^{1}

^{2}

^{*}

^{1}

^{2}

The authors have declared that no competing interests exist.

Conceived and designed the experiments: CMR BTH. Performed the experiments: CMR. Analyzed the data: CMR BTH SJT. Contributed reagents/materials/analysis tools: CMR BTH SJT. Wrote the paper: CMR BTH SJT.

Concerns regarding plastic debris and its ability to accumulate large concentrations of priority pollutants in the aquatic environment led us to quantify relationships between different types of mass-produced plastic and metals in seawater. At three locations in San Diego Bay, we measured the accumulation of nine targeted metals (aluminum, chromium, manganese, iron, cobalt, nickel, zinc, cadmium and lead) sampling at 1, 3, 6, 9 and 12 months, to five plastic types: polyethylene terephthalate (PET), high-density polyethylene (HDPE), polyvinyl chloride (PVC), low-density polyethylene (LDPE), and polypropylene (PP). Accumulation patterns were not consistent over space and time, and in general all types of plastic tended to accumulate similar concentrations of metals. When we did observe significant differences among concentrations of metals at a single sampling period or location in San Diego Bay, we found that HDPE typically accumulated lesser concentrations of metals than the other four polymers. Furthermore, over the 12-month study period, concentrations of all metals increased over time, and chromium, manganese, cobalt, nickel, zinc and lead did not reach saturation on at least one plastic type during the entire 12-month exposure. This suggests that plastic debris may accumulate greater concentrations of metals the longer it remains at sea. Overall, our work shows that a complex mixture of metals, including those listed as priority pollutants by the US EPA (Cd, Ni, Zn and Pb), can be found on plastic debris composed of various plastic types.

Plastic debris litters the aquatic environment globally

Plastic debris is associated with a “cocktail of contaminants” made up of chemical ingredients in the plastic and chemical pollutants sorbed to the plastic from the environment

The presence of organic chemical pollutants on plastic debris is established globally

Plastic debris is composed of several different polymers, and their unique chemical ingredients may make some types of plastic more hazardous than others when their chemical constituents are bioavailable to organisms. For example, polyvinyl chloride (PVC), polycarbonate, polyurethane and polystyrene (PS) are composed of hazardous monomers (e.g., vinyl chloride, bisphenol-A and styrene) and/or contain hazardous additives (e.g. PBDEs, phthalates and lead)

Plastic pre-production pellets, a recognizable component of marine debris, have been used to establish a global association between organic contaminants and plastic discarded at sea

We aimed to measure the concentrations of pollutants over a 12-month period, and achieved this by deploying five of the most common types of mass-produced pre-production plastic pellets

The map shows the three study locations: Coronado Cays, Shelter Island and Nimitz Marine Facility. Figure generated with ArcGIS version 9.3.

On June 1^{st}, 2009 five types of pre-production plastic pellets, PET, HDPE, PVC, LDPE and PP (denoted by recycle codes 1–5, respectively, in the USA), were deployed from docks at three locations throughout San Diego Bay (

At each location, two replicate samples (5 g) of each plastic type were deployed for future collection at the end of five time periods: 1, 3, 6, 9, and 12 months (150 total samples). Each replicate consisted of 5 g of pellets of one plastic type placed in individual Nitex mesh (1.3 mm) bags. Pellets of PET were cylindrical (3 mm long, 2 mm diameter), and pellets of HDPE, PVC, LDPE, and PP were spherical (3 mm diameter). To facilitate removal of large fouling organisms (e.g., tunicates and bivalves), each Nitex bag was placed inside a nylon mesh (10 mm) bag. Replicate samples were deployed by hanging each nylon bag on one of two identical PVC frames suspended from each dock (see

Unless stated, reagents were purchased from Fisher Scientific (Fair Lawn, NJ, USA) and were of analytical grade or better. All plasticware used for sample processing was rinsed five times with each 10% hydrochloric acid (HCl) followed by Millipore Milli-Q water. After drying under a laminar flow hood, clean plasticware was stored in polyethylene bags until use.

Pellets from each field-collected Nitex bag were rinsed in ultrapure water to remove sediment. In a few cases, large fouling organisms were seen attached to individual pellets; these pellets were excluded from chemical analyses. The frequency at which individual pellets were observed to be fouled did not vary noticeably among samples from different deployment times or from different sampling locations. Thirty pellets from each sample were weighed into individual polypropylene centrifuge tubes. Two milliliters (mL) of 20% aqua regia (1 part nitric acid (HNO_{3}) to 3 parts HCl) were added to each sample and centrifuge tubes were screw-capped. Samples were then agitated for 24 hours at room temperature using a Roto-Shake Genie (Scientific Industries, USA) before 1 mL of digested solution was transferred into a clean tube and diluted with 9 milliliters Millipore Milli-Q water for analysis. Virgin pellets of each type were used as laboratory blanks and 0-month controls (n = 3).

The prepared samples were analyzed for Al, Cr, Mn, Fe, Co, Ni, Zn, Cd and Pb by the Interdisciplinary Center for Plasma Mass Spectrometry at the University of California at Davis (ICPMS.UCDavis.edu) using an Agilent 7500CE ICP-MS (Agilent Technologies, Palo Alto, CA). Although concentrations of copper are of widespread interest in urban marine systems, concentrations of copper were inadvertently not measured by the analytical laboratory used in our study. The samples were introduced using a MicroMist Nebulizer (Glass Expansion, 4 Barlow's Landing Rd., Unit 2A Pocasset, MA 02559) into a temperature controlled spray chamber with helium as the Collision Cell gas. Instrument standards were diluted from Certiprep ME 2A (SPEX CertiPrep, 203 Norcross Avenue, Metuchen, NJ 08840) to 0.5, 1, 10, 100, 200, 500, 1000, 2000 and 5000 µg/L respectively in 3% Trace Element HNO_{3} (Fisher Scientific) in 18.2 Megaohm-cm water. A NIST 1643E Standard (National Institute of Standards and Technology, 100 Bureau Drive, Stop 2300, Gaithersburg, MD 20899-2300) was analyzed initially and QC standards consisting of a ME 2A at a concentration of 100 µg/L were analyzed every 12^{th} sample as quality controls. Sc, Y, and Bi Certiprep standards (SPEX CertiPrep) were diluted to 100 µg/L in 3% HNO_{3} and introduced by peristaltic pump as an internal standard.

Average levels of metals measured in laboratory blanks (i.e. time-zero controls that were never deployed in the field) were subtracted from the reported concentrations of metals extracted from pellet samples. Statistical analyses were performed using GMAV (GMAV; EICC, University of Sydney). We decided _{t} = C_{eq}(1-e^{−kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant.

All nine targeted metals were detected and quantified on all polymer types, with the exception of Cd on HDPE (

Concentrations of metals are shown for each of the five plastic types (PET-black, HDPE-hatched, PVC-diagonal stripes, LDPE-white, PP-horizontal stripes) at each of the three sites (CC-Cornado Cays, SI-Shelter Island, NMF-Nimitz Marine Facility) ordered from the back to the front of the bay. Each graph represents one of the 9 targeted metals (Al, Cr, Mn, Fe, Co, Ni, Zn, Cd, Pb) ordered from left to right according to molecular weight. Each bar represents the mean concentration (ng/g) + standard error (n = 2). A non-detect is denoted by nd. ANOVA showed statistically significant differences among plastic types (

For all metals, concentrations accumulated on plastics varied significantly among locations at a minimum of three out of the five time periods. For Al, Cr, Fe and Pb, concentrations were greatest at Shelter Island and Nimitz Marine Facility closer to the mouth of the bay (

We quantified temporal patterns of metal accumulation by each of the five plastic types at each of the three locations in San Diego Bay (

Concentrations of Mn, Co, Ni, Zn and Cd (ng/g of pellets) vs time for each type of plastic at Coronado Cays (CC) where contamination was greatest. Rows represent plastic types PET, HDPE, PVC, LDPE and PP (in order from top to bottom). Columns represent metals ordered from left to right according to molecular weight. Note that vertical axes differ among graphs. Data were fit to the first-order approach to equilibrium model _{t} = C_{eq} (1 − e^{−kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

Concentrations of Al, Cr, Fe and Pb (ng/g of pellets) vs time for each type of plastic at Nimitz Marine Facility (NMF) or Shelter Island (SI) where contamination was greatest. Rows represent plastic types PET, HDPE, PVC, LDPE and PP (in order from top to bottom). Columns represent metals ordered from left to right according to molecular weight. Note that vertical axes differ among graphs. Data were fit to the first-order approach to equilibrium model _{t} = C_{eq} (1 − e^{−kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

For Ni, Zn, Pb, Cr, Mn and Co predicted equilibrium concentrations (C_{eq}) were not achieved for some polymers on either replicate sample during the entire 12-month period at several locations (

Overall, our work shows that a complex mixture of metals is found on plastic debris composed of the five most commonly produced types of plastic

Plastic will be exposed to different mixtures and concentrations of chemical pollutants based upon the location where it has been discarded

Similarly, the length of time plastic remains aquatic debris has implications for management. Our previous work shows that plastic pellets accumulate greater concentrations of organic contaminants the longer they remain at sea _{eq}), and the assumption that metal accumulation patterns will follow the exponential rise to maximum equation, Co is predicted to reach equilibrium on polyethylene at 22 months and Pb at 64 months. Comparing plastic debris to plastic in a controlled laboratory setting has limited applicability to quantifying chemical hazards associated with plastic debris in aquatic habitats. Concentrations of ambient metals will change over time based upon their sources and physical changes in the water column (e.g. temperature). As plastic debris weathers it will gain surface area, generate oxygen groups (increasing polarity)

It is well-known that a wide range of animals ingest plastic debris in nature, including invertebrates

Exposure to metals may cause ecological effects from decreased growth and survival

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

_{t} = C_{eq}(1-e^{-kt}), where C_{t} is the concentration at time t, C_{eq} is the predicted equilibrium concentration, and k is the rate constant. The horizontal dotted line denotes the predicted C_{eq} for each plastic type. Where no equation is given, the model could not be fit to the data and where no horizontal line is given the non-linear regression was not statistically significant (p>0.05).

(TIFF)

(PDF)

(PDF)

The American Chemistry Council donated virgin plastic pellets. We thank Coronado Cays Yacht Club, Nimitz Marine Facility, and Michelson Yachts at Shelter Island for donating dock space. J. Commisso, S. Wheeler, E. Hoh and R. Gersberg assisted with chemical analyses and Z. Schakner, S. Wheeler, M. Colvin, C. Mazloff, J. Barr, M. Moore and S. Celustka assisted in the field. We thank K. Watanabe for assistance with

^{55}Fe-aided assessment of capacity, affinity and kinetics