Synthesis of 14C-labelled polystyrene nanoplastics for environmental studies

Available analytical methods cannot detect nanoplastics at environmentally realistic concentrations in complex matrices such as biological tissues. Here, we describe a one-step polymerization method, allowing direct radiolabeling of a sulfonate end-capped nano-sized polystyrene (nPS; proposed as a model nanoplastic particle representing negatively charged nanoplastics). The method, which produces nanoplastics trackable in simulated environmental settings which have already been used to investigate the behavior of a nanoplastic in vivo in a bivalve mollusc, was developed, optimized and successfully applied to synthesis of 14C-labeled nPS of different sizes. In addition to a description of the method of synthesis, we describe the details for quantification, mass balance and recovery of the labelled particles from complex matrices offered by the radiolabelling approach. The radiolabeling approach described here, coupled to use of a highly sensitive autoradiographic method for monitoring nanoplastic body burden and distributions, may provide a valuable procedure for investigating the environmental pathways followed by negatively charged nanoplastics at low predicted environmental concentrations. Whether the behaviour of the synthetic nPS manufactured here, synthesised using a very common inititator, represents that of manufactured nPS found in the environment, remains to be seen. Nanoplastics are a substantial environmental risk, and it is important to understanding where and how they are released into the environment. Here, a simple methodology is reported for the one-step synthesis of radiolabelled nanopolystyrene that can be used in environmental studies.

P aradigms for assessing the potential environmental impacts of manufactured nanomaterials have been eloquently described 1 . Among such nanomaterials, plastic particles exhibiting at least one dimension below 1 µm (i.e. so-called nanoplastics; NPs) have been reported to have the potential to become the most hazardous component of marine plastic litter 2,3 . NPs are suspected to be released into the environment directly from their addition to a range of commercial products 4,5 and by fragmentation of larger pieces of plastic litter 2,6-9 . The potential impacts of the sizes of particles (i.e. nano-sizes) on their degradation, uptake and biodistribution in organisms have been the subject of several studies 3,10 . Recently, the use of a synthetic 14 Cradiolabeled nanopolystyrene (nPS) either to study NP degradation in water and air 9 or, when coupled with autoradiography, to monitor the biodistribution of an NP in a bivalve mollusc has been described 11 . Although a multistep method for producing 14 C-nPS from cinnamic acid via styrene has been reported 9 , the details of the one-step synthesis of 14 C-nPS used in the biodistribution study 9 have not. Therefore, here we show these details, including the effects of controlling factors on the production of nPS particles with a range of sizes suitable for deployment in simulated environmental studies at low predicted NP concentrations. The resultant method using 14 C-styrene is simple and robust, requiring only minimal reagents, a 20 mL vial sealed with a rubber septum, a magnetic stirrer and an oil bath. We review the advantages of using synthetic radiolabelled particles in environmental experiments.
The NP concentrations to which organisms are predicted to be exposed in natural and technical environments are very low (between ca. 1 pg L −1 and ca 20 μg L −1 ) 12 . It is therefore important that the possible in vivo biodistributions and toxicokinetics of NP are explored using approaches which offer sufficiently low NP limits of detection. However, deployment of the most common analytical methods to study the fate and effects of commercially available NP may mean they are not detectable at environmentally realistic concentrations in complex matrices, such as biological tissues. Most current studies investigating bioaccumulation and effects of NP in organisms such as mussels 13 , oysters 13 , algae 14 , copepods 15 , Daphnia 16 and fish 17 have been carried out with commercial fluorescent or functionalized NP, with experimental exposure concentrations two to seven orders of magnitude higher than those predicted to occur in the environment 12 . For fluorescent NP these high limits of detection are due to factors such as high background signals, poor resolution due to autofluorescence, quenching of fluorescence 18 and leaching of the fluorescent label 19 . These factors limit the applications of commercial NP for providing reliable and quantitative data. However, it is now possible to gather information on the quantitative biodistribution and toxicokinetics of NP particles in environmentally realistic scenarios using radiolabelled 14 C-NP 9,11 .
To produce 14 C-NP of different sizes, we modified and improved the synthesis methods used for non-radiolabelled nPS previously 18,19 . One challenge of the synthesis of radioactive nPS, compared to that of unlabelled material, is the high cost of radioactive styrene monomer. Tian et al. 9 overcame this somewhat by conducting a multistep synthesis of 14 C nPS from labelled cinnamic acid. An alternative, used herein, was to improve the efficiency such that reactions using <10 mg of styrene monomer and very small volumes of reagents (<10 mL) could be conducted. Since the use for environmental studies also requires nPS with minimal-and ideally no co-products (e.g. styrene monomer, surfactants), the final method comprised a one-step polymerisation process using either 2 wt% of surfactant (for small, target 20 nm) or no surfactant (for large, target 250 nm) particles, followed by purification by dialysis. To optimise the method of production with minimal amounts of reagents, we examined the effects of the changes in the concentrations of the most commercially used initiator (here sodium persulfate, KPS), surfactant (sodium dodecyl sulfate, SDS), and of sodium hydroxide. The use of KPS gives a negative surface charge to the particles, which may make them a good model for studies in environmental settings. In fact, such negative charges are likely to occur on particles in environmental media by the sorption of fulvic and humic acids at relevant pH values 20 . As in industrial settings, other initiators could be considered. Indeed, it would be interesting in future to investigate the behaviour of primary NPs with different surface chemistries in environmental contexts using the approach discussed in this study.
Results and discussion Radioabelling strategy. As described in Fig. 1, the reaction was initiated by formation of a KPS radical at 65-75°C in water (Fig. 1a, 1). The presence of such a radical initiated the formation of styrene radicals (Fig. 1a, 2), which created a polymerisation cascade (Fig. 1a, 2).
In this study, nominal 20 nm nPS was formed by the polymerisation of 14 C-styrene in the presence of SDS in an emulsion of water and hexane (7:3), mixed by mechanical shear (i.e. 150 r.p.m.). SDS reduced the surface tension of water, producing nano-sized particles of polystyrene, defined by microemulsions with dispersed domain diameters varying from 1 to 100 nm 21 . Thus, in the presence of SDS, nPS monomers nucleated inside SDS-micelles swollen with the input of monomers 22 , reaching sizes up to 100 nm (Fig. 1b). In the absence of surfactants, the nucleation is thought to occur by the collapsing of water-insoluble polymer chains, forming precursor particles 22 . These unstable precursors (i.e. solidified nanoparticles of 10 nm 23 ) that become monomers of larger particles are believed to aggregate until they produce enough surface charge to be individually stable (Fig. 1c). Such aggregates can then swell by incorporating more monomers, forming mature particles (up tõ 500 nm) 24 . Both sizes of particles are colloidally stabilised electrostatically by the persulfate end groups of the initiator at the surface of the particles 25 .
Syntheses optimisation. Syntheses of both sizes of non-labelled nPS (i.e. 20 and 250 nm) were also performed with 12 C-styrene in order to provide materials for transmission electron microscopy (TEM) measurements. (Safety regulations prevented use of radioactive nPS for TEM investigations.) Results from TEM analyses indicated that the size of the nominally 20 nm polystyrene particles (nPS 20 ) remained fairly constant, even when the concentrations of styrene or SDS were changed substantially (Fig. 2a, c). Particles reached an overall average size of 20.6 ± 4.9 nm (n > 100) over the different styrene concentrations: 24.8 ± 12.8 (n > 100) with changes in KPS and 21.6 ± 3.0 nm (n > 100) with changes in SDS concentrations. Interestingly, at the highest KPS concentration of 50 mM (Fig. 2b), the particles showed a tendency to reach larger sizes (47.5 ± 9.3 nm), which might be due to the high concentrations of initiator, perhaps initiating more polymerisation and increasing the aggregation of nucleus per drop of monomer. Also, according to the results of dynamic light scattering (DLS) measurements, particles formed at low styrene concentrations were estimated at about 70 nm, whereas they were measured at about 20 nm by TEM. Such results show that the particles probably clustered when produced under such conditions.
The larger nPS particles were produced using a technique employed industrially to produce polymeric particles with large diameters (i.e. d > 100 nm) and narrow size distributions 26 . The syntheses of these larger polystyrene particles (nominally nPS 250 ) showed that, over all the synthesis conditions tested, particles reached minimum sizes of 198 ± 43 nm, with 1 wt% styrene, as measured by TEM (Fig. 2d), but the average size also increased with increases in the concentration of KPS, up to 451.9 ± 186.8 nm (Fig. 2e) where the concentration of initiator (i.e. KPS) promoted particle growth. The absence of surfactant in the syntheses of nPS 250 simplified the synthesis (i.e. no nitrogen atmosphere was needed).
Particles purification and characterisation. To avoid potential toxicological effects associated with SDS and unreacted chemicals (e.g. KPS and styrene monomer), purification by dialysis (exclusion size of membrane 30,000 g mol -1 , ca. d = 2 nm 27 ) was performed until the conductivity of the surrounding water remained constant at a value of <1.5 µS cm −1 and 14 C counting reached background levels. Interestingly, keeping a KPS concentration below~15 mM also allowed the preparation of particles with a narrow size distribution (i.e. a range of SD < 30 nm), which is uncommon in systems with no stabilising agent [24][25][26][27][28] .
Besides the potential to produce NP for ecotoxicological studies, the strength of the method lies in its simplicity and robustness: the method corresponds to a one-step synthesis and requires only a 20 mL vial sealed with a rubber septum, a magnetic stirrer and an oil bath.
Following the TEM measurements of unlabelled styrene and optimisation of the reaction conditions (Fig. 2), nPS particles were produced using 14 C-styrene. Based on the activity measured in the reaction medium before and after synthesis and purification, the yields for 'small' and 'large' nPS were estimated at 98% and 99%, respectively. A specific activity of 0.16 MBq 14 C/mgC NP was found after synthesis and the nPS suspensions were stable for weeks (i.e. no deposition was observed). The particles were stabilised by the negative charges of the sulfonic groups (i.e. derived from the initiator). At pH 6 in sodium chloride (5 mM), their ξ-potentials were −129 ± 10 and −83.6 ± 11.9 mV, while their sizes, as determined by DLS performed on a dedicated Zetasizer (Malvern Panalytic®), were 146 ± 29 and 275.9 ± 55 nm, for nominal 14 C-nPS 20 and 14 C-nPS 250 , respectively. These measurements, as expected, almost certainly overestimated the particle sizes. DLS intensity values tend to overestimate particle size and results are influenced by the few large aggregates and/or clusters that likely occur. The TEM images clearly showed that~20 nm particles were indeed produced (Fig. 2), but in the absence of surfactant, several clustered and may have been measured as one by DLS, explaining the systematic offset of these data compared to those of the TEM measurements (Fig. 2).
Tissue distribution, quantification, and recovery. We have shown previously the use of two sizes of 14 C-nPS in an exposure experiment with king scallops (Pecten maximus) with monitoring of uptake by quantitative whole-body autoradiography (QWBA) 11 to localise the labelled nPS. Without wishing to repeat publish those data, but in order to give further details of the measurement methods, we present additional, unpublished images, from the same experiments herein. Tissue sections were collected in duplicate. The autoradiogram produced from a scallop collected after 6 h exposure to waterborne nominal 14 C-nPS 20 is shown in Fig. 3 and illustrates a tissue distribution typical of the duplicates. Most tissues were labelled (including muscles, gonads, mantle, gills, intestine and kidney), but the hepatopancreas exhibited the highest 14 C concentration.
These and previous data 7 indicate the great advantage of QWBA analysis, viz applicability to the study of anatomical structures, including those which may be very difficult to dissect (e.g. due to their small sizes, their location, or fragility). To achieve such quantitative analyses herein, calibration curves were produced according to a standard operation procedure 11,29 using 20 different screens on which were randomly distributed all the scallop and 14 C standard sections (both in CMC gel sections of 50 µm produced by the same procedure using a cryomicrotome). This corresponds to 200 calibration data points. The digital light intensity was plotted against the 14 C concentrations of the standard spots. The limits of detection (LOD; 0.19 Bq g −1 = 0.28 ng g −1 ≅ 6.4 × 10 7 nPS 20 particles g −1 ≅ 3.3 × 10 4 nPS 250 particles g −1 ) and quantification (LOQ; 0.56 Bq g −1 = 0.85 ng g −1 ≅ 1.3 × 10 8 nPS 20 particles g −1 ≅ 6.5 × 10 4 nPS 250 particles g −1 ) of nPS particles were calculated as corresponding to 3 and 10 times the standard deviations (SD) of the intercepts of the calibration curves, respectively. The phosphor screens used in QWBA are 20 times more sensitive than X-ray films and their linear dynamic range extends over four to five orders of magnitude, allowing capture and quantification of both very low and very high activities, in a single exposure 30 . Quantification data from two scallops per aquarium for three different replicates (i.e. six scallops) for 14 C-nPS obtained by QWBA analyses of the different organs (shown in Fig. 3) are given in Table 1. The use of 14 C-labelled nPS allowed a mass balance and indicated that recoveries of 81-84% were possible (Table 2). We suspect that the remaining 16.0-18.6% of the 14 C-nPS had become incorporated into the faeces or pseudofaeces, or both. One could quantify the amount of particles ending in this compartment, or on the aquarium walls, by collecting them and producing wipes of the walls that could be measured by liquid scintillation counting (LSC).
In summary, a method was developed herein which allowed efficient one-step synthesis of radiolabelled sulfonate end-capped plastic nanoparticles (nPS). Particle synthesis conditions from earlier studies of non-labelled nPS 26,31 were improved in order that only small amounts of expensive isotopically labelled monomers were needed. Also, a better control over nPS sizes was obtained and the method provided 14 C-nPS with a specific activity suitable for studies at predicted environmentally relevant concentrations in the parts per billion range 11,12 . As an illustrative example, data showed that waterborne 14 C-nPS were readily taken up by the scallop, P. maximus. Autoradiograms revealed a high accumulation of the radiolabelled nPS in the digestive gland. Using QWBA, the quantification of uptake, elimination and biodistribution of plastic particles proved possible, as published in detail elsewhere 11 . The use of a radiotracer also allowed characterisation of the biodistributions in small and fragile organs in the scallop, such as the kidney, anus and intestine. Such detailed distribution images will be highly valuable for improving the understanding of the physiological processes governing the bioaccumulation of micro-and nanoplastics in marine bivalves and other organisms and may help to guide further research work. For example, fundamental biotic and abiotic mechanisms of degradation and assimilation of NP may be studied at environmentally relevant concentrations.
In a fourth step (Exp 1.4), the effect of SDS concentration (Δ[SDS]) on particles size were tested at 0.5, 1, 2, 3 wt% with the emulsion system fixed at styrene 1 wt%, water 90.20 wt% and KPS 3.6 mM at 70°C.
Two hundred and fifty nanometre nPS (nPS 250 ) synthesis (Exp 2) is a surfactant-free synthesis. nPS 250 was prepared by dropwise addition of styrene into the NaOH/H 2 O system at room temperature and polymerised in a three-necked flask under a pure nitrogen atmosphere to complete conversion at 70°C. The solution turned white with the formation of nPS when synthesis was complete.
In a first step (Exp 2.1), nPS 250 was prepared by surfactant-free emulsion polymerisation adapted from Telford et al. 26 . In the surfactant-free emulsion polymerisations, styrene 20 wt% and Milli-Q water 59 wt% were mixed at room temperature in a 50 mL round-bottom flask. The flask was sealed with a rubber septum. The headspace inside the flask was then quickly purged with nitrogen for 30 s. The flask was immersed in a silicon oil bath at 70°C for 5 min, with stirring. In the meanwhile, initiator solution was prepared separately with KPS 3 wt% (14.6 mM in the final water phase), NaOH 2 wt% (56.7 mM in the final water phase), and Milli-Q water 17 wt% were mixed in a glass vial. Once the roundbottom flask medium reached 70°C, the initiator solution was injected in the round-bottom flask through the rubber septum, using a syringe. The reaction mixture was vigorously stirred at 70°C overnight.
In a second step (Exp 2.2), to assess the impact of the limiting amount of 14 Cstyrene on particle size, the concentration of styrene (Δ[Sty]) was considered. Concentrations tested were 1, 2, 3, 5 and 20%. The system was fixed with NaOH 2 wt%, water 76 wt% and KPS 14.6 mM at 70°C.
Radiolabeling synthesis of 20 and 250 nm nPS (Exp 3) used the optimised parameters from Exp 1 and 2. For comparison both size target and radiolabeling purpose, styrene proportion was decreased at 1 wt% and the volume of synthesis at 10 mL. Both synthesised, the reactions were mixture was vigorously stirred at 70°C overnight.
Radiolabeled 20 nm nPS ( 14 C-nPS 20 ) was prepared (Exp 3.1) by a conventional emulsion system with styrene 1 wt%, SDS 2 wt%, water 90.20 wt%, and KPS 3.60 mM in a 40-mL vial closed with a rubber cap. The vial was purged with nitrogen gas for 30 s and then heated at 70°C. Once 70°C was reached, 14 C-styrene was a dropwise addition with a syringe. The solution turned white with the formation of nPS when synthesis is successful.
Radiolabeled 250 nm nPS ( 14 C-nPS 250 ) was prepared (Exp 3.2) by a surfactantfree emulsion polymerisations. The initiator solution was prepared separately with KPS 3 wt% (14.6 mM in the final water phase), NaOH 2 wt% (56.7 mM in the final water phase), and Milli-Q water 17 wt% were mixed in a glass vial. Styrene 1 wt% and Milli-Q water 59 wt% were mixed at room temperature in a 40 mL vial. The vial was sealed with a rubber septum. The headspace inside the flask was then purged with nitrogen for 30 s. The vial was immersed in a silicon oil bath at 70°C for 5 min, with stirring. In the meanwhile, Once the vial medium reached 70°C, the initiator solution was injected through the rubber septum, using a syringe.
Purification of PS latexes. All chemicals unreacted were systematically removed by ultrafiltration (exclusion size of membrane: 30,000 g mol −1 ) with a dialysis membrane. Water was frequently changed for 48 h, until the conductivity of water remained constant at a value of <1.5 µS cm −1 . Table 1). After completion of the polymerisation reaction, the l hydrodynamic diameters of the lattices were characterised in water at 20°C. The average sizes and ζ-potentials of the polymer particles were measured using a multi-angle Nicomp ZLS Z3000 (Particle Sizing System, Port Richey, FL). DLS measurements give a value called Z-average size (or cumulative mean), which is an intensity mean, and the polydispersity index. The latex dispersions are optimised for the DLS measurements by diluting the crude, concentrated, reaction product with an excess of Milli-Q water until the count rate of the resulting mixture is within the optimal range. The final concentrations were typically about 1 mg L −1 and can be estimated from the concentration of the primary emulsion. Each sample was analysed at least five times.   Fig. 3). The polymers were characterised by FTIR spectroscopy (Alpha FTIR, Brucker) working in Attenuated Total Reflectance (ATR) mode with a DTGS detector. Spectra were recorded with 16 scans at 4 cm −1 resolution, covering the spectral range between 4000 and 600 cm −1 . FTIR measurements were performed to confirm polymerisation success and library spectral matching was used to identify each product.

DLS, surface charge (ζ-potential) (Supplementary
Yield determinations. To assess synthesis yields, all masses were measured before and after synthesis for the cold synthesis. Comparison with the yields of the radiolabeled synthesis were made by measuring radioactivity before synthesis and after the purification step. Scallop exposure design. In a preliminary experiment aimed at verifying the usefulness of radiolabeled nPS for bioaccumulation studies, a group of 12 king scallops (P. maximus) were exposed to a waterborne 14 C-nPS (0.3 MBq mg −1 14 C, t 1/2 = 5700 y, β-emitter) at 5 kBq L −1 (15 µg L −1 14 C) for 6 h. Seawater (20 L) was filtered on a sand filter. Water temperature was maintained at 12-13°C and kept oxygenated by air bubbling. The salinity of the seawater was 32 psu and the pH was stable at 8.0.
Liquid scintillation counting. 14 C activity was measured by an Hidex 300SL counter at the University of Cambridge, Gordon institute facilities (10 min counting time, minimum detectable activity 0.08 Bq ml −1 ). Synthesis yield was assessed by measuring the 14 C activity at the addition of 14 C-styrene and after ultrafiltration purification (cut-off, 8 kDa).
Quantitative whole-body autoradiography. At the end of the exposure period, scallop soft tissues were removed from the shell and used for QWBA. Briefly, tissues were included in a carboxylmethylcellulose gel and frozen in liquid nitrogen. From the resulting block, 20 pairs of 50-μm-thick sections were sampled at −25°C with a Leica CM3600 cryomicrotome. Sections were freeze-dried for 36 h at -25°C, exposed on phosphor screens sensitive to beta radiation for 1 week and scanned with a Typhoon FLA7000 from GE Healthcare to reveal the autoradiograms. Radiolabel concentrations in tissues were quantified from the images obtained with ImageJ as Digital Light Unit (DLU) per mm 2 and transformed into ng/g using a calibration blood spot on each screen.
Sectioning. Cross-contamination was avoided as there was no direct manipulation of organs and tissues and the only instrument coming into contact with them was the microtome knife, which was carefully cleaned with a brush before every section.

Data availability
All data generated or analysed during this study are included in this published article (and its supplementary information files).