (1978), Plenum Press. 2. E.E. McConnoll, J.A. Moore, J.K. Hascman and M.w Harris, Tox. Anul Phorm., 37, 146 (1976). . MONS 213789 Ho * 3 281 3. A. Poland, E. Glover and A S. Kende, J. Biol. Chom., 251, 4936 (1976). 4 . J. Bradlaw, FDA, Washington, USA. Personal Communication, 1976 , 5 . G.R. Higginbotham, A. Huang, D. Firestone, J. Verrett, j. Ress and A.D Campbell, Nature, 220, 702 (1968). 6. 0. Aniline In Chlorodloxins - Origin on Fate, E.H. Blair (Ed.) Ad v. Chem. Series, 120, 126 (1973). 7. H.R. Buser, J, Chromatogr., 114, 95 (1975). 8. H.R. Buser, 11.-P. Bosshardt and c. Rappe, Chemosphere, In press. 9. H.R. Buser and C. Rappe, Submitted to chemosphere. 10. J.O. Levin, C. Rappe and C.-A. Nilsson, Scand. J, Work Environ. Health. 2, 71 (1976) . 11. H.R. Buser and H.-P. Bosshardt, j. Ass. Of fie. Anal. Chem.. 59, 562 (1976) . 12. H.R. Buser, J. Chromatogr., 107, 295 (1975). 13. J.O. Levin, C.-A. Nilsson and K. Andersson, Chenosphere, 1977, 595. 14. A.P. Gray, S.P. Cepa and J.S. Cantrell, Tetrah. Letters, 1975, 2873. 15. A.S. Kende and M.R. oe Camp, Tetrah. Letters, 1975, 2877. 16. J.A. Goldstein, J.D. McKinney, G.W. Lucier, p. Hickman, H. Bergman and J A, Moore, Toxicol, Appl. Pharmacol., 36, 81 (1976). 17. C .-A. Nilsson, K. Andersson, C. Rappe and S.-O. Westermark, J. Chromatoqr., 96, 137 (1974). 18. K. Andersson, H.-P. Bosshardt, H.R. Buser, S. Marklund and C. Rappe in Chlorinated Phenoxy Acids and Their Dioxins, C. Ramel (Ed.) Ecol. Dull. 27, 26 (1978) . 19. R. Stehl and L.L. Lamparski, Science, 197, 1008 (1977). 20. B. Ahllng, A. Llndskog, B. Jansson and G. Sundstrbm, Chemosphere, 1977, 461. HONS 213790