Atmospheric Concentrations and Dry Deposition of Polybrominated Diphenyl Ethers in Southern Taiwan

While polybrominated diphenyl ethers (PBDEs) have been used extensively for decades as flame retardants in a variety of consumer and commercial products, concerns about these substances have risen due to their adverse effects on human health. To the best of our knowledge, no study has reported on the dry deposition flux and velocity of individual PBDEs. This study was undertaken to investigate the dry deposition characteristics of PBDEs in the ambient air of southern Taiwan. The average atmospheric concentrations for total PBDEs (sum of thirty BDEs) ranged from 24.0 ± 1.83 to 102 ± 13.3 pg/Nm. The calculated dry deposition fluxes of total PBDEs were 13.4–60.4 ng/m-day, BDE 209 accounted for over 75% of the total PBDEs. The results showed that particle phase deposition dominated the dry deposition processes for PBDEs, and the same trends have been observed in other semi-volatile organic compounds (SOCs). The dry deposition velocities of individual PBDEs increased along with the number of brominated substitutes, ranging from 0.014 to 0.755 cm/s. Together with the results of the author’s previous work, the deposition flux of total PBDEs could reach three orders of magnitude higher than those of PCDD/Fs (157–544 pg/m-day) and PCBs (89–1010 pg/m-day). Since atmosphere deposition is believed to be the main transfer pathway for SOCs entering food chains, its impact on human exposure to PBDEs is of great importance.


INTRODUCTION
Polybrominated diphenyl ethers (PBDEs) are common additive brominated flame retardants (BFRs) that have been widely used in commercial and industrial products to prevent both the large quantities of PBDEs consumed and their widespread distribution, they have become ubiquitous environmental pollutants, found in air, water, soil, fly ash, sediment, and human tissues (Gill et al., 2004;Hites, 2004;Chao et al., 2007;Artha et al., 2011;Tu et al., 2011;Wang et al., 2011;Tu et al., 2012).Concerns about polybrominated dibenzo-p-dioxins and dibenzofurans (PBDD/Fs) have also increased, because PBDD/Fs can be formed during thermal processing of products containing BFRs (Lai et al., 2007).
Atmospheric deposition has been recognized as an important route for the long-range transportation of air pollutants entering the ecosystem.When semi-volatile organic compounds (SOCs), such as PCDD/Fs, PCBs, and PBDEs, are emitted into the air, they can be partitioned between the gas and particulate phases related to their vapor pressures, temperature, and concentrations of total suspended particulates in the atmosphere (Hoff et al., 1996;Lohmann and Jones, 1998;Tasdemir et al., 2004).Gas/particle partitioning is therefore an important factor in understanding the environmental fate of PBDEs and other SOCs.Recently, several studies describing gas-particle partitioning, as well as the atmospheric PCDD/Fs and PCBs deposition from relevant sources in Taiwan, have been published (Lee et al., 1996;Shih et al., 2006;Lee et al., 2009;Wang et al., 2010d;Huang et al., 2011;Mi et al., 2012), however, there is still very little information on the atmospheric deposition of PBDEs.
This study is an extension of our previous research, which focused on the dry deposition characteristics of PBDEs in the ambient air of industrial, urban, and rural areas in southern Taiwan.The simulated values for the gasparticle partitioning of PBDEs were compared with those obtained by direct measurement.Dry deposition fluxes and velocities of PBDEs were then predicted by calculations based on the model.

PBDEs Sampling
Ambient air samples were collected at four locations, one industrial area, two urban areas, and one rural area located in Tainan city, Taiwan.A total of six ambient samples in each area were collected simultaneously by using a PS-1 sampler (Graseby Andersen, GA) according to the revised EPA Reference Method T09A between November 2010 and May 2011.Particle phase PBDEs were collected on quartz filters, and those in the gas phase were collected in a modified cartridge containing polyurethane foam.A detailed description of sampling sites and procedures, as well as meteorological information for the sampling sites during the periods examined, are given in our previous work (Mi et al., 2012).

Analyses of PBDEs
All the chemical analyses in this study were performed in the Super Micro Mass Research and Technology Center in Cheng Shiu University, certified by the Taiwan EPA for PCDD/Fs sampling and analyses.PBDE analyses were performed following U.S. EPA Method 1614.Each sample was spiked with a known internal standard (10 congeners: 13 C 12  before Soxhlet extraction to monitor the extraction and cleanup procedures.After extraction, the extract was concentrated and treated with sulfuric acid, followed by a series of cleanup and fraction procedures, including a multilayered silica column, an alumina column, and an activated carbon column (Wang et al., 2010a).A high resolution gas chromatograph with a mass spectrometer (HRGC/HRMS) was used for PBDE analyses.The HRGC (Hewlett-Packard 6970 Series gas, CA) was equipped with a DB-5HT capillary column (J&W Scientific, CA) and with a splitless injector.The HRMS (Micromass Autospec Ultima, Manchester, UK) had a positive electron impact (EI+) source.Detailed analytical procedures and instrumental parameters of PBDEs are given in our previous works (Wang et al., 2010a, b, c;Wang et al., 2011).

Gas-Particle Partitioning
Several researchers have used a temperature-dependent partitioning coefficient (K p ) to describe the partitioning of SOCs between gas and particle phases (Mandalakis et al., 2002;Lin et al., 2010;Wang et al., 2010d;Huang et al., 2011).
where K p (m 3 /μg) is a temperature-dependent partitioning constant, TSP (μg/m 3 ) is the total suspended particulate concentration, and F (pg/m 3 ) and A (pg/m 3 ) are the associated particulate and gaseous concentrations of the analyte, respectively.
When the log K p is regressed against the logarithm of the subcooled liquid vapor pressure log P L o , the partitioning constant can be calculated from the slope m r and yintercept of the trend line b r (Yamassaki et al., 1982), as shown in Eq. (2).
The y-intercept of the trend line (b r ) depends mainly on the properties associated with SOC.Complete datasets on the gas-particle partitioning of PBDEs at Guangzhou City in southern China have been reported by Chen et al. (2006), giving values of m r = -0.607and b r = -5.07 with R 2 = 0.632.Due to the lack of measured data for the gasparticle partitioning of PBDEs, same parameters measured by Chen et al. (2006) were adopted in this study.
The sub-cooled liquid vapor pressure P L 0 (Pa) was evaluated from solid vapor pressure P S 0 (Pa) using the following equation (Palm et al., 2002): where F r is the fugacity ratio, T M is the melting point of the BDE congener (K), and T is the atmospheric temperature (K).

Atmospheric Dry Deposition
The atmospheric deposition flux of total PBDEs is a combination of both gas-and particle-phase fluxes, which is given by (5) where F T is the total PBDE deposition fluxes contributed from both gas and particle phases, F g and F p are deposition fluxes of PBDEs contributed by the gas phase and particle phase, respectively, C T is the measured concentrations of total PBDEs in air, V d,T is the dry deposition velocities of total PBDEs, C g and C p are the calculated PBDE concentrations in the gas phase and particle phase, respectively, and V d,g and V d,p are the dry deposition velocities of PBDEs in the gas phase and particle phase, respectively.Dry deposition of gas-phase SOCs is controlled mainly by diffusion, while that of particle-phase is deposited mainly by the gravitational settling.Lee et al. (1996) reported that the dry deposition velocity of gaseous SOCs is fairly constant as compared to that of their particle phases.A value of 0.01 cm/s for dry deposition velocity (V d,g ) of gas-phase polycyclic aromatic hydrocarbon (PAH), proposed by Sheu and Lee (1996) and used by Lee et al. (1996), is adopted in this work for the calculation of the PBDE deposition flux contributed by their gas phases.Because of the lack of measured data for the deposition velocities of PBDEs, the average of 0.8 cm/s for particle-bond PBDEs determined by Su et al. (2007) was also used here for the approximate calculation of particle deposition flux.

PBDE Concentrations in the Ambient Air
The atmospheric concentration of total PBDEs (sum of thirty BDEs) ranged from 67.3 ± 26.5 to 102 ± 13.3 and 24.0 ± 1.83 to 64.4 ± 24.0 pg/m 3 (mean ± standard error) in fall and spring, respectively (Table 1).The highest concentration of total atmospheric PBDEs was found in the urban area among the four sampling areas.Higher atmospheric PBDE concentrations in industrial sites than those in urban and residential areas have been recently reported (Cetin and Odabasi, 2008;Choi et al., 2008), due to leakage from consumer products and industrial facilities that manufacture PBDEs.As noted in our previous work, sampling site I is located near a science-based industrial park dominated by electronics and semiconductor firms (Mi et al., 2012).A much higher level of PBDEs would thus be expected for site I compared to the other sites, although this was not found in the current study.Therefore, the concentrations of atmospheric PBDEs may represent the usage level of PBDEcontaining products (Wang et al., 2011).As shown in Table 1, the higher PBDEs concentration in fall rather than in spring could be attributed to several loss processes, including photolysis, chemical reactivity, wet and dry deposition, and scavenging by vegetation (Duarte-Davidson et al., 1997).Similar trends for other SOCs were also observed in earlier studies (Agree et al., 2004;Huang et al., 2011;Mi et al., 2012).
The results for the atmospheric PBDE levels (the sum of all seven congeners) in the suburban (21-24 pg/m 3 ), urban (32-40 pg/m 3 ), and industrial areas (53-117 pg/m 3 ) of Turkey (Cetin and Odabasi, 2008), in an urban area (100 pg/m 3 , sum of 26 congeners) of Chicago, USA (Hoh and Hites, 2005), and in an urban area (35.3 ± 15.5 pg/m 3 , sum of 30 congeners) of Taiwan (Wang et al., 2011), were all comparable with our findings (Table 2).In Guangzhou City, the largest urban center in southern China, with a dense population and heavy traffic, Chen et al. (2006) observed very high atmospheric levels of the summed 11 congeners (347-577 pg/m 3 ).These results are higher than those reported here, despite the variation among the number of PBDEs detected in both studies.Additionally, the contribution of BDE-209 to the total PBDEs concentration varies from 37 to 100% in these earlier studies, and this is likely to be due to the increased use of deca-BDE formulations in recent years.
Deca-BDE and nona-BDE homologues, the highly brominated-substituted congeners, were the most dominant PBDE congeners in all sampling sites (Fig. 1).BDE-209 was the most abundant congener in all sites (more than 70%), followed by  Similarly, BDE-209 has been shown to be relatively abundant in most ambient air samples in the USA (Hoh and Hites, 2005), Taiwan (Wang et al., 2011), and Turkey (Cetin and Odabasi, 2008).However, the percentage contribution by tetra-BDE and penta-BDE homologues (mainly BDE-47 and BDE-99) in this study was relatively low compared to that found in other works.This is probably due to the change in the formulation of commercial PBDEs from penta-BDE to deca-BDE (Agrell et al., 2004).

Gas-Particle Partitioning of PBDEs
The subcooled liquid vapor pressure (P L o ) and gasparticle partitioning constant (K p ) for individual PBDEs in the ambient air, as well as gas-particle partitioning, were calculated using the regression models mentioned above.The mean partitioning fractions of particle-phase PBDEs 10.7 ± 3.71 14.1 ± 5.97 11.7 ± 2.08 11.5 ± 2.97 7.44 ± 3.41 10.9 ± 0.846 9.60 ± 0.957 5.20 ± 0.446 Σ 9-10 BDE 56.6 ± 25.9 87.9 ± 14.5 89.6 ± 28.6 60.2 ± 20.6 33.4 ± 6.25 46.7 ± 7.07 54.8 ± 23.1 18.8 ± 2.27 Total PBDEs 67.3 ± 26.5 102 ± 13.3 101 ± 28.6 71.1 ± 21.7 40.8 ± 9.67 57.6 ± 6.23 64.4 ± 24.0 24.0 ± 1.83 n: sample numbers, Σ 2-8 BDE: sum of di-to octa-BDE, Σ 9-10 BDE: sum of nona-to deca-BDE.3  and 4. Good agreement is found between the experimentally measured and simulated fractions of particulate PBDEs at all sites.The fractions of particulate PBDEs increased along with the number of brominated substitutes, and the highly brominated-substituted congeners (hepta-to deca-BDE) were mainly in the particle phase (> 60%).This is probably because gas-particle partitioning was controlled significantly by the chemical and physical properties of individual compounds, such as vapor pressure.Similar results were also observed elsewhere (Chen et al., 2006;Cetin and Odabasi, 2008).Additionally, individual PBDEs bound to particles increased slightly with decreasing temperature, and the highest level was observed in fall (temperature ranging from 21.3°C to 23.3°C), while the lowest was in spring (temperature ranging from 24.4°C to 26.8°C).Similarly, relatively high levels of PCDD/Fs and PCBs in the particle phase during cold season were observed in the authors' previous work, and this is mainly caused by atmospheric variables, including temperature, wind speed, and wind direction.
The different percentages of PBDEs found by the two methods ranged from -45.2% (deca-BDE) to 78.1% (hexa-BDE) (Tables 4 and 5).The positive values resulted from the adsorption of gas-phase SVOCs onto the filter surface, whereas the negative values occurred due to volatilization (blow-off) of SVOCs from particles on the filter (Bidleman and Harner, 2000) or breakthrough of the compounds from the gas-phase sorbent (Lee and Jones, 1999).The results show that the influence of the blow-off mechanism was observed mostly from highly brominated-substituted congeners (nonaand deca-BDE) in both seasons.Mandalakis and Stephanou (2007) reported that sampling artifacts (evaporation losses and/or adsorption onto filters) during the collection of particulate PBDEs, as well as the uncertainties inherent in model predictions, may bias the results of both approaches.Slight differences (< 30%) were observed between the measured and simulated particle-bound fractions of tri-and tetra-BDE, and the greatest differences were found for pentaand hexa-BDE.These findings may be attributed to the greater analytical errors found with penta-and hexa-BDE.

Dry Deposition Fluxes of PBDEs
The mean dry deposition fluxes of total PBDEs measured at four sites ranged from 39.0 to 60.4 ng/m 2 -day and 13.4 to 37.6 ng/m 2 -day in fall and spring, respectively (Tables 5  and 6).The results reveal that particle bond deposition contributed more than 99% of total dry deposition flux at all sites, which is similar to the observations of dry deposition flux of PCDD/Fs and PCBs reported in the authors' previous work.This may be due to the higher amounts of particles in the atmosphere of the investigated areas.The dry deposition fluxes of total PBDEs determined from this study were approximately two times higher than those observed at Lake Maggiore in northern Italy (17.6 ng/m 2 -day, sum of eight congeners) (Mariani et al., 2008), and at an industrial urban reference site in southern Sweden (18.2 ng/m 2 -day, sum of nine congeners) (ter Schure et al., 2004).The differences   the different ambient concentrations and meteorological conditions (i.e., temperature, relative humidity, wind speed, and atmospheric stability) (Tasdemir and Holsen, 2005).Additionally, the deposition flux of total PBDEs can be three orders of magnitude higher than those of PCDD/Fs (157 to 544 pg/m 2 -day) and PCBs (89 to 1010 pg/m 2 -day).Since atmospheric deposition is believed to be the main transfer pathway of SOCs into food chains, its impact on human exposure to PBDEs is of great importance.

Dry Deposition Velocities of Individual PBDEs
In order to better understand the dry deposition process, the total deposition velocities of individual PBDEs were calculated, as shown in Fig. 3.The estimated deposition velocities of individual PBDEs ranged from 0.014 cm/s (di-BDE) to 0.748 cm/s (deca-BDE), with a mean of 0.679 cm/s in fall, and from 0.015 cm/s (di-BDE) to 0.755 cm/s (deca-BDE) with a mean of 0.664 cm/s in spring.The calculated deposition velocities of PBDEs increased along with the number of brominated substitutes, and similar patterns of deposition velocities were observed for PCDD/Fs and PCBs in our previous work (Mi et al., 2012).However, no significant seasonal variation was found for the total (gas + particulate phases) deposition velocity of PBDEs.
The total (gas + particulate phases) deposition velocity for BDE-209 calculated in the present work was comparable to that for particulate phases measured in Hong Kong (Li et al., 2010) and urban Guangzhou in China (Zhang et al., 2012), but significantly lower than the values detected for a site near a municipal solid waste incineration and electronics recycling plant in Sweden (ter Schure et al., 2004) and in urban and suburban Izmir in Turkey (Cetin and Odabasi, 2007).The differences among deposition velocities could be attributed to the particle size distribution, sampling site, meteorological conditions, and different measurement techniques (ter Schure et al., 2004).Finally, when compared with the total deposition velocity of other SOCs, the calculated values of PBDEs were in line with those reported for PCDD/Fs (0.031-0.546 cm/s) and PCBs (0.069-3.38 cm/s), as shown in our previous work (Mi et al., 2012).

CONCLUSIONS
This study is an extension of our previous research, which focused on the dry deposition characteristics of PBDEs in the ambient air of southern Taiwan.Higher atmospheric PBDE concentrations in urban sites than those in rural and industrial areas were observed, and these may indicate different usage levels with regard to PBDE-containing products.The fractions of particulate PBDEs increased along with the number of brominated substitutes, and the highly brominated-substituted congeners were found to be predominantly associated with particles at all sites.The calculated deposition flux of total PBDEs was found to decrease as the temperature increased between 13.4 and 60.4 ng/m 2 -day, and BDE 209 accounted for over 75% of the total PBDEs.The high contribution of BDE-209 to total flux may reflect the changes in production and usage patterns from the commercial penta-to deca-BDE mixtures.The deposition velocities of individual PBDEs increased along with the number of brominated substitutes, and ranged from 0.014 to 0.755 cm/s.Similar patterns of velocities have also been observed in our previous work (Mi et al., 2012) for PCDD/Fs and PCBs.
of four different areas are shown in Tables

Fig. 1 .
Fig. 1.Congener profiles of PBDEs in the ambient air of four different areas.
F d,T : Total deposition flux of PBDEs, b P %: The ratio of dry deposition flux contributed by the particle-phase of PBDEs.

Fig. 2 .
Fig. 2. Estimated monthly dry deposition fluxes of PBDEs of four different areas.

Fig. 3 .
Fig. 3. Mean dry deposition velocities of individual PBDEs of four different areas.

Table 1 .
Mean concentrations of PBDEs in the ambient air of four different areas.

Table 2 .
Comparison of PBDE concentrations in the ambient air from different countries.

Table 3 .
Mean partitioning fractions of particle-phase PBDEs in the ambient air of four different areas in fall.

Table 4 .
Mean partitioning fractions of particle-phase PBDEs in the ambient air of four different areas in spring.

Table 5 .
Dry deposition fluxes of PBDEs in the ambient air of four different areas in fall (pg/m 2 -day).

Table 6 .
Dry deposition fluxes of PBDEs in the ambient air of four different areas in spring (pg/m 2 -day).