Environ. Sci. Technol. 2010, 44, 6196–6201
Evaluation of Tree Bark as a Passive Atmospheric Sampler for Flame Retardants, PCBs, and Organochlorine Pesticides AMINA SALAMOVA AND RONALD A. HITES* School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405
Received May 11, 2010. Revised manuscript received June 23, 2010. Accepted June 30, 2010.
To investigate the relationship between the levels of persistent organic pollutants in tree bark (a passive sampler) and those in air and precipitation, tree bark and air and precipitation samples were collected during the same time period at the five U.S. Integrated Atmospheric Deposition Network (IADN) sites located in Great Lakes basin. The concentrations of polybrominated diphenyl ethers, Dechlorane Plus, decabromodiphenyl ethane, polychlorinated biphenyls, DDTs, and chlordanes were measured in these samples. Overall, the pollutant concentrations in tree bark are significantly related to the concentrations of these compounds in the air and precipitation collected where the tree was growing. Generally, the highest tree bark and air pollutant concentrations were observed at urban sites, and the lowest concentrations were observed at remote sites. The overall correlation between bark and atmospheric and precipitation concentrations for all the compounds measured in this study was highly significant (P < 0.0001) over 3-4 orders of magnitude. In addition, bark-air partition coefficients, measured for all the chemical categories in this study, were about 106, which was in good agreement with previously estimated bark-air partition coefficients for corresponding pollutant groups.
Introduction The measurement of persistent organic pollutant concentrations in air is important to assess the atmospheric transport of these compounds. These measurements, however, are often limited by the high cost of conventional high-volume, active air sampling. Passive air sampling is a less expensive alternative to active air sampling; in fact, polyurethane foam (PUF) disk passive samplers are beginning to be widely used and have been shown to be effective (1, 2). One significant disadvantage of these PUF samplers is that they require deployment for several weeks to months. An alternate passive sampler is tree bark, which is always deployed and continuously collects pollutants from the surrounding air. Obtaining tree bark samples is easy, fast, and inexpensive. Tree bark passive sampling is especially advantageous in remote settings, where electrical power is not available to operate conventional high volume samplers. Contaminant screening by tree bark has been used to monitor various inorganic and organic pollutants, such as ammonia (3), heavy metals (3-8), polychlorinated biphenyls * Corresponding author e-mail:
[email protected]. 6196
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
(9-11), polycyclic aromatic compounds (12), polychlorinated dibenzo-p-dioxins (13), organochlorine pesticides (14), and brominated and chlorinated flame retardants (15, 16). Tree bark works well as a passive sampler because of its high lipid content and its large surface area; thus, it is a good passive sampler for pollutants with high octanol-air partition coefficients. Tree bark seems to accumulate both vapor- and particle-phase pollutants from the surrounding air (17) and integrates pollutant levels over a period of several years, depending on the time the bark stays on the tree (18). The goal of this study is to examine the relationship between the levels of selected pollutants in tree bark and their levels in the vapor and particle phases in the surrounding air obtained using high-volume air sampling methods and to examine the relationship between the levels in tree bark and their levels in precipitation falling on the tree. Significant correlations between these concentrations on the one hand and in the tree bark on the other will demonstrate the ability of tree bark to passively sample these three atmospheric media and give some indication about how tree bark accumulates persistent organic pollutants from the atmosphere. The analytes of interest in this study are polybrominated diphenyl ethers (PBDEs), Dechlorane Plus (DP), decabromodiphenyethane (DBDPE), polychlorinated biphenyls (PCBs), and selected organochlorine pesticides, such as DDT and chlordane. Tree bark, air, and precipitation samples were collected at the five U.S. Integrated Atmospheric Deposition Network (IADN) sites located in Great Lakes basin.
Experimental Section Sampling Information. The locations of the U.S. IADN sampling sites are shown in the Supporting Information (Figure S1). There are five U.S. IADN sites: urban sites in Chicago, IL (41.8344°N, -87.6247°W) and Cleveland, OH (41.4921°N, -81.6785°W); rural sites at Sleeping Bear Dunes, MI (44.7611°N, -86.0586°W) and Sturgeon Point, NY (42.6931°N, -79.0550°W); and a remote site at Eagle Harbor, MI (47.4631°N, -88.1497°W). The IADN Web site provides more detailed information on air sampling procedures and site operations (www.msc.ec.gc.ca/iadn). The atmospheric samples discussed here were collected during the period January 1, 2003 to December 31, 2007. All tree bark samples were collected in the winter of 2008. Only white pine trees were used because of their occurrence at all sites. The average tree diameter at 1.5 m height was 89 cm. Although the exact age of the trees was unknown, based on their trunk diameter it was evident that these trees were fully grown. Given that pine bark has a residence time on the tree of about 5 years, bark samples collected in 2008 had been passively sampling the surrounding air during the 2003-2007 period corresponding to the air sampling. Bark samples were collected from four individual trees at each of the five IADN locations. Trees closest to the IADN active air sampler were chosen for bark sampling. A sample of bark (about 100 g) was collected from two different sides of each tree using a precleaned chisel at a height of 1.5 m. To avoid long-term damage to the tree, the cambium was not sampled. The bark sample was wrapped in aluminum foil, sealed in a plastic bag, and kept at ambient temperature until the samples were shipped to the laboratory, where they were kept at -20 °C until extraction. A modified Anderson high-volume air sampler (General Metal Works, model GS2310, modified) was used to collect air samples for 24 h every 12 days at a flow rate giving a total sample volume of about 820 m3. The vapor phase was collected on Amberlite XAD-2 resin (Supelco, Bellefonte, PA; 10.1021/es101599h
2010 American Chemical Society
Published on Web 07/21/2010
VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6197
date
a
EH EH EH EH EH CH CH CH CH CH SB SB SB SB SB CL CL CL CL CL SP SP SP SP SP
ND, not detected.
3/1/08 3/1/08 3/1/08 3/1/08 mean ( std. err. 4/9/08 4/9/08 4/9/08 4/9/08 mean ( std. err. 3/22/08 3/22/08 3/22/08 3/22/08 mean ( std. err. 4/9/08 4/9/08 4/9/08 4/9/08 mean ( std. err. 3/18/08 3/18/08 3/18/08 3/18/08 mean ( std. err.
S NW NNW ENE
W W W W
NE SW SSW ESE
NE SW SSW ESE
S S S S
73 19 33 61
804 809 816 825
32 60 65 87
965 805 483 1127
122 123 126 127
111 139 71.1 76.2
81.3 86.7 59.7 92.7
90.2 90.2 99.2 72.1
87.6 81.3 83.8 91.7
79.0 86.0 108 91.0
distance trunk from direction diameter site site (m) from site (cm) 0.559 1.08 2.09 0.908 1.16 ( 0.329 14.8 14.8 12.7 17.9 15.1 ( 1.08 2.00 0.493 1.25 0.220 0.989 ( 0.400 12.1 11.4 6.48 6.95 9.22 ( 1.46 2.29 1.68 5.75 1.66 2.85 ( 0.979
total PBDEs 0.189 0.260 0.550 0.290 0.322 ( 0.079 2.32 2.67 2.35 3.43 2.69 ( 0.259 0.355 0.170 0.290 0.083 0.225 ( 0.061 2.00 1.63 1.38 1.45 1.61 ( 0.137 0.586 0.289 1.20 0.295 0.592 ( 0.213
BDE-47 0.179 0.317 0.920 0.301 0.429 ( 0.167 3.62 3.59 3.22 5.01 3.86 ( 0.392 0.854 0.160 0.384 0.072 0.368 ( 0.175 3.08 2.98 2.53 2.56 2.78 ( 0.142 0.868 0.445 2.73 0.277 1.08 ( 0.564
BDE-99 0.043 0.069 0.229 0.071 0.103 ( 0.043 0.972 0.906 0.789 1.38 1.01 ( 0.128 0.288 0.058 0.089 0.017 0.113 ( 0.06 1.01 1.04 0.812 0.836 0.925 ( 0.059 0.227 0.128 0.725 0.074 0.289 ( 0.149
BDE-100 0.111 0.172 0.166 0.182 0.158 ( 0.016 5.82 5.14 4.77 4.85 5.14 ( 0.239 0.099 0.048 0.230 0.036 0.103 ( 0.044 3.67 3.39 0.518 0.575 2.04 ( 0.863 0.412 0.689 0.347 0.923 0.593 ( 0.133
BDE-209
DBDPE
0.025 NDa 0.068 ND 0.028 ND 0.056 ND 0.044 ( 0.011 ND 0.331 0.191 0.399 0.298 0.482 0.159 0.637 0.145 0.462 ( 0.066 0.198 ( 0.035 0.025 ND 0.066 ND 0.019 ND 0.008 ND 0.015 ( 0.004 ND 0.632 ND 0.296 0.160 0.093 0.043 0.103 0.056 0.281 ( 0.126 0.087 ( 0.037 3.20 ND 4.08 ND 4.02 ND 3.68 ND 3.74 ( 0.203 ND
DP
4.03 9.91 3.97 4.70 5.65 ( 1.43 13.0 18.3 22.0 13.7 16.8 ( 2.11 5.44 7.44 4.29 5.72 ( 0.92 27.2 7.40 11.4 10.0 14.0 ( 4.49 6.92 6.52 7.60 7.01 ( 0.316
total PCBs
0.135 0.511 0.694 0.465 0.451 ( 0.116 1.41 2.21 3.08 1.40 2.02 ( 0.400 0.669 0.272 0.316 0.148 0.351 ( 0.112 1.69 0.375 0.664 0.835 0.891 ( 0.283 0.820 0.948 1.79 0.883 1.11 ( 0.229
total DDTs
TABLE 1. Locations of Tree Bark Samples and Concentrations (in ng/g Bark) of Total PBDEs, BDE-47, 99, 100, and 209, DP, DBDPE, Total PCBs, Total DDTs, and Total Chlordanes
0.095 0.252 0.289 0.151 0.197 ( 0.045 1.10 2.11 1.52 2.17 1.73 ( 0.255 0.195 0.263 0.119 0.192 ( 0.042 0.562 0.561 0.339 0.648 0.528 ( 0.066 0.160 0.162 0.742 0.355 ( 0.194
total chlordanes
20-60 mesh) held in a stainless steel cartridge, and particles were collected on Whatman quartz fiber filters (QM-A, 20.3 × 25.4 cm). Details of the sampling procedures and site operations can be found elsewhere (19). Precipitation samples were collected using MIC automated wet-only samplers (MIC Co., Thornhill, ON). Details on the performance of these samplers are provided elsewhere (20). Each sampler consists of a 46 × 46 cm stainless steel funnel connected to a 30-cm long by 1.5-cm i.d. glass column (ACE Glass, Vineland, NJ) packed with XAD-2 resin. The sampler is normally covered; it opens when a precipitation event is sensed by a conductivity grid located outside of the sampler. Precipitation flows through the funnel and the XAD-2 column into a large carboy used to measure the total precipitation volume. Both particulate and dissolved organic phase compounds are collected by the XAD-2 column. The funnel and the interior of the sampler are kept at 15 ( 5 °C to melt snow collected in the sampler and to prevent the XAD column from freezing. Precipitation events are integrated over a 1-month period. Sample Analysis. Details on sample preparation and analysis procedures are provided in the Supporting Information. The quality of the data was assured by the analysis of blank samples and spiked reference samples.
Results and Discussion The sampling dates and locations, some details about the trees sampled, and the concentrations of the analytes of interest in each tree bark sample are given in Table 1. The corresponding 25th, 50th, and 75th percentiles, the average concentrations of the analytes, and the number of samples collected in the vapor phase, in the particle phase, in these two phases added together, and in precipitation are shown in Tables S1-S4, respectively. Our data analysis focused on the correlations between the atmospheric and the tree bark concentrations. For example, Figure 1 shows the tree bark concentrations of PBDEs (top), and their atmospheric vapor phase concentrations (middle) as a function of location, and the correlation between these two sets of data (bottom). In both air and bark, the concentrations are highest in the cities of Chicago and Cleveland and lowest at the remote locations in northern Michigan. The vapor and bark concentrations are highly correlated (r2 ) 0.917, P ) 0.010). A similar analysis of the tree bark concentrations vs the atmospheric particle concentrations, the vapor plus particle concentrations, and the concentrations of PBDEs in precipitation gave generally lower, but still significant, correlations (r2 ) 0.804, 0.929, and 0.787; P < 0.045). These correlation analyses were repeated for all of the analytes and all of the phases, and the results are given in Table 2. PBDEs. The correlations between the atmospheric total PBDE concentrations (sum of 11 congeners) and their tree bark concentrations were all significant (r2 > 0.804, P < 0.039), indicating that tree bark responds equally well regardless of the atmospheric phase in which the PBDEs are presented to the tree. The correlation between precipitation and tree bark concentrations is somewhat less than for the atmospheric phases (r2 ) 0.787, P ) 0.045), but it is still significant, indicating the possibility that tree bark is accumulating PBDEs from precipitation. BDE-47, 99, 100, and 209 are the main PBDE congeners in the tree bark samples and comprise around 70-80% of the ΣPBDE (see Figure S2). These congeners are also the dominant congeners in other environmental compartments, such as Great Lakes sediment (21) and atmosphere (22). As indicated in Table 2, there are strong positive correlations between BDE-47, 99, 100, and 209 concentrations in atmospheric vapor phase and in the tree bark (r2 > 0.821, P < 6198
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
FIGURE 1. (A) Total PBDE concentrations in tree bark (in ng/g bark) at the five U.S. IADN sites; the standard errors are shown. (B) Total PBDE concentrations in the atmospheric vapor phase (in pg/m3); the thin black lines represent the median, and the thick red lines represent the mean; the boxes represent the 25th and 75th percentiles; the whiskers represent the 5th and 95th percentiles. (C) Correlation between total PBDE concentrations in air (vapor phase only) and those in tree bark. Site abbreviations: EH Eagle Harbor, SB Sleeping Bear Dunes, SP Sturgeon Point, CL Cleveland, CH Chicago. 0.034). The relationships are weaker for the particle phase, with BDE-99 and 209 showing no significant relationship, but the relationships between the vapor plus particle concentrations and tree bark concentrations are still strong (r2 > 0.806, P < 0.039). The relationship is very weak for these individual PBDE congener concentrations in the precipitation vs tree bark, showing a significant correlation only for BDE209. This latter observation weakens the argument that tree bark is a good passive sampler of PBDEs in precipitation. DP and DBDPE. Dechlorane Plus (DP) is a highly chlorinated flame retardant marketed as a replacement for Dechlorane (aka Mirex), a flame retardant that had been widely used until the 1970s when it was banned because of its toxicity to marine invertebrates (23). In this study, Dechlorane Plus was found in all of the atmospheric and tree bark samples, including the ones from the remote sites. Like PBDEs, the lowest DP mean concentrations (see Table 1) were observed at Sleeping Bear Dunes and Eagle Harbor. However, unlike PBDEs, the highest mean tree bark concentrations for Dechlorane Plus were measured at the rural
TABLE 2. Correlation Coefficients (r2) and Their Probabilities (P) Relating Tree Bark and Air and Precipitation Concentrations for Total PBDEs, BDE-47, BDE-99, BDE-100, BDE-209, DP, Total PCBs, Total DDTs, Total Chlordanes, and All of These Compounds Taken Together (See Figure 1 for an Example of How These Values Were Calculated) vapor
a
total PBDE BDE-47 BDE-99 BDE-100 BDE-209 DP total PCB total DDTe chlordanesf all compd.
vapor + particle
particle
N
r
2
P
N
r
5 5 5 5 5 5 5 5 5 25
0.917 0.844 0.830 0.821 0.883 0.712 0.923 0.790 0.827 0.597
0.010 0.027 0.031 0.034 0.018 NS 0.009 0.044 0.032
