International pellet watch: Global monitoring of polybrominated diphenyl ethers (PBDEs) in plastic resin pellets

Polybrominated diphenyl ethers (PBDEs) were measured in plastic resin pellets collected from 65 beaches across 27 countries worldwide. They were detected at 49 locations at concentrations of the sum of 49 congeners of up to 46 ng/g-pellet and a median of 2 ng/g-pellet. These values are one to two orders of magnitude lower than those of polychlorinated biphenyls (PCBs) (median, 51 ng/g-pellet). This difference can be attributed to lower production of the Penta-BDE technical mixture, which is Environmental Monitoring & Contaminants Research Japan Society for Environmental Chemistry This article is licensed under a Creative Commons [Attribution 4.0 International] license. © 2021 The Authors.


INTRODUCTION
The extensive and widespread use of chemicals has polluted the Earth's surface. Chemical compounds with persistent, bioaccumulative, and hazardous characteristics are of great concern, and some are internationally regulated as persistent organic pollutants (POPs) by the Stockholm Convention (UNEP, 2017). Their distribution is heterogeneous and their concentrations vary widely. Therefore, understanding the spatial and temporal distribution of individual POPs through monitoring is essential for assessing their threat, modeling their behavior, and regulating and evaluating their effectiveness. International Pellet Watch (IPW), launched in 2005, uses plastic resin pellets to observe the global status of POPs (Takada, 2006;Ogata et al., 2009;Takada and Yamashita, 2016).
Plastic resin pellets are the industrial feedstock for all plastic products. A small but significant proportion of pellets is unintentionally spilled during their manufacturing, transportation, and handling. Polyethylene (PE) and polypropylene (PP) pellets are two major polymers that are lighter than water and can be carried by surface runoff, streams, and rivers to the ocean. Owing to their persistent nature, pellets are distributed on beaches worldwide. While floating at sea, the pellets sorb and concentrate hydrophobic organic pollutants from the surrounding water (Mato et al., 2001). In IPW, pellets are collected from beaches and sent to the Laboratory of Organic Geochemistry at the Tokyo University of Agriculture and Technology, where POPs in them are analyzed. The usefulness of IPW as a tool for global POP monitoring was demonstrated by analyzing polychlorinated biphenyls (PCBs) around the world and the correlation of their spatial patterns with those revealed by the traditional monitoring protocol, i.e., Mussel Watch (Ogata et al., 2009;Takada and Yamashita, 2016). Because of low sampling and shipping costs, IPW can be applied worldwide. Uniquely, citizen science can contribute to IPW as non-specialists can easily collect pellets (Thiel et al., 2017;Zettler et al., 2017). The analytical results can be used to inform people of the hazardous nature of microplastics (plastic particles < 5 mm in size), increasing public awareness of plastic pollution (Yeo et al., 2015). The participation of citizens has increased the coverage of monitoring. To date, pellets from 300 locations in 40 countries have been analyzed, and the pollution status of POPs-i.e., PCBs and organochlorine pesticides (dichloro diphenyl trichloroethanes; DDTs and hexa-chlorocyclohexanes; HCHs)-has been revealed (Ogata et al., 2009;Karapanagioti et al., 2011;Mizukawa et al., 2013a;Hosoda et al., 2014;Yeo et al., 2015;Takada and Yamashita, 2016). The target compounds of IPW have been expanded to include polycyclic aromatic hydrocarbons (PAHs) and hopanes Yeo et al., 2017;Alidoust et al., 2021).
This report focuses on polybrominated diphenyl ethers (PBDEs), which are brominated flame retardants that have been widely used since the 1970s. PBDEs are added to electrical equipment, textiles, and other industrial plastic products to suppress flammability. According to UNEP (2010), the total global production of PBDEs from 1970 to 2005 was 1.3 to 1.5 million tons. Depending on the congener composition, three types of PBDE mixtures were commercially produced: Penta-, Octa-, and Deca-BDE technical mixtures. Congener profiles of the technical mixtures have been characterized by La Guardia et al. (2006). Penta-BDE technical mixtures are mainly contain tetra-and penta-BDEs with hexa-and tri-BDEs as minor or trace components. Octa-BDE technical mixtures are a complex mixture of hexa-BDEs to nona-BDEs with deca-BDE (BDE-209) present as a trace component. Deca-BDE mixtures predominantly contain deca-BDE with nona-BDEs as trace components. The major homologs in penta-and octa-BDE mixtures were listed as POPs in 2009 because of their neurotoxicity (Eriksson et al., 2001;McDonald, 2002), thyroid hormone disruption (McDonald, 2002), and immunotoxicity (Frouin et al., 2010). Deca-BDE production has dominated global production (1.1-1.25 million t from 1970 to 2005), whereas the production of penta-BDE mixtures (91,000-105,000 t) and Octa-BDE mixtures (102,700-118,500 t) has been more modest (UNEP, 2010). Deca-BDE can be debrominated to lower-brominated congeners (Söderström et al., 2004;Eriksson et al., 2004;Kajiwara et al., 2008;Kwan et al., 2013aKwan et al., , 2013bMizukawa et al., 2013b), which are toxic; therefore, deca-BDEs were listed as POPs and banned by the Stockholm Convention in 2017.
PBDEs are emitted to the environment via various routes from multiple sources, including industrial wastewater from factories that use PBDEs (e.g., La Guardia et al., 2007), sewage effluent (e.g., Kim et al., 2013), house dust containing PBDEs from the wear of consumer products, and urban runoff (e.g., Melymuk et al., 2014). Facilities that recycle electronic waste (e-waste), dismantling sites (e.g., Luo et al., 2009), and leachate from landfill sites (Li et al., 2012;Kwan et al., 2013a) are also important PBDE sources. Because of their persistence, Environmental Monitoring and Contaminants Research Vol.1, pp.75-90 (2021) PBDEs released into the environment are carried to the ocean, where they can be detected in seawater (Oros et al., 2005;Wurl et al., 2006), biota (Ramu et al., 2007), and sediments (Kwan et al., 2013b). Hence, monitoring PBDEs is important to understand their distribution and fate in marine environments.
The chemical structure and nature (such as hydrophobicity) of PBDEs are similar to those of PCBs. Thus, similar to PCBs, PBDEs are sorbed onto and concentrated in pellets from ambient seawater. PBDEs in plastic pellets have been measured in several countries. Taniguchi et al. (2016) measured PBDEs and other POPs in pellets collected at 41 locations from 15 cities in Brazil and detected PBDEs at 10 locations. Camacho et al. (2019) analyzed plastic debris, including pellets, from the Canary Islands and found PBDEs in all pellet samples. PBDEs were also detected in pellets from central Chile (Pozo et al., 2020). However, the spatial coverage of measurement campaigns is limited, and the global pollution status of PBDEs in pellets has not been reported thus far. Thus, we analyzed pellet samples collected from 65 beaches in 27 countries around the globe to understand the global pollution status of PBDEs and establish criteria for the categorization of the pollution status of PBDEs in pellets. In addition, we analyzed PCBs in the pellet samples and compared them with PBDEs to reveal the unique features of PBDE pollution.

SAMPLE COLLECTION AND TREATMENT
Within the IPW framework, samples are collected by volunteers and the analytical results of each sample are shared with the collectors via e-mail and a map on the IPW website. Beached plastic resin pellets were collected from 65 locations across the globe (Table 1). Sampling locations included large cities in the industrially and economically developed countries of USA, Japan, and Europe; cities of industrially developing nations such as the Philippines, India, and Ghana; and remote, undeveloped areas, such as the Galápagos Islands (Ecuador), Macquarie Island (Australia), the Canary Islands (Spain), Easter Island (Chile), Hulhumae (the Maldives), and Praia de São Lourenço (Portugal). Details of places and dates of collection are listed in Table 1. Samples were collected using stainless steel tweezers or clean hands and wrapped in aluminum foil and sent to the laboratory in Tokyo. Upon arrival, the pellets were kept frozen at − 30°C until sample analysis. The sorting and analytical procedures followed Ogata et al. (2009). Polymers were identified through near-infrared spectrometry (Plascan-WTM OPT Research Inc., Tokyo, Japan) as PE, PP, and others. Yellowing PE pellets with a yellowness of ≥ 40 were analyzed. From each location, five pools of pellets (each comprising five randomly selected pellets) were analyzed to determine the median concentrations of PBDEs and PCBs. To assess PBDE levels of virgin pellets, we also analyzed two types of commercial virgin PE pellets (Nippon-Chuko, Tokyo, Japan and DJC, Yokohama, Japan).

QUALITY ASSURANCE AND QUALITY CONTROL
Extraction efficiency was confirmed by analyzing additional hexane in which the pellets were soaked. The pellets had been already extracted thrice with hexane and we detected significant quantities of BDE209 and BDE47 via normal analysis. The fourth extract revealed < 3% of BDE-209 and < 0.1% of BDE-47, indicating that triple-soaking extraction is sufficient to extract PBDEs from the pellets. The reproducibility of this analytical procedure (i.e., column chromatography and instrumental determination) was confirmed by analyzing four aliquots from a single extract of pellets from Tokyo Bay, Japan, for PCBs and PBDEs, which was prepared for quality assurance and quality control (QA/QC). The relative standard deviations of the concentrations were < 9% for PCBs and < 11% for PBDEs. Recovery was tested by spiking the aliquots of the extracts with concentration-certificated standards; recoveries were 70% to 124% for PCBs and 70% to 144% for PBDEs. PBDE concentrations were corrected against the recovery of isotopically labeled surrogates, which were spiked before cleanup similar to that performed in normal pellet analysis. No recovery correction was performed for PCBs to maintain consistency with PCBs in the IPW samples in IPW website, some of which were determined using GC-ECD without isotopically labeled standards. A signal-to-noise ratio of < 3 (non-significant  (2021) peaks) on solvent injection into GC -IT / MS or GC-ECD was considered as the limit of detection (LOD). The LOD was normally 0.08 ng / g for lower-brominated congeners such as BDE-47 and 0.3 ng / g for higher-brominated congeners such as BDE-209 (Table S1). A procedural blank using only the solvent was run with every set analyzed (five pools). Analytical values < 3 × the corresponding blanks were considered to be below the limit of quantification (LOQ). The LOD and LOQ were usually 0.2 and 0.7 ng / g-pellet for Σ 13 PCBs, and 0.08 and 0.2 ng / g-pellet for Σ 6 PBDEs (Table S2). As pellet blanks, two types of virgin PE pellets were analyzed; no significant PBDE congeners were detected in either pellet.

CONCENTRATIONS AND CONGENER PROFILES OF PBDEs IN PELLETS
Of the 65 collection locations, PBDEs were significantly detected ( > 0.7 ng / g) at 49 sites (Fig. 1). The concentration of total PBDEs (sum of 49 congeners from mono-to deca-BDEs) reached 46 ng / g-pellet (USA #1), with a median of 2.0 ng/gpellet (Table 2). Higher PBDE concentrations were observed in pellets collected from the USA, western Europe, and Ghana. The spatial patterns of PBDE concentrations are discussed in the next section.
The congener profiles at all 65 locations and their concentrations are presented in Fig. S1 and Table S3, respectively. Typical congener profiles found in some countries are shown in Fig. 2. Lower-(mono-to hexa-) brominated congeners were dominant over higher (hepta-to deca-) ones at most places (35 of 51). Of these, BDE-47 (Br 4 ) was the most abundant congener at 25 sites. The prevalence of other relatively abundant lighter compounds, including BDE-99 and 100 (Br 5 ), BDE-49 and 66 (Br 4 ), and BDE-154 and 153 (Br 6 ), was also noted in pellets from Chile (Pozo et al., 2020), Brazil (Taniguchi et al., 2016), and the Canary Islands (Camacho et al., 2019). The congener profiles in pellets were similar to those found in penta-BDE technical mixtures (Fig. S2), and the predominance of tetra-and penta-BDEs (mainly BDE-47 and 99) is consistent with that observed in seawater and bivalves (de Wit, 2002;Bervoets et al., 2004;Bodin et al., 2007;Mizukawa et al., 2009). In other environmental media, the congener profile of pellets collected from Tokyo Bay is more similar to that of dissolved-phase seawater and oyster than that of sediment, in which only BDE-209 was detected (Fig. 3). This suggests that pellets adsorb penta-BDE-derived congeners from the dissolved phase of seawater. Despite much higher production of deca-BDE technical mixtures, BDE-209 was detected in pellets at a limited number of locations. This predominance of lower-brominated congeners accounts for the less availability of BDE-209 for equilibrium partitioning owing to its higher molecular weight and bulky molecular structure (Bao et al., 2011). Di-and tri-brominated congeners (e.g., 15,17,28) were not detected in the pellets but were significantly detected in seawater and oysters (Fig. 3). Less-hydrophobic compounds such as di-and tri-brominated diphenyl ethers may be less sorbed onto hydrophobic sorbent pellets. In the case of organisms, the possibility of debromination to lower-brominated congeners as metabolism can also be considered.
Overall, PBDE congener profiles in pellets were similar to those of penta-BDE technical mixtures, with some contribution from deca-BDE mixtures. In addition, some pellet samples had congeners that are absent in the technical products. For example, BDE-49 was detected in 12 samples. This is probably due to the debromination of penta-BDEs, particularly BDE-99, and is consistent with the tendency for BDE-99 to be less abundant than BDE-47 in the pellet samples, with a BDE-47 :99 ratio of 2.4 ± 0.8; in contrast, BDE-99 was the most abundant conge-ner in the technical products, with a BDE-47:99 ratio of 0.78-0.96 (La Guardia et al., 2006). This discrepancy can be explained by the debromination or selective degradation of BDE-99 through photolytic, microbial, or anaerobic reactions. Debromination is also suggested for higher-brominated congeners in some pellet samples: nona-and octa-brominated congeners, particularly BDE-202, in some pellet samples (e.g., USA #1 and France #2) indicate the debromination of BDE-209.
Majority locations (pellets) did not show similar congener profiles to those of the octa-BDE technical products, which are dominated by BDE-183 (Fig. S2). Among the 325 analyzed pools (65 locations × 5 pools each), only one pool (Fujimae, pool #5 of Japan #7) was found to have a congener profile similar to that of DE-79 which is an octa-BDE mixture (Fig. S3). This may suggest the minimal contribution of octa-BDE mixtures to marine pollution because of PBDEs. However, this is not consistent with the greater volume of production of octa-BDE mixtures (102,700-118,500 t) than that of penta-BDE mixtures (91,000-105,000 t). Less availability for equilibrium partitioning due to the higher molecular weight and bulkier nature of BDE-183 and associated congeners (Bao et al., 2011) or faster photodegradation of BDE-183 (Fang et al., 2008) could explain the lower contribution of octa-BDE mixtures to PBDE pollution in marine environments.
Concentrations of total PBDEs in pellets (0.08-46 ng/g, median of 2.0 ng/g) were one to two orders of magnitude low-Environmental Monitoring and Contaminants Research Vol.1, pp.75-90 (2021) er than those of Σ 13 PCBs (0.15 -2,230 ng / g-pellet, median of 51 ng / g-pellet; Table 2). This seems to be inconsistent with the similar levels of PBDE and PCB production: ~1.3 to 1.5 million t of PBDE from 1970 to 2005 (UNEP, 2010) and ~1.3 million t of PCBs until 1993 (Breivik et al., 2002). However, as PBDEs in pellets were mainly derived from Penta-BDE technical mixtures, and their production (~0.1 million t) was one order of magnitude lower than that of PCBs, lower concentrations of PBDEs than that of PCBs in pellets can be reasonably expected. Similar lower concentrations of PBDEs than that of PCBs were observed in seawater , bivalves (Bayen et al., 2003;Bervoets et al., 2004;Bodin et al., 2007;Mizukawa et al., 2009), mammal tissues (Kunisue et al., 2008), and human breast milk (Malarvannan et al., 2009). Different forms of usage between the two chemicals may also contribute to the lower environmental concentrations of PBDEs: PBDEs are added as flame retardants to long-lived plastic products and textiles and thus are not leached into the environment as easily as PCBs, which are used in insulation oils, lubricants, heating fluids, and pigments. PCBs are thus more easily leached from these products during their use and disposal, resulting in higher environmental concentrations than that of PBDEs.

SPATIAL PATTERN OF PBDEs IN PELLETS
Σ 6 PBDEs are used as frequently detected congeners to spatially compare PBDE concentrations in pellets. They are often measured in other PBDE monitoring studies. BDE-209 and related higher-brominated congeners are not included in the sum because they may originate from recycled plastics and the direct attachment of soil and sediment particles, resulting in the spurious detection of higher-brominated congeners in The results of the assays are shown in the global map (Fig.  4a) and regional maps for Europe (Fig. 4b) and Japan (Fig. 4c). Σ 6 PBDEs ranged from n.d. (usually < 0.08 ng / g-pellet) to 36 ng/g-pellet. The samples were categorized into five pollution levels based on their concentrations. Twenty locations exhibited no significant concentrations of Σ 6 PBDEs ( < 0.2 ng/g-pellet, which is the typical LOQ). This is classified as "no local pollution." The other, higher, concentration levels are categorized based on equal sorting of samples according to Σ 6 PBDEs such that each category contains an equal number of samples; they are classified as "slightly polluted" (0.2 -0.78 ng/g), "moderately polluted" (0.78 -2.6 ng / g), "highly polluted" (2.6-8.5 ng / g), and "extremely polluted" ( > 8.5 ng /g). Ex-tremely polluted samples came from the USA, Belize, Puerto Rico, Ghana, and Spain (Fig. 4). Highly polluted sites were identified in China, France, the Philippines, Ghana, Japan, and Argentina. Moderately polluted sites were in Australia, the UAE, Brazil, and other countries in Europe and Africa (Fig. 4). No local pollution and slightly polluted areas were detected mainly on remote islands and in coastal zones in rural regions (Fig. 4).
These concentrations of PBDEs in pellets and the categorizations are consistent with those previously reported for pellets in Brazil (Taniguchi et al., 2016), Chile (Pozo et al., 2020), and the Canary Islands (Camacho et al., 2019). Taniguchi et al. (2016) measured pellets from 41 locations in São Paulo, Brazil, and reported Σ 6 PBDEs ranging from 0.76 to 2.67 ng/g-pellet. They analyzed pellets from three locations in Guarujá, São Paulo, where pellets were also sampled for our study. At the three locations, they found Σ 6 PBDEs ranging from < 0.76 to 1.61 ng/g, similar to our measurement (0.85 ng/g-pellet). These low levels of PBDE pollution are consistent with the results obtained in crabs and fishes in a neighboring region (Magalhães et al., 2012(Magalhães et al., , 2017. In Chile, Pozo et al. (2020) measured pellets from three sites in San Vicente Bay, where pellets were also sampled for our study; they reported Σ 6 PBDEs ranging from 0.03 to 0.45 ng/g-pellet. All samples were categorized as reflective of "no local pollution" or "slightly polluted." This is consistent with our Σ 6 PBDEs value of 0.18 ng/g in the pellets collected from the same bay. Camacho et al. (2019) measured PBDEs in pellets from four locations in the Canary Islands and reported the sum of eight BDE congeners 47,99,100,154,153,85,183) ranging from 0.5 to 3.1 ng/gpellet. This wide range is similar to our measurement of PBDEs in two locations in the Canary Islands, Σ 6 PBDEs of 0.2 and 3.3 ng/g-pellet. These comparisons support that our measurements reflect the global spatial pattern of PBDE concentrations, although our collection of pellets from each country was limited.
High concentrations of Σ 6 PBDEs were observed in industrialized countries: USA (17-36 ng/g-pellet), France, the Netherlands, and Japan ( < 0.7-8.8 ng/g-pellet). Higher concentrations of PBDEs in pellets collected from USA (Fig. 4a) were anomalous with the measured concentrations of PCBs. PCB concentrations in pellets collected from US coasts were similar to or lower than those collected from the coasts of France, the Netherlands, Italy, and Japan (Fig. 5). This is because the consumption of penta-BDE technical mixtures, which contain tetra-and penta-brominated congeners as the main components, is higher in USA than in Europe and Japan. USA is estimated to constitute 85% of the global consumption of penta-BDE (Abbasi et al., 2015), compared with only 48% of the global historical emissions of PCBs (Breivik et al., 2007). Resulting from the prevalent consumption of penta-BDE technical products in the USA, a greater amount of tetra-to penta-brominated congeners is likely to have been released and accumulated there than in other countries.
Higher concentrations of PBDEs in pellets collected from USA were also consistent with those observed in seawater, sediments, and biota (Oros et al., 2005). Concentrations of BDE-47 in mussels, which have long been used to monitor POPs, obtained from the literature (de Boer et al., 2003 Mizukawa et al. (2009Mizukawa et al. ( , 2013b. Values for Tokyo Bay sediment from Mizukawa et al. (2009). Gray bars: LOD ("not detected"). † Values for the filtrate of surface water of Tokyo Bay from Mizukawa et al. (2021). Gray bars: < LOQ. ‡ This study (Odaiba). Gray bars: < LOQ. § Values for Tokyo Bay sediment from Mizukawa et al. (2009). Gray bars: < LOQ.
Environmental Monitoring and Contaminants Research Vol.1, pp.75-90 (2021) and Hites, 2003;Oros et al., 2005;Johansson et al., 2006;Ramu et al., 2007;Mizukawa et al., 2009;Ueno et al., 2010) are summarized in Table S5. Concentrations of BDE-47 in mussels from USA (1.4 -43 ng / g-dry) were one order of magnitude higher than those from Europe (0.58 -4.96 ng / g-dry), which were subsequently higher than those from Japan (0.02-0.98 ng/g-dry), China, Malaysia, Indonesia, and India (0.002-0.108 ng/g-dry). This order in other environmental media is consistent with that determined using pellets. Widespread PBDE pollution in the air in the Northern Hemisphere (Jaward et al., 2004;Shen et al., 2006) is also consistent with the occurrence of higher concentrations of PBDEs in pellets collected from the Northern Hemisphere (Fig. 4a). These results indicate that the global spatial pattern of PBDE pollution in pellets is similar to those in other environmental media.

PBDE POLLUTION IN DEVELOPING COUNTRIES
Similar high concentrations of PBDEs to those in Europe and USA were observed in Ghana, Puerto Rico, and Belize, where no PBDEs are produced. Moderately high concentrations were observed in the Philippines, Hong Kong, and Brazil, where no PBDEs are produced. These concentrations can be attributed to imported electrical and electronic equipment (EEE) or e-waste imported (officially and unofficially) from    (Lundgren, 2012). Large amounts of old EEE are disposed of and transported to official (and unofficial) sites for disposal or recycling. Official sites have a large capacity to conduct appropriate processes, and the facilities and working environment are properly managed there. However, unofficial sites are unregulated and activities are not conducted under controlled conditions. At such sites, with small capacity and inefficient management (Tsydenova and Bengtsson, 2011), the release of PBDEs that have leached out from the EEE and e-waste into the surrounding environment is of great concern. Sometimes, EEE and e-waste are burned in the open (e.g., Fig. S4), and evaporated PBDEs could be deposited on the surrounding soil and transported to coastal waters in surface runoff.
Extremely high concentrations of Σ 6 PBDEs were observed in Puerto Rico (17 ng / g) and Belize (26 ng / g). Puerto Rico is an unincorporated territory of USA in the Caribbean. The elevated levels of PBDEs in pellets collected from Península La Esperanza, Cataño, in Puerto Rico, might be associated with its location in San Juan Bay (Pérez-Alvelo et al., 2021). This semi-enclosed bay is located in a highly urbanized, commercialized, and industrialized zone that includes maritime cargo port facilities, cruise line terminals, and a nearby airport. Enclosed harbors where shipping and unloading operations occur are significantly contaminated by PBDEs (Zheng et al., 2004). We suspect pollution derived from industrial activities (stemming from its close association with USA) in the Bay area to be a source of the high concentrations of PBDEs. In addition, PBDEs derived from the use of imported consumer products (e.g., electronic and electrical equipment), mainly from USA, could be another pollution source. PBDEs could enter this harbor through improperly disposed-of PBDE-containing products from adjacent land-based sources, such as discharges from wastewater treatment plants, leachates from landfills, stormwater runoff, and riverine input. Belize also imports relatively large amounts of industrial equipment from USA. Shen et al. (2006) studied PBDEs 99,100,153,and 154) in the air of North and Central America using a passive air sampler and reported that concentrations in Belize were the highest among all 40 locations, including those in mainland USA. The high concentration of PBDEs in pellets collected from Belize is consistent with the spatial pattern of atmospheric PBDEs.
The range of Σ 6 PBDEs in Ghana (0.84-16 ng /g-pellet) was the same as or higher than that in Europe ( < LOQ-8.8 ng/g-pellet). Σ 6 PBDEs were significantly higher (13-16 ng/ g-pellet) in Accra than in other areas sampled in Ghana (0.84-6.2 ng/g-pellet). Thus, the specific source of PBDEs is likely to be present in the capital city. In a previous study on PCBs in Accra (Hosoda et al., 2014), the e-waste scrapyard at Agbogbloshie was identified as the source of PCBs in the coastal zone there. Thus, the e-waste scrapyard in Accra is suspected to be an important source of PBDEs in pellets from the Accra coastal zone. The dismantling and dumping of e-waste could contribute to PBDE pollution there.
Pellets collected from the Philippines and Hong Kong exhibited moderate levels of PBDEs, i.e., 7.5 and 7.7 ng/g-pellet, respectively. These are similar to those found in Europe. The reason for these high concentrations could be e-waste. A similar trend was observed in the Asian Mussel Watch: higher concentrations of BDE-47 in mussels were reported from the Philippines (0.21-0.56 ng/g-dry, Ramu et al., 2007) and Hong Fig. 5 Concentrations of Σ 13 PCBs in beached plastic resin pellets. Σ 13 PCBs: sum of concentrations of 101,110,149,118,105,153,138,128,187,180,170,and 206 Fig. 5. Concentrations of Σ 13 PCBs in beached plastic resin pellets. Σ 13 PCBs: sum of concentrations of CB-66, 101, 110, 149, 118, 105, 153, 138, 128, 187, 180, 170, and 206. Kong (0.06 -29.8, Liu et al., 2005;Ramu et al., 2007) than from other tropical Asian countries (0.002 -0.108 ng / g-dry, Ramu et al., 2007) (Table S5). E-waste recycling plants (both official and unofficial) in Manila may be a major source of PBDEs in the Philippines. In addition, leachate from garbage dumping sites could be another important source. Landfill leachate in Manila exhibited higher concentrations of PBDEs among those in Asia (Kwan et al., 2013b). In Hong Kong, trading in e-waste is now banned. However, e-waste was imported and recycled in the unofficial e-waste sector in the past (Shinkuma and Nguyen Thi Minh, 2009) in Hong Kong. In another study, PBDEs were determined in soil samples collected from an e-waste dismantling workshop, an open burning site, and e-waste storage in the new territories of Hong Kong (Lopez et al., 2011). These findings suggest that e-waste activities cause PBDE pollution in those areas, which our results may reflect.
In the background of these case studies, the global generation of e-waste was estimated to be 53.6 million t in 2019, up by 9.2 million t since 2014; it is projected to grow to 74.7 million t by 2030 (Forti et al., 2020). The fate of ~80% of e-waste is uncertain and the resulting environmental impact varies across regions (Forti et al., 2020). Thus, continuous PBDE monitoring through IPW must be performed in these regions.

CONCLUSIONS
High PBDE levels in pellets collected from USA, western Europe, and Japan reflect the legacy of the use of penta-BDE technical products. Elevated concentrations in some industrially developing countries can be attributed to the processing of e-waste. As the global generation of e-waste continues to increase, pollution potentially associated with it should be studied in more detail by testing sediment and air samples together with pellets. For the initial screening of areas potentially polluted by e-waste, IPW and the five-level categorization of PBDEs would provide a valuable benchmark. and The University of Tokyo FSI -Nippon Foundation Research Project on Marine Plastics.