Characterizing the Emissions of Polybrominated Dibenzo-p-dioxins

F. Bjurlid , M. Dam , K. Hoydal , J. Hagberg. Chemosphere 2018 195 ... Nicolas A.O. Morin , Patrik L. Andersson , Sarah E. Hale , Hans Peter H. Arp. J...
0 downloads 0 Views 354KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 1159-1165

Characterizing the Emissions of Polybrominated Dibenzo-p-dioxins and Dibenzofurans from Municipal and Industrial Waste Incinerators LIN-CHI WANG* AND GUO-PING CHANG-CHIEN Department of Chemical and Materials Engineering, Cheng Shiu University, 840 Chengching Road, Kaohsiung 833, Taiwan, ROC

As yet, very little is known about the polybrominated dibenzo-p-dioxins and dibenzofurans (PBDD/Fs) characteristics in the stack flue gases of incinerators. In this study, nine large-scale continuous municipal solid waste incinerators (MSWIs), two small-scale batch MSWIs, and nine industrial waste incinerators (IWIs) were investigated for 2,3,7,8-substituted PBDD/Fs. The elevated PBDD/F concentrations (18.2 pg/Nm3, 4.17 pg TEQ/Nm3) of the IWIs, which were eight times higher than those of MSWIs (2.28 pg/Nm3, 0.557 pg TEQ/Nm3), are accompanied by PCDD/ Fs that are in the same range as those measured from MSWIs (0.0171-1.98 ng I-TEQ/Nm3). The obtained TEQ ratios (in percentage) of the PBDD/F to the PCDD/F concentration in the stack flue gases of the MSWIs (0.72%), batch MWIs (0.18%), and IWIs (5.4%) are useful for the future estimate of PBDD/F emission quantity based on the wellestablished PCDD/F inventory. In addition, a significantly high correlation was found between the PBDD/F and PCDD/F concentrations, and the PBDD/F and PCDD/F congener profiles of the same emission source were similar, indicating a similar formation and substitution mechanism of bromine and chlorine in the combustion system.

Introduction Polybrominated dibenzo-p-dioxins and dibenzofurans (PBDD/Fs) have a chemical structure and physical-chemical properties similar to those of polychlorinated dibenzo-pdioxins and dibenzofurans (PCDD/Fs) and polychlorinated biphenyls (PCBs), and invoke a common battery of toxic responses. Unlike PCDD/Fs and PCBs, PBDD/Fs are much less studied maybe because the PBDD/F congeners cannot be distinguished from the fragment ions of polybrominated diphenyl ethers (PBDE) during GC/MS measurement without a mature separating and identifying method (1). However, concerns about PBDD/Fs have increased because of brominated flame retardants (BFRs), which were extensively used in a variety of materials and miscible with most plastics, including various fabrics, polymers, paints, and furnishings for car interiors. Several potential cases for the release of PBDD/Fs into the environment had been proposed, including formation in the process of manufacturing BFRs, formation during the processing of BFRs with polymer resin, and incinerating waste which contain BFRs * Corresponding author phone: +886-7-7310606 ext. 3921; fax: +886-7-7332204; e-mail: [email protected]. 10.1021/es061801q CCC: $37.00 Published on Web 01/17/2007

 2007 American Chemical Society

(2-3). Nevertheless, the PBDD/F contributions from the above potential cases have never been investigated and measured. The World Health Organization (4) reported that the PBDD/F contents in resins to which PBDEs has been added are mostly high at ppm level. Sakai et al. (3) measured a total of 15 discarded television sets and personal computers manufactured in Japan between 1984 and 1998 for their PBDD/F contents, and PBDD/Fs were found at average levels of 280 000 ng/g in waste television cabinets. Products with BFRs will sooner or later be treated by municipal solid waste incinerators (MSWIs) or metal recycling plants and increase the quantity of brominated material (5). Polybromochlorodibenzo-p-dioxins and polybromochlorodibenzofurans (PBCDD/Fs) have been detected in fly ashes of MSWIs (6-12) and the average contribution was 1%-20% of PCDD/Fs (10-12). Wanke et al. (13) studied the influence of additional bromine input (up to 6-fold) into a pilot MSWI and the results showed that, for the PBCDFs, there is a correlation between the content of bromide in the fly ash (up to 5% by weight) and the PBCDF concentrations in the stack flue gases. Funcke et al. (14) performed experiments and found that the formation of PBCDD/Fs was correlated with the bromine input from the fuel. Vehlow (15) noted a similar discrepancy. Many studies concerning PBDD/Fs focused on thermal treatment experiments in lab-scale quartz reactors (3, 5, 1620); however, the emission cannot directly be extrapolated to full-scale incinerators with waste containing BFRs. As for other research regarding full-scale incinerators (11, 13, 2122), only lower brominated congeners (mono-, di-, and tribrominated) or lower brominated-chlorinated congeners like monobromopolychloroDD/Fs have been investigated, not the 2,3,7,8-substituted congeners. In this study, PBDD/F cleanup and separating procedures suitable for stack flue gas samples were established by conducting elution tests of different columns to optimize the cleanup and fractionation behavior of PBDD/F analyses. Afterward, nine large-scale continuous MSWIs, two smallscale batch MSWIs, and nine industrial waste incinerators (IWIs) were investigated for their 2,3,7,8-substituted PBDD/F and PCDD/F characteristics in the stack flue gases. As yet, very little is known about the PBDD/F characteristics of MSWIs and IWIs. The mass and TEQ ratios of the PBDD/F to the PCDD/F concentration in the stack flue gases of the MSWIs and IWIs are presented and compared. The obtained results are useful not only for assessing their PBDD/F TEQ contributions in Taiwan, but can also be used as a reference for other countries. Furthermore, the congener profiles of PCDD/Fs and PBDD/Fs in the stack flue gases were compared to reveal the relationship between them.

Materials and Methods Elution Test. For PBDD/F cleanup methods which use the macro-aluminum oxide column, mixed acid-alkaline silica gel column, and mini-aluminum oxide column, it has been reported that more than 50% of the PBDEs in a sample remain in the resulting solution for GC/MS-determination (1). In this study, a new modified cleanup method for PBDD/Fs was used, which combines acid silica gel column, acid alumina column, and active carbon columns. Furthermore, we conducted elution tests of different column chromatographies to optimize the cleanup and fractionation behavior of PBDD/F analyses. PBDE standard in a nonane solution was purchased from Wellington Laboratories (Ontario, Canada), including PAR VOL. 41, NO. 4, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1159

TABLE 1. Basic Information Concerning the Nine IWIs I1

I2

company category

petrochemical company

operation type feeding waste

continuous continuous waste plastic mixture, sludge, oily sludge (66 ton/day) heavy oil (165)

auxiliary fuel (L/hr) APCDs bag filter in sequence (175 °C) (operation temperature)

I3

construction petrocompany chemical company

I4 chemical company

9

chemical company

I6 petrochemical company

continuous intermittent intermittent continuous

I7

I8

nuclear power plant

agricultural and chemical company continuous intermittent

I9 nuclear power plant continuous

waste sludge construction (21 ton/day) materials (86 ton/day)

waste rubber and latex (125 kg/hr)

waste plastic (230 kg/hr)

oily low sludge radiation (44 ton/day) waste (24 ton/day)

waste paper bag (150 kg/hr)

low radiation waste (43 ton/day)

diesel (310)

heavy oil (150)

diesel (4.6)

diesel (2.1)

heavy oil (100)

diesel (50)

diesel (60)

diesel (50)

dry scrubber (170 °C) packed-bed scrubber (70 °C)

cyclone (370 °C) wet scrubber (120 °C)

bag filter (190 °C)

quench c hamber (160 °C) bag filter (148 °C)

cyclone (280 °C) electrostatic precipitator (240 °C)

quench chamber (250 °C) bag filter (180 °C)

cyclone (300 °C) bag filter (190 °C)

quench chamber (300 °C) bag filter (200 °C)

stock solution (27 congeners), labeled compound solution (12 congeners), and recovery standard solution (one congener). The internal standard solution and recovery standard solution of PBDD/F were purchased from Cambridge Isotope Laboratories, Inc. (USA), and contained the following PBDD/F congeners: 13C12-2,3,7,8-TeBDD, 13C12-1,2,3,7,8-PeBDD, 13C121,2,3,7,8,9-HxBDD, 13C12-2,3,7,8-TeBDF, 13C12-1,2,3,7,8-PeBDF, 13C12-2,3,4,7,8-PeBDF, and 13C12-1,2,3,4,7,8-HxBDF in nonane solution. Solutions, spiked with 10 µL of PBDD/F internal standard solution (25 pg/µL), 50 µL of PBDD/F PAR stock solution (0.5/2.5 pg/µL), 10 µL of PBDEs internal standard solution (50/100/250 pg/µL), and 20 µL of PBDE PAR stock solution (10/20/50 pg/µL) were used for the elution tests of different column chromatographies Basic Information Concerning the Incinerators. The capacity of each furnace of the nine continuous MSWIs (M1M9) was 300-450 t/day and their air pollution control devices (APCDs) were dry scrubber, activated carbon injection, and bag filter. For the two batch MSWIs (S1 and S2), the capacity of each was about 1 t/h and the APCDs were cyclone, activated carbon injection, and bag filter. As for the nine IWIs (I1-I9), their basic and operational information is described in Table 1. The nine IWIs owned by petrochemical, chemical, and construction companies, and nuclear power plants disposed of waste including waste plastic, rubber, sludge, paper bags, and low radiation waste. Basic Information Concerning the Chlorine and Bromine Content in the Waste. In this study, surveying and using the published data are appropriate for obtaining information concerning the total amount of chlorine and bromine in the waste because very large numbers of samples need to be collected and analyzed to be representative, and handling industrial waste, like low radiation waste, has potential dangers. Taiwan EPA reported that mean organic chlorine was 0.08%-0.25% of waste (23). Consideringthat European MSWIs received almost equal amounts of chlorine from inorganic (salt) and organic sources (24), we extrapolate that Taiwan municipal solid waste (MSW) contains 0.16% to 0.5% chlorine. Data from literature regarding bromine content in MSW from households and small businesses indicate typical bromine content of 0.003%-0.006% by weight of bromine (25). One source (26) reported the typical bromine content in Swedish waste accounted for 0.004% by weight from a study in 1992, but that the level has increased over the past decade. In comparison, normal chlorine level in municipal waste is approximately 0.75% by weight. An investigation of halogens in hazardous and clinical waste reported that the mean chlorine and bromine contents 1160

I5

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 4, 2007

(n ) 12) are 0.47% and 0.014%, respectively (27). In TV casing materials and printed circuit boards, the chlorine contents are 0.021% and 1.1% with the bromine contents at 3.7% and 2.2%, respectively (3). Otherwise, the bromine contents in BFR-plastics ranged from 0.4% to 4.2% (25). In other words, if the industrial wastes consist of 1% BFR-plastics, the bromine content will be 0.004%-0.04% of waste, which is equal to or one order higher than that of MSW. Although the diversity of the industrial waste makes it impossible to predict a probable chlorine or bromine content without conducting real element analyses, the bromine contents of the feeding industrial waste in this study should be higher than that of MSW by consulting the above references. Stack Flue Gas Sampling. All stack flue gas samplings and chemical analyses in this study were carried out by our accredited laboratory for PCDD/F stack flue gas samplings and analyses in Taiwan. All the stack flue gases were sampled during the normal operation periods and for MSWIs, at least 1 month after the start-up procedure to avoid the influence of the memory effect. Five stack flue gas samples were conducted for each continuous MSWI, while three stack flue gas samples were conducted for each batch MSWI and IWI. The sampling time of each stack flue gas sample was 3 h, although in another study (28) it lasted for 8 h. The stack flue gases of the selected incinerators were collected isokinetically according to U.S. EPA modified Method 23. The sampling train adopted in this study is comparable with that specified by U.S. EPA Modified Method 5. Prior to sampling, XAD-2 resin was spiked with PCDD/F surrogate standards prelabeled with isotopes, including 37Cl4-2,3,7,8-TCDD, 13C121,2,3,4,7,8-HxCDD, 13C12-2,3,4,7,8-PeCDF, 13C12-1,2,3,4,7,8HxCDF, and 13C12-1,2,3,4,7,8,9-HpCDF. The recoveries of PCDD/F surrogate standards were 103%-120%, and met the criteria within 70%-130%, revealing no PCDD/F breakthrough occurred. Otherwise, only a few PCDD/F congeners in a few stack flue gas samples in this study are ND (Supporting Information), revealing that a 3-h sampling time is adequate for accurate PCDD/F analyses. Details are similar to that given in our previous work (29). Because no PBDD/F surrogate standards could be purchased, the sampling train collection efficiency check for PBDD/Fs used the recoveries of the corresponding PCDD/F surrogate standards. Stack Flue Gas Analyses. To measure both the PCDD/Fs and PBDD/Fs, after the extraction of the stack flue gas samples, the extracted solution was divided equally into flask A and flask B. All the A flasks were measured for PCDD/Fs individually, but for PBDD/F measurement all the B flasks were combined into one because of the detection limit. That is, one continuous MSWI had data for five PCDD/F and one PBDD/F stack flue gas.

FIGURE 1. Recoveries of PBDD/Fs and PBDEs for different columns. PCDD/F analyses of the stack flue gases followed U.S. EPA modified method 23. Detailed analytical procedures are given in our previous work (29). Compared to PCDD/Fs, which had 17 congeners reported, only 7 of the possible 17 PBDD/F congeners were reported due to lack of standard. The PBDD/F analyses are as follows. After extraction, treatment with concentrated sulfuric acid was used for the first cleanup. For the next cleanup procedure, an acid silica gel column was used. The sample extract, dissolved in 5 mL of hexane, was added to the column with two additional 5-mL rinses. The column was eluted with an additional 50 mL of hexane and the entire eluate was retained. This solution was concentrated to a volume of about 1 mL

using a rotary evaporator. For the next step, an acid alumina column was used. The concentrated extract was transferred from the silica gel column to the top of the acid alumina column and the column was sequentially eluted with 10 mL of hexane followed by 30 mL of dichloromethane/hexane (2/98, v/v). The eluate was discarded. The column was eluted again with 35 mL of dichloromethane/hexane (40/60, v/v). The eluate was collected and concentrated to near dryness by using a nitrogen stream. For the next step, an active carbon column was used for separating PBDEs and PBDD/Fs. It was expected that PBDEs would be well separated from PBDD/Fs on the carbon column due to differences in planarity and absorption properties, as VOL. 41, NO. 4, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1161

TABLE 2. PBDD/F and PCDD/F Concentrations in the Stack Flue Gases of the Nine MSWIs PBDD/Fs

2,3,7,8-TeBDD 1,2,3,7,8-PeBDD 1,2,3,4/6,7,8-HxBDD 1,2,3,7,8,9-HxBDD 2,3,7,8-TeBDF 1,2,3,7,8-PeBDF 2,3,4,7,8-PeBDF total (pg/Nm3) total TEQ (pg TEQ/Nm3) PCDD/Fs mean total (ng/Nm3) mean total TEQ (ng I-TEQ/Nm3) ratio of mass ratio PBDD/Fs to (%) PCDD/Fs (percent) TEQ ratio (%) a

BD: below detection limit.

b

M1

M2

M3

M4

M5

M6

M7

M8

M9

mean

RSD (%)

0.0128 BD 0.809 0.607 0.244 0.115 0.0561 1.84 0.213

0.0172 BD 0.314 0.0784b 0.0544 0.0567 0.107 0.628 0.118

BDa BD 0.165 BD 0.0272 0.0274 0.229 0.449 0.135

0.0543 BD 0.678 BD 0.154 0.144 0.359b 1.39 0.324

0.0468 0.138b 0.671 0.503 0.115 0.275 0.429 2.18 0.473

0.0156 0.107 0.616 0.462 0.843 0.613 1.02 3.68 0.802

0.223 0.0163 0.782 0.0752b 0.111 0.0381 0.306 1.55 0.483

0.0127b 0.296 1.73 0.154b 0.281 0.326 0.554 3.35 0.671

0.188 0.259 0.443 0.332b 1.16 0.569 2.51 5.46 1.79

0.0641 0.102 0.690 0.288 0.332 0.240 0.619 2.28 0.557

128 107 64.6 69.7 119 93.0 123 70.7 93.3

0.337

0.555

0.754

0.866

1.39

2.02

1.32

6.15

3.75

172 234

0.0171 0.0198

0.0256 0.0346 0.0540 0.0675 0.0737

0.194

1.98

0.274

0.55

0.11

0.06

0.16

0.16

0.18

0.12

0.05

0.03

0.16

98.7

1.2

0.60

0.53

0.94

0.88

1.2

0.66

0.35

0.09

0.72

53.0

Below the limit of quantification (LOQ), above the limit of detection (LOD).

TABLE 3. PBDD/F and PCDD/F Concentrations in the Stack Flue Gases of the Two Batch MSWIs PBDD/Fs

PCDD/Fs

ratio of PBDD/Fs to PCDD/Fs (percent)

2,3,7,8-TeBDD 1,2,3,7,8-PeBDD 1,2,3,4/6,7,8-HxBDD 1,2,3,7,8,9-HxBDD 2,3,7,8-TeBDF 1,2,3,7,8-PeBDF 2,3,4,7,8-PeBDF total (pg/Nm3) total TEQ (pg TEQ/Nm3) mean total (ng/Nm3) mean total TEQ (ng I-TEQ/Nm3) mass ratio (%) TEQ ratio (%)

a

BD: below detection limit.

b

S1

S2

BDa 0.402b 0.678 0.276 0.451b 0.577 0.831 3.22

0.516 1.19 3.25 2.14 1.44 0.767 2.15 11.5

0.786

2.91

1.65 0.239

216 10.2

0.20

0.0053

0.33

0.028

Below the LOQ, above the LOD.

is the case with separations of mono-ortho PCBs from PCDD/Fs (30). The column was sequentially eluted with 25 mL of dichloromethane/hexane (40/60, v/v) for PBDEs, followed by 35 mL of toluene for PBDD/Fs. The toluene eluate was collected and concentrated to near dryness by using a nitrogen stream. Prior to PBDD/F analysis, the standard solution was added to the sample to ensure recovery during the analysis process. The more detailed cleanup procedures of PBDD/Fs are provided in the Supporting Information. A high-resolution gas chromatograph/high-resolution mass spectrometer (HRGC/HRMS) was used for PBDD/Fs analyses. The HRGC (Hewlett-Packard 6970 Series gas, CA) was equipped with an Rtx-5MS capillary column (L ) 30 m, i.d. ) 0.25 mm, film thickness ) 0.25 µm) (Restek, PA), and with a splitless injection. The oven temperature program was set according to the following parameters: injector at 290 °C, transfer glass line at 280 °C, initial oven temperature begin at 150 °C (held for 1 min), then increase at 40 °C/min to 220 °C, then increase at 2 °C/min to 240 °C, and finally increase at 10 °C/min to 310 °C (held for 1 min). Helium at 1162

9

20.4

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 4, 2007

a flow rate of 1.2 mL/min was the carrier gas. The HRMS (Micromass Autospec Ultima, Manchester, UK) mass spectrometer was equipped with a positive electron impact (EI+) source. The analyzer mode of the selected ion monitoring (SIM) was used with a resolving power of 10 000. The electron energy and source temperature were specified at 35 eV and 250 °C, respectively. We had participated in the first round of the intercalibration for PBDD/Fs and PBCDD/Fs in 2004. The result revealed that the performance of our method was quite acceptable for PBDD/F analysis. A detailed HRGC/ HRMS grouping and the monitor ions of PBDD/Fs are shown in the Supporting Information.

Results and Discussion Column Chromatography for Separating PBDD/Fs and PBDEs. The elution test results of different column chromatographies are illustrated in Figure 1. In the silica gel column, no separation of PBDD/Fs from PBDEs was observed in 50 mL of hexane elute, and the complete elution of congeners was found at 30 mL of hexane. The recoveries of PBDD/Fs and PBDEs were 68%-83% and 70%-123%, respectively, and met the criteria for PCDD/Fs (within 50%120%). As the silica gel column, the acid alumina column had no separation ability for PBDD/Fs and PBDEs. PBDE and PBDD/F congeners were not found in the 25 mL of dichloromethane/hexane (2/98, v/v) elute, but both completely eluted at 15 mL of dichloromethane/hexane (40/60, v/v). The recoveries of PBDD/Fs and PBDEs were 81%-108% and 71%-120%, respectively. When conducting the real stack flue gas sample, we increased the elution volume of dichloromethane/hexane (2/98, v/v) to 30 mL to reduce the interference completely. For the active carbon column, PBDE congeners were completely eluted at 15 mL of dichloromethane/hexane (40/ 60, v/v). The PBDD/F congeners were not found in this fraction and toluene was used for the second elution. High brominated substituted PBDD/F congeners, like 1,2,3,4/6,7,8HxBDD and 1,2,3,7,8,9-HxBDD, need more toluene to elute, and total PBDD/F congeners could be eluted completely with 30 mL of toluene. The recoveries of PBDD/Fs were 81%100%. By using active carbon column chromatography in the final procedure, we found that PBDEs were well separated from PBDD/Fs in an elution test with reference standards. PBDD/F Concentrations in the Stack Flue Gases of the MSWIs. Because the toxic equivalency factors (TEFs) have

TABLE 4. PBDD/F and PCDD/F Concentrations in the Stack Flue Gases of the Nine IWIs I1 PBDD/Fs

PCDD/Fs

ratio of PBDD/Fs to PCDD/Fs (percent)

a

I2

I3

I4

I5

I6

I7

I8

I9

mean

RSD (%)

0.146b 2,3,7,8-TeBDD 0.0582b 1.05 1.24 1,2,3,7,8-PeBDD 0.0685b 1.54 1,2,3,4/6,7,8-HxBDD 0.176b 1.99 0.368b 0.537b 1.03 1,2,3,7,8,9-HxBDD 0.665b 2,3,7,8-TeBDF 1.36 1.28 4.91 1,2,3,7,8-PeBDF 0.908 1.72 6.54 2,3,4,7,8-PeBDF 2.11 2.31 9.52 5.35 10.4 19.0 total (pg/Nm3) total TEQ 1.41 3.44 4.10 (pg TEQ/Nm3) 0.321 0.409 0.475 mean total (ng/Nm3) mean total TEQ 0.0128 0.0263 0.0364 (ng I-TEQ/Nm3) mass ratio 42 2.6 4.0 (%)

0.0802b 0.222 0.672b 0.503b 0.236b 0.192b 0.794b 0.540b 0.989 3.25 0.822 3.96 0.907 5.21 4.50 13.9

0.0693b 0.616b 0.272b 0.767b 0.777 1.277 3.20 6.98

BDa 4.38 1.39 0.823 BD 2.68 1.21 1.01 0.125 0.456 0.144 0.440 0.749 0.587 0.650 0.702 2.50 41.7 11.3 7.56 1.77 17.7 2.60 4.14 2.32 9.79 1.76 4.13 7.46 77.3 19.1 18.2

173 77.0 134 22.1 175 130 81.2 125

1.11

3.67

2.22

1.58

108

0.631

0.841

1.40

2.66

0.0460

0.0680 0.137

0.71

1.6

TEQ ratio (%)

2.4

5.4

BD: below detection limit.

11 b

13

11

15.8

4.21

4.17

4.16

3.95

1.65

0.235

0.722

0.897

0.242

137

0.50

0.28

1.9

0.48

6.0

225

1.6

0.67

2.2

0.47

5.4

93.5

95.0

Below the LOQ, above the LOD.

not been determined for PBDD/Fs, the I-TEFs of PCDD/Fs were used for the corresponding congeners of PBDD/Fs (4, 31-33) to calculate their toxicity in the stack flue gases. Table 2 lists the PBDD/F and PCDD/F concentrations in the stack flue gas of these nine MSWIs (M1-M9), which are listed in order according to their PCDD/F I-TEQ concentrations. The limit of detection (LOD) in this study was defined as a signalto-noise (S/N) greater than three, while the limit of quantification (LOQ) was defined as an S/N greater than ten. The RSD values of the PCDD/F I-TEQ concentrations of the nine MSWIs (Supporting Information) were in the range 14%60%. The moderate RSD value reveals the great representativeness of operation condition for each tested MSWI. Only two MSWIs (M8 and M9) exceeded the Taiwan large-scale MSWI emission limit of 0.1 ng I-TEQ/Nm3 maybe because of incompletely combustion (the CO reading reached about 90 ppm, which is much higher than 15 ppm during normal condition), revealing that most of the investigated MSWIs operated in good condition. The mean PBDD/F concentration in the stack flue gases of the nine MSWIs was 2.28 pg/ Nm3 (range: 0.449-5.46 pg/Nm3, RSD: 70.7%) and the corresponding mean TEQ concentration was 0.557 pg TEQ/ Nm3 (range: 0.118-1.79 pg TEQ/Nm3, RSD: 93.3%). The 2,3,7,8-substituted PBDD/F concentrations in the atmosphere of Kyoto (n ) 5) and Osaka (n ) 1), Japan, were 0.252.3 pg/m3 (tetra-hexa) (34) and 0.036 pg/m3 (35), respectively, and were close to or one order lower than that in the stack flue gas of the MSWIs, revealing that much more significant PBDD/F emission sources than MSWIs existed in the environment. The Pearson correlation analyses were performed and revealed that the logarithm of the PCDD/F concentration significantly correlated with the PBDD/F concentration of the MSWIs (r ) 0.88, p ) 0.001); the logarithm of the PCDD/F I-TEQ concentration also significantly correlated with the PBDD/F TEQ concentration (r ) 0.96, p < 0.001). The high correlation found between the PBDD/F and PCDD/F concentrations revealed their similar formation mechanism in the combustion system and removal rate by APCDs. The mean ratio (in percentage) of the PBDD/F to the PCDD/F concentration in the stack flue gases of the MSWIs was 0.16% (range: 0.03%-0.55%, RSD: 98.7%) while their corresponding mean TEQ ratio (in percentage) was 0.72% (range: 0.09%-1.2%, RSD: 53.0%). The obtained result can be used for future estimates of PBDD/F emission quantity based on the well estimated PCDD/F inventory. The ratio of

the PBDD/F to the PCDD/F concentration significantly correlated negatively with the log PCDD/F concentration of the MSWIs (r ) -0.60, p ) 0.044 on mass basis; r ) -0.69, p ) 0.020 on toxicity basis), revealing that more PCDD/Fs are generated than PBDD/Fs at poor combustion conditions. The PBDD/F and PCDD/F concentrations in the stack flue gases of the two batch MSWIs (S1 and S2) are listed in Table 3 and were 0.786, 2.91 pg TEQ/Nm3 for PBDD/Fs and 0.239, 10.2 ng I-TEQ/Nm3 for PCDD/Fs, respectively. The TEQ ratios (in percentage) of the PBDD/F to the PCDD/F concentration in the stack flue gases were 0.33% for S1 (PCDD/Fs: 0.239 ng I-TEQ/Nm3) and 0.028% for S2 (PCDD/ Fs: 10.2 ng I-TEQ/Nm3), respectively. PBDD/F Concentrations in the Stack Flue Gases of the IWIs. Table 4 lists the PBDD/F and PCDD/F concentrations in the stack flue gas of the nine IWIs (I1-I9), which are in order according to the PCDD/F I-TEQ concentrations. The mean PBDD/F concentration in the stack flue gases of the nine IWIs was 18.2 pg/Nm3 (range: 4.50-77.3 pg/Nm3, RSD: 125%) and the corresponding mean TEQ concentration was 4.17 pg TEQ/Nm3 (range: 1.11-15.8 pg TEQ/Nm3, RSD: 108%). Compared to the MSWIs (2.28 pg/Nm3; 0.557 pg TEQ/ Nm3), the mean PBDD/F mass and TEQ concentrations of the nine IWIs were both eight times higher. Comparing MSWIs, batch MSWIs, and IWIs with each other, the MSWIs and batch MSWIs possess the same feeding waste composition but different furnace designs and APCDs. As for the IWIs, the diversity of the industrial waste makes them more different from MSWIs and batch MSWIs. Thisconforms with the result that the mean mass and TEQ ratios (in percentage) of the PBDD/F to the PCDD/F concentration of the MSWIs (0.16%, 0.72%) and batch MWIs (0.10%, 0.18%) were similar, whereas the ratios (6.0%, 5.4%) of the IWIs were 38 ()6/0.16) and 7.5 ()5.4/0.72) times higher than those of MSWIs, respectively. The elevated PBDD/F concentrations from the IWIs, which were eight times higher than that of MSWIs, are accompanied by PCDD/Fs that are in the same range as those measured from MSWIs. This may be due to the bromine content in industrial waste being higher than that in MSW. Our previous study (36) found that the composition of the waste feed is the more important factor than the types of furnaces and APCDs on the PCDD/F characteristics in the stack flue gases of incinerators. It seems also true for PBDD/Fs which should have formation mechanisms similar to those of PCDD/Fs. VOL. 41, NO. 4, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1163

FIGURE 2. PCDD/F and PBDD/F congener profiles in the stack flue gases of MSWIs and IWIs. The calculated mass ratio (in percentage) of the PBDD/F to the PCDD/F concentration in the atmosphere of Osaka, Japan, from the research of Ohta et al. (35) was 2.45%, which is very close to the mean mass ratio (1.51%, in percentage) of the PBDD/F to the PCDD/F concentration of IWIs, omitting I1’s high ratio value. This suggests that IWIs may be one of the important PBDD/F emission sources to the atmosphere. Five large-scale fluidized-bed MSWIs with activated carbon injection and a bag filter as APCDs were assessed for the impact of coplanar PCBs on total TEQ emission (37). In 17 stack measurements, the coplanar PCBs contributed on average less than 3% to the total TEQ with a maximum contribution of 7.5% to the total TEQ in one measurement. Consequently, the TEQ contribution of PCBs from MSWIs is higher than that of PBDD/Fs but is on the same level with that of PBDD/Fs from IWIs. PCDD/F and PBDD/F Congener Profiles in the Stack Flue Gases. Figure 2 shows the congener profiles of the seventeen 2,3,7,8 chlorinated substituted PCDD/Fs and the seven 2,3,7,8 brominated substituted PBDD/Fs (mean ( SD) detected from the stack flue gases of the MSWIs and IWIs. The fraction of low chlorinated substituted PCDD/Fs, like 2,3,7,8-TeCDF, 1,2,3,7,8-PeCDF, and 2,3,4,7,8-PeCDF of the 1164

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 4, 2007

IWIs was higher than that of the MSWIs, which is consistent with the results on PBDD/Fs. The 2,3,7,8-TeBDF, 1,2,3,7,8PeBDF, and 2,3,4,7,8-PeBDF of the IWIs are more dominant than those of MSWIs. The PCDD/F and PBDD/F congener profiles of the same emission source were similar, indicating a similar substitution mechanism of bromine and chlorine during formation under the same feedstock.

Acknowledgments We thank the National Science Council in Taiwan for supporting this research work (Grant NSC-93-2218-E-230003).

Supporting Information Available Detailed cleanup procedures, HRGC/HRMS grouping and the monitor ions of PBDD/Fs; furthermore, the range, mean and RSD value of the PCDD/F concentrations in the stack flue gases of each incinerator are also listed. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Ebert, J.; Lorenz, W.; Bahadir, M. Optimization of the analytical performance of polybrominated dibenzo-p-dioxins and diben-

zofurans (PBDD/F). Chemosphere 1999, 39, 977-986. (2) Dumler, R.; Thoma, H.; Lenoir, D.; Hutzinger, O. PBDF and PBDD from the combustion of bromine containing flame retarded polymers: A survey. Chemosphere 1989, 19, 20232031. (3) Sakai, S.; Watanabe, J.; Honda, Y.; Takatsuki, H.; Aoki, I.; Futamatsu, M.; Shiozaki, K. Combustion of brominated flame retardants and behavior of its byproducts. Chemosphere 2001, 42, 519-531. (4) World Health Organization. Polybrominated dibenzo-p-dioxins and dibenzofurans; Environmental Health Criteria 205; International Programme on Chemical Safety (WHO/IPCS): Geneva, Switzerland, 1998. (5) So¨derstro¨m, G.; Marklund, S. Formation of PBCDD and PBCDF during flue gas cooling. Environ. Sci. Technol. 2004, 38, 825830. (6) Scha¨fer, W.; Ballschmiter, K. Monobromopolychloroderivatives of benzene, biphenyl, dibenzofuran and dibenzodioxin formed in chemical-waste burning. Chemosphere 1986, 15, 755-763. (7) Huang, L. Q.; Tong, H.; Donelly, J. R. Characterization of dibromopolychlorodibenzo-p-dioxins and dibromopolychlorodibenzofurans in municipal waste incinerator fly ash using gas chromatography/mass spectrometry. Anal. Chem. 1992, 64, 1034-1040. (8) Watanabe, I.; Sakai, S. Environmental release and behavior of brominated flame retardants - an overview. Organohalogen Compd. 2001, 52, 1-4. (9) Weber, R.; Kuch, B. Relevance of BFRs and thermal conditions on the formation pathways of brominated and brominatedchlorinated dibenzodioxins and dibenzofurans. Environ. Int. 2003, 29, 699-710. (10) Harless, R. L.; Lewis, R. G.; Mc Daniel, D. D.; Dupuy, A. E. Identification of bromochloro dibenzo-p-dioxins and dibenzofurans in ash samples. Chemosphere 1989, 18, 201-208. (11) Chatkittikunwong, W.; Creaser, C. S. Bromo-, bromochloro- and chloro-dibenzo-p-dioxins and dibenzofurans in incinerator fly ash. Chemosphere 1994, 29, 559-566. (12) Sovocool, G. W.; Donelly, J. R.; Munslow, W. D.; Nunn, N. J.; Todeur, Y.; Mitchum, R. K. Analysis of municipal incinerator fly ash for bromo- and bromochloro-dioxins, dibenzofurans, and related compounds. Chemosphere 1989, 18, 193-200. (13) Wanke, T.; Vehlow, J.; Mark, F. E.; Brenner, K. S. The influence of flame retarded plastic foams upon the formation of Br containing dibenzo- p-dioxins and dibenzofurans in a MSWI. Organohalogen Compd. 1996, 28, 530-535. (14) Funcke, W.; Hemminghaus, H. J.; Mark, F. E.; Vehlow, J. PXDF/D in flue gas from an incinerator charging wastes containing Cl and Br and a statistical description of the resulting PXDF/D combustion profiles. Organohalogen Compd. 1997, 31, 93-98. (15) Vehlow, J. Thermal treatment of E+E wastes. In International Seminar on Integrated Solid Waste Management, the joint auspices of Integrated Solid Waste Management Group of the International Energy Agency (IEA-ISWMG) and Japan Waste Research Foundation (JWRF); Kyoto, Japan, Sept 24-25, 1997, 16. (16) Sidhu, S. S.; Maqsud, L.; Dellinger, B.; Mascolo, G. The homogeneous, gas-phase formation of chlorinated and brominated dibenzo-p-dioxin from 2,4,6-trichloro- and 2,4,6tribromophenols. Combust. Flame 1995, 100, 11-20. (17) Sedlak, D.; Dumler-Gradl, R.; Thoma, H.; Vierle, O. Polyhalogenated dibenzo-p-dioxins and dibenzofurans in the exhaust air during textile processings. Chemosphere 1998, 37, 20712076. (18) Wichmann, H.; Dettmer, F. T.; Bahadir, M. Thermal formation of PBDD/F from tetrabromobisphenol A-a comparison of polymer linked TBBP A with its additive incorporation in thermoplastics. Chemosphere 2002, 47, 349-355. (19) So¨derstro¨m, G.; Marklund, S. PBCDD and PBCDF from incineration of waste containing brominated flame retardants. Environ. Sci. Technol. 2002, 36, 1959-1964.

(20) Schu ¨ ler, D.; Jager, J. Formation of chlorinated and brominated dioxins and other organohalogen compounds at the pilot incineration plant VERONA. Chemosphere 2004, 54, 49-59. (21) Wilken, M.; Beyer, A.; Jager, J. Generation of brominated dioxins and furans in a municipal waste incinerator (MWI): Results of a case study. Organohalogen Compd. 1990, 2, 377-380. (22) Lahl, U.; Wilken, M.; Wiebe, A. Polybrominated diphenylether in waste incineration. Mu ¨ ll Abfall 1991, 23, 83-87 (in German). (23) Taiwan EPA. 2006; http://www.epa.gov.tw/main/index.asp. (24) Rigo, H. G.; John, C. A.; Steven, L. W. The relationship between chlorine in waste stream and dioxin emissions from waste combustor stacks. ASME Res. Rep. 1995, 36, 1-1-2-3. (25) Nordic Council of Ministers. Emission measurements during incineration of waste containing bromine; Copenhagen, 2005; ISBN 92-893-1185-1. (26) Swedish Environmental Protection Agency. Fo¨rbra¨nning av ka¨llsorterat hushållsavfall, Rapport; Sweden, 1993; 4192, 2:2. (27) C¸ etin, S¸ .; Veli, S.; Ayberk, S. An investigation of halogens in Izmit hazardous and clinical waste incinerator. Waste Manage. 2004, 24, 183-191. (28) Liu, Y.; Liu, Y. Novel incineration technology integrated with drying, pyrolysis, gasification, and combustion of MSW and ashes vitrification. Environ. Sci. Technol. 2005, 39, 3855-3863. (29) Wang, L. C.; Lee, W. J.; Lee, W. S.; Chang-Chien, G. P.; Tsai, P. J. Characterizing the emissions of polychlorinated dibenzo-pdioxins and dibenzofurans from crematories and their impacts to the surrounding environment. Environ. Sci. Technol. 2003, 37, 62-67. (30) Choi, J. W.; Onodera, J.; Kitamura, K.; Hashimoto, S.; Ito, H.; Suzuki, N.; Sakai, S.; Morita, M. Modified clean-up for PBDD, PBDF and PBDE with an active carbon column-its application to sediments. Chemosphere 2003, 53, 637-643. (31) Schacht, U.; Gras, B.; Sievers, S. Determination of polybrominated and polychlorinated dibenzodioxins and -furans in various environmentally relevant materials. Organohalogen Compd. 1995, 22, 325-334 (in German). (32) Litten, S.; Mcchesney, D. J.; Hamilton, M. C.; Fowler, B. Destruction of the world trade center and PCBs, PBDEs, PCDD/ Fs, PBDD/Fs, and chlorinated biphenylenes in water, sediment, and sewage sludge. Environ. Sci. Technol. 2003, 37, 5502-5510. (33) Ashizuka, Y.; Nakagawa, R.; Tobiishi, K.; Hori, T.; Iida, T. Determination of polybrominated diphenyl ethers and polybrominated dibenzo-p-dioxins/dibenzofurans in marine products. J. Agric. Food Chem. 2005, 53, 3807-3813. (34) Hayakawa, K.; Takatsuki, H.; Watanabe, I.; Sakai, S. Polybrominated diphenyl ethers (PBDEs), polybrominated dibenzop-dioxins/dibenzofurans (PBDD/Fs) and monobromo-polychlorinated dibenzo-p-dioxins/dibenzofurans (MoBPXDD/Fs) in the atmosphere and bulk deposition in Kyoto, Japan. Chemosphere 2004, 57, 343-356. (35) Ohta, S.; Nakao, T.; Nishimura, H.; Okumura, T.; Aozasa, O.; Miyata, H. Contamination levels of PBDEs, TBBPA, PCDDs/ DFs, PBDDs/DFs and PXDDs/DFs in the environment of Japan. Organohalogen Compd. 2002, 57, 57-60. (36) Wang, L. C.; Lee, W. J.; Lee, W. S.; Chang-Chien, G. P.; Tsai, P. J. Effect of chlorine content in feeding wastes of incineration on the emission of polychlorinated dibenzo-p-dioxins/dibenzofurans. Sci. Total Environ. 2003, 302, 185-198. (37) Sakurai, T.; Weber, R.; Ueno, S.; Nishino, J.; Tanaka, M. Relevance of coplanar PCBs for TEQ emission of fluidized bed incineration and impact of emission control devices. Chemosphere 2003, 53, 619-625.

Received for review July 28, 2006. Revised manuscript received November 30, 2006. Accepted December 7, 2006. ES061801Q

VOL. 41, NO. 4, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1165