Article pubs.acs.org/est
Plant Uptake of Atmospheric Brominated Flame Retardants at an E-Waste Site in Southern China Mi Tian,†,‡ She-Jun Chen,*,† Jing Wang,†,‡ Yong Luo,§ Xiao-Jun Luo,† and Bi-Xian Mai† †
State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China ‡ Graduate University of Chinese Academy of Sciences, Beijing 100049, China § Guangdong Forestry Survey and Planning Institute, Guangzhou 510520, China S Supporting Information *
ABSTRACT: Brominated flame retardants (BFRs) were measured in eucalyptus leaves and pine needles as well as the leaf surface particles (LSPs) of the two species at an e-waste site in southern China in 2007−2008. The monthly concentrations of total BFRs in the eucalyptus leaves and pine needles were in range of 30.6−154 and 15.1−236 ng/g dry weight, respectively, and relatively higher concentrations were observed in winter and spring. Correlation analysis of BFR concentrations and comparison of PBDE compositions between the plants and LSPs, air (gaseous and particle-bound phases), and ambient variables were conducted. The results revealed that BFRs in the plants, especially for less brominated BFRs, showed positive relationships with BFRs in the LSPs and negative relationships with the gaseous BFRs and ambient temperature. The PBDE profiles in the plants were similar to the gaseous profile for low brominated BDEs (di- through hexa-BDEs) and to the LSP profiles for highly brominated BDEs (hepta- through decaBDEs). Applying McLachlan’s framework to our data suggests that the uptake of BFRs was controlled primarily by gaseous partitioning equilibrium for compounds with log octanol-air partition coefficients (KOA) < 12 and by particle-bound deposition for compounds with log KOA > 13. Different relationships between the plant/air partition coefficient (KPA) and KOA, which depend on the uptake mechanisms, were observed for polybrominated diphenyl ethers (PBDEs). This paper adds to the current knowledge of the factors and mechanisms governing plant uptake of semivolatile organic compounds with relatively high KOA in the environment.
■
and subsequent partitioning between particles and plant cuticles.8 McLachlan’s framework identifies particle-bound deposition as the dominant uptake process for SOCs with log KOA > 11.4 Despite its importance, however, the mechanism of this process is poorly understood.8,9 In addition, most field studies focused on SOCs between air and plant from off-site aerial sources. By contrast, plant uptake of SOCs from local sources, in particular the particle-bound deposition process, is a pressing area for investigation.8 Brominated flame retardants (BFRs) are commonly used in various products including textiles, furniture, electronics, and cars. These chemicals, in particular polybrominated diphenyl ethers (PBDEs), have received great attention over the past decade because of their ubiquitous presence in the environment
INTRODUCTION The uptake of semivolatile organic compounds (SOCs) from the atmosphere into plants has attracted considerable research interest because it plays an important role in trapping and transferring airborne SOCs to terrestrial ecosystems and affecting their global movement.1−3 McLachlan has developed an interpretive framework identifying three dominant processes (equilibrium partitioning between the vegetation and the gas phase, kinetically limited gaseous deposition, and wet and dry particle-bound deposition) as a function of the octanol-air partition coefficient (KOA) of SOCs.4 Dry gaseous deposition has shown to be the major uptake process for a variety of SOCs including polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), and some polychlorinated dibenzo-p-dioxins/ dibenzofurans (PCDD/Fs).5−7 Numerous studies have been conducted to elucidate the uptake and clearance mechanisms, partitioning equilibrium, or interspecies variability of this process. Plant uptake of SOCs via particle-bound deposition is a process involving particle diffusion, impaction, sedimentation, © 2012 American Chemical Society
Received: Revised: Accepted: Published: 2708
October 16, 2011 January 29, 2012 February 6, 2012 February 6, 2012 dx.doi.org/10.1021/es203669n | Environ. Sci. Technol. 2012, 46, 2708−2714
Environmental Science & Technology
Article
and their adverse effects on humans and wildlife.10,11 BFRs have high log KOA values (8−15) compared to PCBs (7−11), PAHs (6−13), and PCDD/Fs (8−13),12−15 and therefore, are more likely to partition onto atmospheric particles. Nevertheless, there is limited information in the literature on the uptake of airborne BFRs into plant.2,16−18 Electronic waste (e-waste) recycling has proved to be a significant emission source of BFRs to the atmosphere.19 In our previous study we examined the concentrations, composition, and seasonal variations of BFRs in the air at an e-waste site in southern China and substantially high air BFR concentrations (120−19 000 pg/m3) were observed.19 In the present study, BFRs in leaves of two plant species having different morphologies from this e-waste site were measured. Particles deposited on the leaf surface were separated and analyzed for BFRs. Interspecies variability in BFR concentration and composition profiles in these plants and particles were examined. Correlation analysis was conducted between BFRs in plant leaves and in leaf surface particles (LSPs), gaseous phase, and particle-bound phase to provide information on the plant uptake of these compounds from air. The main objectives were to identify the dominant uptake mechanisms by applying a previous predictive model to the field data and to investigate the relationships between the plant/air partition coefficients (KPA) and KOA for these SOCs with relatively high KOA.
negative ionization mode (GC-ECNI-MS). Di- through heptaBDEs, PBEB, PBT, HBB, and BB153 were separated with a DB-XLB (30 m × 0.25 mm i.d., 0.25 μm film thickness) capillary column. For octa- through deca-BDEs, DBDPE, BTBPT, and BB209, a DB-5HT (15 m × 0.25 mm i.d., 0.10 μm film thickness) column was used. Detailed analytical procedures are described in the SI. Quality Control (QC). The QC results for air samples have been given elsewhere and only QC results for plant samples were shown here.19 Procedural blanks were run sequentially with each batch of the field samples. BDE13, 28, 47, 99, 153, 206, 207, 208, and 209 were found in the blanks, but their amounts were all less than 3% of those in the corresponding sample extracts. The concentrations in the samples were blank corrected accordingly. BDE77, BDE181, and 13C-BDE209 were added as surrogates prior to extraction and the recoveries were 105 ± 23.0%, 64.2 ± 11.1%, and 72.4 ± 12.7%, respectively. The recoveries of target compounds in the spiked blanks ranged from 62.5%−142% (standard deviations 2 m above ground level) at the beginning, middle, and end of each sampling session. Both eucalyptus and masson pine are extensively distributed species in southern China and the e-waste site is located in a rural area with a high forest coverage rate of ∼68%.21 The plant samples from the same species and sampling session were pooled for analysis. Meteorological data, such as temperature, humidity, and wind direction and speed, were obtained from the local meteorological administration. Sample Preparation and Instrumental Analysis. Leaves were rinsed with purified water to collect LSPs. Subsample (∼10 g) of pooled leaves was ground with anhydrous sodium sulfate and Soxhlet extracted. The plant extract was mixed with 60 mL concentrated sulfuric acid to remove lipids and then was liquid−liquid extracted with hexane using a Teflon separatory funnel (NalgeNunc, Rochester, NY). The hexane extracts were then further purified on a multilayer alumina/silica column. Extractable organic matter and water content of the leaves was also determined. BFRs, including PBDEs, decabromodiphenyl ethane (DBDPE), 1,2-bis(2,4,6-tribromophenoxy)ethane (BTBPE), 2,3,4,5,6-pentabromoethylbenzene (PBEB), pentabromotoluene (PBT), hexabromobenzene (HBB), and polybrominated biphenyls (PBBs), were analyzed using a gas chromatograph coupled to a mass spectrometer in electron capture
CV /CG = AvGGt /V
(1)
CP = B × TSP × KOACG
(2)
where CV, is the concentration of SOCs in the vegetation, CG and CP are the gaseous and particle-bound concentrations in the bulk atmosphere respectively, A is the surface area of the vegetation (m2), vGG is the mass transfer coefficient describing transport from the atmosphere to the surface of the vegetation (m h−1), t is time (h), V is the volume of the vegetation (m3), B is a constant whch is a very small number (typically 10−12 m3 μg−1), and TSP is the total suspended particle concentration in the air (μg m−3).22 Therefore, CV /CA =
AvGGt /V 1 + B × TSP × KOA
(3)
KPA (CV/CA) negatively correlates with KOA. However, this relationship is not usually significant for compounds with log KOA < 11 because B × TSP is usually with an order of magnitude of 10−11 (B × TSP × KOA ≪ 1). For compounds with log KOA > 11 that are subject primarily to kinetically limited gaseous deposition (B × TSP × KOA ≫ 1), eq 3 simplifies to CV /CA =
AvGGt /V B × TSP × KOA
(4)
In this case, log KPA shows negative linear relationship with log KOA. McLachlan’s interpretive framework is based on an assumption of air-side resistance dominating the uptake via gaseous deposition. For plant-side resistance, n CV /CG = AvGV mKOA t /V 2709
(5)
dx.doi.org/10.1021/es203669n | Environ. Sci. Technol. 2012, 46, 2708−2714
Environmental Science & Technology
Article
Table 1. Summary of BFR Concentrations in Eucalyptus Leaves and Pine Needles (ng/g Dry Weight) and Leaf Surface Particles (LSPs) (ng/g) from the E-Waste Site eucalyptus leaf di-BDEs tri-BDEs tetra-BDEs penta-BDEs hexa-BDEs hepta-BDEs octa-BDEs nona-BDEs deca-BDEs ∑PBDEs PBT PBEB HBB BTBPE DBDPE PBBs ∑BFRs a
pine needle
LSP/eucalyptus leaf
LSP/pine needle
range
mean/median
range
mean/median
range
mean/median
range
mean/median
0.48−15.9 0.15−12.9 0.42−21.5 0.27−8.49 0.50−2.71 0.25−1.59 0.40−2.88 0.76−13.5 6.87−45.2 18.9−103 0.03−3.73 0.04−8.04 0.19−9.94 0.09−1.09 9.14−35.6 0.13−1.90 30.6−154
6.47/5.80 6.50/5.13 9.93/9.40 5.00/5.79 1.29/1.22 0.68/0.55 1.33/1.22 4.53/3.36 28.3/25.6 64.1/65.5 1.40/1.03 2.63/1.47 5.24/4.98 0.55/0.51 23.7/25.2 0.79/0.68 98.4/104
0.33−102 0.52−83.3 0.68−52.0 0.71−38.4 0.48−8.83 0.32−4.83 0.53−9.77 1.79−12.9 15.6−75.1 27.9−383 0.08−20.1 0.08−42.1 0.44−53.5 0.10−4.81 11.2−42.3 0.54−7.94 40.4−546
31.3/20.4 23.4/18.4 20.2/18.6 14.9/12.7 3.68/2.87 2.00/1.62 3.40/2.56 7.13/7.09 41.7/40.5 148/145 4.31/2.03 8.41/3.40 15.8/11.5 1.63/1.47 27.2/27.8 2.26/1.79 207/196
nda−10.8 1.02−23.3 10.8−133 19.2−172 11.4−147 25.6−315 5.27−390 17.6−2784 176−10850 287−14820 1.64−18.6 0.57−16.2 8.38−108 3.36−616 64.4−8433 n.d.−119 373−21 470
4.14/4.78 7.14/5.16 41.0/27.0 67.4/50.6 56.6/36.6 138/77.8 86.3/47.2 727/162 3397/1740 4524/2260 6.96/5.28 4.03/1.83 29.8/21.8 90.0/14.4 2378/988 19.7/3.44 7053/3293
nda−33.4 2.13−49.2 8.02−70.3 12.5−90.9 5.82−78.9 10.0−180 13.2−86.5 122−836 831−6821 1187−7906 1.15−5.84 0.43−5.78 13.7−69.8 5.79−57.0 493−4885 1.15−44.8 1777−11110
17.2/19.6 22.8/25.3 45.8/48.6 54.3/50.0 30.2/26.0 63.6/59.8 59.2/70.7 413/279 2846/1986 3552/2184 3.18/2.75 3.38/3.65 41.7/39.9 24.1/21.4 1842/686 21.0/16.5 5488/2904
Not detectable.
where vGV is the mass transfer coefficient describing transport from the surface of the vegetation to the contaminant reservoir in the plant (m h−1), m and n are constants that are specific for a given plant.4 Therefore, CV /CA =
n AvGV mKOA t /V 1 + B × TSP × KOA
mass.25 The specific leaf area of pine needle (17.2 m2/kg) is 3-fold larger than that of eucalyptus leaf (5.80 m2/kg),26,27 which may be responsible primarily for the interspecies difference in BFR concentration. The average concentrations of BFRs in the LSPs (7050 ng/g for eucalyptus leaves and 5500 ng/g for pine needles, Table 1) were lower than that in the bulk atmospheric particles in the air (17 400 ng/g, normalized to the particle masses, SI Table S1) at this site. The main reason for this was presumably the particle size distribution of BFRs. Large particles are more readily to deposit on leaf surface because of the high deposition velocities, while, like many other SOCs, BFRs may be mainly associated with fine particles.28,29 An earlier study found that the plant uptake of highly polychlorinated PCDD/Fs was dominated by deposition of large particles (>2.9 μm).7 Another interpretation was that the lower BFR concentrations in LSPs were caused by desorption of the chemicals from LSPs into leaves. There were limited data to compare with the BFR concentrations in plants measured in this study. The PBDE concentrations in plant leaves in the present study ranging from 18.9 to 382 ng/g dw were comparable to the PBDE concentrations (12.8−400 ng/g dw) in plant leaves from other e-waste sites in China,30,31 but were much higher than those in spruce needles nearby a sanitary landfill in Ottawa (0.21−9.94 ng/g dw)32 and those in tree leaflitter (5−14 ng/g dw) at a rural site in southern Ontario.16 The results revealed a significant contamination of BFRs in the terrestrial plants at the e-waste site and implied exposure of local people to these chemicals from fruit and vegetable intake. Relationships with BFRs in Air and LSPs and Ambient Variables. The variation of BFRs in the plants depends on the air concentrations and the partitioning between air and plant, in addition to various ambient variables such as wind speed, temperature, and humidity.33 The relationships between concentrations in plant leaves and those in the gaseous phase, atmospheric particles, and LSPs and ambient variables were therefore investigated for individual BFRs separately (Table 2). BFRs in plants showed no clear relationships with the BFRs in the atmospheric particles with the exception of deca-BDE in eucalyptus leaves showing a significantly negative relationship
(6)
For compounds with log KOA > 11, eq 6 simplifies to CV /CA =
n−1 AvGV mKOA t /V B × TSP
(7)
There would be a negative linear relationship between log KPA and log KOA because n is usually less than 1, if vGV is constant for the compounds studied.
■
RESULTS AND DISCUSSION Concentrations. The concentrations and seasonal variations of BFRs in the air (gaseous and particle-bound) have been reported in our previous study.19 The total BFR concentrations ranged from 120 to 19 000 pg/m3 (with an average value of 3810 pg/m3 (SI Table S1). The statistics of BFR concentrations in the plants were summarized in Table 1. BFRs were found in all the eucalyptus leaves and pine needles, with average total BFR concentrations of 98.4 and 207 ng/g dry weight (dw), respectively. Higher air concentrations of BFRs were observed in the summer months, whereas the total BFR concentrations in the plants were higher in winter and spring with monthly variation by a factor of 1−3 during the sampling year (SI Table S2). The concentrations of BFRs in pine needles were significantly higher than those in eucalyptus leaves (p = 0.04). It was evidenced that leaf lipid plays an important role in uptake and storage of SOCs.23 However, the interspecies variability in BFR accumulation in the present study may not be explained by variability in extractable lipid contents (77 mg/g dw for eucalyptus leaves and 82 mg/g dw for pine needles). Similar results have been observed for a variety of SOCs in earlier studies.6,24 The specific leaf area can be regarded as a measure of the surface available for atmospheric SOC uptake per unit of leaf 2710
dx.doi.org/10.1021/es203669n | Environ. Sci. Technol. 2012, 46, 2708−2714
Environmental Science & Technology
Article
Table 2. Pearson Correlation Coefficients (r) for the Plant BFR Concentration versus the BFR Concentrations in Airborne Particles, Gaseous Phase, and Leaf Surface Particles (LSPs) and Temperature (1/T) at the E-waste Site eucalyptus foliage di-BDEs tri-BDEs tetra-BDEs penta-BDEs hexa-BDEs hepta-BDEs octa-BDEs nona-BDEs deca-BDEs PBT PBEB HBB BTBPE DBDPE ∑BFRs a
pine needles
atmospheric particle
gas
LSP
1/T
atmospheric particle
gas
LSP
1/T
−0.229 −0.018 −0.151 −0.282 −0.154 −0.214 0.043 0.052 −0.669a 0.248 −0.220 0.222 −0.446 0.363 −0.448
−0.213 −0.448 −0.694a −0.677a −0.384 −0.329 −0.790a NAb NA −0.480 −0.452 −0.536 0.386 NA −0.520
0.196 0.656 0.611 0.525 0.412 0.465 −0.267 −0.100 −0.112 0.157 0.389 0.576 0.494 0.270 0.361
0.675a 0.561 0.453 0.468 −0.047 −0.060 0.039 −0.233 0.237 0.714a 0.756a 0.507 0.340 0.549 0.438
−0.160 0.093 0.048 −0.177 −0.029 −0.038 −0.153 0.327 0.141 0.075 −0.306 0.313 −0.360 0.605 −0.107
−0.234 −0.409 −0.612 −0.772a −0.661a 0.214 −0.217 NA NA −0.475 −0.590 −0.325 0.361 NA −0.330
0.736a 0.878a 0.914a 0.814a 0.398 0.161 0.519 −0.216 −0.343 0.554 0.807a 0.897a 0.337 0.368 −0.065
0.706a 0.692a 0.650a 0.648a 0.598 0.623 0.566 0.226 0.523 0.660a 0.641a 0.670a 0.713a 0.696a 0.703a
Correlation is significant at the 0.05 level (2-tailed). bNot analyzed.
Figure 1. Composition profiles of PBDEs in eucalyptus leaves, pine needles, leaf surface particles (LSPs), gaseous phase, atmospheric particles, and air at the e-waste site.
(r = −0.67, p < 0.05). Many BFRs in the plant leaves had positive relationships with those in the LSPs, although many of the relationships were not significant. Surprisingly, however, highly brominated BFRs (nona- and deca-BDEs, BTBPE, and DBDPE) showed less significant such relationships in both the species compared to less brominated ones. McLachlan’s interpretive framework predicts that uptake of SOCs with high log KOA values is dominated by particle-bound deposition.4 Thus the weak relationships could result from the high uptake rates of highly brominated BFRs from LSPs into the leaves, which significantly affected the concentrations in LSPs. It has been shown that cuticular permeability increases with KOA of SOCs.23 This finding also suggests that even for less brominated BFRs plant uptake from particle-bound deposition is not neglectable. It is surprising to find negative relationships between the concentrations in plants and gaseous phase for most BFRs (especially less brominated BFRs), although they are not always statistically significant. This could be explained by the exchange of BFRs between air and plant: uptake into plant leaves will reduce the ambient concentrations, while evaporation from plant leaves will increase the ambient concentrations. This explanation is supported by the negative relationships between the plant concentration and ambient temperature for many less brominated
BFRs. Because of the high forest coverage rate in the studied area, the exchange of BFRs between the air and extensive mass of vegetation in this area would have a significant influence on the air levels. It is worthwhile to note that the change in gaseous BFR concentrations was not only a result of exchange between the air and plant but also a combined effect of temperature-driven exchange between the air and various contaminated components such as soil, plant, and e-waste residues stacked in the field, which was also indicated by our previous investigation.19 The plant BFR concentrations showed no relationships with wind speed or relative humidity (SI Table S3). PBDE Congener Profiles. PBDE congeners were dominated by BDE209 which accounted for 60% and 50% on average of the total PBDE burden in eucalyptus leaves and pine needles, respectively (Figure 1). The BDE209 contributions in the plants in present study were similar to those reported in spruce needles (66%) in Ottawa, Canada, tree bark (60%) in North America, and vegetables in southern China (50%).18,32,34 Low proportions of BDE209 were found in tree bark (approximately 15−35%) in Great Lakes basin and plant leaves ( 13, there was also a positive linear relationship (p < 0.06). Interestingly, the log KOA values at which the relationships changed were generally corresponding to those at which the uptake mechanisms transformed. It is therefore very likely that the plant uptake processes are responsible for the changes in relationship between log KPA and log KOA. For SOCs primarily as a result of gaseous equilibrium partitioning, a positive linear log KPA−log KOA relationship is expected as obtained in many previous studies for PCBs. The negative relationships between log KPA and log KOA for SOCs related primarily to kinetically limited gaseous deposition were consistent with the theory of McLachlan’s interpretative framework (see detailed calculations above). Whereas this relationship is often negligible for SOCs with low log KOA (typically
