Formation of PCDDs, PCDFs, and Coplanar PCBs from Polyvinyl

For a more comprehensive list of citations to this article, users are ..... Respiratory Symptoms Necessitating Spirometry Among Soldiers With Iraq/Afg...
0 downloads 0 Views 70KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 1320-1324

Formation of PCDDs, PCDFs, and Coplanar PCBs from Polyvinyl Chloride during Combustion in an Incinerator TAKEO KATAMI Gifu Prefectural Institute of Health and Environmental Sciences, 1-1 Fudogaoka, Naka, Kagamigahara, Gifu 504-0838, Japan AKIO YASUHARA Research Center for Material Cycles and Waste Management, National Institute for Environmental Studies, 16-2, Onogawa, Tukuba, Ibaraki 305-0061, Japan TOSHIKAZU OKUDA Fuji Seiku Kogyosho, Co., Ltd., 6-18, Honmachi, Kano, Gifu 500-8474, Japan TAKAYUKI SHIBAMOTO* Department of Environmental Toxicology, University of California, Davis, California 95616

Exhaust gases from the combustion of poly(vinyl chloride) (PVC), polyethylene (PE), polystyrene (PS), poly(ethylene terephthalate) (PET), and their various mixtures were analyzed for PCDDs, PCDFs, and coplanar PCBs by gas chromatography/mass spectrometry (GC/MS) in order to investigate the role of PVC in these chlorinated compounds. Total amounts of dioxins (PCDDs + PCDFs) found in the samples were 11.7 ng/g PE alone, 1.17 ng/g from PS alone, 25.3 ng/g from PET alone, 448 ng/g from PE with PVC, 140 ng/g from PS with PVC, 126 ng/g from PET with PVC, 824 ng/g from PVC alone under low-CO level, and 8920 ng/g from PVC alone under high-CO level. CO level in high-CO level condition was 880 ppm which was 20 times greater than that in low-CO level condition. Formation of coplanar PCBs ranged from 0.095 ng/g (PE alone) to 77 ng/g (PVC alone under high-CO level). There is a clear correlation between dioxin formation and chloride content. PCDFs composed 80% (PET + PVC) - 98% (PET alone) of the total dioxins formed in the exhaust gases. The results indicate that PVC contributes significantly to the formation of PCDDs, PCDFs, and coplanar PCBs from mixtures of plastics upon combustion.

Introduction Adverse effects of polychlorinated dibenzo dioxins (PCDDs), polychlorinated dibenzo furans (PCDFs), and polychlorinated biphenyls (PCBs) on human health have been known many years. In particular, PCDDs and PCDFs have received much attention recently not only by environmental scientists but also by the public people because they are formed during the combustion of industrial wastes (1). Consequently there is a pressing need to find the formation mechanisms or reaction pathways of these chlorinated chemicals to reduce * Corresponding author phone: (530)752-4523; fax: (530)752-3394; e-mail: [email protected]. 1320

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

their environmental contamination. However, formation mechanisms of these chlorinated chemicals are not yet completely understood because their reaction mechanisms are extremely complex (2, 3). There are some hypotheses regarding the formation-pathways in these chemicals (4, 5). For example, dioxins form from precursors via organic chemical reactions, such as the condensation reaction of two molecules of chlorophenols and the cyclization reaction of polychlorinated biphenyls (6). They are formed from the reaction involving a radical reaction between simple carbon radicals and chloride radicals (2, 7). They are released from polymers with dioxin moieties (8). However, the complete mechanism of dioxin formation is not yet well-known. Most research on dioxin formation has been studied using samples obtained from exhaust gases from combustion of chlorinated materials such as poly(vinyl chloride) (PVC). These samples are generally obtained after secondary combustion or after some means of treatment. Therefore, it is important to collect samples immediately after primary combustion in order to investigate dioxin formation. Formation of dioxins from PVC upon thermal degradation has been reported (9-12). In the present study, various combination of PVC, polyethylene (PE), polystyrene (PS), and poly(ethylene terephthalate) (PET) were combusted in an incinerator made of firebricks. Exhaust gases from the combustion of these plastics were directly collected and analyzed for PCDDs, PCDFs, and coplanar PCBs in order to investigate the role of PVC in dioxin formation during combustion.

Experimental Section Chemicals. Isotope-labeled PCDDs, PCDFs, and coplanar PCBs for internal standards (10 pg/µL n-nonane) were purchased from Cambridge Isotope Laboratories, Inc. (Andover, MS). For the solution of the sampling-spike recovery test, a 1 mL n-nonane solution containing 0.0005 ng/µL each of 13C12-1,2,3,4-T4CDD, 1,2,3,4,7,8-H6CDF, and 1,2,3,4,7,8,9H7CDF solution was prepared. For the solution of the cleanupspike recovery test, a 100 µL n-nonane solution containing 0.005 ng/µL each of 13C12-2,7-D2CDD, 2,3,7-T3CDD, 2,3,7,8T4CDD, 1,2,3,7,8-P5CDD, 1,2,3,6,7,8-H6CDD, 1,2,3,4,6,7,8H7CDD, 1,2,3,4,6,7,8,9-O8CDD, 13C12-2,3,7,8-T4CDF, 1,2,3,7,8P5CDF, 1,2,3,4,7,8-H6CDF, 1,2,3,4,6,7,8-H7CDF, 1,2,3,4,6,7,8,9O8CDF, 13C12-3,3′,3,4′-T4CB, 3,4,4′,5T4CB, 3,3′,4,4′,5-P5CB, 2′,3,4,4′,5-P5CB, 3,3′,4,4′,5,5′-H6CB, 2,3′,4,4′,5,5′-H6CB, and 2,3,3′,4,4′,5,5′-H7CB, was prepared. For the solution of the internal standards, a 2 µL n-nonane solution containing 0.25 ng/µL each of 13C12-1,3,6,8-T4CDD, 1,2,3,7,8,9-H6CDD, and 2,2′,3,4,4′,5,5′-H7CB was prepared. n-Nonane and other solvents for dioxin analysis were bought from Kanto Chemical Co., Inc. (Tokyo, Japan). Instruments. The chloride content in the samples was measured by a TOX-100 Total Organic Halogen Analyzer (Dia Instruments Co., Ltd., Chigasaki, Japan). PVC was combusted by the bomb method, and then the resulting solution was measured by the same instrument for the chloride content. Combustion chamber and flame temperatures were measured by a LK-1200 thermo-couple conductor interfaced to a CT-1310 digital thermometer (Custom Co., Ltd., Tokyo, Japan). Pretreatment for water removal from exhaust gas was conducted by PS-200SCR (Horiba, Ltd., Kyoto, Japan). Continuous measurement of carbon monoxide, carbon dioxide, and oxygen in exhaust gas was performed by a Horiba PG-230 Gas Analyzer (Horiba, Ltd., Kyoto, Japan). The chloride ion concentration in exhaust gases was measured by a Yokogawa IC-7000S Ion Chromatograph (Yokogawa Analytical Systems, Inc., Tokyo, Japan). 10.1021/es0109904 CCC: $22.00

 2002 American Chemical Society Published on Web 02/05/2002

FIGURE 1. Apparatus used for combustion of PVC and related materials. A Hewlett-Packard (HP) model 5890 gas chromatograph (GC) interfaced to Micromass double focus MS (Auto Spec ULTIMA, England) and equipped with a 60 m × 0.32 mm i.d. (df ) 0.2 µm) SP-2331 bonded-phase fused-silica capillary column (Supelco, Bellefonte, PA) for M1CDD-H6CDD isomers and M1CDF-H6CDF isomers or a 30 m × 0.25 mm i.d. (df ) 0.25 mm) DB-5 bonded-phase fused-silica capillary column (J &W Scientific, Folsom, CA) for H7CDD-O8CDD isomers and H7CDF-O8CDF isomers was used. Gas chromatographic oven temperatures were programmed from 130 °C to 190 °C at 20 °C/min and then to 250 °C at 2 °C/min and held for 27 min for the SP-2331 column, or programmed from 130 °C to 280 °C at 10 °C/min for the DB-5 column. The linear velocity of the helium carrier gas was 30 cm/s. The injector temperatures were 250 °C for the SP 2331 column and 280 °C for the DB-5 column at a splitless mode. MS ion source temperatures were 250 °C for the SP-2331 column and 280 °C for the DB-5 column. MS ionization voltage was 35 eV. Materials for Combustion Experiments. A PE sheet (0.3 mm thickness), in which the chloride content was less than 0.0005% (w/w), was purchased from Shinkobe Electric Co., Ltd. (Tokyo, Japan). A PS sheet (0.5 mm thickness), in which the chloride content was less than 0.0005% (w/w), was bought from Nisso Plastic Co., Ltd. (Ibaragi, Japan). A PET sheet (0.05 mm thickness), of which chloride content was less than 0.0005% (w/w), was bought from Toray Co., Ltd. (Tokyo, Japan). A soft PVC sheet (0.2 mm thickness) was bought from Hiroshima Kasei Co., Ltd. (Fukuyama, Japan). Plasticizers in the PVC were removed by washing with hexane twice and with methanol once prior to use. The chloride content of the resulting PVC was 51.3% (w/w). A mixed sheet of PVC with PE, PS, or PET was prepared from one PVC sheet sandwiched with two PE, PS, or PET sheets. The PE and PVC mixture contained 3271 g of PE and 321 g of PVC. The PS and PVC mixture contained 1368 g of PS and 134 g of PVC. The PET and PVC mixture contained 4554 g of PET and 452 g of PVC. Fuel for the subsidiary combustion burner was liquid propane gas (97.6%) containing less than 0.0005% (w/w) chloride and less than 0.0008% (w/w) sulfur. Its total heat generation was 23 710 kcal/m3. Combustion Apparatus. Figure 1 shows the combustion apparatus used for the present study. The volume of the firebrick combustion chamber was 0.22 m3, and the area of the grate was 0.19 m2. The inlet for combustion samples was 0.35 m (height) × 0.40 m (width). The incinerator was equipped with two subsidiary combustion burners (Kato Burner Co., Ltd., Gifu, Japan), with a heat supply of 30 000 kcal/h, to combust samples completely. The firebricks of the inside walls were changed for each experiment in order to

avoid any contamination from a previous experiment. The combustion gases exhausted through a chimney after the dust was removed by a cyclone (960 m3/h) equipped at the outlet of the combustion chamber. The combustion temperatures were measured at the center of the combustion chamber and at the grate. The subsidiary burners were turned on 2 h prior to combustion of the samples in order to maintain a constant temperature. Combustion of Samples. Eight different samples were combusted in an incinerator shown in Figure 1. Sample I: PE (4.442 kg) was combusted alone in 2 h and 32 min. Sample II: PS (8.009 kg) was combusted alone in 2 h and 26 min. Sample III: PET (1.007 kg) was combusted alone in 2 h and 24 min. Sample IV: Mixture of PE and PVC (3.388 kg, total chloride content ) 4.58%, w/w) was combusted in 2 h and 2 min. Sample V: Mixture of PS and PVC (1.5 kg, total chloride content ) 4.58%, w/w) was combusted in 1 h and 51 min. Sample VI: Mixture of PET and PVC (0.565 kg, total chloride content ) 4.63%, w/w). Sample VII: PVC (0.5 kg, chloride content ) 51.3%, w/w) was combusted under low CO level condition in 1 h. Sample VIII: PVC (0.44 kg, chloride content ) 51.3%, w/w) was combusted under high CO level condition in 1 h. Sample Collections for Oxygen (O2), Carbon Monoxide (CO), Carbon Dioxide (CO2), and Chloride Ion in Exhaust Gas. Gas samples were collected at the sampling port located between the combustion chamber and the cyclone. Sample collection for analysis of O2, CO, and CO2 in the exhaust gas was conducted continuously throughout combustion. Sample collection for chloride ion analysis in the exhaust gas was performed by twice drawing exhaust gas for 30 min. Sample Collections for Dioxin Analysis in Exhaust Gas. The exhaust gas samples for dioxin analysis were collected using the apparatus used for a previous study (13). Dust in the exhaust gas was trapped in an in-line silica-fiber thimble filter. The temperature of the filter area was maintained below 120 °C. The exhaust gas was next drawn into three 1 L-impingers connected in series. The first impinger contained 150 mL of hexane-washed distilled water. Dioxin standards were added here to the first impinger for sampling-spike recovery tests. The first impinger was connected to the second impinger, which contained 300 mL of hexane-washed distilled water, and then interfaced to the third empty impinger. The empty impinger was further connected to a column packed with 40 g of XAD-2 resin which was interfaced to a 1 L-impinger containing 250 mL diethylene glycol and an empty impinger connected in series. The impingers were kept at 5 °C in an ice-cooled water bath during sample collections. The exhaust gas was drawn using a diaphagm vacuum pump with the flow rate the same as that of the exhaust gas in the duct (21-23 L/min). The sample collections were conducted from the beginning to the end of combustion. Dioxin Analysis in Samples. Analysis of dioxins in the collected exhaust gas was conducted according to the official method of the Japan Ministry of Health and Welfare using a GC/MS (14). The dust trapped in the silica-fiber thimble filter (approximately 1 g) was washed with a 2 mol/L hydrochloric acid solution (20 mL) and combined with the XAD-2 resin. The resin was extracted for 16 h with toluene (200 mL) using a Soxhleit extractor. The water (500 mL) and diethylene glycol (250 mL) in the impingers as well as the water trapped (trace) in the empty impingers were combined and extracted with 100 mL of dichloromethane twice. The hydrochloric acid solution used to wash the filter was also extracted with 50 mL of dichloromethane twice. Dioxin standards were added here for cleanup recovery tests. After extraction, each extract was condensed by distillation using a rotary flash evaporator, and the combined samples were cleaned with multilayer silicagel chromatography (13). The sample was further cleaned with a 120 mL hexane/dichloroVOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1321

TABLE 1. Conditions and Contents of Incinerator and Concentrations of CO2, CO, and Cl- in the Exhaust Gases

grate temp (°C): chamber temp (°C): av exhaust gas temp (°C) av amt of dry exhaust gas (m3/h) av oxygen concn (%) av CO2 concn (%) av CO concn (ppm)a av Cl - concn (mg/m3N)a amt of exhausted chloride ion (mg/g) a

range av range av

Sample I PE

Sample II Sample III Sample IV Sample V Sample VI PS PET PE + PVC PS + PVC PET + PVC

Sample VII Sample VIII PVC PVC (low CO) (high CO)

925-560 799 744-601 656 533 231 14.5 4.6 77 4 0.39

918-607 749 902-658 738 514 263 13.9 5.5 130 2 0.15

930-874 900 720-611 633 542 298 13.6 5.2 42 1600 610

809-740 769 664-645 656 550 240 13.7 5.2 14 5 2.2

921-576 771 886-667 747 576 255 12.7 5.8 41 420 60

856-624 731 755-639 677 570 242 14.2 4.7 40 170 38

951-844 891 704-673 689 600 230 13.4 5.5 2 92 63

776-720 742 567-514 531 448 267 16.5 3.2 880 1800 550

Relative to 12% oxygen.

methane (1/1) solution using alumina column chromatography. Each sample was condensed using a rotary flash evaporator and the condensed sample was subsequently dissolved into a minimal amount of n-nonane. After 0.5 ng each of 13C12-1,3,6,8-T4CDD and 1,2,3,7,8,9-H6CDD were added to each sample as internal standards for the quantitative analysis of dioxins, the volume of the samples was adjusted to exactly 50 µL with n-nonane. Dioxin standards were added here for the syringe spike test. The samples were analyzed by GC/MS for dioxins.

Results and Discussion The recovery efficiencies of standard dioxins with samplingspike were 89 ( 18.0% for 13C12-1,2,3,4-T4CDD; 91 ( 13.8% for 1,2,3,4,7,8-H6CDD; and 86 ( 11.9% for 1,2,3,4,7,8,9-H7CDF. The recovery efficiencies of standard dioxins with cleanupspike were 91 ( 12.3% for 13C12-2,7-D2CDD; 85 ( 15.5% for 13C12-2,3,7-T3CDD; 88 ( 19.8% for 13C12-2,3,7,8-T4CDD; 88 ( 14.5% for 1,2,3,7,8-P5CDD; 84 ( 15.7% for 1,2,3,6,7,8H6CDD; 91 ( 7.1% for 1,2,3,4,6,7,8-H7CDD; 83 ( 9.0% for 1,2,3,4,6,7,8,9-O8CDD; 84 ( 11.5% for 13C12-2,3,7,8-T4CDF; 92 ( 15.5% for 1,2,3,7,8-P5CDF; 86 ( 19.3% for 1,2,3,4,7,8H6CDF; 82 ( 6.9% for 1,2,3,4,6,7,8-H7CDF; 80 ( 10.5% for 1,2,3,4,6,7,8,9-O8CDF; 77 ( 13.3% for 13C12-3,3′4,4′-T4CB; 77 ( 15.8% for 3,4,4′,5-T4CB; 73 ( 14.1% for 3,3′,4,4′,5-P5CB; 89 ( 15.6% for 2′,3,4,4′,5-P5CB; 70 ( 16.2% for 3,3′,4,4′,5,5′-H6CB; 85 ( 14.1% for 2,3′,4,4′,5,5′-H6CB; and 99 ( 6.4% for 2,3,3′,4,4′,5,5′-H7CB in the present study. Values are the averages of six experiments (n ) 6). The recoveries of both the sampling spikes and the cleanup spikes for PCDDs and PCDFs were satisfactory values of over 80%. However, the recoveries of coplanar PCBs ranged from 70% (3,3′,4,4′,5,5′H6CB) to 99% (2,3,3′,4,4′,5,5′-H7CB). Table 1 shows conditions in the combustion chamber and composition of the exhaust gases. The grate temperatures were over 560 °C because of one subsidiary burner was placed at near the grate (Figure 1) to combust samples completely except in the case of Sample VIII. Its average temperature ranged from 731 °C (for Sample V) to 900 °C (for Sample VII). The other subsidiary burner was placed at the center of the combustion chamber (Figure 1). Therefore the combustion temperature for all samples except for Sample VIII was maintained at over 600 °C throughout the combustion. The temperature of the chamber for Sample VIII was adjusted to lower (531 °C) than those of others in order to obtain high CO concentration-conditions. The average temperatures in the combustion chamber ranged from 531 °C (for Sample VIII) to 747 °C (for Sample IV) during combustion. The average CO2 contents in the exhaust gases were similar among the samples. The average CO concentrations among samples (I-VII), which were 1322

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

combusted under high temperatures, were rather low levels ranging from 2 ppm (Sample VI) to 130 ppm (Sample II). The CO concentration of Sample VIII was adjusted to a high level (880 ppm) in order to investigate the formation difference of dioxins and coplanar PCBs in low (42 ppm, Sample VII) and in high CO concentrations. The Cl- ion concentrations were somewhat proportional to the chloride content of samples combusted. However, the ion concentrations varied over the amount of samples combusted/h. The amount of each sample for combustion was determined according to the optimum combustion conditions. Therefore, the amount of samples combusted ranged from 0.29 kg/h (Sample VI) to 3.3 kg/h (Sample II). Consequently, the average Cl- ion concentrations in mg/g sample were less than 2 when PE (Sample I), PS (Sample II), and PET (Sample III) were combusted alone. On the other hand, when PVC (4.5% as chloride content) was mixed, the average Cl- ion concentrations in mg/g sample were 60 for PE (Sample IV), 38 for PS (Sample V), and 63 for PET (Sample VI). It is interesting that the Cl- ion concentration (610 mg/g sample) in the exhaust gas from PVC combusted under the optimum condition (Sample VII) was higher than that (550 mg/g sample) in the exhaust gas from PVC combusted under the low-temperature condition (Sample VIII). The average chloride ion concentration in the exhaust gases from the three polymers (chlorine content approximately 4.6%) upon combustion was 54 mg/g. When PVC (chlorine content 51.3%) was combusted under high and low CO conditions, chloride ion concentration in the exhaust gases from both conditions was 580 mg/g. These results indicate that the higher chlorine containing materials produced more chloride ions upon combustion. Also, the chloride ion formation was not influenced by the combusting conditions, such as CO concentration, when the grate temperature was over 720 °C. Table 2 shows the results of PCDD, PCDF, and coplanar PCB analysis in the exhaust gases of eight different samples. Total numbers of isomers analyzed were 75 for PCDD, 134 for PCDF, and 13 for coplanar PCB. In the case of PE (Sample I), Cl6-PCDD formed in the greatest amount (0.328 ng/g sample) followed by Cl4-PCDD (0.291 ng/g sample) among the PCDDs formed. Cl1-CDF formed in the greatest amount (5.40 ng/g sample) among the PCDFs formed. The higher the number of chlorine the less PCDF was produced among PCDFs formed. Mono-ortho coplanar PCB formed more than nonortho coplanar PCB did. In the case of PS (Sample II), Cl4-PCDD formed in the greatest amount (0.123 ng/g sample) among PCDDs formed. Also Cl4-PCDF formed in the greatest amount (0.713 ng/g sample) among the PCDFs formed. As was the case with PE, mono-ortho coplanar PCB formed more than nonortho coplanar PCB did. In the case of PET (Sample III), Cl1-PCDD formed in the greatest amount (0.203 ng/g

TABLE 2. Analyses of PCDDs, PCDFs, and Coplanar PCBs in Exhaust Gases from Combusted Substances total concentration of isomers (ng/g sample) Sample V PS + PVC

Sample VI PET+ PVC

Sample VII PVC (low CO)

Sample VIII PVC (high CO)

2.27 3.56 7.54 11.6 14.3 13.4 6.93 2.28 61.9

0.396 1.14 2.43 5.38 3.85 2.83 1.74 0.663 18.4

0.000 0.068 0.187 0.724 2.48 6.02 7.82 8.67 26.0

0.948 3.48 5.11 7.45 11.1 12.3 9.63 3.44 53.5

32.8 71.3 118 58.1 57.9 56.1 29.0 6.49 430

PCDF 88.2 77.7 62.5 64.2 48.5 31.0 11.5 2.24 386

29.5 19.4 23.3 24.8 14.8 8.59 1.72 0.127 122

5.39 3.55 6.73 6.62 16.0 23.7 22.4 16.5 101

225 116 114 98.3 90.4 72.5 41.1 13.6 771

2592 3040 1710 493 290 216 126 25.6 8490

coplanar PCB 0.088 5.30 0.219 9.72 0.307 15.0 25.6 463

0.442 0.792 1.23 142

0.484 1.04 1.52 128

5.78 8.54 14.3 839

34.1 43.5 77.6 9000

Sample I PE

Sample II PS

Sample III PET

M1CDD2a D2CDD10 T3CDD14 T4CDD22 P5CDD14 H6CDD10 H7CDD2 O8CDD1 PCDD total

0.058 0.046 0.072 0.291 0.196 0.328 0.031 0.025 1.05

NMb NMb NMb 0.123 0.046 0.025 0.011 0.009 0.214

0.203 0.123 0.000 0.050 0.038 0.046 0.048 0.059 0.566

M1CDF4 D2CDF16 T3CDF27 T4CDF38 P5CDF28 H6CDF16 H7CDF4 O8CDF1 PCDF total

5.40 2.02 1.12 0.780 0.564 0.705 0.058 0.009 10.7

NMb NMb NMb 0.713 0.170 0.047 0.014 0.008 0.951

18.6 3.78 1.42 0.547 0.238 0.117 0.048 0.023 24.7

nonortho PCBs (4) mono-ortho PCBs (8) total PCBs grand total

0.040 0.055 0.095 11.8

0.077 0.113 0.190 1.36

compound

Sample IV PE + PVC PCDD

a

Number of isomers analyzed.

b

NM: not measured.

sample) among the PCDDs formed. Also, Cl1-PCDF formed in the greatest amount (18.6 ng/g sample), which was the highest level among the dioxins formed from plastics combusted without PVC, among the PCDFs formed. The results suggest that dioxins with a low number of chlorine (1 or 2) form first. As was the case with PE and PS, monoortho coplanar PCB formed more than nonortho coplanar PCB did. A mixture of PE and PVC (Sample IV, Cl% ) 4.58) produced Cl5-PCDD in the greatest amount (14.3 ng/g sample) and a total 61.9 ng/g sample of PCDDs were formed. It produced Cl1-PCDF in the greatest amount (88.2 ng/g sample) followed by Cl2-PCDF (77.7 ng/g sample) among the PCDFs formed. This sample produced the greatest amount of total PCDFs (386 ng/g sample) among the three samples with PVC. The higher the number of chlorine, the less PCDF was produced. In the case of a mixture of PS and PVC (Sample V, Cl% ) 4.58), Cl4-PCDD formed most (5.38 ng/g sample). The amount of PCDFs formed in the samples according to the number of chlorine was Cl1 > Cl4 > Cl3 > Cl2 > Cl5 > Cl6 > Cl7 > Cl8 in PCDF isomers. The PET mixed with PVC produced the most PCDDs (total 26.0 ng/g sample) among the three samples with PVC. The higher the number of chlorine, the more PCDD was produced. The total of Cl1-, Cl2-, and Cl3-PCDD formed was 22% of the total PCDDs, and the total Cl1-, Cl2-, and Cl3-PCDF formed was 59% of the total PCDFs in the samples of PE with PVC (Sample IV) or PS with PVC (Sample V). On the other hand, the total of Cl1-, Cl2-, and Cl3-PCDD formed was only 1% of the total PCDDs, and the total Cl1-, Cl2-, and Cl3PCDF formed was 15% of the total PCDFs in the samples from PET with PVC (Sample VI). A mixture of PE and PVC produced coplanar PCB (15.0 ng/g sample) in an amount 10 times larger than a mixture of PS and PVC (1.23 ng/g sample) or a mixture of PET and PVC (1.52 ng/g sample). When PVC was combusted under high-temperature and low-CO-level conditions (Sample VII), a total 53.5 ng/g sample

of PCDDs was formed, with Cl6-PCDD as the greatest amount (12.3 ng/g sample). This sample formed a total 771 ng/g sample of PCDFs, and the higher the number of chlorine the less PCDF was produced. The total Cl1-, Cl2-, and Cl3-PCDD isomers formed was 18% of total PCDDs, and the total Cl1-, Cl2-, and Cl3-PCDF isomers was 59% of the total PCDFs. These results were consistent with the results from a mixture of PE and PVC or PS and PVC. The total coplanar PCBs formed from this sample was 14.3 ng/g, which was similar to that from PE + PVC (Sample IV). There is a report that a high purity-low molecular weight PVC produced 119 ng/g of total Cl4- - Cl8-PCDD and 1186 ng/g of total Cl4- - Cl8-PCDF upon combustion (15). The total Cl4- - Cl8-PCDD and Cl4- - Cl8-PCDF formations were 43.9 ng/g and 316 ng/g, respectively, in the present study. These results were much less than those reported previously (16), suggesting that dioxin formation can be controlled by combustion conditions. When PVC was combusted under low temperature and with a high CO concentration level (Sample VIII), a total of 429 ng/g PCDDs with Cl3-PCDD as the greatest amount (118 ng/g) and a total 8492 ng/g PCDFs with Cl2-PCDF as the greatest amount (3040 ng/g) were formed. Total coplanar PCB was 77.6 ng/g. The CO level (880 ppm) obtained from this sample was approximately 20 times larger than that from Sample VII (42 ppm). The total Cl1-, Cl2-, and Cl3-PCDD isomers formed was 52% of the total PCDDs, and the total Cl1-, Cl2-, and Cl3-PCDF isomers was 86% of the total PCDFs, suggesting that high levels of dioxins with low numbers of chloride tend to form under incomplete combustion conditions. A double bond was formed from PVC via dehydrochloridereaction and subsequently aromatic hydrocarbons were formed by a ring formation-reaction. Dust (mainly soot) formed under a high CO condition contained many aromatic hydrocarbons (17). It has been hypothesized that these aromatic hydrocarbons yielded dioxins in the presence of VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1323

TABLE 3. Toxicity Equivalency Quantity in Exhaust Gases from Incinerated Substances amount of compound (ng-TEQ/g sample) Sample I PE

Sample II PS

Sample III PET

Sample IV PE + PVC

Sample V PS + PVC

Sample VI PET+ PVC

Sample VII PVC (low CO)

Sample VIII PVC (high CO)

0.073

0.021

0.032

4.7

1.8

2.3

8.3

29

TABLE 4. Pattern Similarity of Cl1- - Cl8-PCDD, PCDF, and Coplanar PCB Formation PS + PVC (Sample V)

PET + PVC (Sample VI)

PVC-low CO (Sample VII)

PVC-high CO (Sample VIII)

0.93 1

0.32 0.27 1

0.95 0.95 0.37 1

0.92 0.86 0.17 0.88

PE + PVC (Sample IV) PS + PVC (Sample V) PET + PVC (Sample VI) PVC-low CO (Sample VII)

hydrogen chloride (18). It is also hypothesized that dioxins formed from chloro benzenes, which were produced from PVC upon thermal degradation (4). In the present study, there were fewer benzofurans with higher numbers of chlorine than benzofurans with lower numbers of chlorine. The aromatic hydrocarbons, such as benzene, biphenyl, phenol, and benzofuran, produced by the imcomplete combustion were chlorinated step by step with chlorines formed from PVC. De Fre and Rymen (18), who investigated the role of HCl presence in the formation of PCDDs and PCDFs from hydrocarbon combustion, found that 45 ng/Nm3 of PCDDs and 499 ng/Nm3 PCDFs were formed from hydrocarbons upon combustion under a low CO concentration (95 ppm). On the other hand, when the same hydrocarbons were combusted under a high CO concentration (2500 ppm), 3741 ng/g of PCDDs and 8190 ng/g of PCDFs were formed. These results also suggest that adjusting the combustion conditions, including the combustion temperature and the CO concentration, could reduce the formation of PCDDs and PCDFs. Table 3 shows toxicity equivalency quantity in the exhaust gases from incinerated substances (16). When PE, PS, and PET were combusted alone. The TEQ of dioxins formed ranged from 0.021 to 0.073 ng-TEQ/g. These values are similar to that obtained from London plane (0.030 ngTEQ/g) combusted in a different incinerator (13). When PVC was added to PE, PS, and PET (chlorine content approximately 4.6%), the TEQ of dioxins formed ranged from 1.8 to 4.7 ng-TEQ/g. These values are consistent with the one (4.25 ng-TEQ/g) obtained from newspapers combusted with PVC in a different incinerator (13). Table 4 shows the pattern similarity of Cl1- - Cl8-PCDD, -PCDF, and -coplanar PCB formations. The pattern similarity (S2) was calculated using the following equation (19):

∑Xa ‚Xb i

S2 (a, b) )

i

i

x∑ x∑ X2ai‚

i

X2bi

i

Similarity patterns between PE + PVC and PVC-low CO and PE + PVC and PVC-high CO were over 0.9. The similarity pattern between PS + PVC and PVC-low CO was 0.95. These results suggest that a constituent of PVC must play an important role in the formation of PCDDs, PCDF, and coplanar PCB in these samples. On the other hand, the similarity patterns between PET + PVC and PVC-low CO and PET + PVC and PVC-high CO were 0.37 and 0.17, respectively. These results suggest that PCDDs, PCDFs, and coplanar PCB formation from PET with PVC upon combustion may have 1324

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

different mechanisms of those of PE with PVC or PS with PVC upon combustion. The results of the present study indicate that upon combustion PVC produces certain amounts of PCDDs, PCDFs, and coplanar PCBs. In particular, upon combustion under low temperatures and low CO concentration conditions, PVC produces high levels of both PCDDs and PCDFs. PVC is difficult to burn. Consequently, if PVC is combusted with other waste materials under poorly controlled conditions, such as low temperature and high CO concentration, PCDDs and PCDFs will probably form in significant levels.

Literature Cited (1) Coulston, F.; Pocchiari, F. Accidental Exposure to Dioxins: Human Health Aspects; Academic Press: New York, 1983. (2) Huang, H.; Buekens, A. Chemosphere 1995, 31, 4099-4117. (3) Fangmark, I.; Stromberg, B.; Berge, N.; Rappe, C. Chemosphere 1994, 29, 1903-1909. (4) Lustenhouwer, J. W. A.; Olie, K.; Hutzinger, O. Chemosphere 1980, 9, 501-522. (5) Tuppurainen, K.; Halonen, I.; Ruokojarvi, P.; Tarhanen, J.; Ruuskanen, J. Chemosphere 1998, 36, 1493-1511. (6) Luijk, R.; Akkemrman, D. M.; Slot, P.; Olie, K.; Kapteijn, F. Environ. Sci. Technol. 1994, 28, 312-321. (7) Stieglitz, L.; Vogg, H. Chemoshere 1987, 16, 1917-1922. (8) Yasuhara, A.; Morita, M. Environ. Sci. Technol. 1988, 22, 646650. (9) Sinkkonen, S.; Makela, R.; Vesterinen, R.; Lahtipera, M. Chemosphere 1995, 31, 2629-2635. (10) Kanters, M. J.; van Nispen, R.; Louw, R.; Mulder, P. Environ. Sci. Technol. 1996, 30, 2121-2126. (11) Wikstrom, E.; Marklund, S. Environ. Sci. Technol. 2000, 34, 604609. (12) Wagner, J.; Green, A. Chemosphere 1993, 26, 2039-2054. (13) Yasuhara, A.; Katami, T.; Okuda, T.; Ohno, N.; Shibamoto, T. Environ. Sci. Technol. 2001, 35, 1373-1378. (14) Japan Ministry of Health and Welfare. Standard method for determination of dioxins in waste management. JMHW, Notification No. 234; 1997. (15) Christmann, W.; Kasiske, D.; Kloppel, K. D.; Partscht, H.; Rotard, W. Chemosphere 1989, 19, 387-392. (16) Van den Berg, M.; Birnbaum, L.; Bosveld, A. T. C.; Brunstro¨m, B.; Cook, P.; Feeley, M.; Giesy, J. P.; Hanberg, A.; Hasegawa, R.; Kennedy, S. W.; Kubiak, T.; Larsen, J. C.; Rolaf van Leeuwen, F. X.; Djien Liem, A. K.; Nolt, C.; Peterson, R. E.; Poellinger, L.; Safe, S.; Schrenk, D.; Tillitt, D.; Tysklind, M.; Younes, M.; Wærn, F.; Zacharewski, T. Environ. Health Perspec. 1998, 106, 775792. (17) Liao, J. C.; Browner, R. F. Anal. Chem. 1978, 50, 1683-1686. (18) De Fre, R.; Rymen, T. Chemosphere 1989, 19, 331-336. (19) Murai, H.; Toyokawa, Y. Jpn. J. Pub. Health 1978, 25, 363-370.

Received for review May 23, 2001. Revised manuscript received September 13, 2001. Accepted December 13, 2001. ES0109904