The levels of all analysed POPs were below the detection limit in the MP polymers placed at the reference station

In the marine environment, they have been reported in coastal waters, surface waters, the water column, deep-sea sediments, estuaries and fjords , including in northern waters . MPs can have a wide range of negative effects on biota, from reduced feeding to immune system alterations . Depending on size and shape, MPs can cross body barriers, e.g. be translocated from the digestive tract to other tissues and organs, and be taken up into cells . MPs have been detected in liver and muscle tissue of both farmed and wild salmonids . Plastic microfibers have also been detected in several organs from blue mussels exposed to concentrations of 2000 particles/L, including mantle and foot . Accumulation of MPs in edible tissues might in turn constitute a health risk for human consumers . Negative effects of MPs on organisms may derive from plastic additives as well as the plastic polymer particles themselves . Additionally, persistent organic pollutants can sorb to MPs due the hydrophobic nature these chemicals, and act as vectors for POPs in the marine environment . For example, several studies have documented sorption of PCBs, PAHs, pesticides and other POPs, on MPs collected from marine environments . POPs sorbed to MPs may be transferred to the organisms after ingestion, increasing and complicating the impact of such particles and pollutants . The role of MPs as vectors of pollutants to organisms has been suggested to be irrelevant compared to dietary or environmental exposure. In addition, MPs could pass through the gastrointestinal tract without leaching chemicals . Nevertheless, translocation of small polluted MPs into cells, tissues and organs of organisms only exposed to low levels of POPs could have a significant impact in such individuals. In such cases, MPs might be a pathway for uptake and bioaccumulation of POPs in marine organisms . Dioxins are chemical substances formed as byproduct during combustion reactions in the presence of chlorine .

In vertebrates, dioxins are carcinogenic and can disrupt the immune, nervous and endocrine systems, and affect reproduction and development . Considering the increasing numbers of wildfires as result of climate change ,hydroponic grow table and that dioxins tend to accumulate in polar regions due to slower degradation in cold areas and the grasshopper effect , more attention should be paid to the impact of dioxins in the environment together with MPs. To our knowledge, there is no record of levels of dioxins adhered to MPs in the environment except for the levels found in charred MPs . Therefore, interaction of MPs with certain POPs such as dioxins should be studied more in depth. Atlantic salmon is one of the most consumed fish species in the world and the commercially most important farmed marine fish species . The levels of pollutants associated with Atlantic salmon farming have long been a concern. Salmon feed contains about 10% of fish oil , which is the main source of POPs to farmed salmon. Analysis of 20 salmon feed samples randomly collected from salmon farms in Norway in 2017 showed concentrations of polychlorinated biphenyls of 3 μg/kg, sum of dioxins and dioxin-like PCBs of 0.6 ng TEQ/kg, and concentrations of polybrominated diphenyl ethers of 0.38 μg/kg . In addition, fish feed also contained traces of organochlorine pesticides such as DDT , toxaphene and endosulfan, and organophosphorus pesticides such as chlorpyrifosmethyl and pirimiphos-methyl. In recent years, replacement of fish oils with plant oils has led to reduced levels of POPs in farmed Atlantic salmon in Norway . However, the use of plant ingredients in aqua feeds has increased the levels of other organic pollutants in fish feed, such as PAHs and pesticides . In Norway, about 1.3 million metric tons of Atlantic salmon is produced annually . For each kg of salmon produced, about 0.5 kg of feces and unconsumed feed pellets are generated . This waste slips through the open-cage net pens and spreads in the environment, depending on local physical, chemical and biological factors. Open fish farms thus represent local point-sources of pollution. Yet little is known about how MPs might contribute to the spreading of POPs near fish farms. Fish farms benefit from the lightweight, strong and flexible plastic in permanent installations. Net pens, ropes, floats and pontoons are some of the framework structure made of plastic polymers such as HDPE, PP, PET or PVC . Abrasion and loss of plastic items inevitably lead to the release of MPs into the environment. A recent study showed the presence of MP polymers such as PP, PE, PVC and PET in sediment, seawater and suspended matter around Atlantic salmon farms of Norway .

In Norway, about 1000 commercial salmonid farm facilities are placed in the marine environment . These are estimated to use a total of 191,799 tons of plastic materials, of which 35,571 tons are nets , 17,201 tons are mooring ropes , 4440 tons are feeding pipes, and 108,405 tons are cage floaters, handrails and walkways . Based on the degradation rates of PE , nylon and PP ropes in the marine environment calculated by Welden and Cowie , the total amount of nets and ropes deployed by fish farms in Norway are estimated to release 3137 tons of MPs/year and 805 tons of MPs/year into the environment, respectively. Feeding pipes used in salmon farming have recently been shown to release an average of 0.25 g/m/day MPs along with the fish feed . Based on this study, feeding pipes used in Norwegian aquaculture are estimated to release 225 tons of MPs/year in the water. In addition, cage floaters are estimated to release 5.4 tons of MPs/year based on the assumption that plastic degrades at an annual rate of 0.5% in the Norwegian marine environment . Based on all the above, fish farming in Norway might realease more than 4172.4 tons of MPs annually into the marine environment. Thus, understanding the potential role of MPs as vectors of pollutants from fish farming is important for countries such as Norway. The objective of this study was to evaluate whether MPs can sorb POPs released from fish feed significantly, and consequently act as potential vectors for these chemicals to the surrounding environment. As a step towards better understanding the impact of different MPs on the environment, we assessed the capacity of HDPE, PP, PET and PVC to sorb POPs. For this purpose, MPs of four different types of polymers commonly found in the environment and detected in the vicinity of salmon farming facilities were placed close to two salmon farms for three months. For comparison, MPs were additionally deployed in two sites not influenced by salmon farming; one low-polluted and one polluted location. Furthermore, a positive control for fish feed pollutants was set up under laboratory conditions. POPs sorbed to MPs were then qualitatively and quantitatively analysed. Blue mussels were collected from a reference site and placed next to the MPs at all stations, with the objective of documenting the pattern of POPs in the environment. This filter-feeder is often used in water monitoring studies as a bioindicator of water pollution . Recently, mussels have also been suggested as bioindicator for marine MP pollution, since they are widely distributed and have been shown to indicate differences between sites with large differences in MP concentrations . Blue mussels are currently used in coastal environmental monitoring programs in Norway as bioindicators for heavy metals, organic pollutants and MPs .

Two replicates of each of the PP, HDPE, PET and uPVC polymers incubated at each site were used for chemical analysis. Frozen samples were treated gently to remove the biofilm on the MPs through an enzymatically and strong alkali driven cleaning process. Each sample was sequentially incubated at room temperature with a mixture 1:20 of Celluclast and 1 mL Viscozyme enzymes in 0.1 M PBS at pH 6.0 for 24 h and with 10% KOH for 6 h at RT. Plastic beads were gently flushed and rinsed with millipore water through a vacuum system to remove the degraded biofilm and were allowed to dry at room temperature for 1 h. Samples were then sonicated three times for 20 min with 5 mL of dichloromethane, HPLC grade, and the supernatant was collected after each sonication step. Extracts were pre-concentrated by a Rotavapor system and analysed using the following HRGC/HRMS methods: USEPA Method 1668B, USEPA Method 8290A and USEPA Method 1614A for polychlorinated biphenyls, dioxins and brominated flame retardants, respectively, using a GC–MS/MS ; and the following GC-ECD method for chlorinates pesticides: USEPA Method 508.1 using a GC–MS . Isotope labelled PCB mixture 77/101/141/178; 1,2,3,4,6,7,8- heptabromodibezofuran and Aldrin 13C12 were added as internal standards. The limit of quantification was set to three times the value of limit of detection .For the harbour, the two fish farms and the positive control, the levels of the majority of pesticides and half of the novel brominated flame retardants analysed were also below the detection limit in all polymer types. Hence, only data for dioxins and PCBs were used in the PERMANOVA analyses. The concentrations of all studied POPs sorbed onto the four MP polymer types are shown in Supplementary Table S7. The levels of POPs sorbed to the MPs were quantitatively assessed to understand whether MPs can bind POPs associated with fish farming. PERMANOVA was used to compare the composition of POPs sorbed to MPs from the different sites and the positive control . The results of the PERMANOVA analysis showed that the composition of POPs in MPs from the harbour was statistically different from the composition of POPs in MPs from FF1 and FF2,flood tray and the positive control, . However, there were no significant differences in the composition of POPs sorbed to MPs from FF1, FF2 and the positive control. To evaluate whether binding affinities of POPs differed among the studied MP polymers, a Kruskal-Wallis test was done for each dioxin congener and each PCB Aroclor . Since this study focuses on the sorption of POPs from fish farming, only data of dioxins and PCBs sorbed to MPs from the two fish farms and the positive control were used. Positive control was included in this analysis because no differences were found in the previous analysis in the composition of POPs of these MPs with the ones of the fish farms. Data of dioxins and PCBs sorbed onto the MPs placed at the harbour and the reference station were excluded because different cocktail of pollutantsmight result in different sorption affinities.

Chemical properties of the analysed dioxins and PCBs are shown in Supplementary Table S8. This study suggests that MPs can sorb POPs associated with Atlantic salmon farming, and documents that such pollutants have polymerspecific binding affinities. Polymer type and pollutant type, in addition to background pollution in the water, are therefore determining factors that should be considered when assessing the potential role of MPs as vectors of pollutants from aquaculture in the marine environment. In this study, four types of MP polymers were either incubated for three days with fish feed or placed in the sea for three months next to two marine fish farms, in an urban harbour and in a non-polluted fjord. MPs collected from the two fish farms, the harbour and the lab control sorbed dioxins, PCBs and some brominated flame retardants , whereas pesticides were barely detected. In contrast, MPs placed in the non-polluted fjord did not have quantifiable levels of any of the analysed POPs. It is well known that MPs can sorb pollutants in the environment. At the beginning of this century, plastic pellets sampled in waters from all five continents were reported to be polluted with PCBs and pesticides . Since then, several studies have reported the presence of POPs bound to MPs in the marine environment . The main mechanisms by which MPs sorb POPs include hydrophobic and electrostatic interactions, although multiple mechanisms are often cooccurring in complex natural environments. Chemicals characterised by high hydrophobicity, which are defined by high octanol/water partition coefficients , are expected to be more easily sorbed by plastic particles .