Occurrence And Occupational Risk Of Polybrominated Diphenyl Ethers (PBDEs) And Dechloran Plus (DP) In A Formal E-Waste Recycling Plant, Northwest China

To prevent widespread diffusion of toxic chemicals from e-waste recycling industry in southern and eastern China, a scaled e-waste recycling industry is being relocated to northwestern China. The present study examined the levels of several chemicals in a typical e-waste recycling plant in north western China. In the first phase of our field sampling campaign, we collected total 100 PM1.0, PM2.5, PM10 and gas phase samples about PBDEs and DP at three sampling sites. The ambient mean concentrations of ∑9PBDEs and ∑DP in particle and gas phase during the whole sampling period followed a sequence of indoor>outdoor>background. The highest level of ∑9PBDEs and ∑DP in particle phase were found in PM2.5 at the indoor site were 1978.58 pg/m3 and 155.50 pg/m3, respectively. The highest concentration of ∑9PBDEs in gas phase also appeared at the indoor site at 7.33 pg/m3, followed by the outdoor site (4.10 pg/ m3), and the background site (0.70 pg/m3). DP concentrations in gas phase were 0.63 pg/m3 at the indoor site and 0.10 pg/m3 at outdoor site, respectively. BDE-209 was the dominant congener in all particles at the indoor site. PBDEs and DP were mainly adsorbed to the particulate phase, especially in PM2.5. The inhalation exposure risk assessment combined with the particle size distribution suggested that PM2.5-bounded PBDEs and DP exhibited the highest inhalation risk and deposition flux in the alveolar region and had the largest relative contribution to health risks.


Introduction
China has the world largest electronic waste (e-waste) recycling industry [1]. About 80% of e-wastes were exported to Asia, and nearly 90% of which was shipped to China legally or illegally [2]. China itself has been one of the largest consumers of appliances in the world, yielding huge amount of e-wastes. As a result, the e-waste recycling industry gained rapid development in China, such as Guyui in Guangdong Province in southern China and Taizhouin Zhejiang Province in eastern China, which were known as the e-waste recycling capitals in China and the globe [3,4]. The considerable amount of e-waste recycling and the lack of modern techniques in e-waste recycling in those e-waste recycling plants in southern and eastern China caused serious toxic pollution in the environment. On the one hand, given shallow level of groundwater, runoff due to high precipitation rate, and abundant surface water [5], and toxic chemicals in those e-wastes recycling sites are readily diffused and poured into the water environment. On the other hand, many such chemicals also undergo long-range atmospheric transport due to their persistence and volatility [6]. Extensive investigations to e-waste recycling industry in southern and eastern China revealed serious adverse effects of toxic chemicals on human health [6][7][8][9].
To prevent and remediate the environmental pollution and reduce human exposure risks from e-waste recycling in southern and eastern China, one of the important strategies taken by Chinese Environmental Administration was to relocate e-waste recycling industry to northwestern China. Featured by arid and semi-arid climate, lower precipitation, and higher elevation as compared to eastern and southern China, both groundwater and surface water resources are scarce in this part of China, which reduce the risk of environmental contamination from e-waste recycling activities. However, given the poor ecological environments, northwestern China are uniquely susceptible to atmospheric pollution [10]. Low precipitation (~100-300 mm/year), barren lands, and sparse vegetation coverage reduce considerably the elimination of contaminants from the air through precipitation washout and vegetation uptake. Concerns have been raised for the impacts of the future development of e-waste recycling industries on water resources by toxic chemicals which might worsen originally very limited water resources (only about 22% of the national mean), such as the Yellow River, the second largest river in China and the main source of water supply for the local and vast downstream regions in central and east China. Hence, there is an urgent need to assess the potential While extensive field sampling studies have been carried out in many e-waste recycling sites in southern and eastern China, there are almost no field measurements ever conducted to examine the ambient pollution levels of PBDEs and DP as well as other persistent toxic chemicals releasing from e-waste dismantling industries in northwestern China, the future home of e-waste recycling industry in China. In this sense, the present study aimed to fill knowledge gaps in the influences of e-waste recycling activities on the dry and harsh environment and to provide a scientific support to the development of environmental friendly e-waste recycling industry in northwestern China, which is presently unclear. The major objectives of the present study are: (1) to measure the ambient air levels of PBDEs and DP as typical chemicals emitted from a formal e-waste recycling plant in the arid environment in northwestern China; (2) to examine compositional profiles, gas-particle partitioning, and dry deposition of the target chemicals; and (3) to assess potential human exposure risks to PBDEs and DP released from this formal e-waste recycling plant.

Study location
The sampling study was conducted in a formal e-waste dismantling plant located in Gaolancountry (36°20′N, 103°57′E), Lanzhou City (Figure 1), Gansu Province ( Figure 1) in northwestern China. The area of Gaolan is 2476 km 2 with the population about 180,000. This plant was constructed under the guidance of the Chinese Ministry of Environment Protection (MEP) as a typical e-waste recycling plant for the future development of this type of industry in northwestern China, which was constructed to follow the regulations and the environmental protection law of China. There were 140 dismantling workers were employed in the plant. Three sampling sites were set to measure the pollution levels and compositional profiles of PBDEs and DP. Among these sites, the one site was set within indoor site (the E-Waste Dismantling Workshop, EWDW), and another one was at outdoor site (the Top Of Office Building, TOB), and the third one was set at a Rural Area (RA), 5 km up wind of thee-waste recycling plant ( Figure 1). A detailed description of sampling sites is provided in the Supporting Information (SI).

Sample collection
Gas phase and particle phase including PM 1.0 , PM 2.5 , and PM 10 were collected at the indoor (EWDW) and outdoor (TOB) sites from 17 to 26 November 2016. 30 samples (PM 1.0 , PM 2.5 , PM 10 ) and 10 samples about gas phase were collected in EWDW and TOB, respectively. At the RA site, only gas phase (10 samples) and PM 2.5 (10 samples) were sampled. Two high volume air samplers (Grase by Anderson) were used to collect PM 2.5 , PM 1.0 , respectively, and another brand high-volume air sampler (Tianhong Intelligent Instrument Plant, Wuhan, China) was used to sample PM 10 . PBDEs and DP in particle phase were extracted from these PM samples. Polyurethane Foam (PUF) cartridges with 6.5 cm in diameter × 7.5 cm in thickness (a density of 0.030 g/m 3 ) were also used simultaneously to collect gas phase PBDEs and DP at the three sampling sites. The sampling flow rates at different sampling points (EWDW, TOB, RA) are 200L/ min, 300L/min, 250L/min respectively. Particulate-associated contaminants were isolated from the atmosphere by drawing air through a quartz fiber filter (QFF, 20.3 cm × 25.4 cm, Whatman) for approximately 23 h of each day for 10 days.
Before sampling, PUF was soxhlet extracted with acetone and n-hexane for 72 h and dried in vacuum desiccators, and QFF was baked at 450°C and stored in sealed desiccators. After sampling, loaded QFF was wrapped with pre-baked Al foils and sealed with double-layer polyethylene bags. PUF cartridges were stored in solvent-cleaned glass jars with Al foil-lined lids. The samples were transported to the laboratory and stored at -20°C. PUF and QFF were spiked with 13 C-PCB 141 as surrogate standards, and then were extracted using the Soxhlet-extractor with 200 ml mixture of acetone/hexane (1:1) for 72 h. The extract was rotary-evaporated to approximately 2 ml by a rotary evaporator, and then solvent-exchanged to n-hexane. Concentrated extracts were cleaned and fractionated on acid/basic/neutral multilayer silica gel/alumina columns, then eluted with 70 ml solvent mixture (hexane: Dichloromethane of 1:1). The final extracts were solvent-exchanged into hexane and concentrated to 200μl under a gentle N 2 stream.
Sample analysis was performed with a Shimadzu Model 2010 Gas Chromatograph (GC) coupled with a Model QP 2010 Mass Spectrometer (MS) (Shimadzu, Japan) using Negative Chemical Ionization (NCI) in the Selected Ion Monitoring (SIM) mode. A 1μl extract of sample was injected automatically in split less mode. The initial oven temperature was maintained at 110°C for 5 min, programmed at a rate of 20℃/min to 200°C, held for another 4.5 min, then 10°C/ min to 310°C, and then held for another 15 min. The determinations of BDE-28, -35, -47, -99, -100, -153, -154, -183 except for BDE-209 were performed by GC/MS-QP2010 equipped with a DB-XLB column (30 mx0.25 mm, 0.25μm film thickness). For BDE-209, a CP-Sil 13 CB (12.5mx0.25mm, 0.2µm film thickness) capillary column was used. The rest details of experimental steps were described in previous reports [6,28].

QA/QC
The Quality Assurance/Quality Control (QA/QC) measures of analysis process were performed including the simultaneous operation of QC control samples. One field QFF blank sample, one laboratory blank sample, one standard spiked blank sample, and one standard spiked matrix sample were performed for each batch of twelve samples for quality control. All samples were spiked with 13 C-PCB-141 to assess the extraction efficiency and clean-up procedures. The recovery rates of the target compounds in the blank spike were less than 5%. To verify the recovery of target contaminants in each sample, recovery indicators such as PCB198, PCB30, PCB209 and TCMX were added. The recoveries of indicators in all samples ranged from 65% to 110%.

Gas-particle partitioning model
PBDEs are Semi Volatile Organic Compounds (SVOCs), and could be partitioned between the vapor and particle phases in the air. We examined the partition behaviors of measured PBDEs between the vapor and particle phases based on the sampled concentrations of PBDEs in gas phase and PM 10 . To better understand the environmental behavior of PBDEs emitted from an e-waste recycling plant in northwestern China under unique geographical environment and climate, and to provide scientific support to development of the e-waste dismantling plants, we calculated the gas-particle partitioning of target chemicals using Junge-Pankow partition/adsorption model described in SI.

Inhalation exposure risk assessment model
Generally, inhalation is one of the primary path ways of human exposure to PBDEs [29]. Flame retardants associated with airborne particles would pose a potential health risk to the exposed workers [30]. In this study, occupational workers exposure to PBDEs via inhalation was assessed based on monitored PBDEs concentrations and the exposure/ingestion factors recommended by the U.S.EPA. To evaluate the differences of human health risks of target pollutants with different particle sizes, International Commission on Radiological Protection (ICRP) model was used for inhalation exposure risk assessment in this study [4]. ICRP model was used to calculate the deposition flux of target pollutants in various organs of the respiratory system. The respiratory deposition flux was used as the daily intake and was substituted into the risk assessment model to calculate the hazard quotients. The ICRP model and the respiratory exposure risk assessment model are described in SI and the assessment results are elaborated below.

Result and Discussion
PBDEs Levels of PBDEs concentration: 9 PBDE congeners were detected in gas and particle phase, as shown in Table S1 of SI and figure 2. Table  1 presented PBDEs concentration levels at different e-waste dismantling sites over the world, collected from literatures. As seen, PBDE concentrations differed greatly in different studies due to the difference in dismantling processes and complex emission mechanisms of PBDEs. In contrast to ambient PBDEs levels sampled in these field studies, especially in those places in China, the PBDE air concentrations collected in this study was lower, likely due to the adoption of more strict pollution control measures and new techniques in this regular-waste recycling following the guidance of Chinese MEP.
At EWDW site, the concentrations of ∑ 9 PBDEs were 137.66±32 pg/m 3 in PM 10 , 1978.58±140 pg/m 3 in PM 2.5 , and 506.24±115 pg/m 3 in PM 1.0 , respectively. The ∑ 9 PBDEs concentrations were 39.98±10 pg/ m 3 for PM 10 , 383.5±54 pg/m 3 for PM 2.5 and 268.88±97 ng/m 3 for PM 1.0 at the outdoor of TOB site. At the RA site, ∑ 9 PBDEs concentrations were 24.53pg/m 3 in PM 2.5 . The results indicated that PBDEs were more easily absorbed to fine particles. At the TOB site, the ambient BDE-47, BDE-99, and BDE-153 levels in PM 1.0 were higher than that in PM 2.5 .
Comparing the sampling results at EWDW and TOB (Figure 3  were higher than that at the EWDW, especially in PM 1.0 . However, these five congeners are the part of the industrial pentabromodiphenyl ether as shown in figure 2. Where C is air concentration (gas+particle) (ng/m 3 ), e is the emission rate (g/day), ∆y and ∆z are a horizontal fetch (distance away from a sampling site) and the height of sampler above a ground surface, and u is a mean wind speed (m/s) at the sampling region. Outdoor TOB site was selected to estimate emission from indoor EWDW. According to the data from Kestrel 4500 Portable climate meter, the average wind speeds were range from 0.6-0.76 m/s. In rural conditions, ratios of height to fetch vary from as small as 1:10 in unstable conditions to as large as 1:500 in stable cases [36]. Under low wind conditions and considering the neutral atmospheric stability, we selected a ratio of about 1:100 to calculate urban cases [37]. Since the sampler was mounted at the roof of a five-story office building, the sampling height was chosen as 20 m. This yields the horizontal fetch (radius of the point source) of 200 to 10000 m in the neutral atmospheric boundary-layer. Available data from open published article were collected to compare PBDE emission from e-waste dismantling operations between our studies and other field studies. Results are presented in table S6.
The EWDW site was regarded as a key PBDEs emission source to the TOB site. The estimated PBDEs daily emissions at the EWDW range from0.081 to 5.15 g/d, which were much lower than those emission sources associated with e-waste burning operations in southern China [3,38,39]. Nevertheless, lower emissions were also observed in Taizhou    Detailed results are presented in Table S6. Comparing with the PBDE emissions caused by the electronic waste dismantling in southeastern China, low emission occurring in the formal e-waste dismantling plant in Gaolan is very likely attributed to the control of e-waste burning.
Congener profiles of PBDEs: Figure 2 shows that the composition of PBDE congeners differs greatly in the proportions of the different particles and the gas phase at the three sampling sites. In the gas phase, the proportion of lower-brominated congeners was greater than that of higher-brominated congeners in both indoor and outdoor environment. As vapor pressures of PBDEs increase with decreasing molecular weight and degree of bromination [41], lower-brominated PBDE congeners with great vapor pressure will easily be transferred to the gas phase. However, particle phase BDE-209 accounted for 50%-72% of total PBDEs at the EWDW site and therefore was the most abundant congener at this site. This is consistent with the fact that the Deca-BDE mixture is one of the most frequently used flame retardants across the world and China, especially in electronic/electric products [3,42,43]. Nevertheless, no other components of commercial Deca-BDE, such as BDE-206, 207, and 208, were detected in the present study. At the TOB site, BDE-47 and BDE-99 were the dominant congeners in PM 1.0 , and BDE-183 was the most abundant congener in PM 2.5 . The composition and proportion of PBDEs in PM 1.0 at the TOB site were similar to that of commercial Penta-BDE. As aforementioned, the printed wiring board was produced decades ago, and commercial Penta-BDE products prevailed during that period. Complicated emission mechanism of PBDEs as well as dismantling processes might release other components of these two PBDE mixtures of commercial flame retardants, which were not discerned in this field study.
Gas-particle partitioning of PBDEs: The relative abundances of PBDEs at the sampling sites are shown in figure S1. The particle phase PBDE congeners were dominant at the EWDW and TOB sites, accounting for over 90% and 83% of ∑ 9 PBDEs concentration, respectively. As SVOCs, the distribution of atmospheric PBDEs partitioning between gas and particle phase is an important factor controlling their environmental fate, such as migration, degradation, sedimentation, and human ingestion. Figure 4 shows the log-log plots of K P versus P L for PBDEs measured in the present study. The mean ambient temperature averaged over each sampling period ranged from -13°C to 9°C. log P L of PBDEs were calculated using log P L =m/T + b, where m and b are related to physical and chemical properties of PBDEs which were obtained from literature [44]. Linear correlations obtained from r 2 values between log Kp and log P L of PBDEs ranged from 0.626 to 0.797. All m values were greater than -0.6, ranging from -0.352 to -0.280, indicating that PBDEs congeners were dominated by the adsorption of organic matter onto particulate matter. Our results seem to differ from Chen et al. [11], likely due to significant change in the air temperature which is a dominate factor to influence on P L . In their case, the air temperature ranged from 27°C to 32°C whereas in our case, the air temperature was below 0°C. Lower temperature favors deposition of gaseous BFRs top articles [3]. There are uncertainties in the estimation of K p and P L , partly from the replacement of TSP by PM 10 in K p calculation. Other ambient conditions such as precipitation and temperature, chemical reaction, evaporation of PBDEs from particle, and the non-equilibrium of gas/particle partitioning could all cause uncertainties.
As shown in figure 5, the J-P model tended to underestimate the sorption of high-bromine PBDEs. While ɸ values of lower-brominated PBDEs were overestimated by the J-P model at the EWDW and TOB site, agreeing with other studies [6,11,45]. The low-brominated PBDEs at the two sampling sites were mainly present in the particle phase compared with the gas phase. The main reason is likely that the effect of temperature on lower-brominated congeners was greater than higher-brominated congeners. Given that the mean temperature during the sampling period was below 0°C, the volatilization of lower-brominated congeners was limited to some extent. DP Concentration levels of DP: DP is globally ubiquitous as a potential candidate of POPs. As can be seen from table S2, two isomers of DP were detected in both gas phase and particle phase at the three sampling sites. The concentrations of ∑DP (sum of syn-and anti-DP) at the EWDW site were 89.28pg/m 3 in PM 10 , 155.5 pg/m 3 in PM 2.5 , and 67.92 pg/m 3 inPM 1.0 , which were mostly in particle phase with only 0.63 pg/m 3 in gas phase. The result confirmed that the recycling of e-waste was an important source of DP. At the TOB site, concentrations of ∑DP were 6.74 pg/m 3 in PM 10 , 5.9 pg/m 3 in PM 2.5 , and 13.8 pg/m 3 in PM 1.0 , respectively. It was much lower than that at the EWDW site, as summarized in table S2. The concentrations of ∑DP at the RA site were 1.53pg/m 3 in PM 2.5 and 0.14pg/m 3 in gas phase. These values were also lower than that measured at a background site of rural area [46]. At a background site lower DP was entirely due to source proximity [18,47].
Isomer profiles of DP: Generally, the fraction of anti-DP to the total DP isomers ( Figure 6) has always been used to predict the environmental transport and fate of the two structural isomers [48].  The difference of solubility characteristics of DP isomer suggests that the environmental behavior of these two isomers is also different [49]. Two stereoisomers (syn-DP and anti-DP) of DP are mixed in a ratio of 1:3 in commercial products, namely anti-DP accounts for about 75% of ∑DP [50]. We calculated the ratio of anti-DP to summed anti-and syn-DP, defined by fanti=anti-DP/(anti-DP+syn-DP

Deposition flux of inhaled PBDEs and DP in the respiratory tract
The deposition fluxes of PBDEs and DP with different particle sizes in various organs of the respiratory system were presented in tables S3-S5. The mean deposition fluxes of Σ 9 PBDEs and DP with different particle size in different organs of human respiratory system were in sequence of Head Regions (HR)> Alveolar Regions (AR)> Tracheobronchial Regions (TR), is in line with the results obtained by Zhang et al, [40]. The total deposition fluxes of target pollutants with different particle sizes follow a sequence of PM 2.5 >PM 1.0 >PM 10 , indicating that fine particles tend to pose higher health risks than coarse particles due to greater capability of absorbing toxic chemicals of fine particles. Figure 7 illustrates relative contributions of individual PBDE congener with different particle size in the alveolar regions of human respiratory system at the EWDW and TOB sites. Contaminants entering the alveolar area can cause direct harm to the human body. At the EWDW site, deca-BDE congener (BDE-209) exhibited the largest deposition flux, showing 611.41 pg/h in PM 2.5 , 64.83 pg/h in PM 1.0 , and 28.49 pg/h in PM 10 . In addition, the deposition flux of tribromo congeners accounted for 17.5%-23.1% of the total flux. At the TOB site, the lower-brominated congeners (Tribromo to pentabromodiphenyl ether) were associated with smaller particle size prone to deposit in the alveolar region. Decabromodiphenyl ether was shown to be prone to lower-brominated congeners in the environment by de-bromination, and lower-brominated congeners were more toxic and persistent [6,54,55]. Hence, the toxicity risk caused by lower-brominated congeners to occupational workers should not be overlooked.
Inhalation exposure risk to PBDEs and DP subject to different particle size Daily inhalation intake of PBDEs and DP: Figure 8 displays the respiratory exposure and Hazard Quotient (HQ) values of the target pollutants combined with the particle size distribution. At the EWDW site, the total daily inhalation intake of ∑ 9 PBDEs and∑DP were 9.35 pg/kg/d and 6.50 pg/kg/d for PM 10 ; 130.27 pg/kg/d and 10.24 pg/kg/d for PM 2.5 ; and 15.32 pg/kg/d and 2.04 pg/kg/d for PM 1.0 , respectively. At the TOB site, the total daily inhalation intake of ∑ 9 PBDEs and ∑DP were 2.35 pg/kg/d and 0.40 pg/kg/d for PM 10 , 22.72 pg/kg/d and 29.13 pg/kg/d for PM 2.5 , and 4.03 pg/kg/d and 0.9 pg/kg/d for PM 1.0 , respectively. Respiratory exposure is not only determined by the concentration of the contaminant, but also by the particle size distribution. As a result, PBDEs and DP in PM 2.5 exhibited the biggest exposure dose at both sampling sites.
Overall, the daily inhalation intake at the EWDW site was greater than that at the TOB site. The inhalation exposure dose of different congeners of PBDEs and DP were associated significantly with particle size distribution. At the EWDW, the main contribution of total exposure does in PM 10 to the inhalation intake was DP, followed by BDE-209. In PM 2.5 and PM 1.0 , BDE-209 and BDE-35 were the dominant congeners in the inhalation exposure dose. At the TOB site, BDE-209 and BDE-28 in PM 10 were the main contributors to the inhalation exposure dose whereas DP and BDE-183 in PM 2.5 were the dominant contributors, and BDE-47 and BDE-99 in PM 1.0 were the main contributors to the inhalation exposure dose.
Hazard quotients and exposure risk of PBDEs and DP: HQ is the ratio of the potential exposure to a substance and the level at which no adverse effects are expected. If the HQ value is greater than 1, then adverse health effects are possible. In our study, the HQ of target pollutants in PM 2.5 was the largest at the EWDW and TOB sites with the values for ∑ 9 PBDEs and ∑DP at the EWDW were 3.68x10 -4 and 1.00x10 -6 , respectively. The HQ of the ∑ 9 PBDEs and ∑DP at the TOB was 1.80x10 -4 and 2.28×10 -6 , respectively. The HQ values related to  the inhalation exposure of the target pollutants at the EWDW and TOB were lower than the critical value (1.0) specified by the US EPA, indicating that occupational workers had a relatively low risk of respiratory exposure to PBDEs and DP. Generally, the total hazard quotients of PBDE in respiratory system exposure at the EWDW were greater than that at the TOB. But HQ of DP at the EWDW was lower than that at the TOB. The main contribution of the target contaminants to the HQ at the EWDW and TOB was lower-brominated congeners. BDE-35 in PM 2.5 yielded the highest HQ at the EWDW, followed by BDE-47 and BDE-28 in PM 2.5 , respectively. BDE-99, BDE-183, and BDE-47 in PM 2.5 led to the highest HQ at the TOB. Although the assessment results showed a relatively low risk of respiratory exposure in our case, the amount of contaminants deposited and accumulated in the lungs increased which enhanced exposure risk of the occupational workers to the target contaminants. The metabolism and transformation mechanisms of individual congener of PBDE in humans were unknown yet.

Conclusion
The results presented in the present study revealed that PBDEs and DP containing in the obsolete electronic products were emitted into both indoor and outdoor ambient air environment during e-waste dismantling process. Comparing with other field measurement studies conducted in southern and southeastern China, the levels of PBDEs and DP collected in our field sampling campaign were low in ambient air, which is likely due to strict control measures taken in this formal e-waste recycling plant in our case. Such control measures are essential for protecting the rigorous environment in northwestern China from the relocation of e-waste recycling industry from well-developed southern and eastern China. The gas-particle partitioning calculation indicated that PBDE congeners and DP were mainly adsorbed to particle phase, especially to fine particles. BDE-209 was the predominant PBDE congener at the indoor of EWDW site, and BDE-47 and BDE-99 were dominant congeners at the outdoor of TOB site. The inhalation exposure risk assessment combined with the deposition fluxes of target pollutants with different particle sizes onto various organs of the respiratory system suggested that PM 2.5 -bounded PBDEs and DP exhibited the highest inhalation risk and deposition flux in the alveolar region which would has the largest relative contribution to health risks. It is worthwhile to point that the present study only focused on the risk assessment of respiratory exposure. Further research is needed to examine the influences of other exposure pathways such as skin exposure and dietary exposure on the health risks of occupational workers in this formal e-waste recycling industry in northwestern China.  Sampling site detail kEWDW: kEWDW is a dismantling workshop mainly based on manual dismantling. CRT TVs, discarded desktop computers, washing machine enclosures, etc. were dismantled here. Computer host was placed on the workbench, and manually opened the side cover, and the data cable, power supply, hard drive, fan, and motherboard were removed. LCD TVs, computers, and laptops were manually opened at the bottom of the machine, and parts such as housings, circuit boards, and LCD screens were manually dismantled. The washing machine was transported from the material storage area to the dismantling area by a cart, and the washing machine was manually placed on the dismantling operation table, the bottom cover of the machine was opened, and the relevant parts were manually taken.
TOB: TOB is at the roof of a five-story office building located 20 meters downwind from the dismantling plant. The office building mainly involves civilian activities. We set it as a sampling point for collecting PBDE gas and particle phase concentrations in outdoor environments.
RA: RA is located in a rural village 5 km upwind away from the e-waste dismantling plant. Several households are located, and the rest of the areas are farmlands. RA site is considered as the background site.

Emission and Simplified Gaussian Model
Outdoor TOB site was selected to estimate emission from indoor EWDW. According to the data from Kestrel 4500 Portable climate meter, the average wind speed were range from 0.  Table S6. Target Table S1: Average concentrations of PBDEs in particle phase and gas phase at three sampling sites (pg/m 3 ).

Particle partitioning and adsorption
The Junge-Pankow partition/adsorption model (J-P model) is the most commonly used method to estimate the atmospheric gas-particle partitioning of SVOCs associated with aerosols [7], defined by where F and A are the particulate and gas-phase concentrations, respectively, and TSP is the total concentration of suspended particulate (μg/m 3 ). Since the sample of atmospheric particulate collected in this study was PM 10 , the TSP in equation (1) is replaced by PM 10 . So the formula (1) has the following deformation: Lee and Tsay (1992) [8] replaced TSP with PM 10 for gas particle distribution studies. Similar approach has been also used in other studies [9,10]. It should be noted that Kp calculated from TSP and PM 10 are different. We shall follow this approach because PM 10 concentrations were directly measured in the present study. It should be noted that the replacing TSP by PM 10 might yield bias in the computation of Kp [9].
Pankow [11] has derived a log-linear relationship between K p and P L , defined by: where P L is the analytic's sub cooled liquid vapor pressure, m is the slope, b is the intercept. In general, the slope m should be close to -1 [11]. In reality, even in an equilibrium state m r value could be deviated from -1 [12,13]. Under certain conditions, the value of m can be regarded as an indicator for adsorption or absorption when determining the gas-particle partitioning of SOCs [12]. The temperature dependent P L values of PBDE congeners were calculated using the regression parameters by [14].
The Junge-Pankow adsorption (P L -based) model has been conventionally used to predict the fraction of SOCs in particles, which is related to the sub cooled liquid vapor pressure (P L ) of the compound and the particle surface area per unit volume of air (θ), given by: The constant c (Pa cm) is related to the heat of condensation and chemical surface properties. An empirical value of 17.2 Pa cm is often used [15]. θ=1.1×10-5 cm 2 /cm 3 for urban air and 4.2-35×10 -7 cm 2 /cm 3 for rural air [7].
Only the PBDE congeners with detectable concentrations in a given sample were accounted for in our study. Based on the value of m and b [16] and the temperature during sampling period, we estimated P L value. P L value for BDE-209 was not available as BDE-209 was present only in the particle phase in most cases. The P L value of BDE-209 was predicted based on literature [17] in a QSPR model. Figure 6 Compares the percent (ɸ×100%) of PBDEs in the particle phase predicted by the Junge-Pankow adsorption model (using c=17.2 Pa cm and θ=1.1×10 -5 cm 2 /cm 3 ) with the average measured values. Each measured ɸ value was calculated by the amount of the chemical in the particle phase (F) divided by the total amount of the chemical (F+A) [18,19].

ICRP Model
To assess the deposition flux of inhaled PBDEs and DP in the human respiratory system, this study uses the simplified ICRP model provided by the International Commission on Radiological Protection (ICRP) [20,21]. Inhaled particulate matter is presumed mainly in the nasal cavity of the human respiratory system (Head Regions; HR; including nose, mouth, pharynx and larynx), Tracheobronchial Region (TR) and Alveolar Region (abbreviated) AR). The formula for calculating the Sedimentation Fraction (DF) of different pollutants in different organs of the human respiratory system is as follows: where, D j,i is the deposition flux of the particles in the i-th particle size range at the j-th part of respiratory system; D j,i is the deposition Figure S1: Composition of PBDEs and DP in gas and particle phase.