Environmental Pollution 209 (2016) 155e163

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Interplay of metals and bromine with dioxin-related compounds concentrated in e-waste open burning soil from Agbogbloshie in Accra, Ghana Takashi Fujimori a, b, Takaaki Itai c, *, Akitoshi Goto c, Kwadwo A. Asante c, d, Masanari Otsuka c, e, Shin Takahashi c, f, Shinsuke Tanabe c a

Department of Global Ecology, Graduate School of Global Environmental Studies, Kyoto University, Katsura, Nisikyo-ku, Kyoto, 615-8540, Japan Department of Environmental Engineering, Graduate School of Engineering, Kyoto University, Katsura, Nisikyo-ku, Kyoto, 615-8540, Japan Center for Marine Environmental Studies (CMES), Ehime University, Bunkyo-cho 2-5, Matsuyama, Ehime 790-8577, Japan d CSIR Water Research Institute, PO Box AH 38, Achimota, Accra, Ghana e Ehime Prefectural Institute of Public Health and Environmental Science, 8-234 Sanban-cho, Matsuyama 790-0003, Japan f Department of Environmental Conservation, Ehime University, Matsuyama 790-8577, Japan b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 September 2015 Received in revised form 12 November 2015 Accepted 19 November 2015 Available online xxx

Open burning of electronic waste (e-waste) releases various metals and organohalogen compounds in the environment. Here we investigated the interplay of metals (Cu, Pb, Zn, Fe, Co, and Sr) and bromine (Br) in the formation of dioxin-related compounds (DRCs), including polychlorinated dibenzo-p-dioxins/ furans (PCDD/Fs) and dioxin-like polychlorinated biphenyls (DL-PCBs), as well as non-regulated DRCs such as polybrominated dibenzo-p-dioxins/furans (PBDD/Fs) and their monobrominated PCDD/Fs in soils sampled from open burning e-waste sites at Agbogbloshie in Accra, Ghana. The predominant DRCs were PBDFs, PCDFs, PCDDs, and DL-PCBs. Statistical analyzes, X-ray absorption spectroscopy, and the PCDF/ PCDD ratio suggested possible formation paths of PCDD/Fs and DL-PCBs by catalytic behaviors of copper chlorides (CuCl, CuCl2, and Cu2(OH)3Cl) and thermal breakdown of polyvinyl chloride. Predominant formation of brominated furans may be derived from electron transfer from intermediates of PBDE to copper, Cu(II) / Cu(I). Lead chloride also contributed to generate DRCs and may become highly bioaccessible through the open burning of e-waste. The main zinc species (ZnCl2 and ZnS) suggested a possible relationship to generate DRCs and specific zinc source such as tire burning. Cu, Pb, Zn, and Br contained in various e-wastes, wires/cables, plastics, and tires strongly influenced generation of many DRCs. © 2015 Elsevier Ltd. All rights reserved.

Keywords: e-waste Open burning Metals Dioxin-related compounds XAFS

1. Introduction Although open burning facilitates the easy recovery of precious metals from electronic waste (e-waste) and reduces the waste mass, both toxic heavy metals and organohalogen compounds (OHCs) are released to the surrounding environment (Wong et al., 2007). Residual soil has been shown to be severely contaminated by toxic metals (including Pb, Cu, and Zn) (Wong et al., 2007) and dioxin-related compounds (DRCs) (Tue et al., 2013), including polychlorinated dibenzo-p-dioxins/furans (PCDD/Fs) (Leung et al., 2007; Zennegg et al., 2009; Wong et al., 2007) and dioxin-like

* Corresponding author. E-mail address: [email protected] (T. Itai). http://dx.doi.org/10.1016/j.envpol.2015.11.031 0269-7491/© 2015 Elsevier Ltd. All rights reserved.

polychlorinated biphenyls (DL-PCBs), (Wong et al., 2007) as well as non-regulated DRCs such as polybrominated dibenzo-p-dioxins/ furans (PBDD/Fs) (Zennegg et al., 2009) and their mixed brominated/chlorinated homologues (PXDD/Fs) (Zennegg et al., 2009) subsequent to open burning activities. The possible pathway for the generation of DRCs was proposed to be the burning of plastic products containing polyvinyl chloride (PVC) (Leung et al., 2007; Tue et al., 2013; Sepúlved et al., 2010; Wong et al., 2007) and brominated flame retardants (BFRs) (Weber and Kuch, 2003; Sepúlved et al., 2010; Tue et al., 2013; Weber and). Particularly, copper electrical wiring coated with PVC containing chlorine may contribute to the formation of PCDD/Fs through the thermochemical catalytic role of copper (Sepúlved et al., 2010; Wong et al., 2007). In Europe, similar open burning of cables for the

156

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

reclamation of metals was carried out frequently in the 1980s (Quaß et al., 2004). Therefore, contamination by PCDD/Fs has been reported since early times (Van Wijnen et al., 1992). Field studies using samples from these metal reclamation sites have also hypothesized the catalytic roles of copper to generate PCDD/Fs and PCBs (Harnly et al., 1995; Huang et al., 1992; Nie et al., 2012; Van Wijnen et al., 1992). Several laboratory-scale studies have simulated the burning of plastics and e-wastes. Researchers reported the generation of DRCs by thermal decomposition and the combustion of PVC (PCDD/Fs as the target analyte) (Cheng and Liang, 2000; McNeill et al., 1998; Theisen et al., 1989; Wang et al., 2002), BFRs [PXDD/Fs €derstro € m and Marklund., 2002) and various decomposition (So products (Barontini et al., 2005)], wires/cables (PCDD/Fs) (Gullett et al., 2007), and circuit boards [PCDD/Fs (Duan et al., 2011; Gullett et al., 2007), PBDD/Fs (Duan et al., 2011; Lai et al., 2007), and various decomposition products (Barontini et al., 2005)]. The catalytic pyrolysis of PVC was observed by the addition of metal chlorides such as CuCl2 and ZnCl2 (Cheng and Liang, 2000). The formation of PCDD/Fs was promoted by the combustion of PVC with copper chloride (Wang et al., 2002). Despite the existence of various laboratory studies, the field studies focusing on the formation mechanisms of DRCs are few in number (Liu et al., 2013). If number of the soil from open burning sites increased, the correlation result could be used to discuss the influence of metals on the formation of DRCs by open burning. In addition, the chemical forms of metals in residual soil after open burning could clarify the thermochemical interactions of target metals with other elements. X-ray absorption fine structure (XAFS) technique was applied to the residual samples derived from various thermal processes. These X-ray spectroscopic studies revealed the chemical speciation of copper (Takaoka et al., 2005a, 2005b; Wei et al., 2001), lead (Fujimori et al., 2013a; Funatsuki et al., 2012; Shah et al., 2009; Struis et al., 2009), and zinc (Takaoka et al., 2005c; Shoji et al., 2002; Struis et al., 2004; Fujimori et al., 2011) in various residual samples. Takaoka et al. (2005b) reported a positive relationship between the content of copper chloride hydroxide (CuCl2$3Cu(OH)2) and an increasing ratio of PCBs and chlorobenzenes in the municipal solid waste incineration (MSWI) fly ash. However, there was no report combined chemical forms of heavy metals with concentrations of DRCs using same field samples at e-waste open burning site. In the current study, we investigated the interplay between metals (Cu, Pb, Zn, Fe, Co, and Sr) and bromine (Br) with DRCs in soil sampled from e-waste open burning sites from Agbogbloshie in Accra, Ghana. By extending the dataset through measuring the concentrations of various DRCs such as PCDD/Fs, DL-PCBs, PBDD/Fs, and monobrominated polychlorinated dibenzo-p-dioxins/furans (MoB-PCDD/Fs), we attempted to understand the formation mechanisms of the DRCs by a combined approach of statistical analyzes and XAFS technique. Interrelationships and the grouping of parameters were assessed by Pearson's correlation analysis, principal component analysis, and hierarchical clustering. We identified chemical forms of Cu, Pb, and Zn in representative soils from open burning sites using X-ray absorption near-edge structure (XANES) spectroscopy. 2. Materials and method 2.1. Sampling of soil/ash mixtures from e-waste open burning sites Ten soil/ash mixtures (E1eE10) were collected from the Agbogbloshie market in August 2010. The site detail was described in supplementary section. During our sampling campaign, we observed that the burning of e-waste occurred mainly in two

locations (Fig. S1). The soil/ash mixture samples (0e2 cm depth) were collected using a stainless steel auger. Five subsamples within 1 m2 were sampled at each point to form one composite soil/ash mixture. To minimize chemical change after sampling, the samples were stored in a freezer in the laboratory of the Council for Scientific and Industrial Research (CSIR), Water Research Institute in Accra, following which samples were transported to Japan by air preserved through freezing using gel ice. Subsequently, the samples were registered to the Environmental Specimen Bank (es-Bank) at the Ehime University and stored at
25  C. 2.2. Measurement of DRCs DL-PCBs, PCDD/Fs, PBDD/Fs, and MoB-PCDD/Fs were identified and quantified by gas chromatography/high resolution mass spectrometry (GC-HRMS) using the isotope dilution method operating in electron impact (EI) ionization with the corresponding 13C12labeled congeners (see supplementary section for details on the pretreatment and instruments). The calculation of the limits of detection followed the Japanese Industrial Standards for measurements of PCDD/Fs and DL-PCBs (Japanese Industrial Standard, 2005). Quality assurance and quality control for our analytical methods have been confirmed by an intercalibration study on DRCs and PBDEs using an air-dried sediment sample (Takahashi et al., 2006). Toxic equivalents (TEQs) for PCDD/Fs and DL-PCBs were calculated using the World Health Organization toxic equivalency factors (WHO-TEFs), whereas TEQs for PBDD/Fs and MoB-PCDD/Fs were estimated based on WHO-TEFs of similarly substituted PCDD/Fs as the WHO and United Nations environmental program recommend the use of similar interim TEFs for brominated/chlorinated congeners in human health risk assessment (van den Berg et al., 2013). 2.3. X-ray absorption spectroscopy All XAFS measurements were conducted in the synchrotron light source beamlines in the Photon Factory, Tsukuba, Japan. The Kedge XAFS spectra of Cu, Zn, and the L3-edge of Pb were measured using hard X-ray beamlines BL9C and 12C employing a Si (111) double crystal monochrometer. The XANES spectra of reference materials were measured by the transmission mode, whereas the soil/ash mixture samples were measured by the fluorescence mode using Lytle detector. Details of XAFS measurement was appeared in supplementary section. We conducted a linear combination fit (LCF) of the XANES spectrum to determine the major species using REX 2000 ver 2.5.5 software (Rigaku, Japan). The residual value, R value ¼ S(XANESmeasd
XANEScalcd)2/S(XANESmeasd)2, was used to evaluate the LCF for the experimental spectra. If the LCF result showed <5% fit, we remove the result from the analysis. 3. Results and discussion 3.1. The pollution levels of toxic metals and DRCs According to our previous report, we selected seven elements (Fe, Co, Cu, Zn, Br, Sr, and Pb) detected in all soil samples to discuss their interplay with DRCs as shown in Table 1 (Itai et al., 2014). Median concentrations of Cu, Zn, and Pb showed levels of 2,000, 4,200, and 1100 mg/kg, respectively, which were comparable with soil concentrations from similar e-waste open burning sites at Guiyu, China (Wong et al., 2007). Wide ranges in the concentrations of toxic metals and bromine were found: 53e22,000 (Cu), 220e17,000 (Zn), 100e14,000 (Pb), and 18e1500 (Br) mg/kg. The median concentrations (and TEQ concentrations) of

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

157

Table 1 The concentration of elements and dioxin-related compounds (DRCs) in soils from e-waste open burning sites. E1 Concentrations of elements (mg/kg)a Fe 15,000 Co 4.0 Cu 170 Zn 270 Br 72 Sr 100 Pb 600 Concentrations of DRCs (pg/g dw) SPCDDs (Cl4-Cl8) 1,800 SPCDFs (Cl4-Cl8) 2,100 SDL-PCBs 3,200 SPBDDs (Br4-Br8) 46 SPBDFs (Br4-Br8) 20,000 SMoB-PCDDs (BrCl3-BrCl7) 270 SMoB-PCDFs (BrCl3-BrCl7) 860 PCDF/PCDD 1.2 PBDF/PBDD 440 MoB-PCDF/MoB-PCDD 3.2 TEQ concentrations of DRCs(pg-TEQ/g SPCDDs 12 SPCDFs 22 SDL-PCBs 4.1 SPBDDs 0.020 SPBDFs 77 SMoB-PCDDs 2.7 SMoB-PCDFs 0.74 Total TEQ 120

E2

E3

E4

E5

E6

E7

E8

E9

E10

35,000 9.0 4,800 17,000 330 130 3,100

27,000 6.0 22,000 19,000 1,500 140 9,800

20,000 N.D. 3,700 1,300 44 84 1,600

38,000 14 380 1,200 74 130 360

26,000 8.0 55 220 55 110 100

21,000 65 240 16,000 870 31 160

17,000 4.0 2,000 13,000 730 150 11,000

10,000 4.0 11,000 7,100 580 140 14,000

19,000 5.0 53 240 18 110 380

120,000 400,000 47,000 1,400 3,800,000 25,000 240,000 3.3 2,700 10

6,600 11,000 3,400 120 83,000 220 6,800 1.7 690 31

1,700 5,300 7,500 32 83,000 230 2,600 3.1 2,600 11

520 510 1,900 1.3 2,600 89 710 1.0 2,000 8.0

990 3,800 1,500 490 6,600 150 970 3.8 13 6.5

62,000 360,000 83,000 4,000 3,100,000 12,000 210,000 5.8 780 18

44,000 240,000 24,000 840 930,000 5,000 160,000 5.5 1,100 32

1,900 4,700 640 2.8 17,000 130 330 25,000

30 120 11 6.8 340 1.9 4.3 510

18 64 11 0 190 0.19 1.9 290

2.9 6.7 1.8 N.A. 12 0.12 0.19 24

25 46 5.1 16 35 3.4 1.0 130

1,400 4,100 560 3.4 15,000 120 260 21,000

790 2,600 400 4.2 3,000 34 270 7,100

70,000 200,000 42,000 3,000 370,000 17,000 38,000 2.9 120 2.2 dw) 700 2,400 380 0.64 1,700 61 30 5,300

Guiyu, China

Ref.

1,400e14,000 550e5,300

Wong et al. (2007) Wong et al. (2007)

860e7,000

Wong et al. (2007)

860 970 1,400 3.7 3,800 22 430 1.1 1,000 20

15,000 6,000 22,000e450,000 19,000 3,900,000 4,000 1,300,000 0.40 210 330

Leung et al. (2007) Leung et al. (2007) Wong et al. (2007) Zennegg et al. (2009) Zennegg et al. (2009) Zennegg et al. (2009) Zennegg et al. (2009)

5.7 11 1.7 N.A. 16 0.32 0.22 35

54 75 7.0e880

Leung et al. (2007) Leung et al. (2007) Wong et al. (2007)

94,000 (CALUX)

Zennegg et al. (2009)

N.D. not detected, N.A. not analyzed a These values of E1-E10 were quantified in previous our study (Itai et al., 2014).

chlorinated DRCs such as SPCDDs, SPCDFs, and SDL-PCBs were 4200 (28), 8200 (93), and 5500 pg/g (11 pg-TEQ/g), respectively. Soil E3 displayed the highest level of pollution by SPCDDs (120,000 pg/g) and SPCDFs (400,000 pg/g), which was higher than soil concentrations in similar open burning sites at Guiyu, China (Leung et al., 2007), as shown in Table 1. The highest concentration of SDL-PCBs was found in soil E8 (83,000 pg/g). PCDF/PCDD ratios (median, 3.0) displayed higher values in the cases of a severe contamination by PCDFs (r ¼ 0.691, p < 0.05). High PCDF/PCDD ratios in E8 and E9 (5.5e5.8) were similar to values found in ash and soil from metal recovery facilities (Harnly et al., 1995; Kuykendal et al., 1989) and open burning sites (Harnly et al., 1995). Formerly operated metal recoveries and open burning sites normally processed copper wires and scraps (Harnly et al., 1995; Huang et al., 1992; Nie et al., 2012; Van Wijnen et al., 1992). Therefore, soils E8 and E9 were thought to have a similar formation path of chlorinated DRCs with former thermal metal (particularly, copper) reclamation activity. In addition, the pyrolysis of PVC influenced the selective formation of PCDFs (McNeill et al., 1998). The mechanism of PVC breakdown proposed by McNeill et al. (1998) is also one of the possible paths for the generation of PCDFs during e-waste burning. Through the measurement of brominated DRCs (PBDD/Fs), we found extremely high concentrations of PBDFs in soil of e-waste open burning sites. Median concentrations and TEQ values of SPBDFs displayed 83,000 pg/g and 260 pg-TEQ/g, respectively. In contrast, SPBDDs showed the lowest-level median concentration of 300 pg/g (1.7 pg-TEQ/g) among all DRCs measured in the current study. The concentration of SPBDFs (and the concentrations of TEQ) in soils E3 and E8 was 3,800,000 (17,000) and 3,100,000 pg/g (15,000 pg-TEQ/g), respectively, which was higher than that in other soils by one or two orders of magnitude and were comparable with previous data of soils at open burning sites (Zennegg et al., 2009).

The predominant formation of PBDFs rather than PBDDs (median PBDF/PBDD ratio ¼ 890) indicated a similar order of PBDF/ PBDD ratio to that of soils from similar e-waste open burning sites at Guiyu, China (PBDF/PBDD ratio ¼ 210 in Table 1) (Zennegg et al., 2009), from the thermal degradation/pyrolysis residues of high impact polystyrene blended by polystyrene, polybutadiene, decabromodiphenyl ether (Deca-BDE), and antimony oxide (Sb2O3) (PBDF/PBDD ratio > 200) (Luijk et al., 1991), and from an accidental fire residue from a plastic TV-case that may contain various BFRs (PBDF/PBDD ratio ¼ 295) (Zelinski et al., 1993). The addition of the metal oxides Sb2O3 and Fe2O3 influenced the condensation of DecaBDE to PBDFs (Dumler et al., 1990; Lenoir et al., 1994; Luijk et al., ~ o et al. (2014) recently reported the thermal de novo 1991). Ortun formation of PBDD/Fs by the addition of copper bromide (CuBr2) to activated carbon. Such thermal formation mechanisms of PBDD/Fs involving plastics, BFRs (particularly, PBDEs), and metal compounds occurred possibly during e-waste open burning. The PBDFdominant profile was also reported in samples from e-waste recycling (Tue et al., 2013), technical PBDE formulations (Hanari et al., 2006), PBDE-containing TV cases and circuit boards (Duan et al., 2011; Sakai et al., 2001; Watanabe and Sakai, 2003). Therefore, PBDFs can be generated from PBDE precursors in various ewastes under mild thermal condition (Duan et al., 2011; Weber and Kuch, 2003) or natural sunlight (Kajiwara et al., 2008). SMoB-PCDDs and -PCDFs in soils of e-waste open burning sites had median concentrations of 250 and 4700 pg/g, respectively. The median TEQ concentrations of SMoB-PCDDs and -PCDFs were 3.1 and 3.1 pg-TEQ/g, respectively. The highest concentrations were found in soil E3 that had higher SMoB-PCDDs (25,000 pg/g) and lower SMoB-PCDFs (240,000 pg/g) than soils from e-waste open burning sites from China (Zennegg et al., 2009) as shown in Table 1. The concentration of MoB-PCDFs was 11 times higher than that of MoB-PCDDs based on the median MoB-PCDF/MoB-PCDD ratio, and one-order higher (330) than the ratio from the soil samples of other

158

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

€ derstro € m and e-waste open burning sites (Zennegg et al., 2009). So Marklund (2002) reported that when BFRs were combusted with artificial municipal solid waste (containing paper, plastics, organic matter, metal, and chlorine), monobrominated trichlorinated dibenzo furans were generated at a concentration 15e35 times higher than moBr triCl dibenzo-p-dioxins.

3.2. Statistical analysis The linkage of the spatial distributions of DRCs with that of toxic and alkaline earth metals and bromine was analyzed by the hierarchical clustering method (Ward, 1963) which was described in supplementary section. Dendrogram (Fig. 1) shows that Fe and Co separated from other elements to shared group I. These two metals did not link with pollution patterns of all DRCs. Another large grouping of the dendrogram consisted of three subgroups IIa (Sr, SPBDDs, and SDL-PCBs), IIb (Cu, Pb, SPCDFs, SPBDFs, and SMoBPCDFs), and IIc (Br, Zn, SPCDDs, and SMoB-PCDDs). Therefore, Cu and Pb show similar pollution patterns derived from e-waste open burning with halogenated furans. The high similarity in pollution pattern between DRCs and Cu and Pb suggests a common pollution source by open burning of specific e-waste containing BFRs, copper, and lead such as wires/cables and circuit boards, which is consistent with actual open burning activities. Component scores by principle component analysis (PCA) from the correlation matrix containing DRCs and elements (Fig. 2) indicated that PC1 and PC2 explained 65.4% and 14.9% (total, 80.3%) of the original dataset, respectively. PC1 showed group I as a negative value and group II as a positive value (Table S1). Particularly, subgroup IIb displayed the highest positive PC1 values as shown in Fig. 2. Pb and Cu are the representatives of high-risk toxic metals in soil from open burning sites (Itai et al., 2014; Wong et al., 2007). In addition, we found PBDFs and PCDD/Fs as major dioxin-like toxins among all DRCs measured in the current study (ref. eq. (1)). Thus, PC1 may be interpreted as the “impact of e-waste open burning.”

ΣPCDDs

IIc

ΣMoB-PCDDs Zn Br ΣPCDFs ΣMoB-PCDFs

3.3. The chemical forms of copper, lead, and zinc

Cu

II

ΣPBDFs Pb

Minor component PC2 of subgroup IIb, SPCDDs, SPBDDs, SDLPCBs, and SMoB-PCDDs was not varied and was concentrated at around the zero value. In contrast, Sr, Br, Zn, and Co in other (sub-) groups showed positive and negative PC2 values. A highly negative PC2 was found in only Sr in the soil samples from open burning sites that showed comparable concentrations with the average upper continental crust (Wedepohl, 1995). Br, Zn, and Co showed relatively high positive PC2 values. PC2 implied “specific sources of such pollutants” except major e-waste open burning, although it was difficult to simplify the PC2 character. Pearson's correlation analysis revealed significant relationship among Cu, Pb, and many DRCs (Fig. 3). Among seven elements, copper significantly showed the strongest positive correlations with the total concentration of PCDDs, PCDFs, PBDFs, and MoBPCDFs (p < 0.001) as shown in Fig. 3. Detailed correlation coefficients are shown in Table S2. However, PBDDs was not correlated with Cu. Lead showed significant positive correlations with total the concentration of MoB-PCDFs (p < 0.001) and PCDFs (p < 0.01). Lead also correlated with the total concentration of PCDDs, PBDFs, and DL-PCBs (p < 0.05). Zinc and bromine were included in the same subgroup IIc (ref. Figs. 1 and 2) and we suggested a specific source except for major pollutants such as Cu, Pb, PCDFs, PBDFs, and MoB-PCDFs. Correlation analyzes revealed that Zn and Br are related with some DRCs. Zinc showed a significant positive relationship with the total concentration of PCDDs (p < 0.01), MoB-PCDDs (p < 0.01), PBDDs (p < 0.05), PCDFs (p < 0.05), and DL-PCBs (p < 0.05) as shown in Fig. 3. Bromine displayed significantly (p < 0.01) positive correlations with brominated furans i.e., PBDFs and MoB-PCDD/Fs. Bromine was one of the key indicators of brominated DRCs because various e-waste containing BFRs possibly related with the Br content in soils of e-waste open burning sites. However, the formation of PBDDs did not relate with Br. In addition, Br significantly correlated with chlorinated DRCs such as total the concentrations of PCDDs (p < 0.01) and PCDFs (p < 0.05). The burning of e-waste did not contribute iron, cobalt, and strontium to the soil based on comparison of their concentrations (Table 1) with that of the average upper continental crust (Wedepohl, 1995). All DRCs showed no correlation with Fe. There was no relationship between Co and all DRCs. Sr showed no correlations with PCDDs and brominated DRCs. Therefore, Fe, Co, and Sr had a weak potential for the thermochemical formations of PBDFs and PCDDs that were serious toxicants and MoB-PCDD/Fs during e-waste open burning. Strontium correlated with the total concentration of PCDFs (p < 0.05) and DL-PCBs (p < 0.05).

IIb

ΣDL-PCBs ΣPBDDs

IIa

Sr Fe

I

Co 5

Fig. 1. The hierarchical clustering of elements and dioxin-related compounds (DRCs) in soils from e-waste open burning sites (n ¼ 10). Red and white diamonds indicate toxic and alkaline earth metals, respectively. The white square represents halogen. The black circle represents the total concentration of DRCs. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Because Cu was the most significant causative toxic metal, we selected six soil samples from open burning sites (E2, E3, E4, E7, E8, and E9) that showed a wide-range of concentrations of toxic compounds among all soils (ref. Table 1) for characterization. Cu Kedge XANES spectra of these soils were categorized into three types A (E2, E4), B (E3, E8, E9), and C (E7) as shown in the left panel of Fig. 4. The linear combination of XANES spectra of reference Cu compounds fitted well to measured spectra. Major components were identified as 42e44% cuprospinel (CuFe2O4) in type A, 38e45% cuprous chloride (CuCl) in type B, and 49% copper carbonate (CuCO3) in type C (right, Fig. 4). Copper is expected to occur within e-wastes and wires/cables as the metallic form or oxide by weathering before burning. However, copper chlorides (CuCl, CuCl2, and Cu2(OH)3Cl) were identified in five soils after burning, indicating that a specific chlorine source in e-wastes such as PVC thermo-chemically interacted with copper. CuCl, CuCl2, and Cu2(OH)3Cl were also identified in fly ash at a post-combustion

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

159

0.2

PC2 (14.9%)

1

Co

Subgroup IIc Zn

0.5

ΣPBDDs

Br

Group I

ΣPCDFs ΣDL-PCBs

-0.1

-0.5

Cu ΣMoB-PCDFs

-0.2

Sr

ΣPCDDs ΣPBDFs

0

Fe

0

ΣMoB-PCDDs

0.1

Pb

Subgroup IIb

Subgroup IIa -1 -0.5

-0.3 0 0.5 PC1 (65.4%)

1

0.7

0.8

0.9 1 PC1 (65.4%)

1.1

Fig. 2. The principle component analysis of elements and dioxin-related compounds (DRCs) in soils of e-waste open burning sites (n ¼ 10). Explanation of symbols is as described in Fig. 1.

Cu Pb

PCDDs PBDDs MoB-PCDDs

Zn Br Fe

PCDFs PBDFs MoB-PCDFs

Co Sr

DL-PCBs p < 0.001 p < 0.01 p < 0.05

Fig. 3. Pearson's correlation analyzes between elements (Cu, Pb, Zn, Br, Fe, Co, and Sr) and dioxin-related compounds (DRCs) using the dataset of soils of e-waste open burning sites as shown in Table 1 (n ¼ 10). The total concentration of DRCs was used in the analysis. Red bold, red normal, and black dashed lines indicate positive correlation with significance at p < 0.001, p < 0.01, and p < 0.05, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

zone in a MSWI (Takaoka et al., 2005a, 2005b; Tian et al., 2009). In general, PCDD/Fs and PCBs concentrated in MSWI fly ash and copper chlorides played a role of a catalyst (Fujimori et al., 2009). The thermal characterization studies of copper reported that the oxychlorination cycle of copper involving CuCl, CuCl2, and Cu2(OH)3Cl catalyzed the chlorination of carbon (Takaoka et al., 2005a) and that reduction from CuCl2 to CuCl influenced direct chlorination (Fujimori and Takaoka, 2009). From previous findings using the MSWI process which has big differences compared with e-waste open burning, the open burning of e-waste in addition facilitated such mechanisms by copper chlorides to generate chlorinated DRCs (Fujimori et al., 2013b), strongly supported by the existence of copper chlorides in soils of burning sites derived from Cu K-edge XANES analyzes. Three copper chlorides, CuFe2O4, and cupric acetate (Cu(CH3COO)2) were fitted as part of copper chemical forms in 4e6 soils. No fitting result was treated as 0% of such chemical forms and we analyzed correlations between the amount of these copper chemicals and DRCs. Although the sample number was only six (n ¼ 6), the significant positive correlations of CuCl with the total

concentration of PCDFs (p < 0.05) and MoB-PCDFs (p < 0.01) was evident as shown in Table S3. CuCl showed the highest positive correlation coefficients with the total concentration of DL-PCBs (r ¼ 0.69, p ¼ 0.128) and PBDFs (r ¼ 0.77, p ¼ 0.072) among copper chemicals. CuCl2 also displayed the highest positive correlation coefficients with the total concentration of PCDDs (r ¼ 0.73, p ¼ 0.103) and MoB-PCDDs (r ¼ 0.76, p ¼ 0.078). Cu2(OH)3Cl displayed relatively high positive r values in PCDD/Fs, PBDFs, and MoB-PCDD/Fs. In contrast, there was no clear correlation of CuFe2O4 and Cu(CH3COO)2 with DRCs. Therefore, the amount of copper chlorides influenced the formation of many DRCs during the open combustion of e-waste. Particularly, the remaining quantity of CuCl significantly related with the formation of total and dioxinlike chlorinated furans (PCDFs and MoB-PCDFs). PCDDs, DL-PCBs, and MoB-PCDDs were possibly correlated with CuCl, although we will need to confirm significant correlations by increasing sample numbers. It is suggested that CuCl is the reduction product of CuCl2 or Cu2(OH)3Cl via a thermal process, following which the chlorination of carbon progresses, which is consistent with the experimental results reporting formation mechanisms of chlorinated DRCs described above in detail (Kajiwara et al., 2008; Takaoka et al., 2005a). According to relationships between copper compounds and PBDDs, all copper compounds did not have any linkage with PBDDs. In contrast, PBDFs had a possible positive correlation with the amount of CuCl. We have already discussed the significant influence of CuCl to generate chlorinated furan structures. A recent mechanistic study reported that the formation of PBDFs via the ring-closure of PBDEs was the most acceptable pathway (Altarawneh and Dlugogorski., 2013), implying that electron transfer from the intermediate structure of PBDEs to copper i.e., Cu(II)/Cu(I) occurred. In addition, there were possible other pathways of PBDD/Fs (especially, PBDFs) from direct structural arrangement of BFRs or from thermal decomposition precursors such as bromophenols (Altarawneh and Dlugogorski, 2014a, b). Three soils E3, E8, and E9 showed the highest levels of contamination by DRCs and were categorized in type B (mainly occupied as CuCl) through Cu K-edge XANES characterization. We focused on the severest pollution of Pb (9800e14,000 mg/kg) in the type-B soils (ref. Table 1). By applying Pb L3-edge XANES technique to these soils, we found that PbCl2 comprised 51e56% of the major chemical form of Pb as shown in Fig. 5. Lead originally occurs as a metallic state in various e-wastes or oxidative forms by weathering. Chlorination of lead was considered to have occurred via the thermal degradation of plastic materials such as PVC. PbCl2 was reported to promote the thermochemical formation of chlorinated

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

Normalized absorption ( - )

160

CuCl2 CuCl2

CuCl CuCl

Cu2(OH)3Cl Cu2(OH)3Cl

CuSO4 CuSO4

E7

Cu Cu

Cu(CH3COO)2 Cu(CH3COO)2

CuCO3

CuSiO3·H2O CuSiO3·H2O

CuFe2O4 CuFe2O4

CuCO3 CuCO3 E9 E8 Cu3(PO4)2 Cu3(PO4)2 E3 CuCl E7 E4 E2 E9 CuFe2O4 E8

Cu(OH)2 Cu(OH)2

CuO CuO

E3 E4 E2

8960

9000 9040 Photon energy (eV)

0

50 100 Ratio of copper forms (%)

Fig. 4. The chemical forms of copper in soils from six representative e-waste open burning sites. Left, Measurement spectra of Cu K-edge XANES. Solid and dashed lines are measurement and reference spectra, respectively. Colored circle indicates fitting spectrum by the linear combination of references. Right, Ratio of copper chemical forms.

Normalized absorption ( - )

Pb5(PO4)3Cl

PbCl2 PbCl2 Pb5(PO4)3Cl Pb5(PO4)3Cl

Pb3(CO3)2(OH)2 E9 E8 E3 E9 PbCl2

Pb Pb Pb3(CO3)2(OH)2 Pb3(CO3)2(OH)2

Pb E8

E3

13040

13070 13100 Photon energy (eV)

0

50

100

Ratio of lead forms (%)

Fig. 5. The chemical forms of lead in the representative soil samples from three e-waste open burning sites. Left, Measurement spectra of Pb L3-edge XANES. Solid and dashed lines are measurement and reference spectra, respectively. Colored circle indicates fitting spectrum by the linear combination of references. Right, The ratio of lead chemical forms.

aromatics under coexistence condition with CuCl2, using model MSWI fly ash (Fujimori et al., 2013a). Statistically significant correlations between Pb and DRCs give support to the promotion of chlorinated DRCs formation by Pb (ref. Fig. 3). Therefore, PbCl2 may play a role as a promotive factor to generate chlorinated DRCs during open burning of particular e-wastes containing Pb, Cu, and chlorine. A unique Cu K-edge XANES was measured in soil E7 (Fig. 4). No copper chlorides were identified in E7. The XANES experiment indicated that this soil contained smaller amount of Cu (240 mg/kg) than other soils. The open burning of wires/cables or other Cucontaining materials was not thought to be a common occurrence. Lead concentrations also showed a low value (160 mg/kg). On the other hand, the high concentrations of Zn were evident in E7 (16,000 mg/kg in Table 1). We selected type B (E3, E8, and E9) and

specific Zn-rich type C (E7) to characterize chemical forms by using the Zn K-edge XANES technique. According to the shape of the spectrum and LCF, Fig. 6 indicates that the dominant chemical forms of zinc were 44e51% ZnCl2 in E3 and E8, 38% carbonate (ZnCO3) in E9, and 33% ZnS in E7. All four soils contained zinc chloride (ZnCl2) and sulfide (ZnS). Zinc was chlorinated by thermal degradation of plastics (particularly, PVC) as well as Cu and Pb. During thermal treatment simulated MSWI fly ash, ZnCl2 had the potential to generate chlorinated aromatics, which has been reported previously (Fujimori et al., 2009, 2011). However, the coexistence of ZnCl2 with copper catalyst (CuCl2) resulted in decreasing chlorinated aromatics (Fujimori et al., 2011). ZnCl2 may balance the formation of chlorinated DRCs depending on the distribution of ZnCl2 and CuCl2. A relative high ratio of the sulfide form was only found in Zn as

Normalized absorption ( - )

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

161

ZnCl2 ZnCl2

ZnS ZnS

E7

ZnAl2O4 ZnAl2O4

ZnFe2O4 ZnFe2O4

ZnS

ZnO ZnO

Zn(OH)2 Zn(OH)2

E9

ZnCO3 ZnCO3

ZnC2O4 ZnC2O4

ZnCO3 E8 E3

E7 E9

ZnCl2 E8 E3

9640

9670

9700

Photon energy (eV)

0

50 Ratio of zinc forms (%)

100

Fig. 6. The chemical forms of zinc in the representative soil samples from four e-waste open burning sites. Left, Measurement spectra of Zn K-edge XANES. Solid and dashed lines are measurement and reference spectra, respectively. Colored circle indicates fitting spectrum by the linear combination of references. Right, The ratio of zinc chemical forms.

the specific chemical form among three toxic metals (Cu, Pb, and Zn). We observed open burning of car tires in the study site as shown in Fig. S2. The small-angle neutron scattering method revealed ZnS forms in styrene-butadiene rubber that is represented in tire materials through vulcanization (Ikeda et al., 2009). We suggest that the open burning of tires mainly contributed to zincrich soil E7 and other soils were partly affected by contamination by burning residue of tires. 3.4. Toxicological implications As compared to the other regions in the world polluted by DRCs, the TEQ levels observed in the current study were relatively high. The TEQ concentrations of SPCDDs, SPCDFs, and SDL-PCBs in soil E3 were the highest values, namely 1,900, 4,700, and 640 pg-TEQ/g, respectively. Subtotal 7240 pg-TEQ/g of PCDD/Fs and DL-PCBs in soil E3 exceeded the soil guideline (1000 pg-TEQ/g) which was historically selected in many countries (Paustenbach et al., 2006) by a factor of over seven-fold. The remaining three soil samples from open burning sites had a similar order of the TEQ levels of chlorinated DRCs (3480 in E2, 6060 in E8, and 3790 pg-TEQ/g in E9). TEQ concentrations of SPBDFs in soils E3 and E8 were 15e17 times higher than the guidelines and occupied 68 and 71%, respectively, in total TEQ concentrations (Table 1). Therefore, PBDFs concentrated in soils from e-waste open burning sites are major compounds influencing human health via the similar toxic mechanism of 2,3,7,8-tetra chlorinated dioxin (TCDD). We found a lower contribution of PBDDs to the toxic effects because of small TEQ levels (<6.8 pg-TEQ/g, SPBDDs). As compared with PCDD/Fs, the TEQ concentrations of PBDD/Fs were four-times higher. Simulated open burning of circuit board resulted in considerably higher TEQ concentrations of PBDD/Fs than PCDD/Fs (>100-fold) because bromine in circuit boards generally exceeded chlorine concentrations (Duan et al., 2011; Gullett et al., 2007). In the case of Agbogbloshie at Accra, Ghana, workers burned mainly wires/cables containing PVC. PCDD/Fs concentrations showed relatively higher TEQ concentrations subsequent to open burning at the site of the current study. Considering all results, the contribution of DRCs to TEQ concentrations in the current study had the following rank: PBDFs (median percentage of contribution ¼ 57%,

n ¼ 10) > PCDFs (26%) > PCDDs (11%) > DL-PCBs (3.9%) > MoBPCDFs (0.78%) > MoB-PCDDs (0.55%) > PBDDs (0.017%) (1). The summation of PBDFs and PCDD/Fs occupied 94% among total TEQ concentrations of DRCs measured in the current study. A dioxin-responsive chemically activated luciferase gene expression (DR-CALUX) assay revealed an extremely high level dioxin-like toxicity of only persistent halogenated aromatics (94,000 pg CALUX-TEQ/g) within the soil from a similar e-waste open burning site in China as shown in Table 1 (Zennegg et al., 2009). As compared with our results of total TEQs (24e25,000 pg-TEQ/g), there exists other unknown DRCs contributing to dioxin-like toxicity. Unaccounted dioxin-like activity was also reported in settled house dust samples from e-waste recycling sites (Tue et al., 2010). Itai et al. (2014) previously reported that Pb was the most hazardous among various metal(loid)s in this site. According to the bioaccessibility test by a study using 1 M HCl, 61e86% of Pb was found to be in a bioaccessible form. Pb metal and chlorinated pyromorphite (Pb5(PO4)3Cl) had relatively very low bioaccessibility (Kumpiene et al., 2008; Rasmussen et al., 2013). The subtotals of Pb þ Pb5(PO4)3Cl were calculated as 12e49% (Fig. 5) in this study. Thus, bioaccessible Pb was estimated at 51e88% by linear combination analyzes of Pb L3-edge XANES, which was consistent with a previous experiment using 1 M HCl (Itai et al., 2014). PbCl2 mainly contributed to bioaccessible Pb, suggesting that change of Pb chemical form to PbCl2 via e-waste open burning increased its bioaccessibility. 4. Conclusions Soil samples from the sites were severely polluted by toxic metals and DRCs (PCDD/Fs, PBDD/Fs, and MoB-PCDD/Fs). The summation of PBDFs and PCDD/Fs occupied 94% among total TEQ concentrations of DRCs measured in the current study. The PCDF/ PCDD ratio indicated a selective formation of PCDFs over PCDDs. The predominant formation of PBDFs rather than PBDDs was also found. Excess formation of MoB-PCDFs (i.e., high MoB-PCDF/MoBPCDD ratio) may be a common trend pertaining to the burning of e-waste containing BFRs. Statistical analyses suggested that Cu and Pb significantly related with many DRCs in the soil samples from e-waste open

162

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163

burning sites. Zinc and bromine were derived from a specific source except for major pollutants. We characterized these three heavy metals in representative soils from open burning sites using the XANES techniques and discussed the impact of chemical forms on the formation mechanisms of DRCs. Identification of copper chlorides (CuCl, CuCl2, and Cu2(OH)3Cl) indicated that a specific chlorine source in e-wastes such as PVC thermo-chemically interacted with copper. Copper chlorides significantly influenced the formation of many DRCs during the open burning of e-waste. Chlorination of lead (PbCl2) was mainly identified via the thermal degradation of plastic materials such as PVC. PbCl2 may play a role as a promotive factor to generate chlorinated DRCs. Representative soils from e-waste open burning contained zinc chloride (ZnCl2) and sulfide (ZnS). Acknowledgments This study was supported in part by Grants-in-Aid for Scientific Research (A: 25257403) from the Japanese Society for the Promotion of Science (JSPS), and the Environment Research and Technology Development Fund (3K133001) from the Japanese Ministry of the Environment. The authors greatly acknowledge the financial support provided by a Grant-in-Aid for a Young Scientist (A) from JSPS (No. 26701012) and the Grant for Environmental Research Projects from the Steel Foundation for Environmental Protection Technology (No. 13C-29 (2013) and 14C-30-23 (2014)). We thank Hiroshi Nitani (BL12C and 9A) for support with XANES measurements at KEK-PF (Proposals 2009G632). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2015.11.031. References Altarawneh, M., Dlugogorski, B.Z., 2013. A mechanistic and kinetic study on the formation of PBDD/Fs from PBDEs. Environ. Sci. Technol. 47, 5118e5127. Altarawneh, M., Dlugogorski, B.Z., 2014a. Mechanism of thermal decomposition of tetrabromobisphenol A (TBBA). J. Phys. Chem. A 118, 9338e9346. Altarawneh, M., Dlugogorski, B.Z., 2014b. Thermal decomposition of 1,2-Bis(2,4,6tribromophenoxy)ethane (BTBPE), a novel brominated flame retardant. Environ. Sci. Technol. 48, 14335e14343. Barontini, F., Marsanich, K., Petarca, L., Cozzani, V., 2005. Thermal degradation and decomposition products of electronic boards containing BFRs. Ind. Eng. Chem. Res. 44, 4186e4199. Cheng, W.H., Liang, Y.C., 2000. Catalytic pyrolysis of polyvinylchloride in the presence of metal chloride. J. Appl. Polym. Sci. 77, 2464e2471. Duan, H., Li, J., Liu, Y., Yamazaki, N., Jiang, W., 2011. Characterization and inventory of PCDD/Fs and PBDD/Fs emissions from the incineration of waste printed circuit board. Environ. Sci. Technol. 45, 6322e6328. Dumler, R., Lenoir, D., Thoma, H., Hutzinger, O., 1990. Thermal formation of polybrominated dibenzofurans and dioxins from decabromdiphenylether flame retardants: influence of antimony (III) oxide and the polymer matrix. Chemosphere 20, 1867e1873. Fujimori, T., Takaoka, M., 2009. Direct chlorination of carbon by copper chloride in a thermal process. Environ. Sci. Technol. 43, 2241e2246. Fujimori, T., Takaoka, M., Takeda, N., 2009. Influence of Cu, Fe, Pb and Zn chlorides and oxides on formation of chlorinated aromatic compounds in MSWI fly ash. Environ. Sci. Technol. 43, 8053e8059. Fujimori, T., Tanino, Y., Takaoka, M., 2011. Role of zinc in MSW fly ash during formation of chlorinated aromatics. Environ. Sci. Technol. 45, 7678e7684. Fujimori, T., Tanino, Y., Takaoka, M., 2013a. Thermochemical behavior of lead adjusting formation of chlorinated aromatics in MSW fly ash. Environ. Sci. Technol. 47, 2169e2176. Fujimori, T., Takigami, H., Takaoka, M., 2013b. Organochlorines in surface soil at electronic-waste wire burning sites and metal contribution evaluated using quantitative X-ray speciation. J. Phys. Conf. Ser. 430, 012094. Funatsuki, A., Takaoka, M., Oshita, K., Takeda, N., 2012. Methods of determining lead speciation in fly ash by X-ray absorption finestructure spectroscopy and a sequential extraction procedure. Anal. Sci. 28, 481e490. Gullett, B.K., Linak, W.P., Touati, A., Wasson, S.J., Gatica, S., King, C.J., 2007. Characterization of air emissions and residual ash from open burning of electronic wastes during simulated rudimentary recycling operations. J. Mater. Cycles

Waste Manag. 9, 69e79. Hanari, N., Kannan, K., Okazawa, T., Kodavanti, P.R.S., Aldous, K.M., Yamashita, N., 2006. Occurrence of polybrominated biphenyls, polybrominated dibenzo-pdioxins, and polybrominated dibenzofurans as impurities in commercial polybrominated diphenyl ether mixtures. Environ. Sci. Technol. 40, 4400e4405. Harnly, M., Stephens, R., McLaughlin, C., Marcotte, J., Petreas, M., Goldman, L., 1995. Polychlorinated dibenzo-p-dioxin and dibenzofuran contamination at metal recovery facilities, open burn sites, and a railroad car incineration facility. Environ. Sci. Technol. 29, 677e684. Huang, C.-W., Miyata, H., Lu, J.-R., Ohta, S., Chang, T., Kashimoto, T., 1992. Levels of PCBs, PCDDs and PCDFs in soil samples from incineration sites for metal reclamation in Taiwan. Chemosphere 24, 1669e1676. Ikeda, Y., Higashitani, N., Hijikata, K., Kokubo, Y., Morita, Y., Shibayama, M., et al., 2009. Vulcanization: new focus on a traditional technology by small-angle neutron scattering. Macromolecules 42, 2741e2748. Itai, T., Otsuka, M., Asante, K.A., Muto, M., Opoku-Ankomah, Y., Ansa-Asare, O.D., Tanabe, S., 2014. Variation and distribution of metals and metalloids in soil/ash mixtures from Agbogbloshie e-waste recycling site in Accra, Ghana. Sci. Total Environ. 470e471, 707e716. Japanese Industrial Standard (JIS) K0312, 2005. Method for Determination of Tetrathrough Octachlorodibenzo-p-dioxins, Tetrathrough Octachlorodibenzofurans and Dioxin-like Polychlorinatedbiphenyls in Industrial Water and Waste Water. Kajiwara, N., Noma, Y., Takigami, H., 2008. Photolysis studies of technical decabromodiphenyl ether (DecaBDE) and ethane (DeBDethane) in plastics under natural sunlight. Environ. Sci. Technol. 42, 4404e4409. Kumpiene, J., Lagerkvist, A., Maurice, C., 2008. Stabilization of As, Cr, Cu, Pb and Zn in soil using amendments e a review. Waste Manag. 28, 215e225. Kuykendal, W.B., Lamason II, W.H., Miles, A.J., Keating, M.H., 1989. Ash data from combustion sources: results of tier 4 of the National Dioxin study. Chemosphere 18, 1227e1234. Lai, Y., Lee, W., Li, W., Wang, L., Chang-Chien, G., 2007. Inhibition of polybrominated dibenzo-p-dioxin and dibenzofuran formation from the pyrolysis of printed circuit boards. Environ. Sci. Technol. 41, 957e962. Lenoir, D., Zier, B., Bieniek, D., Kettrup, A., 1994. The influence of water and metals on PBDD/F concentration in incineration of decabromobiphenyl ether in p246olymeric matrices. Chemosphere 28, 1921e1928. Leung, A.O.W., Luksemburg, W.J., Wong, A.S., Wong, M.H., 2007. Spatial distribution of polybrominated diphenyl ethers and polychlorinated dibenzo-p-dioxins and dibenzofurans in soil and combusted residue at Guiyu, an electronic waste recycling site in southeast China. Environ. Sci. Technol. 41, 2730e2737. Liu, M., Huang, B., Bi, X., Ren, Z., Sheng, G., Fu, J., 2013. Heavy metals and organic compounds contamination in soil from an e-waste region in South China. Environ. Sci. Process. Impacts 15, 919e929. Luijk, R., Wever, H., Olie, K., Govers, H.A.J., Boon, J.J., 1991. The influence of the polymer matrix on the formation of polybrominated dibenzo-p-dioxins (PBDDs) and polybrominated dibenzofurans (PBDFs). Chemosphere 23, 1173e1183. McNeill, I.C., Memetea, L., Mohammed, M.H., Fernandes, A.R., Ambidge, P., 1998. Polychlorinated dibenzodioxins and dibenzofurans in PVC pyrolysis. Polym. Degrad. Stab. 62, 145e155. Nie, Z., Zheng, M., Liu, G., Lv, P., Zhang, B., Su, G., et al., 2012. A preliminary investigation of unintentional POP emissions from thermal wire reclamation at industrial scrap metal recycling parks in China. J. Hazard. Mater. 215e216, 259e265. ~ o, N., Conesa, J.A., Molto , J., Font, R., 2014. De novo synthesis of brominated Ortun dioxins and furans. Environ. Sci. Technol. 48, 7959e7965. Paustenbach, D.J., Fehling, K., Scott, P., Harris, M., Kerger, B.D., 2006. Identifying soil cleanup criteria for dioxins in urban residential soils: how have 20 years of research and risk assessment experience affected the analysis? J. Toxicol. Environ. Health Part B Crit. Rev. 9, 87e145. €ker, G., 2004. The European dioxin air emission inventory Quaß, U., Fermann, M., Bro project e final results. Chemosphere 54, 1319e1327. nier, M., Levesque, C., MacLean, L.C.W., Rasmussen, P.E., Beauchemin, S., Che Marro, L., et al., 2013. Canadian house dust study: lead bioaccessibility and speciation. Environ. Sci. Technol. 47, 4959e4965. Sakai, S., Watanabe, J., Honda, Y., Takatsuki, H., Aoki, I., Futamatsu, M., Shiozaki, K., 2001. Combustion of brominated flame retardants and behavior of its byproducts. Chemosphere 42, 519e531. Sepúlved, A., Schluep, M., Renaud, F.G., Streicher, M., Kuehr, R., Hagelüken, C., Gerecke, A.C.A., 2010. Review of the environmental fate and effects of hazardous substances released from electrical and electronic equipments during recycling: examples from China and India. Environ. Impact Assess. Rev. 30, 28e41. Shah, P., Strezov, V., Nelson, P.F., 2009. X-ray absorption near edge structure spectrometry study of nickel and lead speciation in coals and coal combustion products. Energy Fuels 23, 1518e1525. Shoji, T., Huggins, F.E., Huffman, G.P., Linak, W.P., Miller, C.A., 2002. XAFS spectroscopy analysis of selected elements in fine particulate matter derived from coal combustion. Energy Fuels 16, 325e329. €derstro €m, G., Marklund, S., 2002. PBCDD and PBCDF from incineration of wasteSo containing brominated flame retardants. Environ. Sci. Technol. 36, 1959e1964. Struis, R.P.W.J., Ludwig, C., Lutz, H., Scheidegger, A.M., 2004. Speciation of zinc in municipal solid waste incineration fly ash after heat treatment: an X-ray absorption spectroscopy study. Environ. Sci. Technol. 38, 3760e3767. Struis, R.P.W.J., Nachtegaal, M., Mattenberger, H., Ludwig, C., 2009. The fate of lead

T. Fujimori et al. / Environmental Pollution 209 (2016) 155e163 in MSWI-fly ash during heat treatment: an X-ray absorption spectroscopy study. Adv. Eng. Mater. 11, 507e512. Takahashi, S., Sakai, S., Watanabe, I., 2006. An intercalibration study on organobromine compounds: results on polybrominated diphenylethers and related dioxin-like compounds. Chemosphere 64, 234e244. Takaoka, M., Shiono, A., Nishimura, K., Yamamoto, T., Uruga, T., Takeda, N., et al., 2005a. Dynamic change of copper in fly ash during de novo synthesis of dioxins. Environ. Sci. Technol. 39, 5878e5884. Takaoka, M., Yamamoto, T., Shiono, A., Takeda, N., Oshita, K., Matsumoto, T., Tanaka, T., 2005b. The effect of copper speciation on the formation of chlorinated aromatics on real municipal solid waste incinerator fly ash. Chemosphere 59, 1497e1505. Takaoka, M., Yamamoto, T., Tanaka, T., Takeda, N., Oshita, K., Uruga, T., 2005c. Direct speciation of lead, zinc and antimony in fly ash from waste treatment facilities by XAFS spectroscopy. Phys. Scr. T115, 943e945. Theisen, J., Funcke, W., Balfanz, E., Konig, J., 1989. Determination of PCDFs and PCDDs in fire accidents and laboratory combustion tests involving PVCcontaining materials. Chemosphere 19, 423e428. Tian, S., Yu, M., Wang, W., Wang, Q., Wu, Z., 2009. Investigating the speciation of copper in secondary fly ash by X-ray absorption spectroscopy. Environ. Sci. Technol. 43, 9084e9088. Tue, N.M., Suzuki, G., Takahashi, S., Isobe, T., Trang, P.T.K., Viet, P.H., Tanabe, S., 2010. Evaluation of dioxin-like activities in settled house dust from vietnamese ewaste recycling sites: relevance of polychlorinated/brominated dibenzo- p -dioxin/furans and dioxin-like PCBs. Environ. Sci. Technol. 44, 9195e9200. Tue, N.M., Takahashi, S., Subramanian, A., Sakai, S., Tanabe, S., 2013. Environmental contamination and human exposure to dioxin-related compounds in e-waste recycling sites of developing countries. Environ. Sci. Process. Impacts 15, 1326e1331. van den Berg, M., Denison, M.S., Birnbaum, L.S., DeVito, M.J., Fiedler, H., Falandysz, J.,

163

et al., 2013. Polybrominated dibenzo-p-dioxins, dibenzofurans, and biphenyls: inclusion in the toxicity equivalency factor concept for dioxin-like compounds. Toxicol. Sci. 133, 197e208. Van Wijnen, J.H., Liem, A.K.D., Olie, K., Van Zorge, J.A., 1992. Soil contamination with PCDDs and PCDFs of small (illegal) scrap wire and scrap car incineration sites. Chemosphere 24, 127e134. Wang, D., Xu, X., Zheng, M., Chiu, C.H., 2002. Effect of copper chloride on the emissions of PCDD/Fs and PAHs from PVC combustion. Chemosphere 48, 857e863. Ward, J., 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58, 236e244. Watanabe, I., Sakai, S., 2003. Environmental release and behavior of brominated flame retardants. Environ. Int. 29, 665e682. Weber, R., Kuch, B., 2003. Relevance of BFRs and thermal conditions on the formation pathways of brominated and brominated-chlorinated dibenzodioxins and dibenzofurans. Environ. Int. 29, 699e710. Wedepohl, K.H., 1995. The composition of the continental crust. Geochim. Cosmochim. Acta 59, 1217e1232. Wei, Y.L., Yang, Y.W., Cheng, N., 2001. Study of thermally immobilized Cu in analogue minerals of contaminated soils. Environ. Sci. Technol. 35, 416e421. Wong, M.H., Wu, S.C., Deng, W.J., Yu, X.Z., Luo, Q., Leung, A.O.W., Wong, C.S.C., Luksemburg, W.J., Wong, A.S., 2007. Export of toxic chemicals e a review of the case of uncontrolled electronic-waste recycling. Environ. Pollut. 149, 131e140. Zelinski, V., Lorenz, W., Bahadir, M., 1993. Brominated flame retardants and resulting PBDD/F in accidental fire residues from private residences. Chemosphere 27, 1519e1528. Zennegg, M., Yu, X., Wong, M.H., Weber, R., 2009. Fingerprints of chlorinated, brominated and mixed halogenated dioxins at two e-waste recycling sites in Guiyu/China. Organohalogen Compd 71, 2248e2252.