Article pubs.acs.org/est
Surface Charge Controls the Fate of Au Nanorods in Saline Estuaries Justina M. Burns,† Paul L. Pennington,‡ Patrick N. Sisco,§ Rebecca Frey,† Shosaku Kashiwada,∥ Michael H. Fulton,‡ Geoffrey I. Scott,‡ Alan W. Decho,⊥,# Catherine J. Murphy,§,# Timothy J. Shaw,†,# and John L. Ferry*,†,# †
Department of Chemistry and Biochemistry, University of South Carolina, Columbia, South Carolina 29208, United States National Centers for Coastal Ocean Science, Center for Coastal Environmental Health and Biomolecular Research, Charleston, South Carolina 29412, United States § Department of Chemistry, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United States ∥ Department of Life Sciences, Toyo University, Itakura, Oura, Gunma 374-0193, Japan ⊥ Department of Environmental Health Sciences, University of South Carolina, Columbia, South Carolina 29208, United States # Nanocenter at the University of South Carolina, Columbia, South Carolina 29208, United States ‡
S Supporting Information *
ABSTRACT: This work reports the distribution of negatively charged, gold core nanoparticles in a model marine estuary as a function of time. A single dose of purified polystyrene sulfonate (PSS)-coated gold nanorods was added to a series of three replicate estuarine mesocosms to emulate an abrupt nanoparticle release event to a tidal creek of a Spartinadominated estuary. The mesocosms contained several phases that were monitored: seawater, natural sediments, mature cordgrass, juvenile northern quahog clam, mud snails, and grass shrimp. Aqueous nanorod concentrations rose rapidly upon initial dosing and then fell to stable levels over the course of approximately 50 h, after which they remained stable for the remainder of the experiment (41 days total). The concentration of nanorods rose in all other phases during the initial phase of the experiment; however, some organisms demonstrated depuration over extended periods of time (100+ h) before removal from the dosed tanks. Clams and biofilm samples were also removed from the contaminated tanks post-exposure to monitor their depuration in pristine seawater. The highest net uptake of gold (mass normalized) occurred in the biofilm phase during the first 24 h, after which it was stable (to the 95% level of confidence) throughout the remainder of the exposure experiment. The results are compared against a previous study of positively charged nanoparticles of the same size to parameterize the role of surface charge in determining nanoparticle fate in complex aquatic environments.
■
INTRODUCTION Nanomaterials enter the environment through a manifold of processes, including mechanical fracturing/weathering of minerals, biosynthesis, and the degradation of anthropogenic materials.1−4 Aqueous nanoparticle suspensions are generally thermodynamically unstable toward aggregation and precipitation and can be precipitated from aqueous solution by exposure to common anions, clays, or dissolved organic matter.5−7 Recent laboratory and model ecosystem data suggest this instability drives their reactive attenuation (i.e., loss as a result of abiotic molecular processes) during transport and may ultimately limit the risk that they pose.8−16 The manner of reactive attenuation is a function of the type of nanomaterial and the environmental compartment through which the material passes.17,18 It includes processes that degrade nanostructures (e.g., photolysis, oxidative dissolution, etc.) or processes that extend or incorporate nanostructures (e.g., aggregation, adsorption to larger materials, entrainment, etc). It © 2013 American Chemical Society
is notable that these processes are controlled by the nature of the nanomaterial surface in the sense that the surface coating enables intersurface interaction as well as interfacial electron and mass transport.19−21 This implies that building observations of diverse nanoparticle fates into a predictive model of their environmental impacts will require surface characterization as well as information about the size, morphology, and composition of the particle. High urban populations in the coastal areas of the U.S. make saline estuary environments of particular concern for emerging contaminants. Transport of anthropogenic materials to the coastal oceans through estuaries is often mediated by chemical and physical processes associated with mixing fresh water with Received: Revised: Accepted: Published: 12844
July 1, 2013 September 25, 2013 October 21, 2013 October 21, 2013 dx.doi.org/10.1021/es402880u | Environ. Sci. Technol. 2013, 47, 12844−12851
Environmental Science & Technology
Article
Figure 1. Transmission electron micrograph image of the gold nanorods used in the study.
seawater.22−24 Several studies have observed that nanomaterial suspensions are prone to precipitate in conditions of rapidly changing ionic strength, particularly as a consequence of exposure to polyvalent ions.18,25−29 Because estuaries are also the habitat for many commercially and ecologically important shellfish and finfish, they could also be a critical geographical point of nanomaterial entry to the marine food web for filter feeders and bottom feeders.23,30−32 For example, positively charged gold nanorods in model estuarine ecosystems are readily taken up by Mercenaria mercenaria, a species of clam used as a human food source, and also by photosynthetic biofilms.12,31 Crassostrea virginica, a common species of oyster, is known to take up silver and titania nanoparticles.31,33,34 Viruses, which are negatively charged naturally occurring nanoparticles, are frequently mechanically taken up by several species of shellfish during feeding.35−37 Daphnia magna has also been shown to uptake and depurate certain nanomaterials, including fullerenes, gold nanospheres, and carbon nanotubes.38−40 Biofilm communities are also effective at sequestering nanoparticulate materials of synthetic or biogeochemical origin.12,41−43 Au can occur in four different oxidation states, Au(0) (elemental) or Au(I), Au(II), or Au(III).44 The latter three are strong oxidants that are prone to rapid disproportionation or are thermodynamically unstable in the presence of common seawater constituents, including dissolved organic matter and halides. However, Au nanomaterials that are stabilized against precipitation can be stable against oxidation and dissolution for years in aqueous solution.44,45 The chemical stability and low environmental background of Au(0) make it an ideal nanomaterial probe for tracing general nanoparticle fates in biomedical monitoring and model environmental systems.45 This application is particularly important for environmental compartments, such as saline estuaries, where changing ionic strength, oxygen levels, and microbial activity could result in conflation of physical mass balance measurements with oxidative or biodegradation of more reactive nanomaterials. In this study, we report the outcome of a series of replicate (n = 3) nanoparticle exposures in complex mesocosm models of
Spartina-dominated estuarine environments that include multiple trophic levels. Negatively charged, polystyrene sulfonate (PSS)-stabilized gold-core nanorods (48.3 × 9.8 nm) were prepared and purified using well-established methods for excluding residual Au ions.46−48 Gold nanorods were spiked in the aqueous portion of the mesocosms, and their distribution between seawater, natural sediments, cultured biofilms, mature cordgrass, juvenile northern quahog clam, mud snails, and grass shrimp was monitored for approximately 11 days. The singlespike approach was chosen to parameterize episodic environmental releases, e.g., a spill or sudden runoff event. At the end of this period, surviving clams and biofilms were divided between the nanoparticle-treated environments and clean seawater for a 30 day depuration study. Initial nanoparticle concentrations were set to 3.42 × 107 particles mL−1 to correspond to the range of naturally occurring viral particle concentrations in pristine waters.49,50 The low dose was chosen to ensure environmental relevance based on the assumption that the viral loading of natural seawater parameterized the probable low end of summed aqueous biogeochemically and anthropogenically sourced nanoparticles. Therefore, the study was conducted at the conservative end of the naturally occurring nanoparticle range. This study also serves as a parallel to work that the authors reported previously with net positively charged gold nanorods of the same dimensions; a comparison of the outcomes of both provides a test of the hypothesis that net surface charge characteristics are determinant for the environmental distribution of nanomaterials.12
■
EXPERIMENTAL SECTION Materials. All purchased materials were used as received. All acids (Optima grade) were obtained from Fisher Scientific. Chloroauric acid (HAuCl 4·3H 2O), sodium borohydride (NaBH4), silver nitrate, ascorbic acid, poly(sodium-4-styrenesulfonate), and poly(diallyldimethylammonium chloride) (PDADMAC) were obtained from Aldrich (99%). Hexadecyltrimethylammonium bromide (CTAB) (99%) was obtained from Sigma Chemicals. Gold and iridium standards (99.9+%) 12845
dx.doi.org/10.1021/es402880u | Environ. Sci. Technol. 2013, 47, 12844−12851
Environmental Science & Technology
Article
Figure 2. Series of replicate estuarine salt marsh mesocosms (volume, 366 L each) modeling the edge of a tidal marsh creek maintained in a greenhouse. Each tank contained natural seawater, sediment, biofilms, S. alterniflora, P. pugio, Cyprinodon variegatus, I. obsoleta, and M. mercenaria and received a single dose of negatively charged gold nanorods (48.3 × 9.76 nm). The exposure period was 11 days with monitoring throughout. Aerial portions of S. alterniflora are visible, growing in the sediment trays.
the original synthesis were volumetrically diluted in excess of 1018 times during this process, resulting in dissolved Au(III) concentrations well below the detection limit for this study (5.96 pg kg−1). Mesocosm Experiments. Salt marsh mesocosms were maintained in a greenhouse at the Center for Coastal Environmental Health and Biomolecular Research (CCEHBR) in Charleston, SC (Figure 2). Three separate mesocosms were constructed and charged with seawater, sediment, flora, and fauna in accordance with procedures outlined in National Oceanic and Atmospheric Administration (NOAA) Technical Memorandum NOS NCCOS 62.54 The systems were made from natural, unfiltered seawater collected at Cherry Point Boat Landing on Wadmalaw Island, SC (salinity determined by conductivity and adjusted to 20 ppt by the addition of deionized water). A periodically submerged sediment tray was in the primary tank with an attached reservoir for water storage (isolated by a screen) to simulate a semi-diurnal tidal cycle (high tide at 10:00 and 22:00 h and low tide at 4:00 and 16:00 h).54 The “tides” were driven by submersible pumps set with timers. Sediment trays were charged with intertidal estuarine sediments (top 2−4 cm) that were collected from a reference site located on Leadenwah Creek (32° 38.848′ N, 080° 13.283′ W), Wadmalaw Island, SC. Sediments were sieved with a 3 mm sieve, homogenized, and dispensed into sediment trays. Four sediment trays were placed into each tank and elevated ∼5 cm to allow for drainage at low tide. Each sediment tray contained ∼12.3 kg of sediment. A shallow sediment tray (1 cm) was set on the floor of the primary tank to serve as a Mercenaria mercenaria habitat (100 juveniles). Sediments were planted with Spartina alterniflora (cordgrass), and the primary tank was seeded with 50 Ilyanassa obsoleta (mud snails) and 100 Palaemonetes pugio (grass shrimp) collected wild from the
were acquired from High Purity Standards (Charleston, SC). Solutions were made in Barnstead E-pure (18 MΩ cm−1) water. Nanoparticles. Gold nanorods (48.3 × 9.8 nm; relative standard deviation in the aspect ratio of 5.5%; Figure 1) were grown, purified, and characterized according to published procedures developed in the Murphy group.47,48 Particles were purified before use by repeated centrifugation at 4000 rpm. At the conclusion of each centrifugation step, the supernatant was discarded and a 100 μL pellet of particles was retained. This pellet was resuspended in 150 mL of ASTM-grade water. The process was repeated 2 additional times. Resuspended nanorods were present as single particles and stable in solution.47,48 The rods were then derivatized with PSS as follows: aliquots (30 mL) of as-prepared gold nanorods were added to centrifuge tubes and centrifuged for 20 min.51,52 A 100 μL pellet of gold nanorods was retained at the bottom of the centrifuge tubes, and the supernatant was slowly removed without disturbing the pellet. The pellet was dispersed in 15 mL of ASTM-grade water. A total of 1.5 mL of 10 mM NaCl and 3 mL of PSS (10 mg/mL) stock solution were added to the suspension simultaneously. After 30 min, the excess polymer in the supernatant fraction was removed by centrifugation for 20 min. The resulting pellet was dispersed in 15 mL of 18 MΩ deionized water, and the process was repeated 2 more times with PDADMAC first then again with PSS. This triple coating of gold nanorods has been shown to reduce potential cytotoxicity.53 The charge on the nanorods (−53.50 ± 1.25 mV; n = 10; pH 7; no added ionic strength adjuster) was measured using a ZetaPALS ζ potential analyzer from Brookhaven Instruments Corporation. Each nanorod was composed of ∼500 000 gold atoms. The final concentration of purified nanorod stock solution was 4.16 × 10−10 M nanorods (2.08 × 10−4 M total Au). Residual Au(III) ions from 12846
dx.doi.org/10.1021/es402880u | Environ. Sci. Technol. 2013, 47, 12844−12851
Environmental Science & Technology
Article
shell material was analyzed. Once the samples were mineralized, they were heated to dryness and redissolved in HNO3 (Optima grade). Samples were spiked with an Ir internal standard and diluted with 2% HCl (Optima grade) for immediate analysis on the HR-ICP−MS.58 Previous work with marine samples indicated that there were no statistically significant losses of total Au during this process.12
Leadenwah Creek reference site. Microscope slides were hung from a line across the top of the tanks to culture easily removable biofilms. The mesocosms (seawater, sediments, and Spartina) were established and allowed to equilibrate for approximately 3 months before exposure. All fauna were added to the mesocosms 5−10 days before the start of exposure. One tank was monitored continuously (n = 50 per day) for several water-quality parameters, including temperature (°C), pH, dissolved oxygen (mg L−1), and salinity, while for the remaining tanks, these parameters were measured on sampling days. Salinity was held constant with a collective average of 20 ppt. The other parameters varied diurnally in accordance with daytime heating and photosynthetic activity; however, these differences were within the established norms for this system (see Figures S1−S4 of the Supporting Information).54−56 Exposure Studies. Each of the mesocosms was dosed independently. Nanorod stock solution (50.0 mL) was introduced by means of gravity feed through a Teflon capillary into the inlet line of each mesocosm during the rising tide cycle to ensure the best possible mixing. The addition required less than 1 min. Previous work has shown that elemental gold [Au(0), the oxidation state of Au in this study] is stable against oxidative dissolution in aerated seawater, ensuring that measured gold was particulate in this study.44,57 The initial theoretical particle loading in the aqueous phase of each 366 L tank housing each mesocosm unit was 3.42 × 107 particles mL−1 (5.60 μg of Au kg−1), assuming for the definition that all particles were instantaneously dispersed in the aqueous phase only. Experiments were conducted from mid-October to the end of November, 2009. Depuration Studies. After 11 days of exposure in the spiked mesocosms, some of the remaining clams and biofilm slides were removed and placed in clean tanks of seawater (10 L tanks, Cherry Point Landing, vide supra). The seawater in these tanks was replaced with new, unspiked seawater once a week for 4 weeks. Samples of the clams, biofilms, and water were taken on a weekly basis over 30 days in parallel with comparative samples from the spiked mesocosms. Samples were withdrawn immediately prior to water replacement. Sampling and Measurement. Samples were withdrawn at various time intervals throughout the experiment and frozen in a −80 °C freezer for later workup and analysis.12 All samples were drawn with n = 3/mescosm; all standard deviations on Au measurements refer to pooled standard deviations for the three replicate mesocosms. Aqueous, sediment (top 2 cm), and organic phases of the mesocosm were monitored for Au throughout the initial course of the exposure experiment (266 h), which was sufficient time to detect partitioning and mortality but not multi-generational effects in the test organisms. Over the time period of these experiments (41 total days for combined exposure and depuration studies), no overt test organism mortality was observed. Water samples were diluted in aqua regia and spiked with an iridium external standard for immediate analysis on a ThermoFinnigan Element XR high-resolution inductively coupled plasma mass spectrometer (HR-ICP−MS).58 Sediment samples were dried, finely ground, and weighed into a Teflon digestion vial. All dried sediment and organic samples (i.e., plant material, biofilms, etc.) were digested through a series of Optima-grade concentrated HF−HCl−HNO3 additions dispensed using an acid-washed Teflon bottle top dispenser (Seastar Chemicals, Inc.) and heated at 140 °C in Teflon vessels on a customized heated reaction block from J-Kem Scientific. No calciferous
■
RESULTS Water-quality parameters were constant across all mesocosms over the course of the experiment. Gold concentration in the seawater rose rapidly after nanorod addition from (4.76 ± 0.16) × 10−2 μg kg−1 to a measured maximum of 1.10 ± 0.22 μg kg−1 in the aqueous phase (Figure 3). The measured maximum gold
Figure 3. (A) Concentration of Au in seawater in the salt marsh mesocosms monitored over the course of the 11 day experiment and the following 30 days of depuration studies. Error bars correspond to the pooled standard deviations for three replicate samples drawn from each of the three mesocosms (n = 9). (B) (Inset) aqueous concentration of gold stabilized within 50 h of nanorod addition.
was 20% of the theoretical maximum value (vide supra), suggesting that Au nanoparticles initially partitioned from the aqueous phase very quickly, on the time scale of minutes. Total gold in suspension fell continuously from this maximum over the course of the next 50 h, after which it was maintained near the detection limit for the remainder of the experiment. Sediment samples (top 2 cm) were also analyzed for gold. The gold concentration in the top 2 cm of the sediment plateaued within 20 h and then remained constant at a mean of 14.80 ± 0.50 μg kg−1 for the remainder of the experiment (see Figure S5 of the Supporting Information). Water and sediment combined accounted for 99+% of the total mass of the mesocosm systems and were significant “sinks” for nanoparticles, responsible for 79.4% of the total mass of particles recovered (Table 1). The aqueous phase was the only compartment continuously in contact with all other phases during the experiment. Therefore, a descriptive index for the environmental fate of Au nanoparticles was developed on the basis of normalizing the per mass concentration of Au in a given phase over that in the aqueous phase at a fixed time. The index, a concentration factor (Cf), was operationally defined as the ratio of the concentration of gold (μg kg−1) in the measured phase over the concentration in the aqueous phase (μg kg−1) at the time of measurement (t = 11 days).12 Biofilms are significant sinks for naturally occurring and synthetic aquatic nanomaterials.8,42,59 Natural microbial populations in the mesocosms colonized multi-species biofilms on suspended glass slides that were periodically removed and 12847
dx.doi.org/10.1021/es402880u | Environ. Sci. Technol. 2013, 47, 12844−12851
Environmental Science & Technology
Article
Table 1. Distribution of Gold in Estuarine Mesocosms after Aqueous Introduction phase sea water sediment biofilm S. alterniflora, aerial S. alterniflora, stems S. alterniflora, roots P. pugio, grass shrimp I. obsoleta, snail M. mercenaria, juvenile clams
wet mass in mesocosm (g)
average percent water
[Au] (μg kg−1) 0 daysa
366000d 49140d 1009e 450e 450e 600e 15.6d 5.5e 10e
100 30.9 ± 5.27 93.6 ± 2.54 54.6 ± 12.2 69.0 ± 8.38 90.0 ± 12.2 52.1 ± 20.0 57.5 ± 12.2 88.6 ± 2.79
0.05 ± 0.15 3.66 ± 1.59 7.00 ± 1.29 1.96 ± 3.52 2.47 ± 1.58 1.18 ± 4.20
