Introduction

The quantity of discarded electronic equipment (e-wastes) has been rapidly increasing around the world as a consequence of newer generations and constant upgrades of electronics (Robinson 2009). It was estimated that 50 million tons of e-wastes are produced annually (Ni et al. 2009). Massive quantities are transported to developing countries. Approximately 80 % of e-waste collected in the western USA was exported to Asia, 90 % of which were sent to China coastal areas through illegal imports for recycling (Schmidt 2006).

Bromine flame retardants (BFRs) are commonly used in electronic products to meet fire regulations. Polybrominated diphenyl ethers (PBDEs) were the most extensively used BFRs in recent decades. Commercial penta- and octa-BDEs have been found to pose high health risks to humans and wildlife (Hites 2004), which resulted in them being banned in some regions of the world (EU 2003, Renner 2004). However, commercial deca-brominated diphenyl ether (BDE) mixtures were also banned in the European Union and parts of the USA, although they are still widely used in many countries (Betts 2008). The restrictions and bans for the use of PBDEs have caused manufacturers to seek non-PBDEs BFRs. For instance, decabromodiphenyl ethane (DBDPE) is now used as an alternative for deca-BDE in various products. Because DBDPE has been recently detected in the environment and wildlife (Kierkegaard et al. 2004, Shi et al. 2009), it should receive more attention.

In addition to BFRs, metal is also an important hazardous substance that is released from e-waste recycling. Previous studies at e-waste recycling sites in Asia have revealed hazardous metal pollution in various environmental media including the air, soil, water, sediments and food (Chen et al. 2011, Li et al. 2011, Luo et al. 2011, Wong et al. 2007a, b, c, Zheng et al. 2013). Metals have great potential to accumulate in vital organs of the human body. Pb and Cd are associated with a number of adverse health effects such as cardiovascular, nervous system, blood and bone diseases (Jarup 2003). Excessive Cu may lead to toxic effects in humans (Brewer 2010).

Among the various environmental media, dust has been found to have high levels of BFRs from different regions of the world (Harrad et al. 2008, Wilford et al. 2005). Metal concentrations in dust were also at high levels at e-waste sites (Wong et al. 2007a, c). Dust is a significant medium and exposure pathway of toxic substances. Significant association was found between the PBDE concentration in breast milk and dust, indicating the prominent role of dust in exposure (Wu et al. 2007). The concentration of Pb in human blood was also strongly associated with the Pb concentration in dust (Lanphear et al. 1997). BDE-209 and Pb were found in the blood of e-waste workers and toddlers living in e-waste areas. The amounts were 3436 ng/g lipid weight and 100 μg/L for the workers and toddlers, respectively (Huo et al. 2007, Wen et al. 2008).

Taizhou has been involved in e-waste recycling for decades. In recent years, a large fraction of e-waste recycling shifted to remote areas. Simple household recycling workshops run by families are the most common form. High concentrations of BFRs and metals in this area have been reported in environment media and the human body (Jiang et al. 2014, Tang et al. 2010, Zhang et al. 2011). However, e-waste workshops have a steady accumulation of BFRs and metals due to the recycling activities, which cause adverse effect on human body, and could be the point source for the surrounding environment. Given their potential impact on human health, particularly on young people, understanding dust BFRs and metals as well as the associated exposure of people in workshops is urgent. Therefore, the present study aimed to investigate the characterization of heavy metals and BFRs in dust directly collected from the workshops, and the on-site human exposure of these two types of contamination was also assessed. Such a survey is also needed for local governments to make more strict regulations.

Materials and methods

Research area

A large composite sample of dust was procured in October 2013 from family-run workshops involved in abandoned recycling for decades from remote areas of Wenling City, southeast of Zhejiang Province, China. The location and the information of the sampling sites are shown in Fig. 1 and Table S1. Site E was chosen as the reference site and was located in a village far from other sites and without any e-waste recycling.

Fig. 1
figure 1

Map of the study area and sampling sites

Full size image

Sample collection and preparation

Multiple ground surface dust samples both inside and outside of the workshops and reference site (indoor and outdoor dust mentioned later) were collected using plastic brushes and dustpans with a gentle sweeping motion. A total of 47 composite samples were collected, 24 indoor and 23 outdoor. The mass of each sample ranged from 15 to 30 g. After each sampling, the brushes and dustpans were cleaned with paper towels. All samples were stored in paper bags, which were previously heated at 50 °C overnight to remove volatiles after collection and sealed in polyethylene bags (Ziploc) at <4 °C for transport to the laboratory. All of the samples were air-dried at room temperature to a constant weight and then sieved (<75 μm) prior to chemical analyses.

Chemical analysis

BFRs analysis

For the measurement of PBDE and DBDPE, each sample (0.5 g) was spiked with the surrogate standards 13C12-PCB-141 and13C12-BDE-209 and extracted with 100 ml of acetone/dichloromethane (1:1, v/v) for 8 h in a soxhlet apparatus (Buchi, Switzerland). The extracts were condensed to near dry and solvent-exchanged to 1 mL of hexane. The cleanup was conducted on a multi-layer silica gel column, and 40 mL of n-hexane/dichloromethane (1:1, v/v) was added to elute the target substances. The fractions were concentrated to 1 mL. A known amount of injection internal standard, 13C12-PCB-208, was added before instrumental analysis. Detailed information of the analysis conditions for the BFRs is shown in Supplementary Material.

Heavy metals analysis

Each sample (0.250 g) was digested according to the USEPA Method 3052, with modifications. An acid mixture of HNO3, HCl and HF (3:1:1, v/v/v), along with the samples, was added into polytetrafluoroethylene vessels for microwave digestion (Ethos A, Milestone, Italy). After digestion, the extracts were filtered into 25-mL volumetric flasks. The concentrations of Co, Cr, Mn, Cu, Ni, Pb, Sb, Zn and Cd were determined by inductively coupled plasma mass spectrometry (ICP-MS) (NexION300, PerkinElmer, USA).

Quality control and quality assurance

Each sample was duplicated for accuracy and reproducibility. The recoveries of 13C-PCB-141 and 13C-BDE-209 were in the range of 88–107 and 78–98 %, respectively, for dust samples. A procedure blank and a solvent blank were run for each batch of ten samples. No target substances were detected. The limit of detection (LOD) was defined as a signal/noise (S/N) ratio of 3, and it ranged from 0.1 to 4.5 ng/g for PBDEs and 4.5 ng/g for DBDPE.

To confirm the accuracy of the analytical methods for metals, a certified reference material GSS-5 (Institute of Geophysical and Element Prospecting, People’s Republic of China) was analyzed. The results were in good agreement (98–108 %) with certified values.

Enrichment factors for metals

The enrichment factor (EF) is defined by the following equation:

$$ \mathrm{E}\mathrm{F}\mathrm{s}=\frac{{\left({\mathrm{C}}_x/{\mathrm{C}}_{\mathrm{Mn}}\right)}_{\mathrm{sample}}}{{\left({\mathrm{C}}_x/{\mathrm{C}}_{\mathrm{Mn}}\right)}_{\mathrm{background}}} $$
(1)

where C x is the concentration of metals in dust and background, and CMn is the concentration of Mn. The background metal concentration in Zhejiang Province was obtained from CNEMC (1990). Mn was abundant in the crust and had a small concentration range in this study, so it was selected as the reference element in the EF calculations.

Enrichment factor: EF <2 minimal enrichment, EF =2–5 moderate enrichment, EF =5–20 significant enrichment; EF =20–40 very high enrichment; and EF >40 extremely high enrichment (Li et al. 2010).

Statistical analysis

All statistical analyses were performed using SPSS 16.0 for Windows (SPSS, Inc.). For data less than the LOD, a value of one half of the LOD was assigned. The probability distribution tests through the Kolmogorov-Smirnov test for all the data were normal. ANOVA tests were used for statistical comparisons between more than two groups. Pearson regressions were used to test the correlations between contaminants. PCA is an effective analytical tool to minimize the set of original variables and extract a small number of latent factors or principal components for analyzing the relationships between observed variables and samples. Statistical significance was set at p < 0.05.

Exposure and risk assessment

Risk assessment with regard to exposure to contaminated dust by ingestion was performed to estimate the non-cancer toxic (chronic) risk. The exposure doses were calculated according to the equation as follows (USEPA 1997):

$$ \mathrm{ADD}=\frac{\mathrm{C}\times \mathrm{IngR}\times \mathrm{E}\mathrm{D}\mathrm{F}}{\mathrm{BW}} $$
(2)

Non-cancer toxic risk was determined by calculating the hazard quotient (HQ),

$$ \mathrm{H}\mathrm{Q}=\frac{\mathrm{ADD}}{\mathrm{RfD}} $$
(3)

If HQ ≤1, noncancerous effect are unlikely, whereas for HQ >1, adverse health effects might be possible. HQ >10 is considered to be a high chronic risk. The detailed definition and parameters for these two equations are listed in the Supplementary Material.

Results and discussion

BFRs and metal concentrations in dust

The mean and range of the concentrations for BFRs and metals are shown in Table 1 and Table S2. Among the five sampling sites, the highest mean concentrations of PBDEs in both the indoor and outdoor dust were at site A at 46,600 and 9000 ng/g, respectively. However, with the exception of nona- and deca-BDE, samples from site D provided evidence of a significant emission source of low brominated PBDE to the environment. The sum of the eight PBDE congeners (BDE-28, 47, 100, 99, 154, 153, 183 and 203) was 14,400 and 6250 ng/g for the indoor and outdoor samples, respectively. The mean concentration in indoor samples from sites A to D was 27,400 ng/g, which is comparable to those detected in workshops from the Guiyu and Taizhou area; the values of which were 14,800 and 30,700 ng/g, respectively (Leung et al. 2011, Ma et al. 2009). The mean concentration in outdoor samples was 7060 ng/g, which is higher than those from e-waste areas from south China (Leung et al. 2011). DBDPE was also detected in 43 of the 47 dust samples. The highest concentrations were detected in the samples from site A, which were 9200 and 1130 ng/g for the indoor and outdoor samples, respectively. The mean value (4800 ng/g) of the in indoor samples from sites A to D was higher than those detected in e-waste storage facilities (1300 ng/g) and workshops (3030 ng/g) from south China (Muenhor et al. 2010, Zheng et al. 2011). A summary of the BFRs concentrations compared with other studies is shown in Table S3. In general, the BFR concentrations in our study were much higher than those reported in non-recycling areas (Ali et al. 2011, 2013, Fromme et al. 2014, Takigami et al. 2009), indicating that the high level of BFRs contamination resulted from e-waste recycling. For the reference site, BDE-209 was the major component, and the mean values were 226 and 103 ng/g for the indoor and outdoor samples, respectively. This was expected due to the still common use of commercial deca-BDE in China. The BFR concentrations in indoor samples were significantly higher than those in the corresponding outdoor samples (p < 0.001).

Table 1 Mean concentrations of BFRs (ng/g) and metals (mg/kg) in the dust from five sampling sites
Full size table

For metal concentrations, soil guidelines (Table S4) were applied to evaluate the extent of contamination due to the lack of guidelines in dust. Of the nine metals investigated in the indoor dust samples, the Cu and Pb levels were extremely elevated. The highest concentration of Pb was found in samples collected from site D, and the mean value was 17,230 mg/kg. It was an order of magnitude lower than those detected in workshops in southeastern China, but it was comparable for Cu concentration (Leung et al. 2008). The Cu and Pb concentrations exceeded the New Dutch List action value (VROM 2001) by 11–69 and 5–54 times, respectively. The Co and Cr concentrations were all below the action value. The Zn and Ni concentrations exceeded the action value, with the exception of two samples of Zn and all samples of Ni at site B. The mean Cd concentrations were below the action value. However, all Cd concentrations significantly exceeded the Chinese guidelines, even for grade III (industrial activity) (SEPA 1995), with a range from 3.3 to 20.5 mg/kg. The Sb concentrations were all above the action value, with the exception of two samples, and the highest concentration was 244 mg/kg.

For the outdoor dust, the metal concentrations were significantly lower than the corresponding indoor samples (p < 0.01). The Cu and Pb concentrations in the outdoor samples were above the New Dutch List action value, with the exception of two samples for Pb, and the highest concentrations were 1790 and 6960 mg/kg, respectively. They were higher than those from non-recycling areas (Han et al. 2006, Lu et al. 2010). The mean Sb concentration exceeded the action value by 1.8 times, and the concentration ranged from n.d. to 112 mg/kg. The Cd concentrations in the outdoor samples were also above the less stringent Chinese grade III value, with one sample not detected. The Co, Cr, Ni and Zn concentrations were within the action value or with little impact.

Mn was also analyzed in both the indoor and outdoor samples. Moreover, Cu, Pb, Ni, Zn and Cd were also detected in samples from the reference site. In contrast to the recycling area, the metal concentrations in the outdoor samples were higher, indicating the influence of other human activities. A summary of the metal concentrations compared with other studies is shown in Table S5.

Profiles, relationships and debromination for BFRs

In general, deca-BDE accounted for the major proportion of the total PBDE in samples from sites A, B and C, with 80.8, 82.5 and 78.6 % for indoor and 73.6, 69.9 and 67.3 % for outdoor dust, respectively. The profiles were similar to those observed in studies elsewhere (Cao et al. 2013, Wang et al. 2010). This was mostly attributable to the high use of commercial deca-BDE in older electronic products around the world. However, the composition of PBDE congeners detected in samples from site D could be commonly found in commercial penta-BDE products. The concentrations were significantly higher than other sites (p < 0.01). BDE-47 and BDE-99 attributed 30.7 and 35.7 % for indoor samples and 24.4 and 26.5 % for outdoor samples, respectively. The characteristics of PBDE homologues in dust were highly affected by the types of e-wastes and specific recycling activity. Heating or burning can lead to a high concentration of low brominated congeners in the workshops (An et al. 2011), which might cause the high percentage of penta-BDE in the dust from site D.

As shown in Fig. 2, similarities were found in the composition of PBDE congeners between the indoor and outdoor samples. However, the elevated total percentage of tri- through hepta-BDE in the outdoor dust from the former three sites suggested that less brominated PBDE congeners were more likely to escape the workshop than the higher ones due to their relatively high vapour pressures (Jones-Otazo et al. 2005) and depositing along with fine particles (Wei et al. 2009). On the contrary, a decline of this percentage was observed in the composition at site D. It was due to the long-range atmospheric transport (LRAT) of PBDEs (Melymuk et al. 2012), especially for low brominated congeners after heating and exhaustion. A previous study confirmed that tri- through hexa-BDE presented significant proportions in the atmosphere of rural sites away from the e-waste areas (Tian et al. 2011). Moreover, the mean concentrations of octa- through deca-BDE, except for BDE-206 in outdoor samples, were higher than their corresponding indoor concentrations at site D. This implied that the decline of this percentage was also driven by the high contamination of high brominated PBDEs from other emission sources at this sampling site.

Fig. 2
figure 2

Congener profiles of PBDEs in the indoor and outdoor dust from four sampling sites

Full size image

A significant and positive association between BDE-209 and DBDPE (r 2 = 0.907, p < 0.001) was found in the linear correlation examination of all dust samples (Fig. S1). This implied similar emission sources of BDE-209 and DBDPE and was consistent with the use of DBDPE as an alternative to deca-BDE. The mean ratio of DBDPE to BDE-209 (0.21) was higher than the calculated values obtained from e-waste storage facilities (0.04) but lower than those found in workshops from south China (0.32) (Muenhor et al. 2010, Zheng et al. 2011). The presence of DBDPE indicated that there were e-wastes containing “novel” BFRs in our study areas.

Figure 3a shows the ratios of BDE-208:BDE-183 and BDE-208:BDE-209. It was reported that the former ratio in the octa-BDE and the latter ratio in the deca-BDE formulation was under 0.45 and 0.08 %, respectively (La Guardia et al. 2006). In fact, these two ratios in our dust samples from e-waste workshops were all substantially exceeded and implied the occurrence of debromination of BDE-209 to BDE-208. Deca-BDE was confirmed to undergo photo degradation under light to form nona-BDE (Bezares-Cruz et al. 2004). This decomposition may occur during large-scale and long-term e-wastes recycling activities. However, it is also possible that debromination may have occurred during the manufacturing of PBDE-containing products (Chen et al. 2010). In addition to BDE-208, the other nona-BDEs (BDE-206 and BDE-207) in our study were also considerable. This suggested that there were different degradation products for deca-BDE to nona-BDE (Söderström et al. 2004).

Fig. 3
figure 3

Comparison of PBDE congener ratios in the indoor and outdoor dust with those in commercial BDE products

Full size image

Moreover, penta-BDE (BDE-99 and BDE-100) was confirmed to further debrominate to tetra-BDE (BDE-47) under solar irradiation (Söderström et al. 2004). A comparison of the BDE-47 to (BDE-100 + BDE-99) ratios between the indoor and outdoor samples is exhibited in Fig. 3b. The standard deviation of this ratio in outdoor dust was 0.15 in comparison with 0.06 for indoor dust. In addition, these ratios in the outdoor samples deviated more from those in two commercial penta-BDE products. To some extent, these results suggested that the PBDE in outdoor dust went through more physiochemical processes, such as debromination and volatilization. However, further studies are required to fully understand the degradation of these brominated congeners in the indoor and outdoor dust.

EFs, contamination profiles comparison and statistical analysis

The EF values calculated by Eq. (1) are plotted in Fig. 4. The EFs for Co and Cr in both indoor and outdoor samples were all lower than 1, indicating a predominant source of crustal soils rather than other anthropogenic sources (Liu et al. 2003). Elements Ni (2.47) and Zn (2.63) in indoor dust, with mean EF values higher than 2, had moderate enrichment and were derived from e-waste recycling activities. However, these two metals in the outdoor samples were lower than 2, with the exception of two samples for Ni. The mean EFs of element Sb were 6.40 and 3.70 for indoor and outdoor samples, respectively. Elements Cu, Pb and Cd in the outdoor samples showed “significant enrichment” with mean EF values of 11.6, 19.6 and 16.2, respectively. However, the values for these three metals in the indoor samples were significantly higher (p < 0.01) than other metals. The EFs of Cd were within the grade “very high enrichment” with a mean value 23.0. Cu and Pb had “extremely high enrichment”, with mean EF value 56.5 and 64.8, respectively.

Fig. 4
figure 4

The enrichment factors of metals in the indoor and outdoor dust samples

Full size image

Figure S2 shows the comparison of the profiles for PBDEs in indoor dust samples and e-waste. For DBDPE, the comparison was not conducted due to the lack of general data in e-waste. The mean percentage from sites A to C (9.84, 3.25, 80.6 and 6.27 % for penta-, octa-, deca-BDE and other congeners, respectively) was similar to those detected in e-waste (5.5, 9.1, 80.8 and 5.4 %, respectively) (Li et al. 2013). For site D, the penta- BDE composition was similar to that in two commercial products, which indicated that the e-waste in the workshops was the source of PBDEs in the dust. However, the metal profile in the dust was different from that in typical e-waste (Fig. S3), indicating the complex mechanism of the metals in the dust emitted from e-waste. The percentages of Mn and Pb in the dust were higher than those in general e-waste. However, it was in line with our selecting Mn as the reference metal due to its high concentration in the crust. Pb was a highly enriched metal, and its calculated content in general e-waste was underestimated.

The differences and similarities of the BFRs profiles in the dust were examined by PCA, which implied a composition characteristic of their sources and subsequent environmental behaviour. Two clusters are exhibited in Fig. 5a. The higher molecular weight PBDE congeners and DBDPE had high loadings on PC2 and lower loadings on PC1 for in the indoor samples. Figure 5b shows that the PCA result for outdoor samples was similar to the indoor result. However, octa- and nona-BDE had higher loadings on PC1 than those in indoor PCA result, indicating the similar trend to the less brominated PBDEs emitted from the workshops.

Fig. 5
figure 5

Loading plots of the principal component analysis (PCA) for BFRs and metals in the (a), (c) indoor and (b), (d) outdoor dust samples

Full size image

The metals in the dust were also examined by PCA. The first two principal components explained 66.0 and 22.8 % of the total variability of the metals in indoor dust, respectively (Fig. 5c). Through PCA, the metals were classified into two clusters. Group Cu, Pb and Sb had high loadings on PC2, and group Co, Cr, Mn, Ni, Zn and Cd had high loadings on PC1, indicating a common source and behaviour of the metals within groups from e-waste recycling activities. Figure 5d shows the first two principal components of the data from the outdoor dust samples, and the result of which was similar to the PCA for the indoor samples. In fact, Pb, Sb, Cu and Cd all had positive scores on PC2. A previous study confirmed that the principal component axis could be interpreted as the indicator of pollution or crust (Fujimori et al. 2012). Thus, combined with the EF calculations in Fig. 4, Pb, Sb, Cu and Cd were concluded to be the pollution metals emitted from e-waste recycling, whereas Co, Cr, Ni, Zn and Mn were more likely affected by the crust in outdoor samples.

The relation between BFRs and metals in the dust samples was examined by a single PCA (Fig. S4). Three clusters were found in the PCA for indoor dust. Pb and Sb were in the same cluster with the low molecular weight PBDEs. Whereas for the PCA for outdoor dust, two clusters, “metals” and “BFRs,” indicated no relation between these two contaminants.

Moreover, Pearson’s correlation coefficients are presented in Table 2 for the BFRs and metals for all dust samples. Significant positive correlations (in italics) were found for some pairs for metals and BFRs respectively. However, strong correlations between metals and BFRs were only found in Pb and Sb with penta-BDE and Pb with octa-BDE. This was consistent with the result of single PCA for the indoor dust we discussed and might indicate similar characteristics of emission source for Pb, Sb and low brominated commercial BDE during e-waste recycling.

Table 2 Pearson correlation coefficient matrix between BFRs and metals
Full size table

On-site exposure and health assessment

To perform a preliminary evaluation of human exposure to dust ingestion, we assumed 100 % absorption of dust intake, which is in line with other studies (Jones-Otazo et al. 2005). To be more acceptable for on-site (recycling work place) exposure and health assessment, the exposure duration fraction (EDF) in our study was assigned as the time people spent in workshops rather than the common calculated parameters used in most previous studies (Leung et al. 2008, Wang et al. 2010), which might result in an overestimation. Moreover, the mean concentration derived from the four sampling sites was used in the assessment.

The calculated hazard quotients of the BFRs and metals for adults and toddlers are shown in Table 3. The average exposures of penta-, octa-, deca-BDE and DBDPE via dust ingestion for both adults and toddlers were several orders of magnitude lower than their corresponding RfD values. The HQ range of the three PBDE congeners in this study suggested the likelihood of the low health risk of PBDEs exposure. For DBDPE, acute toxicity of such alternative flame retardants was low (Harju et al. 2008), whereas little is known about its chronic toxicity. Therefore, more toxicological research is required on account of the exposure dose via dust.

Table 3 Hazard quotient for BFRs and metals in the dust
Full size table

The HQ for metals were all lower than one, which indicated no adverse effects, except for Pb; the values of which were 2.27 and 9.52 for adults and toddlers, respectively. Notably, the value for Pb was close to 10 for toddlers, suggesting the high chronic risk of Pb. For Sb and Cu, the HQ was also considerable, with values 0.14 and 0.13 for adults and 0.59 and 0.55 for toddlers, respectively. However, there are adverse health effects for toddlers when using the 95th percentile concentrations, 1.00 and 1.97 for Cu and Sb, respectively.

Our results highlight the human exposure to BFRs and metals via dust during workshop and work time. However, the evaluated exposure was only an indication of the estimates. This was due to the small number of dust samples analyzed, the substantial inter-individual variation depending on the time spent in workshop and the quantity of dust ingested.

Conclusions

The uncontrolled e-waste recycling activities conducted in family-run workshops in recent decades have resulted in notable contamination of BFRs and heavy metals in house dust and the corresponding outdoor dust. We provided evidence that e-waste workshops could be regarded as the point source of BFRs and some metals because they could be emitted from workshops with the dust. We found that the composition of BFRs varied with type of e-wastes and recycling activities. BDE-209 had a strong correlation with DBDPE. The occurrence of PBDE debromination was also observed in all dust samples. The EF calculations showed a strong enrichment of heavy metals, especially for Cu, Pb and Cd. The source of elements Co, Cr, Ni and Zn in outdoor dust was affected by the crust. Significant positive correlations were found for Pb and Sb with PBDEs. The exposure through dust ingestion for BFRs was at low risk, whereas it was notable that toddlers were at high chronic risk of Pb. Further studies on BFRs and heavy metals in the air and particle phase are required to elucidate the human exposure in family-run e-waste workshops.