Environ. Sci. Technol. 2005, 39, 3799-3804
Elucidation of Degradation Mechanism of Dioxins during Mechanochemical Treatment YUGO NOMURA,* SATOSHI NAKAI, AND MASAAKI HOSOMI Strategic Research Initiative for Survival Paths, Graduate School of Technology, Tokyo University of Agriculture and Technology, 2-24-16 Naka, Koganei, Tokyo 184-8588, Japan
Model dioxin compounds 4-chlorobiphenyl (4CB), octachlorodibenzo-p-dioxin (OCDD), and octachlorodibenzofuran (OCDF) were degraded by a mechanochemical (MC) process that involved milling with calcium oxide by use of a planetary ball mill. The degradation of 4CB produced mainly chloride ions and biphenyl, with the chlorine removal efficiency reaching about 100%. Biphenyl was transformed into terphenyls, quaterphenyls, cyclohexylbenzene, and bicyclohexyl through polymerization and hydrogenation reactions. Measurements of chloride ions after MC treatment of OCDD and OCDF showed about 100% dechlorination of both compounds; tetra- to heptachlorinated dibenzo-pdioxins/furans (T4-H7CDD/Fs) were detected only at trace levels, and no other chlorinated organic compounds were observed. The residue after MC treatment was gray in color, indicating the possibility of carbonization, but the presence of amorphous graphite could not be confirmed.
Introduction Persistent organic pollutants (POPs) are a well-known threat to the environment, and the Stockholm POPs Convention declared that waste containing POPs such as dioxins and polychlorinated biphenyls (PCBs) should be decomposed in an environmentally sound manner. The Japanese dioxins lawswhich defines dioxins as polychlorinated dibenzo-pdioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and coplanar polychlorinated biphenyls (co-PCBs)sestablished dioxin environmental quality standards for ambient air, water, soil, and sediment. In particular, countermeasures include detoxification and decomposition for dioxincontaminated soil and sediment when concentrations exceed environmental quality standards of 1000 and 150 pg TEQ/g, respectively. We have focused on investigating mechanochemical (MC) treatment of contaminated wastes because it is a noncombustion technology that requires no heating process or offgas treatment that we expect will be more acceptable to the Japanese public than conventional heating processes. In MC reactions, the compounds to be treated are milled with chemicals such as calcium oxide (CaO) by means of a planetary ball mill. Organochlorine compounds such as 1,1,1trichloro-2,2-bis(p-chlorophenyl)ethane (DDT; 1), chlorinated benzenes (2-5), poly(vinyl chloride) (PVC; 6, 7), dioxins (8), and PCBs and pentachlorophenol (5, 9) have been * Corresponding author phone: (81) 42-388-7070; fax: (81) 42381-4201; e-mail:
[email protected]. 10.1021/es049446w CCC: $30.25 Published on Web 04/02/2005
2005 American Chemical Society
successfully degraded by MC treatment. Since MC reactions occur in the solid phase, it is reasonable to expect that dioxins could be degraded in soil or fly ash, and in fact Shinme (8) recently demonstrated that MC treatment with CaO can degrade dioxins in fly ash. To investigate the feasibility of MC treatment of organochlorine compounds, the degradation pathway must be determined. However, no quantitative data on the degradation behaviors of dioxins, such as chlorine balance, have been reported. The present study elucidates the degradation pathway of dioxins by MC treatment of three model compounds, 4-chlorobiphenyl (4CB), octachlorodibenzo-pdioxin (OCDD), and octachlorodibenzofuran (OCDF), by means of qualitative and quantitative analysis of organic and inorganic degradation products.
Materials and Methods Chemicals and MC Reactor. CaO (purity 99%; Kanto Chemicals Co., Japan) was used as the additive for MC treatment, and 4CB (purity, 99%; Tokyo Kasei Co., Japan), OCDD, and OCDF (purity 99%; Wako Pure Chemical Industries, Co., Japan) were used as model compounds for co-PCBs, PCDDs, and PCDFs, respectively. Prior to use, CaO was treated at 800 °C for 2 h. The MC reactor was a planetary ball mill (Pulverizette-7, Fritsch, Germany) with two stainless steel pots (45 cm3) in which seven steel balls (15-mm diameter) were arranged. The pots are situated on a rotating disk, which enables the pots and the disk to rotate in opposite directions. MC Treatment. Figure 1 shows a schematic diagram of the MC treatment process. Under atmospheric pressure, the respective model compounds and CaO were added to the pots at weight ratios of 1:20 for the treatment of 4CB and 1:200 for each treatment of OCDD and OCDF. The planetary ball mill was operated at 700 rpm for 15-min intervals, with a 15-min cooling period after each interval. Analytical Method. A portion of each milled mixture was collected from the pots and agitated in distilled water at 80 °C for 30 min by use of a magnetic stirrer, and then the suspension was subjected to ultrasonic treatment for 30 min. The supernatant was filtered through a glass fiber filter (GF/ F, Whatman), and the residue was extracted two more times in the same manner. The filtrates were analyzed for chloride ions by means of ion chromatography (IC 7000, Yokogawa Analytical Systems Co., Japan) to confirm dechlorination of the model compounds. As a control experiment, a hot-water extract of a nonmilled 4CB and CaO mixture was also analyzed for chloride ions. A second portion of each milled mixture was extracted with toluene (purity 99.5%; Wako Pure Chemical Industries, Co., Japan) by (i) reciprocal shaking at 200 rpm for 4 h for analysis of organic compounds or (ii) Soxhlet extraction for 16 h for analysis of PCDD/Fs. The extracts were concentrated individually in a rotary evaporator and then evaporated to 100 µL under a gentle stream of nitrogen. Finally, 1 µL of each sample was analyzed on (i) a capillary column (HP5-MS, Agilent, USA) with detection by low-resolution gas chromatography/mass spectrometry (LR-GC/MS; HP 5973 MS selective detector, Agilent, Germany) to identify and quantify 4CB and the degradation products of the model compounds or (ii) capillary column (SP2331, Spelco, USA; DB17, J&W, USA) with detection by high-resolution gas chromatography/mass spectrometry (HR-GC/MS; JMS-700, JEOL, Japan) to quantify hexa-tetraand hepta-octachlorodibenzodioxins and -furans, respectively. In the analysis of PCDD/Fs, 13C-labeled PCDD/Fs were VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3799
FIGURE 1. Schematic of MC treatment with a planetary ball mill.
FIGURE 3. LR-GC/MS chromatogram of the degradation products obtained by MC treatment of 4CB for 2 h.
FIGURE 2. Chlorine balance during MC treatment of 4CB.
TABLE 1. Chlorine Removal Efficiency Based on Analyses of Chloride Ions after MC Treatment of OCDD/F for 2 h
OCDD OCDF
chlorine removal (%)
first extracted fraction (%)
second extracted fraction (%)
third extracted fraction (%)
99.3 99.9
66.6 56.0
31.1 42.3
1.57 1.59
used as internal standards. The dioxins were analyzed according to the standard manual in Japan (10). Degradation products were identified by spike tests and by comparing their mass spectral patterns with patterns stored in the U.S. National Institute of Standards and Technology (NIST) mass spectral library (v. 2).
Results and Discussion Dechlorination. The MC treatment of 4CB with CaO was performed for 8 h, during which time the amounts of chloride ions and remaining 4CB were measured; the percentage of chlorine removal was calculated from the measured amount of chloride ions and the amount of chlorine initially added as 4CB. As the MC treatment proceeded, the chlorine removal ratio increased to 100%, and after 2 h, the 4CB concentration was below the detection limit (Figure 2). Since no chloride ions were detected in the hot-water extract of the nonmilled 4CB and CaO mixture (data not shown), these results confirmed the complete removal of chlorine from 4CB by the MC treatment. In addition, the chlorine balance over the course of the MC treatment of 4CB (Figure 2) indicates that no other organochlorine compounds were produced. For the MC treatment of OCDD/F, 10 mg of each compound was used so that the amount of chloride ions would be sufficient for detection by ion chromatography. The chlorine removal efficiencies from OCDD/F after 2 h of treatment were calculated on the basis of the amount of chloride ions recovered after three extractions (Table 1). As was the case for 4CB, the MC treatment removed 100% of the chlorine from OCDD/F. Note that this is the first study to demonstrate chlorine removal efficiencies of PCDD/Fs by measuring the amount of chloride ions produced during degradation. No remaining dioxins or other organochlorine 3800
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005
FIGURE 4. LR-GC/MS chromatogram of the degradation products obtained by MC treatment of biphenyl for 1 h. compounds were detected, which confirms the complete dechlorination of OCDD/F. Degradation Products. GC/MS analysis of the degradation products of 4CB after 2 h of MC treatment revealed that biphenyl, cyclohexylbenzene, terphenyl, and quaterphenyl were the main products (Figure 3). Their presence was confirmed by comparing the mass spectrum patterns with patterns in the U.S. NIST database and by spike tests, with one exception: the presence of quaterphenyls was confirmed solely on the basis of their fragment patterns because only the para isomer (4-quaterphenyl) is commercially available. The detection of biphenyl confirmed that chlorine was removed from 4CB by MC treatment, and the production of cyclohexylbenzene, terphenyls, and quaterphenyls indicated that hydrogenation and polymerization reactions occurred. To determine the degradation pathways of 4CB, biphenyl was subjected to MC treatment in the same manner as for 4CB; the amount of biphenyl added to each pot was equal to that of 4CB in terms of molar ratio. Cyclohexylbenzene, terphenyl, and quaterphenyl were the main degradation products of biphenyl (Figure 4), as for 4CB. Bicyclohexyl was also produced, as confirmed by spike tests (data not shown). To determine the pathway of bicyclohexyl formation, we subjected cyclohexylbenzene to MC treatment with CaO and found that bicyclohexyl was produced (data not shown). Bicyclohexyl was not detected during MC treatment of 4CB, possibly because the amount produced was too small. These results, along with the chlorine balance during MC treatment of 4CB and biphenyl (Figure 2), suggested that biphenyl was the major degradation product of 4CB and that the biphenyl produced the additional degradation products; 100% conversion of 4CB to biphenyl was impossible. Because previous
FIGURE 5. Time dependence of the molar ratio of degradation products during MC treatment of 4CB.
FIGURE 6. Time dependence of the molar ratio of degradation products during MC treatment of biphenyl. researchers have reported that radicals are produced by electron transfer in MC reactions (11-14), it is natural to hypothesize that biphenyl radicals would be produced as major intermediates by means of chlorine abstraction from 4CB. The identified degradation products of 4CB and biphenyl were quantitatively analyzed by LR-GC/MS via the selectiveion monitoring mode. Figures 5 and 6 show the molar ratio of each degradation product with respect to the initial amount of 4CB or biphenyl. The working curve for quantifying quaterphenyls was obtained by analyzing various amounts
of 4-quaterphenyl as a reference compound. During MC treatment of biphenyl, bicylohexyl was not quantified because the amount produced was too small. Biphenyl was the major degradation product of 4CB (Figure 5). Note that quaterphenyls, terphenyl, and cyclohexylbenzene were produced during the degradation of both 4CB and biphenyl, and the changes in their molar ratios with time were similar in terms of concentration and preference of the production of further degradation products. The results confirmed that degradation of 4CB by MC treatment mainly produced biphenyl. Two degradation pathways occur for biphenyl: polymerization and cleavage of the bond between the benzene rings. These were confirmed by the formation of terphenyl and quaterphenyl. Several reports indicate that polymerization occurs during MC treatment (1, 11, 15, 16), particularly during treatment of organochlorine compounds such as DDT (1). Comparisons between the structures of polymerized products and the structures of their respective parent compounds suggest that this polymerization reaction might proceed via radicals. The chlorine balance during degradation of 4CB (Figure 2) indicates that biphenyl radicals could be produced by chlorine abstraction, and subsequent polymerization of two biphenyl radicals would produce quaterphenyls. Similarly, benzene radicals might be produced by cleavage of the bonds between the phenyl groups. In fact, MC treatment of 4-quaterphenyl produced biphenyl and terphenyls (data not shown). Since the amount of terphenyls formed was less than that of quaterphenyls during MC treatment of 4CB (Figure 5), biphenyl radicals might be produced in larger quantities than benzene radicals. Accordingly, terphenyls may be derived from quaterphenyls, although they could also be produced by polymerization of biphenyl and benzene radicals. Formation of cyclohexylbenzene and/or bicyclohexyl during MC treatment indicates that hydrogenation reactions occurred, and this possibility raises the question of the identity of the hydrogen donors. Possible hydrogen donors include (i) H2O in the additive reagent and/or in the atmosphere and/or (ii) the organic compounds in the mill,
FIGURE 7. Comparison of fragment patterns of cyclohexylbenzene (a) and a degradation product of deuterated biphenyl (b) occurring at the retention time of cyclohexylbenzene. VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3801
FIGURE 8. Minor degradation products obtained by MC treatment of p-terphenyl. (The compounds mentioned in the text are shown in boldface type.)
including the target compounds and their degradation products. Although the reagent CaO was heated at 800 °C prior to use, trace amounts of water might be absorbed on the CaO during transfer or storage. The existence of such a hydrogen donor must be verified in future research. To determine whether organic compounds acted as the hydrogen donors, a deuteride of biphenyl (biphenyl-d10) was subjected to MC treatment with CaO. The MC treatment of biphenyl-d10 produced many degradation products having LR-GC/MS elution profiles (data not shown) similar to those observed for MC treatment of bicyclohexyl, cyclohexylbenzene, terphenyls, and quaterphenyls (Figure 4). For instance, the fragment pattern of the peak occurring at the retention time for cyclohexylbenzene resembles the pattern of cyclohexylbenzene (Figure 7). Note, however, that the available mass spectrum library fragment pattern does include a deuteride of cyclohexylbenzene. In addition, each fragment mass of this product is heavier than that of cyclohexylbenzene. If a deuteride of cyclohexylbenzene was produced by complete addition of deuterium, the mass of its parent ion should agree with that of a deuteride of biphenyl plus six deuterium atoms. However, the mass of the parent ion was slightly smaller than the mass of a deuteride of cyclohexylbenzene (cyclohexylbenzene-d16), which indicates that the existence of small amounts of hydrogen as impurities in the deuteride reagent resulted in this product. Thus, it is likely that organic compounds participated in the reaction as hydrogen donors, though 100% conversion was not observed. Identification of the degradation products of the deuteride
FIGURE 9. Proposed pathways for degradation of 4CB by MC treatment (*, not detected; **, detected but not confirmed by spike test). 3802
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005
FIGURE 10. Time dependence of the amounts of OCDD and its degradation products during MC treatment.
FIGURE 11. Time dependence of the amounts of OCDF and its degradation products during the MC treatment.
remains as an important task for elucidating the hydrogen donation mechanism. For both 4CB and biphenyl (Figures 5 and 6), the total molar ratio of confirmed products decreased with increasing milling time, which indicates that the amounts of unidentified degradation products increased during MC treatment, providing mineralization did not occur. One possible explanation for this might be the production of high-molecular-weight compounds that exceeded the maximum measurable mass of the LR-GC/MS 550 system. Since the amounts of terphenyls and quaterphenyls formed, that is, polymerized products, were small compared to the amounts of the parent compounds and since the amounts decreased toward the end of MC treatment, it is reasonable to expect that the amounts of such undetectable polymerized products would be quite small. We therefore surmise that there were numerous lowmolecular-weight compounds that were unquantifiable or undetectable by means of LR-GC/MS. In fact, MC treatment of p-terphenyl produced various minor degradation products (Figure 8), which supports this hypothesis. The occurrence of 4-isopropylbiphenyl, sec-butylbiphenyl, 1,4-diphenylbutane, and 4-pentylbiphenyl indicates that the benzene ring(s) of p-terphenyl were cleaved by MC treatment. Accordingly, this suggests that MC treatment of organic compounds may also result produce lower-molecular-weight compounds by cleavage of C-C bonds. Reports that methane and ethane are produced by MC treatment of chlorobenzenes with CaH2 or CaO (2, 4) support our hypothesis. We propose that MC treatment of 4CB leads primarily to biphenyl radicals by means of chlorine abstraction (see Figure 9). The biphenyl radicals can be reduced by hydrogencontaining compounds in the reaction mixture or by other hydrogen sources to yield biphenyl in about 40% yield (Figures 5 and 6, 1 h). Small amounts of reduced cyclohexyls and bicyclohexyls are produced in a similar manner. Other reaction products include terphenyls, quaterphenyls, and low-molecular-weight compounds. During MC treatment of OCDD and OCDF with CaO, more than 95% of the OCDD and OCDF degraded rapidly, in the first 6 min (Figures 10 and 11); tetra-, penta-, hexa-, and heptachlorodibenzodioxins and -furans were produced and subsequently degraded in both systems. Also note that the amounts of these products peaked at the same time. Such degradation behavior suggests that a consecutive dechlorination reaction did not occur. The position of chlorine in the molecule may influence chlorine removal properties: removal of one Cl may make removal of another Cl much easier (even immediate) or slower. The amounts of formed hepta-tetrachlorodibenzodioxins and -furans were quite small compared to those of the parent compounds, OCDD and OCDF, and further degradation of these lower-chlorinated compounds progressed slowly after 15 min. Another possible degradation pathways of OCDD/F may be cleavage of the dibenzodioxin/furan structure; that is, MC treatment of OCDD/F produced intermediates, such as chlorinated
phenols and quinones, resulting from cleavage of the dibenzodioxin/furan structure. Small amounts of hepta- to tetrachlorodioxins and -furans may result from some other hydrogen donor in the system, such as water adsorbed on the CaO during transfer and storage. The residues in the pots during MC treatment of 4CB, OCDD, and OCDF were gray in color. As the MC treatment proceeded, the color deepened and finally turned black. We therefore expected graphite to be the final degradation product. To verify this possibility, we subjected the residues to X-ray diffraction analysis, but we could not confirm the presence of graphite. If graphite were present in an amorphous form, it would not be detected by X-ray diffraction analysis, which requires a regular crystal arrangement. In fact, when we milled commercial graphite with CaO, the graphite peaks in the X-ray diffraction pattern vanished (data not shown). Although the final product has not yet been identified, the possibility nevertheless exists that an organic compound was carbonized. The identification of the final products is essential for understanding the mechanism of the MC degradation of dioxins and for determining the mass balance during the process. By measuring the amounts of chloride ions produced, we have demonstrated that MC treatment completely dechlorinated dioxin model compounds (4CB, OCDD, and OCDF). These results show that mechanochemical treatment is an effective noncombustion technology for dioxin degradation that does not produce organochlorine compounds as reaction products.
Literature Cited (1) Hall, A. K.; Harrowfield, J. M.; Hart, R. J.; Maccormick, P. G. Mechanochemical reaction of DDT with calcium oxide. Environ. Sci. Technol. 1996, 30, 3401-3407. (2) Loiselle, S.; Branca, M.; Mulas, G.; Cocco, G. Selective Mechanochemical dehalogenation of chlorobenzenes over calcium hydride. Environ. Sci. Technol. 1997, 31, 261-265. (3) Cao, G.; Doppiu, S.; Monagheddu, M.; Orru, R.; Sannia, M.; Cocco, G. Thermal and Mechanochemical self-propagating degradation of chloro-organic compounds: The case of hexachlorobenzene over calcium hydride. Ind. Eng. Chem. Res. 1999, 38, 3218-3224. (4) Tanaka, Y.; Zhang, Q.; Saito, F. Mechanochemical dechlorination of trichlorobenzene on oxide surfaces. J. Phys. Chem. B 2003, 107, 11091-11097. (5) Birke, V.; Mattik, J.; Runne, D. Mechanochemical reductive dehalogenation of hazardous polyhalogenated contaminants. J. Mater. Sci. 2004, 39, 5111-5116. (6) Zhang, Q.; Saito, F.; Shinme, K.; Masuda, S. Dechlorination of PVC by a Mechanochemical treatment under atmospheric condition. J. Soc. Powder Technol. Japan 1999, 36, 468-473. (7) Saito, F. Mechanochemical effects by grinding and its application to material processing. Ceramics 1999, 34, 844-847. (8) Shinme, K. J. Resour. Environ. 2000, 120-121. (9) Pizzigallo, M.; Napola, M.; Spagnuolo, P.; Ruggiero, P. Influence of inorganic soil components and humic substances on the mechanochemical removal of pentachlorophenol. J. Mater. Sci. 2004, 39, 5455-5459. VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3803
(10) Standard Manual for Determination and Analysis of Dioxins in Waste Treatment Process, Waste Management Division, Water Supply and Environmental Sanitation Department, Environmental Health Bureau of the Ministry of Health and Welfare, Japan, 1997 (11) Hasegawa, M.; Sato, M.; Sakuyama, S.; Kimata, M. The Mechanochemical polymerization of methyl methacrylate initiated by the grinding of quartz and quartz glass. Soc. Powder Technol. Japan 1997, 34, 10-16. (12) Vivatpanachart, S.; Nomura, H.; Miyahara, Y.; Kashiwabara, H.; Sakaguchi, M. ESR study of the motion of spin trapped radicals in polyethylene in relation to the morphological changes induced by ball-milling. Polymer 1981, 22, 263-267. (13) Kuzuya, M.; Niwa, J.; Yamauchi, Y.; Kondo, S. Mechanochemical reactions of plasma-induced surface radicals on polyethylene. J. Photopolym. Sci. Technol. 1994, 7, 315-318.
3804
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005
(14) Ikoma, T.; Zhang, Q.; Saito, F.; Akiyama, K.; Tero-Kubota, S.; Kato, T. Radicals in the Mechanochemical dechlorination of hazardous organochlorine compounds using CaO nanoparticles. Bull. Chem. Soc. Jpn. 2001, 74, 2303-2309. (15) Oprea, C. V.; Avran, I.; Avran, R. Mechanochemisch ausgeloste plymerisationsreaktionen I. Angew. Makromol. Chem. 1978, 68, 1-15. (16) Hasegawa, M.; Kimata, M.; Kobayashi, S. Mechanochemical polymerization of styrene initiated by the grinding of quartz. J. Appl. Polym. Sci. 2001, 82, 2849-2855.
Received for review April 13, 2004. Revised manuscript received February 21, 2005. Accepted March 7, 2005. ES049446W
