Environ. Sci. Technol. 2008, 42, 6431–6436
Emerging Contaminants in Car Interiors: Evaluating the Impact of Airborne PBDEs and PBDD/Fs MANOLIS MANDALAKIS,† E U R I P I D E S G . S T E P H A N O U , * ,† YUICHI HORII,‡ AND KURUNTHACHALAM KANNAN‡ Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, GR-71003, Heraklion, Greece, and Wadsworth Center, New York State Department of Health and Department of Environmental Health Sciences, School of Public Health, State University of New York at Albany, Albany, New York 12201-0509
Received December 7, 2007. Revised manuscript received February 14, 2008. Accepted February 20, 2008.
Air samples from automobile cabins were collected and analyzed for polybrominated diphenyl ethers (PBDEs), polybrominated dibenzofurans (PBDFs), and polybrominated dibenzo-p-dioxins (PBDDs). The concentration of total PBDEs (ΣPBDE; sum of 19 congeners) varied from 0.4 to 2644 pg m-3, with a median of 201 pg m-3, while BDE 47, 99, and 209 collectively accounted for 70 ( 30% of ΣPBDE concentrations. Multiple linear regression analysis revealed that ΣPBDE concentration was significantly influenced by vehicle’s age and interior temperature. More specifically, ΣPBDE decreased over time and increased with the rise of temperature. The daily inhalation intake of PBDEs during commuting was estimated to range from 0.5 to 2909 pg day-1 (median 221 pg day-1) and contributed 29% of the overall daily exposure to PBDEs via inhalation. When excluding BDE 209, a lower contribution was calculated for this source (18%), but this was still comparable with residential exposure (22%). The levels of PBDD/Fs were generally below the limits of detection and only in one case were hepta-BDFs positively detected at a concentration of 61 pg m-3. This study demonstrates that car interiors, especially when new, contain high levels of airborne PBDEs and represent a potential route of human exposure via inhalation.
Introduction Humans are continuously exposed to a plethora of toxic agents resulting from a variety of sources, such as vehicle exhaust emissions, industrial discharges, dietary sources, and cigarette smoke (1). Over the last decades, indoor air quality has emerged as a worldwide public health issue. This concern was raised because most people spend approximately 90% of their time indoors (2) and therefore are exposed to indoor gases and particles. In addition, the levels of various hazardous pollutants are usually greater in indoor than outdoor air (3). For these reasons, the U.S. EPA ranked indoor air pollution among the top five environmental risks to public health (4). * Corresponding author e-mail:
[email protected]; phone: +30-2810-545028; fax: +30-2810-545078. † University of Crete. ‡ Wadsworth Center, New York State Department of Health and Department of Environmental Health Sciences. 10.1021/es7030533 CCC: $40.75
Published on Web 03/26/2008
2008 American Chemical Society
Next to homes and offices, people in modern societies spend a considerable amount of time (about 5.5%) in automobiles (2) while commuting from home to work and during shopping, recreation, or travel activities. The passenger compartment of vehicles is recognized as an important indoor microenvironment where people are exposed to a variety of harmful substances, such as volatile organic compounds (VOCs) (5–7) and phthalic acid esters (phthalates) (8), emitted from construction materials and finishes (e.g., plastics, wood, leather, textiles, glues, sealants). Recently, high levels of polybrominated diphenyl ethers (PBDEs) were measured in car interiors (8, 9), and a concern was raised about the exposure of passengers and drivers to these compounds. PBDEs constitute an important class of brominated flame retardants commonly added to a wide variety of consumer products. Their production started in the 1960s and the global sales reached 70 000 t by 2001 (10). The market demand for PBDEs has been dominated by three major industrial formulations (penta-, octa-, and deca-BDE mixtures), which are mixtures of tetra- to deca-BDE congeners at various proportions. PBDEs have been widely applied to automobile fabrics, electronic enclosures, armrests, wire insulation, floor coverings, and other plastic parts (8). Due to the frequent exposure of cars to direct sunlight, excessive temperatures are usually reached in the cabins, especially during the summer months, which in turn can cause volatilization of chemical substances from the interior surfaces. Air temperature extremes of 89 °C and dashboard temperatures of up to 120 °C have been reported in a previous study (11). Moreover, a strong temperature-dependent increase in the concentrations of VOCs in car interiors has been reported (5–7). In conjunction with high temperatures, the transmission of solar radiation through glass windows can induce photochemical reactions and the production of degradation byproducts (7), which may be more harmful than their precursors. It is interesting to note that the labile decabrominated diphenyl ether, which is the main component of the widely produced deca-BDE formulation, can be readily photodegraded to more toxic tetra- and pentabrominated diphenyl ethers and brominated dibenzofurans (PBDFs) (12, 13). Information on the occurrence of brominated flame retardants in automobile interiors is sparse. PBDEs were recently reported in dust and windshield film samples collected from privately owned vehicles (8). However, that study was restricted to the analysis of a limited number of composite samples taken from different car models of various manufacturers. To the best of our knowledge, air concentrations of PBDEs have only been measured in car trunks by the use of passive samplers (9). Nonetheless, the levels of PBDEs in car trunks may be substantially different from those experienced by drivers and passengers. In addition, passive sampling devices may not collect suspended particulate matter (14) and therefore the previous data (9) would represent the gas-phase concentrations alone. Furthermore, factors controlling the levels of airborne PBDEs in car interiors have not been investigated to date. The present study is aimed at assessing the levels of airborne PBDEs in automobile interiors and to investigate the effect of interior temperature and car age on PBDE levels. In addition, the levels of airborne PBDFs and polybrominated dibenzo-p-dioxins (PBDDs) were determined for the first time. To achieve these objectives, a considerable number of air samples were collected on the passenger’s side of various cars operating under normal driving conditions. The results were also used to ascertain whether inhalation of air while VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6431
commuting by private vehicles contributes significantly to the overall exposure of humans to airborne PBDEs.
Materials and Methods Air Sampling. Sampling of automobile interiors was conducted using a custom-made portable low-volume air sampler. This consisted of a sorbent cartridge filled with a polyurethane foam (PUF) plug (length 4 cm, diameter 6 cm, density 0.032 g cm-3) and a 12-V battery-powered pump equipped with an air volume-measuring device. The PUF cartridge was deployed at the breathing zone in the middle of the automobile cabins, while the battery and the pump were placed in car trunks. A small piece of silicone tubing was used to connect the PUF holder with the pump inlet. The pump was operated at a flow rate of 4.5 L min-1 and the sampling was performed continuously for 48 h. Approximately 13 m3 of air was collected at each deployment. The sampler did not include a filter media at the front of the cartridge because this would cause a decrease in the sampling rate. With this configuration, the PUF sorbent collected gaseous compounds and a significant fraction of the suspended particles (especially those of large size). Particle collection characteristics of polyurethane foam have been previously investigated for plugs of the same density and of proportional dimensions with those used in the present study (15). This filtration media was demonstrated to efficiently collect particles larger than 4 µm, while particles of lower diameter exhibited substantial penetration (higher than 50%) (15). Previous breakthrough tests conducted in our laboratory showed that the size of the PUF plug (110 cm3) was large enough to enable efficient trapping of gaseous PBDEs (>90%) even after 240 m3 of air sampling at 30 °C (Table S1, Supporting Information). Forty-one air samples were collected between February 2006 and July 2007 from thirty-three cars owned by residents of the city of Heraklion, Greece. The investigated vehicles corresponded to 29 different models produced by 15 different automobile manufacturers in Europe, Japan, and the United States. All vehicles were purchased between 2000 and 2007 and their ages varied from 1 month to 5.6 years (average age 1.5 ( 1.4 years) (Table S2, Supporting Information). During sampling, the automobiles were not immobilized and the owners were allowed to operate them under normal city driving conditions. Detailed information regarding the frequency of opening of windows and the use of airconditioning or ventilation systems was not available from the drivers. In a recent study, personal samplers (operated at 2 L min-1) were used to measure and compare the concentrations of PBDEs in personal air and residential indoor air (16). To ensure that the measurements would be representative of personal exposure, Allen et al. (16) conducted air sampling only when residents were present. However, this sampling approach could not be followed in our case. Due to the low flow rate of commercial personal samplers (2 L min-1) and the relatively short time drivers spend in cars (∼ 90 min per day), air sampling for over 55 days would be required to collect 10 m3 of air. For this reason, air sampling in vehicles continued even when the drivers were absent. Considering that most of the time the vehicles were parked, our measurements are more likely to represent the levels of PBDEs in vehicles with closed windows. However, it should be pointed out that driver’s window might be open during commutes (or air ventilation might be occasionally used) and therefore the levels of airborne PBDEs that are actually experienced by the drivers could be in some cases lower than those measured by our 48-h air samples. In these terms, the results of the present study are considered to represent the upper limit of the levels of PBDEs to which drivers may be exposed. 6432
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
A temperature sensor was placed near the PUF holder to monitor temperature inside the car cabin. During sampling, the average interior temperature ranged between 16 and 45 °C. To investigate the temporal changes and temperature dependency of PBDE concentrations, two new cars (cars 26 and 27; Table S2, Supporting Information) were sampled for a period of about 1.5 years. The first deployment was conducted within three months after the delivery of the new car and additional four samples were sequentially collected at regular time intervals (approximately every 4 months). Before sampling, PUF plugs were washed with a water/ detergent solution, rinsed with tap water, Milli-Q water, and acetone, and then extracted with dichloromethane/hexane (1:1) using an accelerated solvent extraction system (ASE 300, Dionex, Sunnyvale, CA). The extracted plugs were dried in a vacuum desiccator and finally sealed in clean, solventrinsed amber glass jars. After each sampling, the PUF plugs were collected, resealed in the glass jars, and stored at -18 °C until analysis. Sample Analysis. Procedures for the extraction and purification of PBDEs are described in detail elsewhere (17) and are briefly described below. Each PUF plug was spiked with 13C-labeled internal standards (13C-BDE 15, 28, 47, 99, 153, 154, 183, and 209; AccuStandard Inc., New Haven, CT) and extracted with dichloromethane/hexane (1:1) using the ASE system. The extract was reduced in volume and treated twice with concentrated H2SO4. The organic layer was separated from H2SO4, evaporated to 0.5 mL, and applied to a glass column packed with 40% acidic-silica gel at the top and silica gel at the bottom (∼0.5 g of silica gel in total). The sample was eluted with 8 mL of dichloromethane and the volume was reduced to 0.5 mL. The solvent was exchanged to hexane and the sample was applied to a glass column packed with 1 g of silica gel activated at 300 °C for 3 h. The column was eluted with 6 mL of hexane, 4 mL of 20% dichloromethane in hexane, and 4 mL of 50% dichloromethane in hexane. The second fraction, which contained PBDEs, was reduced in volume and transferred into a glass vial. This was further evaporated to approximately 5 µL under a gentle stream of N2, and 20 µL of a recovery standard solution (13C-labeled BDE 139) was added. The analysis of PBDEs was conducted on an Agilent 6890 gas chromatograph equipped with a cool on-column injector and interfaced with an Agilent 5973 mass spectrometer operating in electron impact ionization and selected ion monitoring (SIM) mode (17). A total of 19 di- to deca-BDE congeners (BDE 15, 32, 36, 17, 25, 39, 28, 35 + 20, 62, 49, 47, 66, 100, 99, 154, 153, 183, and 209) corresponding to 18 chromatographic peaks were regularly detected in air samples and quantified using the internal standard method. After quantification, the fraction of PBDEs was pooled with the third fraction obtained from the silica gel column, evaporated, and analyzed for PBDDs and PBDFs. Previous elution experiments conducted in our laboratory indicated the collection of PBDDs and PBDFs in these two fractions. The analysis of PBDD/Fs was conducted using a Thermo Finnigan Trace GC Ultra gas chromatograph/MAT95XP high-resolution mass spectrometer (HRGC-HRMS). Measurements were carried out at a resolution of R > 9000–10 000 (10% valley). Ion source temperature was kept at 280 °C. A DB-5MSITD capillary column coated with 5% phenyl methyl polysiloxane (30 m length × 0.25 mm i.d., 0.25 µm film thickness; J&W Scientific, Folsom, CA) was used for the separation of the analytes. Helium was the carrier gas, at a flow rate of 1 mL min-1. The GC was equipped with an autosampler, and injections were performed in the splitless mode at a temperature of 270 °C. The temperature of the oven was programmed from 100 °C (2 min hold) to 160 °C at a rate of 15 °C min-1, then ramped at a rate of 10 °C min-1 to 295 °C (6 min hold), and then increased at 9 °C min-1 to 300 °C, with
a final hold time of 9 min. The mass spectrometer was operated in electron impact (35 eV energy and 500 mA ion current). Ions were monitored at the two most intensive ions of the molecular ion cluster (18). Although our study was primarily focused on providing qualitative information about the presence or absence of airborne PBDD/Fs in vehicle interiors, a semiquantitative estimation of these compounds was made by using external PBDD/F standards (EDF-5059; Cambridge Isotope Laboratories, Andover, MA). In total, 34 out of 41 air samples, exhibiting the highest levels of PBDEs, were selected for the analysis of PBDD/Fs. Quality Control/Quality Assurance. To prevent any possible photolysis of PBDEs, special care was taken to avoid the exposure of samples to light during storage and analysis. All the sample preparation and treatments were conducted under reduced light conditions and all glassware was wrapped in aluminum foil. Method recoveries were assessed by spiking four precleaned PUF disks with a mixture of PBDE congeners and analyzing them in the same way as samples. The average recovery was 60% for 13C-labeled BDE 209 and higher than 90% for the rest of the congeners. The recoveries of the surrogate standards spiked into samples were also calculated. The average recoveries of 13C- labeled BDE 15, 28, 47, 99, 153, 154, and 183 were 101 ( 13%, 104 ( 15%, 106 ( 13%, 104 ( 14%, 95 ( 12%, 96 ( 9%, and 82 ( 12%, respectively. 13Clabeled BDE 209 was poorly recovered (less than 20%) in 7 of the 41 samples and the quantification of the native congener was not possible. For the rest of the samples, the average recovery of 13C-labeled BDE 209 was 70 ( 27%. Reported concentrations of individual BDE congeners have been corrected for the percent recoveries of the internal standards. Three laboratory blanks and one field blank were analyzed to assess method detection limits for PBDEs and PBDD/Fs. The levels of individual BDE congeners in blank samples were very low and in several cases they were lower than the detection limit of the instrument. In general, BDE 47, 99, and 209, which were found at trace levels (7-84 pg per congener), collectively accounted for 72-82% of the total PBDEs (ΣPBDE; sum of 19 congeners) in blanks. Comparison of total PBDEs in field (138 pg) and laboratory blanks (97-180 pg) showed small differences and the same was also evident for individual BDE congeners. Method detection limits (MDLs) of the target compounds were calculated as the mean of the four blanks plus three times the standard deviation. MDL values for BDE 209, 47, and 99 ranged from 42 to 106 pg per congener. For those congeners not detected in blanks, MDLs were derived from the corresponding instrumental detection limit that ranged between 0.3 and 26 pg. By summing up the MDLs of all BDE congeners, the estimated method detection limit for ΣPBDE was 320 pg and the corresponding value on a concentration basis was approximately 25 pg m-3 (considering an average sampling volume of 13 m3). All sample concentrations of PBDEs were blank corrected. The detection limits of PBDFs were high due to the potential interference of some 13C-labeled PBDE standards spiked into the samples. For PBDFs, a detection limit of 30, 150, and 300 pg was estimated for the sum of tetra-, penta-, and hexa-BDF congeners, respectively (i.e., 2.3, 11, and 23 pg m-3, respectively), while a higher value of 1500 pg was assigned for the sum of hepta- and octa-BDF congeners (i.e., 115 pg m-3). Similar values were also applied for the tetra- to octa-BDD homologue groups. Statistical Analysis. All statistical analyses were performed with STATISTICA 6.0 software package (StatSoft, Tulsa, OK). Prior to calculating descriptive statistics (mean, standard deviation, median, and geometric mean), nondetectable concentrations of individual BDE congeners were replaced with half the method detection limit. Log-transformed concentrations of PBDEs were tested for normal distribution
TABLE 1. Summary Statistics for the Air Concentrations of Individual BDE Congeners (pg m-3) and ΣPBDE Measured in Vehicle Cabins (N = 41) congener
range
mean
BDE 15 nd-59 3.1 BDE 32 nd-7.2 0.7 BDE 36 nd-4.3 0.5 BDE 17 0.1-27 3.0 BDE 25 nd-17 1.4 BDE 39 nd-21 1.7 BDE 28 0.2-143 12 BDE 35 + 20 nd-3.6 0.4 BDE 62 nd-96 3.4 BDE 49 nd-53 5.9 BDE 47 nd-625 91 BDE 66 nd-31 2.7 BDE 100 nd-113 17 BDE 99 nd-567 79 BDE 154 nd-48 7 BDE 153 nd-68 11 BDE 183 nd-155 11 BDE 209 nd-1053 150 ΣPBDE 0.4-2644 396
SDa
median
GMb
%DFc
9.2 1.4 1.0 6.1 3.7 4.9 30 0.8 15 11 139 6.0 27 132 11 17 32 229 574
0.7 0.1 0.1 0.8 0.2 0.1 2.1 0.1 0.6 2.3 44 1.0 5 10 2.4 2 1.7 104 201
1.3 0.2 0.1 1.0 0.2 0.2 2.3 0.1 0.4 1.8 29 0.8 4.9 16 1.3 1.9 2.4 44 114
41 83 71 100 88 88 100 71 71 83 76 76 59 73 51 63 51 73 100
a SD denotes standard deviation. b GM denotes geometric mean. c %DF denotes % detection frequency.
using the D’Agostino-Pearson omnibus normality test. Data derived from long-term measurements in two cars were combined and statistically analyzed to evaluate the effect of vehicle age and temperature on the concentrations of airborne PBDEs. The relationship between logarithmtransformed concentration and the specific variables was evaluated by forward stepwise multiple linear regression analysis using the following equation: lnC ) aA + bT + I
(1) -3), A is the vehicle
where C is the interior concentration (pg m age (months), T is the interior temperature (°C), I is the intercept term, and a, b are the regression coefficients for vehicle age and interior temperature, respectively. Regression coefficients deduced from MLR analysis were considered to be statistically significant if p e 0.05.
Results and Discussion PBDE Concentrations in Vehicle Cabin Air. Information regarding individual samples and the concentrations of airborne PBDEs measured in automobile interiors is provided in Table S2 (Supporting Information). Descriptive statistics for the levels of individual BDE congeners are summarized in Table 1. The concentrations of most of the individual congeners were log-normally distributed (D’Agostino-Pearson normality test, p > 0.05). Therefore, the arithmetic means were substantially different from the median and geometric mean. Due to the log-normal distribution, the medians and geometric means provide better estimates of the central tendency than arithmetic means. PBDEs were detected in all of the air samples analyzed and the median concentrations of the individual congeners ranged between 0.1 and 104 pg m-3. In most samples, BDE 47, 99, and 209 were the most abundant congeners. These three congeners were detectable in 31 of 41 samples and, on average, their sum collectively accounted for 70 ( 30% of the total PBDE concentrations. The high levels of BDE 209 in the air samples analyzed are indicative of the emissions from materials containing the deca-BDE formulation. The relative proportions of specific BDE congeners were also examined to evaluate the presence of other commercial mixtures. In 16 cars, the concentration ratio of BDE 47 to BDE 99 varied between 0.7 and 1.2, closely resembling the ratios previously reported for VOL. 42, NO. 17, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6433
FIGURE 1. Time-course of (a) BDE 47 to BDE 99 ratio and (b) ΣPBDE concentration (pg m-3) in the air of two vehicle cabins. Dashed line in the upper panel represents the ratio in commercial penta-BDE formulations (DE-71 and Bromkal 70-5DE) (19). DE-71 and Bromkal 70-5DE penta-mix formulations (0.6–1.0) (9, 19, 20). However, much higher BDE 47/BDE 99 ratios, ranging from 1.5 to 12.6, were observed for 12 cars. A more detailed examination of the temporal measurements conducted for 2 cars indicated that BDE 47/BDE 99 ratio can be strongly influenced by the age of the vehicles. In both cases, BDE 47/BDE 99 was 0.66 during the first 3 months, clearly indicating the presence of penta-BDE mixtures in car interiors, but it increased to 4.7 (car 27) and 12.6 (car 26) after 1.5 years (Figure 1a). This observation could be attributed to the progressive photodecomposition of the highly brominated congeners, including BDE 209 (12, 13), to less brominated compounds. In a previous study, Bezares-Cruz et al. (21) demonstrated the formation of BDE 47 and BDE 99 from the photolysis BDE 209, and it was pointed out that BDE 47 would be the end product of this process. The air levels of ΣPBDE varied over 4 orders of magnitude, from 0.4 to 2644 pg m-3, with a median of 201 pg m-3 and a geometric mean of 114 pg m-3.The arithmetic mean of ΣPBDE was 396 pg m-3, but this was heavily influenced by the contribution of 5 cars in which the total concentration was higher than 1380 pg m-3. To the best of our knowledge, the only other study available for comparison is one that measured gaseous PBDEs in car trunks using passive samplers (9). Air concentrations of total PBDEs reported in the earlier investigation (9) ranged from 11 to 8184 pg m-3, with an arithmetic mean of 709 pg m-3, which was also heavily influenced by three extreme values. The respective median was 41 pg m-3, approximately five times lower than our value. However, this comparison is limited by differences in the sampling methodologies and sampling conditions (differences in vehicle age, temperature, ventilation, etc.). In a more recent study, high levels of atmospheric PBDEs (140 to 2330 pg m-3 of total PBDEs in particle and vapor phase) were measured in the outdoor air near an automotive shredding/ metal recycling facility (22). This observation further sub6434
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 17, 2008
stantiates the extensive use of PBDE-containing materials in vehicles and their interiors. In a previous investigation, high PBDE levels were also observed in dust and on interior surfaces of various automobiles in the United States (8) and, based on these results, a ranking of different companies was conducted. A comparison of PBDE results for various car models was not attempted in our study because the measurements in various vehicles were not performed under standardized and controlled conditions (e.g., vehicle age, interior temperature, ventilation rate). In general, PBDE concentrations varied widely within a manufacturer/model (Table S2, Supporting Information). For instance, ΣPBDE concentrations measured in three Peugeot cars (cars 20, 21, and 22) varied from 9.4 to 564 pg m-3, and in three Ford cars (cars 4, 5, and 6) varied from 126 to 2644 pg m-3. Although a part of the large intramodel variability may arise from differences in modelspecific parameters (e.g., cabin volume, type and total area of interior plastic surfaces), other differences in vehicle conditions during sampling (temperature, ventilation, or use of air-conditioning) may also contribute to the observed differences in ΣPBDE concentrations. PBDD/Fs Concentrations in Vehicle Cabin Air. PBDFs are byproducts of photodecomposition of PBDEs (13). In addition, Hanari et al. indicated the presence of PBDFs as impurities in several commercial PBDE formulations (18). Despite these facts, the levels of airborne PBDFs measured in vehicles were generally below the limits of detection. Similarly, no tetra- to octa-brominated dibenzo-p-dioxins were detected in any of the samples analyzed. An explanation for the lack of detection of PBDFs was the high method detection limit, due to the interference of some 13C-labeled PBDEs spiked into the samples. Therefore, further studies should separate PBDEs from PBDFs for the analysis of the latter. In one sample (car 26), hepta-BDFs were positively detected. Based on a semiquantitative estimation using the external standard calibration method, the concentration of total hepta-BDFs was 61 pg m-3. Overall, our results suggest that PBDFs and PBDDs are not important constituents of air in car interiors, but further investigations using more sensitive analytical methodologies and sampling techniques will be required to quantify these contaminants at trace levels. Influence of Vehicle Age and Interior Temperature on Air Concentrations of PBDEs. Several studies have attempted to elucidate the factors affecting the levels of airborne contaminants in indoor environments. A recent investigation demonstrated that formaldehyde levels in residential indoor air were dependent on the age of the house, type of construction, ventilation rate, and indoor temperature (23). Similar studies conducted in vehicles indicated that the concentrations of VOCs declined over time, but increased with a rise in interior temperature (5–7). However, the factors influencing the levels of airborne PBDEs in vehicles (and other indoor microenvironments) remain poorly understood. The effect of age on ΣPBDE concentrations was initially investigated based on ΣPBDE concentrations found in all of the vehicles studied. For this purpose, the data were divided into four categories based on the age of vehicle at the time of sampling (0-6, 7-12, 13-24, and g25 months). ΣPBDE exhibited a gradual decrease from newer (up to 6 months old; 760 ( 740 pg m-3) to older vehicles (older than g25 months; 130 ( 180 pg m-3) (Figure S1, Supporting Information) but this was not statistically significant due to the large variability of concentrations within each group. Similarly, data were divided into two equivalent subsets based on interior temperatures (15-25 °C and 26-45 °C). The average ΣPBDE concentration was elevated at higher temperatures (510 ( 630 pg m-3) than at lower temperatures (230 ( 450 pg m-3), but this difference was also not statistically significant.
To better evaluate the effect of age and interior temperature on PBDE concentrations, the results from the sequential measurements in two cars (cars 26 and 27) were examined. In both vehicles, ΣPBDE concentration was the highest during the first 3 months after their delivery and decreased exponentially over a 1-year period (Figure 1b). This observation is most likely related to the decreasing emissions of PBDEs from the interior surfaces with age and the gradual dilution of their concentrations with outdoor air (i.e., due to prolonged ventilation). At the end of these measurements, a moderate increase of PBDE concentrations was observed, possibly due to the increase of temperature during the collection of the last samples. The log-transformed ΣPBDE concentrations from the two cars were combined and further examined by multiple linear regression (MLR) analysis (eq 1). Logarithmic concentration of ΣPBDE provided statistically significant correlations with both vehicle age (p < 0.01) and temperature (p ) 0.04). The regression coefficient for temperature was positive indicating that an increase of this variable caused an increase of PBDE levels. In contrast, the regression coefficient associated with vehicle age was negative and described the net reduction of PBDEs over time. Based on this value, a half-life of 2.5 months was estimated for the clearance of airborne ΣPBDE from private vehicles. Considering that various BDE congeners may behave differently due to differences in their physicochemical properties, and that the latter are highly dependent on the degree of bromination, the MLR analysis was also conducted for the three main homologue groups of PBDEs (tetra-, penta-, and deca-BDEs). The total concentration of each homologue group was significantly correlated with vehicle age (p < 0.05), while temperature was proved to be a significant predictor only for tetra- (p ) 0.05) and penta-BDEs (p < 0.01). Temperature did not exert a significant influence on the concentration of deca-BDE (BDE 209), and stepwise forward MLR analysis led to the elimination of this variable. Based on the regression coefficients associated with vehicle age, half-lives of 3.3, 1.5, and 5.3 months were estimated for tetra-, penta-, and deca-BDEs, respectively. Implications for Human Exposure via Inhalation. The concentrations of PBDEs measured in vehicle cabins and values reported for other indoor environments (home, workplaces) and outdoor air were used to estimate daily intake of PBDEs via inhalation and to evaluate the magnitude of PBDE exposure while commuting in cars. Daily exposure was calculated using the following algorithm:
∑ exposure ) [(C
WFW) + (CHFH) + (COFO) + (CVFV)]RR
(2)
where Σexposure is the daily adult human exposure via inhalation (pg day-1), RR is the mean respiration rate of adults (m3 day-1), CW, CH, CO, and CV are the ΣPBDE concentrations (pg m-3) in workplace, home, outdoors, and in vehicles, respectively, while Fw, FH, FO, and FC are the respective percent of time spent in each one of these environments. A respiration rate of 20 m3 day-1 (9) was applied, while estimates of adult time-activity patterns previously published by Klepeis et al. (2) were used for these calculations. According to that study, the average percentage of time spent in homes, workplaces (and other indoor locations), cars, and outdoors is 68.7, 18.2, 5.5 and 7.6%, respectively. The median ΣPBDE concentrations previously measured by our research group in homes, workplaces, and outdoor air in Greece (17) were applied in the present calculations. Table S3 (Supporting Information) summarizes ΣPBDE concentrations reported for occupational settings (computer/ electronics shops, computer rooms, Internet cafes, public offices, and furniture stores), houses, and outdoor air in Greece, and the corresponding estimates of exposure via
TABLE 2. Results of Stepwise Forward Multiple Linear Regression Analysis for the Effect of Vehicle Age and Temperature on the Logarithmic Concentration of ΣPBDE and of Three Different Homologue Groups of PBDEs age homologue group intercept tetraBDEs pentaBDEs decaBDEs ΣPBDE
regression coefficienta
temperature regression p-level coefficienta p-level
3.61
-0.21 ( 0.07
0.02 0.10 ( 0.04
3.48
-0.47 ( 0.05
