Water Research 158 (2019) 392e400 lable at ScienceDirectContents lists avaiWater Research journal homepage: www.elsevier .com/locate/watresBiofilm-enhanced adsorption of strong and weak cations onto different microplastic sample types: Use of spectroscopy, microscopy and radiotracer methods Mathew P. Johansen a, *, Tom Cresswell a, Joel Davis a, Daryl L. Howard b, Nicholas R. Howell a, Emily Prentice a, c a Australian Nuclear Science and Technology Organisation (ANSTO), Locked Bag 2001, Kirrawee DC, NSW, 2232, Australia b Australian Synchrotron, 800 Blackburn Road, Clayton, Victoria, 3168, Australia c University of Technology Sydney, 15 Broadway, Ultimo, NSW, 2007, Australiaa r t i c l e i n f o Article history: Received 5 October 2018 Received in revised form 29 March 2019 Accepted 13 April 2019 Available online 15 April 2019 Keywords: Microplastic Biofilm Adsorption Kd Radionuclides Microscopy* Corresponding author. E-mail address: Mathew.johansen@ansto.gov.au (M https://doi.org/10.1016/j.watres.2019.04.029 0043-1354/Crown Copyright © 2019 Published by Elsa b s t r a c t The adsorption of metals and other elements onto environmental plastics has been previously quantified and is known to be enhanced by surface-weathering and development of biofilms. However, further biofilm-adsorption characterisation is needed with respect to the fate of radionuclides. This study uses spectroscopy, microscopy and radiotracer methods to investigate the adsorption capacity of relatively strong and weak cations onto different microplastic sample types that were conditioned in freshwater, estuarine and marine conditions although marine data were limited. Fourier-transform infrared spec- troscopy confirmed that surface oxidation chemistry changes induced by gamma irradiation were similar to those resulting from environmental exposures. Microscopy elemental mapping revealed patchy bio- film development, which contained Si, Al, and O, consistent with microbial-facilitated capture of clays. The plasticsþbiofilm of all sample types had measurable adsorption for Cs and Sr radiotracers, suggesting environmental plastics act broadly as a sink for the key pervasive environmental radionuclides of 137Cs and 90Sr associated with releases from nuclear activities. Adsorption onto high-density polyethylene plastic types was greater than that on polypropylene. However, in most cases, the adsorption rates of all types of plasticþbiofilm were much lower than those of reference sediments and roughly consistent with their relative exchangeable surface areas. Crown Copyright © 2019 Published by Elsevier Ltd. All rights reserved.1. Introduction The increasing abundance of plastics in global aquatic systems (Barnes et al., 2009; Turner and Holmes, 2015), including weath- ered derivative fragments of various sizes and chemical composi- tion (Cole et al., 2011; Gregory, 2009), has led to consideration of plastics as a pervasive media that influences the fate and behaviour of aquatic contamination (Ivar do Sul and Costa, 2014; Wang et al., 2016). Similar to other particulate media, plastics are of interest as vectors for transfer of contaminants among ecosystem compart- ments including living organisms (Browne et al., 2007; Farrell and Nelson, 2013; Rochman et al., 2013; Tanaka et al., 2013) although the flux to organisms from this pathwaymay be low relative to that.P. Johansen). evier Ltd. All rights reserved.from water and food uptake (Bakir et al., 2016; Koelmans et al., 2016; Ziccardi et al., 2016). Such transfer is often facilitated or influenced by contaminant-to-particulate adsorption and under- standing the role of plastics as an agent of transfer requires basic data collection on such adsorption as well as insights into the abiotic and biotic processes influencing adsorption capacities. Upon entry into the aquatic environment, bacteria, algae, fungi, and protozoa colonise plastic surfaces relatively rapidly (Lobelle and Cunliffe, 2011). Microorganism colonisation on plastics differs from that of autochthonous substrates (Zettler et al., 2013) and also differs among plastic types (Eich et al., 2015). This colonisation forms a basis for adherence of additional organic and inorganic suspended matter to form biofilms (Mora-Gomez et al., 2017; Wetzel, 1993), which are known to accumulate metals (Dranguet et al., 2017; Eich et al., 2015; Lavoie et al., 2012; Lehtola et al., 2004; Zettler et al., 2013) typically more so than the underlying plastics (Ashton et al., 2010). M.P. Johansen et al. / Water Research 158 (2019) 392e400 393Radionuclides released into aquatic systems are among a suite of contaminants of concern some of which are anthropogenically- sourced and pervasive such as plutonium isotopes, 90Sr and 137Cs (Fisher and Chen, 2011; Stark et al., 2017; Whicker, 1982). Approximately 950 PBq of 137Cs was deposited in world oceans in past decades from nuclear weapons testing (Buesseler, 2014), with additions of ~85e100 PBq from the Chernobyl, and ~12e15 PBq from the Fukushima accidents (Aoyama et al., 2016; Johansen et al., 2015). Because radioactive and stable isotopes of the same element tend to behave chemically similarly in environmental and biolog- ical processes (Whicker, 1982), it is probable that radionuclides are adsorbing to, and being transported by microplastics on a global scale. There are, however, sparse investigations of radionuclide interactions with environmental plastics. Tazaki et al. (2015) detected elevated 137Cs on the surfaces of plastics recovered from a freshwater lake contaminated by fallout from the Fukushima accident. Some limited data on adsorption of radionuclides on one type of microplastic was provided in a precursor to this paper (Johansen et al., 2018). Lacking are basic studies on the adsorption of key radionuclides on various environmental plastic types and data that relates such adsorption to surface weathering and biofilm development. The objective of this study was to quantify the adsorption onto differing types of environmentally-conditioned aquatic micro- plastics of relatively strong (Cs) and weak (Sr) cations representing common radioactive contaminants (137Cs and 90Sr isotopes). We varied plastic composition (high-density polyethylene, HDPE and polypropylene, PP), physical form (microspheres and 50 mm thick strips), as well as aquatic ecosystem type (freshwater, estuarine and marine). The study also aims to develop the use of radiotracers for estimating bulk adsorption along with the application of multiple spectroscopy and microscopy methods to provide insights on sur- face weathering and biofilm development. Such nuclear techniques that have been identified as valuable tools in environmental microplastics research (Lanctôt et al., 2018). 2. Materials and methods Three plastic sample types were considered: HDPE 100 mm mi- crospheres (90e106 mmdiameter, 0.96 g cc1 density, sourced from Cospheric LLC), and 50 mm thick strips that were microtomed from field-collected HDPE and PP samples. The nominal surface area per gram of the spheres and strips were approximately 625 and 440 cm2 g1 respectively, neglecting surface topography. The ~100 mm HDPE microspheres may provide for a standardised basis of plastic type and size to which future studies may compare. Some limited data on the HDPE microspheres were reported in Johansen et al. (2018). The 50 mm “microstrips,” of HDPE and PP plastic types, were added to the present study as many plastics appear in fibrous forms. As the previous studies showed biofilm-enhanced adsorp- tion is complex, it was desired to study the microspheres and microstrips of varying plastic forms and types using multiple radiotracer adsorption, spectroscopy and microscopy methods. The study design sought to gain information on freshly manu- factured plastic surfaces vs those weathered through environ- mental exposure. The freshly-manufactured HDPE microspheres were pre-conditioned by gamma-irradiation (1.17MeV, 60Co source) to simulate the surface properties of environmentally exposed plastics. Attenuated Total Reflectance Fourier-transform Infrared Spectroscopy (ATR-FTIR) spectra were collected using a Bruker Alpha FTIR with the platinum ATR module. Measurements on the microspheres, taken before and after irradiation, showed an increase in absorbance at 1712 cm1 (Fig. 1). This peak, within the characteristic carbonyl stretch (1780-1680 cm1), was alsomanifest on the external surfaces of comparative HDPE environmentally-weathered field samples. In contrast, the peak is absent in the spectra from freshly-exposed (no environmental weathering) in- ternal surfaces of the same plastics (Fig. 1). Our results are similar to those achieved byMartinez-Romo et al. (2015) in that an irradiation protocol was successful in imitating a major component of the oxidation chemistry that occurs on HDPE as a result of long-term environmental exposure. In order to develop biofilms over a range of aquatic environ- ments, plastics were deployed at sites in freshwater (Cs< 0.1, Sr¼ 91 mg L1), estuarine (Cs¼ 0.65, Sr¼ 4190 mg L1, salinity measured at high water of 23 ± 3mg L1), and marine (35mg L1 salinity). These sites were associated with the Georges River sys- tem, which has a catchment containing natural bushland, resi- dential and light industry, and flows to the southern Pacific Ocean on the eastern coast of New South Wales, Australia. At each site, duplicates of <1.5 g (dry weight) of plastics were deployed within nylon mesh bags with 10 mm apertures. The mesh bags were housed within 250ml polycarbonate centrifuge tubes, which pro- vided protection from disturbance by large organisms while allowing circulation of water and suspended sediment via numerous 3mm circular openings. Subsamples of the deployed plastics were gathered after 19 days, rinsed to remove unattached material, visually inspected to confirm the absence of inadvertently-captured material, and mixed with OCT compound (Tissue-Tek, Sakura Finetek Europe B.V.). The compound with plastics was frozen in liquid nitrogen, cryosec- tioned into 25e50 mm thin sections and thenmounted onto 200 nm silicon nitride windows (1.5 1.5mm). These thin sections allowed interrogation of the freshly-microtomed inner plastic material as well as the outer edges that had been exposed to sunlight andwater in conditions conducive to the development of biofilms. Sub- samples of these plastics were analysed by ATR-FTIR as described above. The sectioned plastics were also imaged using X-ray Fluo- rescence Microscopy (XFM at 16.5 KeV) at the Australian Synchro- tron (Paterson et al., 2011), to create elemental maps of the deposits that had developed on the outer microplastic surfaces during deployment. Wholemicrospheres, with surface biofilm intact, were analysed using scanning electron microscope-energy dispersive X- ray spectroscopy (SEM-EDS) coupled with electron backscattered imaging to determine elemental composition and surface charac- teristics. Approximately 50 Å of carbon was evaporated onto the surfaces under vacuum to prevent charging. SEM analysis was performed using a Zeiss Ultra Plus with an attached Oxford In- struments X-Max 80mm2 silicon drift detector (SDD) X-ray microanalysis system at an accelerating voltage of 15 kV. Oxford Aztec software was used to perform the semi-quantitative elemental analysis on the particles after quant optimisation using copper. All results are in weight percent using processing option “all elements analysed (Normalised),” and are relative to the ele- ments detectable by SEM-EDS. The adsorption capacity of two different cations was accom- plished using 134Cs and 85Sr radiotracers and the remaining plastic samples which were environmentally-conditioned for a total of 142 days (26 February - 18 July 2016). The Cs and Sr radiotracers differ in their relative adsorption strengths (freshwater-sediment parti- tioning coefficient geometric means of >1E04 and <1E03 respec- tively and referred to as relatively strong and weak for the purposes of this paper). After the 142-day deployment period, the freshwater and estuarine samples were retrieved for radiotracer testing. While the marine plastics had been sub-sampled at 19 days and used for FTIR and XFM testing, they had subsequently been removed from the field by persons unknown and were not available for the adsorption experiment. Approximately 10 kBq of 134Cs and 85Sr were added to duplicate 50ml polycarbonate tubes with filtered (0.45 mM) host water from each deployment site and knownmasses 394 M.P. Johansen et al. / Water Research 158 (2019) 392e400 M.P. Johansen et al. / Water Research 158 (2019) 392e400 395 Table 1 Retention of 134Cs and 85Sr tracers on plastics with biofilm after 48 h of exposure to radiotracers following 142 days of deployment in freshwater and estuarine conditions. Retained tracer (Bq Kd (mg l1) g1)a mean min max mean SD min max Freshwater HDPE microspheresb 134Cs 6800 6437 7163 80.3 1.6 79.2 81.4 85Sr 829 730 928 6.9 1.3 6.0 7.8 HDPE strips 134Cs 5689 4786 6592 41.9 16.3 30.4 53.4 85Sr 748 375 1122 6.1 4.5 2.9 9.3 PP strips 134Cs 1502 1201 1802 8.4 2.5 6.6 10.2of each plastic type. As most adsorption of metals on microplastics appears to occur in the approximate first 30 h (Holmes et al., 2012; Turner and Holmes, 2015), the experimental tubes were gently agitated for 48 h, after which the solution was separated from plastics via vacuum filtration (0.45 mm glass fibre). Subsamples of the filtrate were analysed via gamma spectrometry, and the activity concentration for each radionuclide was compared with that from four control treatments that did not contain plastics. Calculations of the distribution coefficients (Kds) provided the ratio of the mass activity density of the solid phase to the volu- metric activity density of the liquid phase, expressed in units of ml g1 (Environment Agency UK, 2005; IAEA, 2010). Additional details on materials and methods can be found in the Supplementary information.85Sr 129 ND3 343 1.6 1.5 ND3 2.7 Estuarine HDPE microspheresb 134Cs 1644 512 2776 9.3 9.2 2.7 15.8 85Sr 623 194 1053 5.0 4.9 1.5 8.4 HDPE strips 134Cs 2433 1892 2975 14.0 5.5 10.1 17.9 85Sr 1619 1227 2012 14.0 5.9 9.8 18.2 PP strips 134Cs 647 593 702 3.6 0.5 3.2 3.9 85Sr 355 291 420 2.9 0.8 2.3 3.5 a The mass of the plasticsþbiofilm used here reflects damp (not completely dry) conditions in order to maintain algae and other living components of the biofilm as viable during the experiment. b Kd data for HDPE microspheres from Johansen et al. (2018). ND refers to non- detection (see text).3. Results and discussion 3.1. Retention of radiotracers on differing types of plasticsþbiofilm The adsorption of 134Cs and 85Sr radiotracers onto the environmentally-conditioned plasticsþbiofilm was measurable for all plastic types. Our adsorption data for Cs and Sr, alkali and alkaline-earth metals, (K 1ds of 1e81ml g ) are comparable to those for the range of transition metal adsorption on virgin and weath- ered plastics (Kds <1e221ml g1) determined previously (Holmes et al., 2012; Turner and Holmes, 2015). The greatest observed adsorption value was for 134Cs on HDPEmicrospheres in freshwater conditions (Table 1), and 134Cs adsorption in freshwater was significantly greater than for estuarine conditions for all plastic types, (p-value¼ 0.03, Mann-Whitney U test, one-tailed). 134Cs adsorption in estuarine/marine conditions is known to be inhibited by other cations, primarily Kþ, competing for adsorption sites (Evrard et al., 2015). Similar to our results, greater adsorption in freshwater, as compared with marine conditions, was observed in recent studies with Cd, Co, Ni, and Pb, which also used tracer- adsorption testing methods (Holmes et al., 2012; Turner and Holmes, 2015). Themean Sr Kds ranged from 1.0 to 14.0mg L1 in contrast to the higher Cs range of 3.6e80.3mg L1. While the 85Sr adsorption was typically less than that of 134Cs, it was, however, measurable for all conditions and plastic types (Table 1). Sr was notmarkedly different between freshwater and estuarine conditions. The least of the observed mean adsorption values was for Sr associated with PP 50 mm strips in freshwater conditions (129 Bq g1, Kd of 1.0mg L1). A single non-detect value for one of the PP subsample replicates, in freshwater, is reported in Table 1 (minimum for the Freshwater- PP-85Sr combination). It indicates the adsorption of 85Sr onto microplastics in that particular subsample was small and generally within the same range as the variation that occurred in adsorption to study-container walls. However, the non-detect was for only one subsample, and the mean of that sample type/condition indicated positive adsorption. Among plastic types, mean adsorption on PP strips was less than that of HDPE samples by factors of 2e9. Rochman et al. (2014) measured adsorption of a range of metals onto environmental PP that was both lower and higher than that of HDPE depending on conditions. However, no directly comparable previous data for adsorption of radionuclides onto different types of plastic were available. As a comparative baseline, we estimated adsorption on unweathered, non-biofilm PP surfaces to be approximatelyFig. 1. Characteristic Infrared Spectroscopy spectra obtained from the internal (A) and ext polymer of the HDPE environmental sample; the microspheres before (E) and after (F) th interpretation of the references to colour in this figure legend, the reader is referred to the1.0e1.6% that of the weatheredþbiofilm conditioned plastics (using the experimental range of solution activity concentrations and nominal area cm2 of our PP ~50 mm of this study along with a 6E- 04% solution cm2 adsorption rate of 137Cs onto clean PP slides at neutral pH reported in Eichholz et al. (1965)).3.2. Biofilm coverage and composition on plastic surfaces In several previous studies, the authors attributed most adsorption of cationic metals to surface-attached matter, as well as surface modification by photooxidation (Ashton et al., 2010; Holmes et al., 2012; Turner and Holmes, 2015). Themicroscopy data of this study revealed higher proportions of Si, Al, and O co-located in heterogeneous patches consistent with the accumulation in biologically-facilitated deposits (Fig. 2). These patchy areas were interspersed with mostly bare, or sparsely-colonised areas that still had adsorbed mineral components, but at much lower concentra- tions (e.g. Al reduced from24% to 12% in freshwater and 21%e13% in estuarine conditions for biofilm-dominated vs mostly bare areas, see supplemental). Using a SEM-EDS summary long-count (12 h) on an estuarine-conditioned plasticþbiofilm surface, O was most abundant (~47%, which was in part due to plastic) followed by Si (~21%), Al (~14%) Fe (~9%) with <2% of Ca, Cl, K, Mg, Mn, Na, P, S, and Ti. This element-abundance profile is consistent generally with that of clays, most of which are dominated by SiO2, Al2O3 with lesser amounts of Fe2O3, and often include trace elements such as Ti (IAEA, 2013). Fe and Ti were present in surface deposits of the study samples, primarily in discrete concentrations that were typically located within the same deposits containing SieAleO (SEM-EDS and XFM data, Figs. 2 and 3). Visual inspection confirmed these patches includedernal (B) polymer of the PP environmental sample; the internal (C) and external (D) e irradiation protocol. The characteristic carbonyl stretch is highlighted in red. (For Web version of this article.) 396 M.P. Johansen et al. / Water Research 158 (2019) 392e400 Fig. 2. SEM-EDS elemental mapping on the curved surface of a single 100 mm HDPE microsphere with biofilm and adhered mineral elements after deployment in estuarine conditions. Colours are artificial to help distinguish among different elements. BSE refers to backscattered electron image which provides an overall rendering of the sample surface. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Fig. 3. Bromine, iron, sulphur and Compton scatter XFM maps of thin-sectioned (25e50 mm) multiple 100 mm HDPE microspheres after deployment in estuarine conditions.agglomerated masses dominated by frangible weakly bound fine- grained material (consistent with clays) along with occasional crystalline-structured grains and fibrous organic material. Some adhered discrete particles had variable elemental makeup that suggests anthropogenic origin. The results provide insights into the elemental composition of the biofilms, however quantification of the total surface area added by the biofilm among plastic types and deployment conditions has yet to be achieved as it requires a sys- tematic method of measuring biofilm composition alongwith three dimensional structure at