Indoor and Outdoor Concentrations of Polybrominated Diphenyl Ethers on Respirable Particulate in Central and Southern Taiwan

High levels of fine particulate (PM2.5) in indoor and outdoor air have globally threatened human health and environment. There are still few studies which concern on the emerging persistent organic pollutants like polybrominated diphenyl ethers (PBDEs) bound on PM2.5. The aim of this study was to investigate PBDEs in PM2.5 in various outdoor (metropolis, industrial, and rural areas) and indoor (library, rail station, hospital, supermarket, department store, and office) environments. PM2.5-bound PBDEs was analyzed by high resolution gas chromatography/high resolution mass spectrometry after PM2.5 was collected. Mean levels of PM2.5-bound Σ14PBDEs were 79.0 and 116 pg m in outdoor and indoor air, respectively. Compared to other outdoor locations, the industrial sites, Taixi (169 pg m) in particular, has the highest PM2.5-bound PBDEs levels which might be attributed to nearby industrial activities and indoor to outdoor migration behaviors. For indoor air, PM2.5-bound PBDEs mean concentrations (libraries, rail stations, department stores, offices, hospitals, and supermarkets) were found to be 357, 35.3, 50.2, 73.2, 59.2, and 124 pg m, respectively. The high indoor PM2.5-bound PBDEs levels found in libraries are heavily affected by the presence of indoor electronic equipment or other consumer products. Similarly, this is also true for supermarkets which merchandise electronic consumer products. Although the abundant congener of deca-BDE consisted of 74.7% and 48.03% of Σ14PBDEs in the indoor and outdoor air, respectively, nona-BDEs predominantly contributed 11.6% in the indoors while triand tert-BDES contributed 11.3% and 16% in the outdoors. Higher brominated PBDEs are more likely due to their emission from electronic surfaces while lower brominated PBDEs are products of photochemical degradation. Other factors affecting both the indoor and outdoor air PM2.5-bound PBDE homologue levels such as migration behaviors might also be considered.


INTRODUCTION
Polybrominated diphenyl ethers (PBDEs) are a class of emerging persistent organic pollutants which ubiquitously exist in both the aerosol indoors and outdoors.They are also a common ingredient in brominated flame retardants (BFRs) which are widely used on a variety of products such as electronic appliances, textiles, paints, etc. (Wang et + These authors are equally contributed. al., 2007;Qi et al., 2014).PBDEs have been regarded as a major concern due to their ubiquitous existence in the environment (e.g., form the source to the surrounding air) and their harmful health effects.They are also highly resistant to biological, chemical, and physical degradation, and tend to bio accumulate easily due to their lipophilic property (Cincinelli et al., 2014;Tung et al., 2014).PBDEs negatively affect the endocrine and reproductive functions of humans most especially in females (Hood, 2006;Shy et al., 2012) and cause neurodevelopmental delays after neonatal ingestion of breast milk with high PBDE level (Lin et al., 2010;Chao et al., 2011;Shy et al., 2011;Gascon et al., 2012;Shy et al., 2012).The use of commercial PBDE technical formulations such as penta-BDE, octa-BDE, and deca-BDE have been restricted and voluntarily discontinued in several parts of the world including Europe and in some locations in USA (Besis and Samara, 2012).However, the use of deca-BDE in Taiwan has not yet ceased as of present and remains available in the manufacture of its many consumer products (Chao et al., 2014).
Due to fine particle deposition, PBDEs bind on particulate matters with aerodynamic diameter less than 2.5 µm, PM 2.5 , respectively (Mandalakis et al., 2009;Li et al., 2015a).Numerous studies have been conducted to investigate PM 2.5 levels due to the possibility of its relatively small particles being deposited more deeply into the lungs causing respiratory ailments and diseases, and other negative health problems (Zhang et al., 2013;Guo et al., 2015a;Li et al., 2015b;Shi et al., 2015;Gao et al., 2016).People with known history of cardiopulmonary problems face a higher risk of mortality due to the increasing PM 2.5 concentration in the atmosphere and so is the prevalence of respiratory problems in normal people (Dominici et al., 2006;de Oliveira et al., 2012;Xing et al., 2016).
Particle-bound PBDEs are difficult to eliminate in the atmosphere and tend to have a longer residence time and are easily transported even at a longer distance (Xu et al., 2016).PBDEs easily diffuse from the surface of BFRcoated products present in the indoor environment and unto PM 2.5 particles via volatilization or abrasion (Rauert and Harrad, 2015).The Σ 15 PBDEs mean concentrations in indoor PM 2.5 in the office environment in Shanghai, China was found to be 51.8 pg m -3 (Xu et al., 2016).In five Hong Kong kindergartens, PBDEs concentrations in indoor PM 2.5 are 0.10-0.64ng m -3 , respectively (Deng et al., 2016).PBDEs can exist in indoor PM 2.5 and outdoor PM 2.5 with its transport relative to both.The presence of PM 2.5 in the indoor environment is mainly attributed to both outdoor and indoor sources, through ventilation systems and infiltration (Weschler, 2004;Meng et al., 2005;Lim et al., 2011).In general, indoor air PBDE concentrations have been found to be much higher compared to the outdoor air PBDE concentrations due to the presence of home electronic devices and a higher risk of exposure due to majority of the people preferring the indoor environment (Wilford et al., 2004;Hazrati and Harrad, 2006).Indoor occupational exposure of PBDE is also prominent in human adults especially those working in the e-waste recycling and dismantling factories (Julander et al., 2005;Ma et al., 2009;Muenhor et al., 2010;Gou et al., 2016a).If only fine particles are considered, PBDE concentrations are also observed to be higher in indoor PM 2.5 compared to the outdoor PM 2.5 possibly due to several factors such as weak indoor air circulation, slow degradation, low ventilation rate from indoors to outdoors, and high pollutant emission from consumer products (Stapleton et al., 2008;Lv et al., 2015;Gou et al., 2016a;Xu et al., 2016).
Exposure routes for PBDEs include dust ingestion, dietary intake and inhalation of particulates for both adults and young children.Previous studies showed that humans, particularly the younger population, are more likely to be exposed in the indoor environment than in the outdoor environment (Chao et al., 2014;Shy et al., 2015;Gou et al., 2016a).A risk assessment study conducted by Gou et al. (2016b) showed that school-age children are more likely to be exposed to PBDEs from the indoor environment particularly in their homes compared to their indoor classrooms.The study on the effects of airborne PBDEs on human health through exposure in the indoor environment is still limited considering that the inhalation of these particulates is just a minor exposure route compared to dietary intake and indoor dust digestion as observed in most of PBDEs risk-assessment studies.PM 2.5 is known to cause adverse health effects especially on the human cardiorespiratory system but PM 2.5 -bound PBDEs health effects are still unknown and yet to be confirmed.In the present study, levels of PBDEs in PM 2.5 were investigated in various indoor and outdoor environments located in central and southern regions of Taiwan.

Chemicals and Reagents
The 14 PBDE congeners standard solution including 47,99,100,153,154,183,196,197,203,206,207,208 and 209 was purchased from Cambridge Isotope Laboratories (Andover, MA, USA).The internal standard of 8 13 C-labeled PBDEs 47,99,153,183,197,207,and 209) was from Wellington Laboratories (Guelph, Canada).Sodium sulfate, alumina oxide, potassium oxalate, and silica gel of the highest grade were obtained from Merck (Darmstadt, Germany).

Air Sampling and Collection
The study design in the present study was focused on the screened survey of PBDE concentrations on PM 2.5 particulate in outdoor air from the different types of the areas (i.e., industrial, metropolis, and rural areas) and in indoor air from various types of the microenvironments in Taiwan.The regulated air pollutants were collected in the indoor and outdoor air based on the standard methods from Environmental Protection Administration in Taiwan (TEPA).Twenty-eight outdoor air samples were gathered from seven TEPA air monitoring sites between May 5 and May 8 or between October 2 and October 5. Indoor air samples were obtained from 24 indoor environments including 3 liberies, 3 rail stations, 3 department stores, 3 offices, 6 hospitals, and 6 supermarkets from May to October in 2013.The sampling processes and analytical methods were followed by announcement of TEPA standard methods.The TEPA standard methods used in the present study were as follows: PM 2.5 (NIEA A205.11C),PM 10 (NIEA A206.10C), carbon dioxide (CO 2 ) (NIEA A448.11C), carbon monoxide (CO) (NIEA A421.12C), ozone (O 3 ) (NIEA A420.11C), total volatile organic compounds (TVOCs) (NIEA A732.10C), formaldehyde (HCHO) (NIEA A705.11C), total bacterial count (TBC) (NIEA E301.11C), and total fungal count (TFC) (NIEA E401.11C).The PM 2.5 outdoor samplers were PQ200 Ambient Air Particulate Samplers form BGI by Mesa Labs (New Jersey, USA).The PM 2.5 indoor samplers, Model 200-Personal Environmental Monitor TM (PEM TM ), were purchased from MSP Corporation (Minnesota, USA).To avoid the amount of PM 2.5 -bound PBDEs below detection limits, the 4 outdoor air samples were pooled from the continuous 4-day sampling samples and the 3 indoor air samples in the same category were pooled together.Prior to PBDEs analysis, the PM 2.5 samples were kept at -20°C after the condition and weighting of the glass fiber filters.

Extraction, Cleanup, and Analysis
The extraction, cleanup procedures, and analysis of airborne PBDEs in the present study were followed by our previous publication with minor modification (Chao et al., 2014;Shy et al., 2015).Briefly, pre-labeled isotopes and identifiable surrogate standards were spiked in the filters to evaluate the PBDEs loss during the sampling process prior to aerosol sampling indoors and outdoors.Owing to monitoring of the extraction and cleanup processes, the internal standards were spiked into the samples and well mixed with toluene before aerosol samples were extracted with toluene for 24 h in a Soxhlet extractor.After extraction, the extracts were then concentrated, treated with concentrated sulfuric acid, and passed through a multicolumn system installed with acid silica, alumina, and activated carbon columns.The eluate solution was gathered, evaporated, and concentrated to near dryness by using a gentle stream of gaseous nitrogen prior to transference into a vial.A total of 50 µL of 13 Clabeled BDE-139 was added to each eluate as an internal recovery standard after the clean-up and prior to injection to minimize the possibility of loss.The elute of PBDEs in a vial was re-dissolved by toluene.The final extract was reduced in volume to 0.2 mL under a stream of nitrogen.
PBDEs in the final extract were determined by highresolution gas chromatography with high-resolution mass spectrometry (HRGC/HRMS) (Hewlett-Packard 6970 Series gas/Micromass Autospec Ultima) using a positive electron impact (EI+) source in the selected ion monitoring (SIM) mode with a resolving power of 10,000.A DB-5HT column (L = 15 m, i.d.= 0.25 mm, film thickness = 0.1 µm) (J&W Scientific, Folsom, CA) was installed on HRGC in splitless mode at 280°C with constant helium flow of 1 mL min -1 .The temperature program of HRGC oven stably maintained at 100°C in the first 4 min, steadily increase to 200°C at a rate of 40 °C min -1 from 100 to 200°C, hold at 200°C for 3.5 min, rapidly increase to 325°C at a rate of 10 °C min -1 , and continued at 325°C for the last 2.5 min.The electron energy and source temperature were specified at 35 eV and 250°C in HRMS, respectively.The two most abundant isotope masses were measured for each component.Quantification was performed using internal/external standard mixtures via the isotope-dilution method.The US EPA Method 1614A of analytical quality assurance and quality control (QA/QC) was followed.Prior to air sampling, PUF cartridges were spiked with PBDE surrogate standards prelabeled with isotopes to obtain the recoveries of PBDEs surrogate standards within 82-121% of acceptable QA/QC limits (i.e., 70-130%).The limits of detection (LODs) and quantification (LOQs) were defined as the amount at which the signal-to-noise (S/N) ratios were higher than 3 and 10, respectively.The LODs for the 13 PBDE congeners (BDE-28 to -208) ranged from 0.288 to 49.0 pg g -1 and the LOD of BDE-209 was 314 pg g -1 .The analysis of the PBDE labeled internal, precision and recovery (PAR), and surrogate standards all met the relevant standards.Laboratory blanks were analyzed for each batch of 10-12 samples.The total amounts of PBDEs in the field and laboratory blanks were extremely low (mostly negligible) compared with those of the real samples.The isotopic ratios of at least two characteristic ions for each sample were consistent with theoretical values to within a deviation of 15%.Calibration mixtures with isotopically labeled internal standards were tested in the quantification of the target compounds.

Statistical Analysis
Measurements of airborne PBDEs below the limits of detection (LODs) were set to zero.The difference in level of airborne PBDE between indoor and outdoor samples was examined by the Mann-Whitney U test.Differences were considered to be significant when the p value was less than 0.05 or at the 95% confidence level.The Statistical Product and Service Solutions (SPSS) software, version 12.0, was used in the present study.

RESULTS AND DISCUSSION
The indoor and outdoor air quality in the present study was shown in Table 1.Median levels of PM 10 , PM 2.5 , and CO were 54.8 µg m -3 , 24.4 µg m -3 , and 0.600 ppm in the indoor air and 35.8 µg m -3 , 18.7 µg m -3 , and 0.330 ppm in the outdoor air, respectively.In the indoor air quality (IAQs) from various microenvironments including hospitals, rail stations, department stores, offices, and supermarkets, median concentrations of CO 2 , O 3 , HCHO, and TVOCs were investigated as 706, 0.0278, 0.0421, and 0.0254 ppm, respectively.Measurements of most regulated IAQ pollutants were beneath Taiwanese IAQ standards except for TVOCs in a department store and HCHO in a supermarket.For bioaerosol in the indoor environment, TBC (median: 730 CFU m -3 ) and TFC (439 CFU m -3 ) concentrations met the Taiwanese IAQs standards.Our IAQ values in the present study were comparable to our previous report (Hsu et al., 2015).In comparison with outdoor PM 10 , indoor PM 10 levels are higher.The same phenomena is also observed in PM 2.5 levels.Several studies revealed that indoor PM including coarse, fine, and submicron particulate possibly originated from outdoor sources (Chen et al., 2016;Raysoni et al., 2016;Wang et al., 2016).The presence of PM 2.5 and PM 10 in the indoor environment might be mainly attributed to the indoor equipment, occupant activities, and outdoor air including building structure, ventilation system, air exchange rate, and ambient conditions (Tippayawong et al., 2009;Madureira et al., 2012;Hassanvand et al., 2014;Chen et al., 2016;Raysoni et al., 2016;Wang et al., 2016).A close indoor and outdoor relations for PM 2.5 was found during unoccupancy and occupancy while the only weak correlation appeared to be that of PM 10 during occupancy in primary schools (Chen et al., 2016).Othman et al. (2016) reported a lower indoor PM 10 level due to low occupant activities during office hours.Aside from occupancy, the presence of many consumer products like printers in the indoor environment also affects PM 10 levels as reported by several studies (Lee et al., 2001;He et al., 2007;Kagi et al., 2007). 2 Monitoring of PM 10 , PM 2.5 , and CO in the present study was obtained from the ambient monitoring sites between May 6 and May 8 or between October 2 and October 5 in 2013.
Table 2 shows the PM 2.5 -bound Σ 14 PBDEs mean concentrations in different outdoor locations in central and southern Taiwan which were grouped under metropolis areas, industrial areas, and rural areas.The mean concentrations of PM 2.5 -bound Σ 14 PBDEs in the metropolis areas which includes Xitun, Douliu, and Tainan are 74.3,76.3, and 53.0 pg m -3 , respectively.For industrial areas like Mailiao and Taixi, the observed mean concentrations are 54.5 and 169 pg m -3 .Lastly, the rural areas which include Hengchun and Meinong, has mean concentrations of 74.3 and 51.0 pg m -3 , respectively.However, in the present study the mean levels in rural areas are relatively similar to that of the metropolis areas (urban) and the highest mean concentrations of Σ 14 PBDEs in PM 2.5 was observed to be in industrial areas particularly in Taixi.Only few studies investigated PM 2.5 -bound PBDEs in outdoor air (Table 3).Compared with PM 2.5 -bound ΣPBDEs in the previous studies (Dong et al., 2015;Liu et al., 2016), our values in Taiwanese outdoor air were comparable to outdoor air levels in the metropolis areas in China, but still 60 to 100-fold higher than the background levels of atmospheric PM 2.5 -bound PBDEs in East China Sea (Li et al., 2015).Although outdoor PM 2.5 -bound Σ 12 PBDEs in the ambient air of Valencia, Spain were presented in extremely low magnitudes, the high molecular weights of PBDEs like octa-BDEs, nona-BDEs, and deca-BDEs were not analyzed in Beser's report (Beser et al., 2014).PBDEs particle distribution in the atmosphere was found to be associated with fine particles of diameters smaller than 0.49 µm as reported by Besis et al. (2015) in traffic and urban background sampling sites located in Thessaloniki, northern Greece.In a ten-year-ago study (Deng et al., 2007), outdoor mean Σ 22 PBDEs in PM 2.5 concentrations from e-waste recycling sites (industrial) in Guiyu, Southeast China to that of the urban sites in Guangzhou and Hong Kong, South China were analysed wherein it showed that Guiyu has the highest atmospheric mean PM 2.5 -bound PBDEs levels (16.6 ng m -3 ) compared to the latter two (33.8-372pg m -3 ) and other urban sites from all over the world with an estimated exposure through inhalation of 133 and 332 ng day -1 for both adult and children.Although the surveillance of ambient air PBDEs (n = 180) were done in metropolis (Hong Kong and Guangzhou) and heavily contaminated industrial areas (Guiyu) (Deng et al., 2007), we have concerns regarding with Deng's report (Deng et al., 2007) due to two reasons as follows: (1) the HRGC/LRMS with EI mode was used to detect PBDE congeners.The resolution and sensitivity is not good enough to distinguish the neighbour peaks, and (2) high molecular weights of PBDEs from 196 to 209 were not measured in their study.Most PBDE congeners from octa-BDEs to deca-BDE are possible dominating congeners among total PBDEs in the outdoor air.
Many atmospheric studies investigated PBDEs in outdoor air.Harrad and Hunter (2006) found that PBDEs favour higher ratios in air as compared in the soil (47:99 ratio) and that its main source is the indoor environment.Rural sites in Ontario, Canada have been reported to have a mean Σ 21 PBDEs in total air concentrations of range between 6 and 85 pg m -3 which are still lower as compared in urban sites (Gouin et al., 2005).In addition, a study conducted in Izmir, Turkey on PBDE concentrations of outdoor and indoor organic films on window glasses showed that industrial sites have the highest PBDE concentrations compared to that of the offices, laboratories, and urban, suburban, and rural homes (Cetin and Odabasi, 2011).Syed et al. (2013) suggested that industrial activities played a key role in the PBDE distribution in areas located near industrial sites.Combustion process in Taiwanese industrial locations (Wang et al., 2011) as well as dismantling and dumping of e-wastes (Deng et al., 2007) are listed as reasons for PBDEs emission in industrial sites (O'Driscoll et al., 2016).
As shown in Table 4, the indoor PM 2.5 -bound Σ 14 PBDEs mean concentrations for different indoor locations which include libraries, rail stations, department stores, offices, hospitals, and supermarkets are 357, 35.3, 50.2, 73.2, 59.2, and 124 pg m -3 , respectively.The highest Σ 14 PBDE level in PM 2.5 was found in the indoor air of library.Many libraries have already been modernized and incorporated with printers and computers and many other electronic systems which may have contributed more to the indoor PBDEs level.A positive correlation (p < 0.001) was found between the number of electronic appliances and polyurethane foam Table 2. PBDE concentrations on PM 2.5 particulate in outdoor aerosol from central and southern Taiwan (pg m -3 ).
The distributions of outdoor and indoor PBDE homologues (di-BDE, tri-BDE, tetra-BDE, pentaBDE, hexa-BDE, hepta-BDE, octa-BDE, nona-BDE, and deca-BDE) are shown in Fig. 1.Deca-BDE has the highest distribution levels in both the indoor and outdoor locations at 74.7% and 48.03%. ) levels (11.6%) is significantly high in the indoor locations while tri-and tert-BDE (11.3% and 16%) in the outdoor locations.Many studies have already reported that deca-BDE (BDE-209) is predominant in both the indoor and outdoor environments (Stapleton et al., 2005;Gevao et al., 2006;Chen et al., 2008;Gou et al., 2016a).But still, BDE-209 is higher in the indoor environment as compared to that in the outdoor environment.Similarly for indoor environments, BDE-209 was found to be predominant in air in Taiwanese residential homes among all 14 PBDE congeners as reported by Shy et al. (2015).The opposite is true in Korean elementary schools where the predominating congener was BDE-47 (Lim et al., 2014).BDE-209 in the outdoor is mainly attributed to indoor BDE-209 (Gou et al., 2016a).
A high percentage of nona-BDE was also observed in a study conducted on e-waste workshops in China in which debromination of deca-BDE to nona-BDEs through exposure to natural sunlight was inferred to be the possible cause (Xu et al., 2015).However, with the lower vapor pressures of nona-and deca-BDEs, they are mostly particle-bound which makes them harder to degrade photochemically (Raff and Hites, 2007).Various equipment containing PBDEs found in the indoor environment are main sources of PBDE emission into the air via volatilization (Tung et al., 2014).The existence of these higher brominated PBDEs in the indoor air is probably attributed to their migration behaviours from outdoors to indoors.Although the indoor air is a significant source of PBDEs emission into the outdoor environment, the individual sources for each indoor and outdoor PBDEs might be different.A high percentage of tri-and tetra-BDEs are accounted in the outdoor locations.An opposite trend was observed in the study conducted by Gou et al. (2016a) wherein the tri-BDE is higher in the indoor than in the outdoor.Ding et al. (2016) reported higher BDE-28 levels in the indoor air than BDE-47 while the outdoor air showed higher BDE-47 levels compared to BDE-28.
Possibly, the presence of BDE-28 and BDE-47 in the indoor and outdoor air is attributed to photochemical degradation by natural sunlight and migration behaviours from indoors to outdoors of the congeners brought about by their low vapor pressure characteristics which makes them easily evaporate from products (Ding et al., 2016).

CONCLUSIONS
The industrial locations were found to have the highest outdoor PM 2.5 -bound PBDEs levels particularly in Taixi located at the industrial area.This is affected mainly by industrial activities such as combustion or petrochemical production around the area and also the migration of indoor PM 2.5 -bound PBDEs into the outdoor environment.Our study found the highest indoor PM 2.5 -bound PBDEs in the libraries, mainly, due to low air exchange rate, low efficiency of ventilation, the volatilization of PBDEs from the surface of electronic equipment (e.g., computer and printers) which are already common in the present day libraries.Deca-BDE and nona-BDEs predominated in the indoor locations.This is primarily attributed to their dispersion from the indoor equipment and migration behaviors from the outdoor environment.Tri-, tetra-, and deca-BDE levels were high in the outdoor areas.Photochemical degradation or debromination is also a possible source of lower brominated PBDEs in both the indoor and outdoor environments.Migration behaviors from the indoor to the outdoor air and vice versa of the PM 2.5 -bound PBDEs might affect the indoor and outdoor PM 2.5 -bound PBDEs levels.E-020-001-MY3) from the Ministry of Science and Technology (MOST), Taiwan.We thank Prof. Lin-Chi Wang and Prof. Gou-Ping Chang-Chien for the assistance in analyzing PBDEs.We also acknowledge our team members in National Pingtung University of Science and Technology (NPUST), Mr. Wen-Kai Wu, Mr. Chi-Wei Wu, and Dr. Ding-Yan Lin, and in Kun Shan University, Mr. Yao-Kai Liu and Mr. Jhy-Ching Chuang.

DISCLAIMER
The authors declare no conflicts of interest with regard to this study.

Fig. 1 .
Fig. 1.The composition of PBDEs on the fine particulate (PM 2.5 ) in the indoor and outdoor aerosol.

Table 1 .
Air quality in indoor and outdoor air.Monitoring of Indoor air quality in the present study is regulated by Taiwanese EPA.
The sampling time was from May 5 to May 8 in 2013.2Thesampling time was from October 2 to October 5 in 2013.

Table 3 .
Summary of the current data on PM 2.5 -bound PBDE levels in outdoor air.

Table 4 .
Concentrations of PBDEs on PM 2.5 fine particulate in the indoor aerosol from various environments (pg m -3 ).Three PM 2.5 samples from the same style of the indoor environment were pooled together.

Table 5 .
Summary of the current data on PM 2.5 -bound PBDE levels in indoor air.