Introduction

Engineered nanoparticles (NPs) are now produced and used worldwide and across an increasingly broad range of applications. The apparent ubiquity of NPs increases their probability to come in contact with living organisms, therefore, raising concerns on the potential effects of these emerging pollutants. The anticipated adverse biological effects would likely occur at the “NP-organism” interfaces, with subsequent implications at both the organism and ecosystem levels (Nel and Xia 2006; Auffan et al. 2008; Kim et al. 2010a, b; Mostafa et al. 2011; Chen et al. 2011; Fang et al. 2011; Gao et al. 2011; Ma and Lin 2013).

Currently, metallic iron-based remediation technologies are taking advantage of the peculiar properties of NPs to remediate environmental compartments contaminated with organic (e.g., chlorinated organic solvents) and/or inorganic (e.g., arsenic, chromium and mercury) pollutants (Cornell 1996; Dixit and Hering 2003; Ponder et al. 2000; Cao and Zhang 2006; Sun et al. 2006; Li and Zhang 2006; Olegario et al. 2009; Vernon and Bonzongo 2014). In particular, nano-zero-valent iron particles, referred to later herein as Fe0-NPs, have been suspended in aqueous slurries for injection in subsurface environments, in the direct vicinity of polluted sections of aquifers (Huang et al. 2008; Johnson et al. 2009). Fe0-NPs have also been evaluated in wastewater treatment for nitrogen removal through chemical reduction of nitrate (Shin and Cha 2008; Hwang et al. 2012), as well as for phosphate removal through chemical precipitation (Shin and Cha 2008). For such applications, it has become a common practice to use stabilizing agents to obtain ideal NP suspensions and maximum sought effects (Pisanic 2007, Zhao et al. 2016). Example stabilizing agents commonly used to prepare iron-based NPs suspensions include polyethylene glycol (PEG) (Pisanic 2007, Gupta and Curtis 2004; Zhang et al. 2004), polyethyleneimine (PEI) (Chertok et al. 2010; Yang et al. 2011) in biomedical research, and polyvinylpyrrolidone (PVP) in environmental remediation studies (Saleh et al. 2008; Phenrat et al. 2006; Sakulchaicharoen et al. 2010). Besides the intentional introduction of NPs to the environment for remediation purposes, it is anticipated that the manufacturing, use, and disposal of NP-based products would likely result in environment pollution as well. Once introduced to natural systems, NPs could undergo transformations through both biotic and abiotic reactions. However, the impacts of such reactions on the bioavailability and potential toxicity of NPs remain largely unknown (Ma and Lin 2013). Accordingly, one potential issue related to the environmental application of nanotechnology which remains poorly addressed in current literature is the fate and the biological impacts of NP-pollutant complexes formed during remediation processes. In this study, a comparative approach is used to investigate the biological effects of three iron-based NPs (Fe3O4, γ-Fe2O3, and Fe0) on a primary producer and a primary consumer. The primary producer used in the different dose-exposure experiments is Raphidocelis subcapitata, previously known as Pseudokirchneriella subcapitata, and will be referred to in this paper as R. subcapitata. Daphnia magna was used as example primary consumer. Tested NPs were used individually in bioassays as (i) water-suspended NPs, (ii) PVP-suspended NPs, and (iii) PVP-suspended NP-arsenic complexes.

Materials and methods

The following is a brief description of the different methods used in this study. A detailed description of used experimental methods and analytical techniques can be found in our earlier publication (Gonzalo et al. 2014).

Materials

Fe3O4 and γ-Fe2O3 NPs used in this study were purchased from Inframat® Advanced Materials LC (Connecticut, USA) with particle size ranging from 20 to 50 nm as reported by the manufacturer. The metallic iron (Fe0) NPs used were purchased from Quantum-Sphere (California, USA), and had an average particle size of 40 nm. PVP was purchased from Sigma-Aldrich Chemical Co. (Missouri, USA). Arsenic pentoxide (As2O5) purchased from Alfa Aesar (Massachusetts, USA) was used to prepare solutions of different arsenate (ASV) concentrations. The freshwater green algae, R. subcapitata was obtained from Microbiotests, Inc. (Denmark), while the freshwater invertebrate D. magna was purchased from “DAPHTOXKIT-FTM Magna” (Ghent, Belgium). All chemicals were reagent grade and used without further purification. NP suspensions were freshly prepared in 18.3 MΩ Milli-Q SP ultrapure water and characterized prior to use in all experiments.

Preparation and characterization of iron-based NP suspensions

Stock suspensions of Fe3O4, γ-Fe2O3, and Fe0 NPs were prepared by suspending 50 mg of each of the three iron-based NPs in 50 mL of ultrapure water, followed by a 10 min ultrasonication using a Sonicator 3000 (Misonix, New York, USA). In addition to water-suspended NPs, Fe3O4, γ-Fe2O3, and Fe0 NP suspensions were also prepared in a 5 wt% PVP aqueous solution and dispersed by ultrasonication as described above. The particle size distribution of the prepared NP suspensions and the ζ-potential of suspended NPs were determined using the Zetasizer Nano Instrument (Malvern Zetasizer). All measurements were carried out at 25 °C. About 1 mL of the sample suspension was transferred to a light-scattering cell and the temperature controlled within ±0.02°C. The analyses were started 10 min after the sample container was placed in the DLS instrument to allow the temperature to equilibrate. DLS measurements were conducted by performing 500 rounds of measurements in particle number mode. The ζ-potential was derived from electrophoretic mobility by applying the Henry equation and the Smoluchowski approximation (Gregory 2006). Finally, the specific surface area (SSA) was calculated from Eq. 1 (Macé et al. 2006), and where ρ is the density (g/cm3) and d is the mean diameter (cm) of the suspended iron-based NP.

$$SSA = \frac{{\pi \times d^{2} }}{{\rho \times \frac{\pi }{6} \times d^{3} }}.$$
(1)

Arsenic sorption onto iron-based NPs suspended in either R. subcapitata or D. magna-specific growth media

Fe3O4, γ-Fe2O3, and Fe0 NPs initially suspended in 5 wt% PVP aqueous solutions were added to growth media specific to R. subcapitata (Table SI-1) and D. magna (Table SI-2). The produced NP-containing growth media were then spiked with As, and then mixed continuously during the adsorption experiment using a roto-shaker®. The sorption capacity of each tested iron-based NP was determined in culture media without organisms by equilibrating As- and PVP-coated NP in a 0.01/1 ratio (m/m). This specific mass ratio was chosen based on preliminary tests conducted to ensure that >95 % of As initially present in solution was adsorbed onto NPs. Finally, after an equilibration time of 24 h, the prepared suspensions were centrifuged at 10,000 rpm and the supernatant removed and analyzed for both As and Fe concentrations by inductively coupled plasma mass spectrometry (ICP-MS).

Effect of iron-based NPs and arsenic on aquatic organisms

Pilot in situ studies of the remediation of contaminated groundwater reported in the literature use a broad range of nano-zero-valent iron concentrations, from values as low as 1.9 g/L (Zhang 2003), to up to 30 g/L as reported in a review by Mueller et al. (2012). For this laboratory study, the biological responses of the selected model organisms to NP-As complexes were investigated using dose-exposure experiments with As concentrations ranging from 10 μg/L (the minimum acceptable level for drinking water in the US) to 900 μg/L, and those of NPs ranging from 1 to 90 mg NP/L. These concentrations were selected based on the results of sorption and toxicity presented herein in Figs. 1 and 2. With regard to iron-based NPs, concentrations ranging from 1 to 90 mg NP/L were used to minimize (i) the interferences with algal growth resulting from dark slurry-like culture media and (ii) the concentration-dependent effect of Fe dissolution on optical density measurements (Gonzalo et al. 2014). The toxicity response was expressed as percent of either growth (R. subcapitata) or mobility (D. magna) inhibition. Experimental details for each of the two tested organisms are outlined below.

Fig. 1
figure 1

Arsenic (As) removal efficiencies of PVP-coated Fe3O4, Fe2O3, and Fe0 nanoparticles suspended individually in aqueous culture media used for growth of a the freshwater green algae, R. subcapitata, and b the freshwater invertebrate, D. magna. Arsenic concentrations were measured after a 24-h equilibration period (n = 3). The initial concentrations of each of the tested NPs and As in the culture media were 100 mg/L and 1000 μg/L, respectively. The horizontal dashed line corresponds to the US-EPA’s action limit concentration of 10 μg As/L

Full size image
Fig. 2
figure 2

Biological responses of R. subcapitata (a) and D. magna (b) grown in culture media containing As concentrations ranging from 0.01 to 0.9 mg/L. Results correspond to the average of triplicates for each tested concentration (error bars not visible when too small). Results are expressed as percent growth inhibition for the algae or percent mobility inhibition for the invertebrate

Full size image

Growth inhibition experiments using R. subcapitata

The freshwater green algae R. subcapitata was grown using a method adapted from the US-EPA protocol (USEPA 2002) in 250 mL flasks (see Table SI-1 in the support information document) maintained at 28 °C under constant light conditions (ca. 65 µmol photons m2s−1) on a rotatory shaker setup at 120 rpm. For the different dose-exposure studies, only freshly prepared Fe3O4, γ-Fe2O3, and Fe0 NP suspensions with and without PVP coating were used. As described above, the tested concentrations of NPs ranged from 0 (control sample) to 90 mg NP/L. Experiments exposing the test organisms to the iron-based NP suspensions were carried out in 6-well microplates placed on a GFL-3005 orbital shaker. Before exposure to the NPs, cultures were washed once and resuspended in the culture media to obtain a final optical density (OD) of 0.1 using a Hitachi U-2000 spectrophotometer and measuring absorbance at 750 nm (OD750 nm). Iron-based NPs were added to the culture-containing algal cells to achieve the desired NP concentrations. The dosed cultures were incubated for up to 72 h on a rotary shaker at 120 rpm and at 28 °C. For each treatment, experiments were performed in triplicates (n = 3), with serial dilutions as described in the standard OECD TG 201 method. Growth inhibition was used as endpoint for the algal-based toxicity assays, based on chlorophyll-a measurements. Bright-field micrographs were taken using an Olympus BH-2 microscope equipped with a Leica DC 300F digital camera to establish the presence or absence of physical NP-organism interactions.

Assessment of reactive oxygen species (ROS) generation

Reactive oxygen species (ROS) produced by R. subcapitata were assessed using the fluorescent dye 2′,7′-dichlorofluorescein diacetate (H2DCFDA) obtained from Invitrogen Molecular Probes (Eugene, OR, USA). This technique is based on the fact that the intracellular oxidation of H2DCFDA produces 2,7-dichlorofluorescein (DCF), a fluorescent compound that serves as indicator of the presence of hydrogen peroxide and other ROS, such as hydroxyl and peroxyl radicals (Gonzalo et al. 2014). Briefly, A 10 mM H2DCFDA stock solution was freshly prepared in DMSO under dim light conditions to avoid degradation. Prior to analysis, the algal cells were incubated for 30 min at room temperature in culture media containing a final H2DCFDA concentration of 10 µM. As a positive control for ROS formation, a 3 % H2O2 aqueous solution (v/v) was used. Fluorescence was monitored on a Synergy HT multimode microplate reader (BioTek, USA) with excitation and emission wavelengths of 488 and 530 nm, respectively. Results were normalized for differences in cell numbers, by measuring the chlorophyll content and expressed as arbitrary fluorescence units (AFU) per μg of chlorophyll-(a + b).

Daphnia magna-based bioassays

Daphnia magna assays were based on the OECD standard procedure (OECD 2004). Briefly, hatching of the ephippia was initiated 3 days prior to the start of the acute toxicity tests in a synthetic growth media (Table SI-2 in support information document). Only young D. magna (age <24 h at the start of the test) were exposed to the different iron-based NP suspensions for a period of 48 h. At time 24 and 48 h after the beginning of the exposure period, individuals that were not able to swim after a sample agitation of 15 s were considered immobilized. Controls were not spiked with NPs, and all tests were conducted in triplicates. In addition to immobilization, any abnormal behavior or physical appearance such as organism external coating with particles or NP accumulation in the gut were noted. Bright-field micrographs were taken to confirm the presence, if any, of physical contact between NPs and organisms.

Results and discussion

Physical characteristics of iron-based nanoparticles used

Physical characteristics of NPs influence the extent of reactive site availability and have been linked to the biological responses of exposed organisms (Dabrunz et al. 2011). With regard to particle size, there is a general tendency of a severe adverse biological impact associated with organism’s exposure to nano-size particles as compared to aggregated and large particles when studied under similar conditions (Dabrunz et al. 2011; Zhu et al. 2010). In this study, PVP was selected as stabilizing agent due to its reported nontoxic effect on several model aquatic organisms including those tested in this study (Oberdorster et al. 2007), as well as for its strong affinity for metals, which is driven primarily by the presence of pyrrolidone functional groups (Malynych et al. 2002).

Data on mean hydrodynamic diameter, ζ-potential, and SSA of both PVP-coated and water-suspended iron-based NPs are presented in Table 1. General trends of obtained results are as follows: NPs not stabilized by PVP show large hydrodynamic diameters, small SSA, and a net surface charge ranging from −24 to −30 mV. In contrast, PVP-coated NPs are characterized by much smaller hydrodynamic diameters and higher SSA, with a ζ-potential of about −18 mV due to charge masking by the coating polymer. In fact, PVP improves NP stability in aqueous suspensions through steric hindrance which occurs when PVP adsorbs onto the NPs (Sakulchaicharoen et al. 2010). The implication is minimization/prevention of particle–particle attraction, and therefore, limited particle aggregation (He and Zhao 2007; Sakulchaicharoen et al. 2010). In contrast, water-suspended iron-based NPs aggregated to form relatively large colloidal particles which precipitated out of solution rapidly.

Table 1 Selected physical characteristics of Fe3O4, Fe2O3, and Fe0 nanoparticles suspended in either plain water or in an aqueous solution of polyvinylpyrrolidone (PVP)
Full size table

Arsenic sorption onto iron-based nanoparticles suspended in R. subcapitata and D. magna growth media

Figure 1 shows the results of As removal onto PVP-coated Fe0, Fe3O4, and γ-Fe2O3 NPs suspended in the two tested growth media. In the culture media used for the growth of R. subcapitata, As removal efficiency decreased in the following order: PVP-Fe0 > PVP-Fe3O4 > PVP-Fe2O3 (Fig. 1a). However, under these specific experimental conditions, only PVP-Fe0 and PVP-Fe3O4 NPs were able to decrease the dissolved As concentration from the initial level of 1000 μg/L (or 1 ppm) down to 4 and 10.25 µg/L, respectively. Comparatively, As sorption by PVP-Fe2O3 NPs was less efficient, with 73.87 µg/L left in solution after a 24-h equilibration period. On the other hand, while general trends of As sorption by NPs suspended in D. magna’s growth medium mimicked those obtained with the algal culture medium, As sorption efficiency dropped significantly (Fig. 1b). In fact, based on differences in the ionic strength, one would expect NPs in the algal culture medium (with higher concentrations of Ca2+ and Mg2+) to undergo particle aggregation, and therefore, to result in a lower As sorption capacities as compared to the D. magna’s growth medium. Additionally for pH ranging between 7 and 8.5, typical of these growth media, the prevalent arsenic species should be \(HAsO_{4}^{2 - }\) under such oxygenated experimental conditions. Accordingly, it is likely that higher concentrations of Ca2+ and Mg2+ in the algal growth medium affect the net surface charge of NPs to favor sorption of the above As oxyanion species. Overall, these results point to the importance of solution chemistry specific to each growth medium on the ability of the tested iron-based NPs to sequester As from aqueous phase.

Figure 2 shows the responses of R. subcapitata (Fig. 2a) and D. magna (Fig. 2b) to increasing concentrations of As in growth media without NP addition. First, R. subcapitata exposed to As showed no toxic response in the culture medium containing As at a final concentration of 0.01 mg/L. However, a noticeable adverse biological response appears as the concentration of As reaches concentrations ≥0.25 mg/L, producing significant toxicity to algal cells, with a plateau-type response as As concentrations increased from 0.25 to 0.90 mg/L. Second, D. magna’s response to increasing As concentrations was different, in that, (i) organism immobilization was observed even in culture media with only 0.01 mg As/L, and (ii) the number of immobilized organisms exceeded 50 % when As concentration reached values ≥0.1 mg/L. Further increases in As concentration resulted in 100 % mobility inhibition (Fig. 2b).

Differences in the extent of biological responses expressed by R. subcapitata and D. magna could be dictated by several parameters including (i) the difference in the solution chemistry of each culture medium and its impact on As speciation and bioavailability, and (ii) differences in exposure pathways specific to each of these two model organisms. Algal cells can uptake As through protein channels earmarked for phosphorus, and once inside the cell, As can interfere with oxidative phosphorylation, interact with glucose and gluconate, and inhibit the activity of several enzymes (Hughes 2002). Unlike the tested green algae, the invertebrate D. magna is a particle feeder and is therefore exposed by a combination of ingestion of As adsorbed onto NPs and passive diffusion of dissolved As into cells. Previous studies have shown that as a chalcophile, arsenic can attach to sulfhydryl groups of molecules (Oremland and Stolz 2005) and result in toxicity (Sharma and Sohn 2009).

Biological responses of R. subcapitata to the iron-based NPs in the absence and presence of PVP and arsenic

Using iron-based NPs, three different sets of dose-exposure studies were conducted with R. subcapitata as the test organism and the obtained results are presented and discussed below.

First, using suspensions prepared in Nanopure® water, the biological impacts of Fe3O4, γ-Fe2O3, and Fe0 NPs were investigated. The results are presented in Fig. 3. In this graph, the horizontal line corresponds to growth obtained from control culture media (i.e., no NP added and no growth inhibition), which is used as reference. Accordingly, positive values correspond to growth inhibition while negative values point to growth stimulation. For the range of NP concentrations tested (0–50 mg/L), and taking into account the error bars on plotted data, it is obvious that these NPs have very little to no impact on R. subcapitata under these specific experimental conditions. In both cases (e.g., growth inhibition or stimulation), measured values were within 10 % of the reference or control values.

Fig. 3
figure 3

Effects of increasing concentrations of water-suspended Fe3O4, Fe2O3, and Fe0 NPs on the growth of R. subcapitata. Results are average of triplicate experiments and are expressed as percent growth inhibition using control media (no nanoparticles added) as reference corresponding to 0 % growth inhibition. Negative values indicate a tendency toward growth stimulation

Full size image

Based on data published in peer-reviewed literature, it is known that noncoated and water-suspended metallic Fe-NPs can adhere to cells and induce a reductive decomposition of functional groups in proteins, which in turn disrupts cell membrane functions due to the strong reductant character of iron (EH° (Fe2+/Fe0) = −0.447 V). In addition, since the change in redox potential is strongly dependent on the dose of metallic Fe-NPs in the culture media, direct interactions between NPs and cells should lead to an increasing trend of toxic response along the dose gradient. However, this is not the case in data obtained in this study. Trends of biological responses obtained here are likely dictated by a combination of factors including (i) the significance of physical contact between cells and NPs (Sadiq et al. 2011a, b), (ii) NPs agglomeration induced by algal-derived extracellular proteins rich in cysteine, those reported for bacteria cells (Moreau et al. 2007), and (iii) toxicity mitigation attributable to algal organic matter adsorbed onto NP surfaces through electro-steric hindrance (Chen et al. 2011; Lee et al. 2008). Overall, the result is a nonlinear dose–response relationship characterized by an irregular pattern as growth stimulation and inhibition alternate along the NP concentration gradients. Such nonlinear dose–response relationships, especially at low level exposures, are rather common and have been investigated in recent NP toxicity studies (Lavicoli et al. 2014; Gonzalo et al. 2014).

Second, the biological responses of R. subcapitata exposed to PVP-coated iron-based NPs were assessed in the absence and presence of As. The obtained results are presented in Fig. 4. The recorded growth inhibition trends from these dose-exposure experiments are as follows: First, an increase in growth inhibition is observed with increasing NPs concentrations when algal cells are exposed to two of the PVP-coated iron oxide NPs (Fe3O4, γ-Fe2O3), and regardless of either the presence or absence of As adsorbed to the NPs (Fig. 4a, b). For the two tested PVP-coated Fe-oxide NPs, the presence of As resulted in increased toxicity, but differences were not always statistically different when compared to biological responses obtained in the absence of As. In contrast to the Fe-oxide NPs, PVP-coated Fe0-NPs had a very limited adverse impact on exposed algal cells (Fig. 4c). Although the induced growth inhibition/stimulation were negligible (<10 %) along the tested Fe0-NP concentration gradient, the use of PVP as stabilizing agent increased NP dispersion within the culture medium, resulting in a slight increase in growth inhibition trend. The presence of As produced a pattern which mimicked those reported earlier in Fig. 3, likely due to differences in NP net surface charges after sorption of As.

Fig. 4
figure 4

Comparative study of the effects of polyvinylpyrrolidone (PVP)-coated iron-based NPs in the absence and presence of As on R. subcapitata: a PVP-coated Fe3O4, b PVP-coated Fe2O3, and c PVP-coated Fe0

Full size image

Overall, several parameters can be considered when explaining the above-described biological responses. Looking at the physical parameters, the use of PVP as a stabilizing agent improves NP dispersion and therefore the prevalence of nano-size particles in the suspension. This would limit the extent of particle aggregation in the culture medium, and therefore, increase the physical contact between NPs and algal cells over the short duration of bioassays conducted in this study. Microscopic observations of samples taken from these NP-treated algal cells showed the presence of heterogeneous aggregates made of algal cells and NPs (Figure SI-1 in support information document), suggesting that the recorded toxic responses were partly favored by direct physical contact between NPs and cells.

The above interpretation of the data holds true only for the two Fe-oxide NPs, but not for Fe0-NPs. Results from previous research do suggest that the high toxicity of polymer-stabilized NPs is linked to the high reactivity of nano-size particles and their ability to produce reactive oxygen species (ROS) (Nel and Xia 2006; Dabrunz et al. 2011). Fe3O4 NPs have a higher ratio of Fe(II)/Fe(III), higher redox reactivity, and dissolution rates than γ-Fe2O3-NPs, followed by Fe0-NPs (Sidhu 1981; Gorski et al. 2009). Furthermore, Fe3O4 NPs are significantly more effective in producing hydroxyl radicals than γ-Fe2O3 NPs at the same nanoparticle total surface and reaction volume ratio (Voinov et al. 2010). Compared to Fe0-NPs, the Fe-oxide NPs exhibit a higher catalytic efficiency in both hydrogen peroxide decomposition and hydroxyl radical production (Chen et al. 2012; Wu et al. 2014). Furthermore, at least one study investigated the reduction of metal oxide nanoparticles to zero valence forms, revealing that a zero-valent metal could potentially be bio-produced from corresponding metal oxide nanoparticles present in algal cultures (Gong et al. 2011). Therefore, soluble organic compounds produced by R. subcapitata may play a role in the reduction of Fe-oxide with adverse impact on algal growth.

Levels of oxidative stress within R. subcapitata growing in media containing iron-based NPs at a final concentration of 90 mg/L were measured as a function of time. The results are presented in Fig. 5. They show that the increase in oxidative stress corresponds to the increase in ROS production (data are normalized to the control data represented by the horizontal line set at 0 %). These results show an overall temporal increase in oxidative stress for the first 48 h in the Fe-oxide NP-treated culture media, followed by a decrease in measured % oxidative stress values (Fig. 5a, b). These results help explain the toxicity of Fe-oxide NPs to R. subcapitata presented earlier in Fig. 4a, b. ROS produced through chemical processes (Auffan et al. 2008; Pisanic 2007) would result in toxic effects, a well-studied and common paradigm involved in the adverse biological responses induced by several engineered NPs (Rodea-Palomares et al. 2012). Also, ROS injuries can induce the activation of apoptotic signaling pathways and ultimately cell death (Xiu et al. 2010). Our results on the production of ROS within algal membranes show that signs of oxidative stress are barely detectable in Fe0-NP-treated algae (Fig. 5c), and no significant toxic response was expressed (Fig. 4c). These trends are likely due to the formation of large aggregates by Fe0-NPs, a phenomenon which would negatively impact the previously reported formation of hydroxyl radicals by dissociative recombination of H3O+ catalyzed by the Fe0/FeII redox couple (Zhaunerchyk et al. 2009; Kim et al. 2010a, b).

Fig. 5
figure 5

Time-dependent induction of antioxidant and intracellular reactive oxygen species in R. subcapitata during exposure to polyvinylpyrrolidone (PVP)-coated iron-based NPs used at a final suspension concentration corresponding to 90 mg Fe/L. a PVP-coated Fe3O4, b PVP-coated Fe2O3, and c PVP-coated Fe0

Full size image

Toxicity assay using the aquatic invertebrate D. magna

The results of D. magna-based toxicity bioassays are presented in Fig. 6. They show the biological responses in terms of percentage (%) of organisms immobilized as NP concentrations increased from 1 to 90 mg/L. For all tested iron-based NPs, the aqueous suspensions prepared without PVP had no toxicity effect on D. magna. This result is explained by lack of organism’s exposure, which is limited under these specific experimental conditions by the removal from solution of aggregated NPs through sedimentation. This was obvious even by simple visual observations. In contrast, cultures treated with PVP-coated iron-based NPs induced mobility inhibition in all cases. Together, these results confirm that beside the chemical composition of NPs, the nano-size, reactivity, and dispersion stability of NPs are key parameters in the toxicity responses of exposed organisms (Dabrunz et al. 2011; Warheit et al. 2007).

Fig. 6
figure 6

Comparative study of the effects of polyvinylpyrrolidone (PVP)-coated iron-based NPs in the absence and presence of As on D. magna: a Fe3O4, b Fe2O3, and c Fe0

Full size image

As particle feeders, D. magna ingest particles, and previous studies have shown that food ingested by D. magna can have a residence time that ranges from about 3 to 60 min (Zhu et al. 2010). However, iron-based NPs ingested by D magna do not constitute a food source and their residence time could depend on a combination of factors including the size of ingested nanoparticles and rates of elimination. For instance, the accumulation of large aggregated NPs in the gut could affect elimination rates. In addition to NP ingestion by D. magna, Fe-based NPs have high binding affinity toward organic surfaces, resulting in a steady growing layer of iron-oxide coating on exposed organisms through binding onto exoskeleton (Dabrunz et al. 2011). This is illustrated by the bright-field micrographs (Figure SI-2), and is a confirmation of direct contact NP organism. The formation of NP coatings onto D. magna’s exoskeleton could increase the weight of the organisms and result in physical resistance during swimming, hence, impacting mobility (Dabrunz et al. 2011).

In line with these studies, the acute test primarily on PVP-coated and water-suspended Fe3O4, γ-Fe2O3, and Fe0 NPs with and without As adsorb onto NPs showed a significantly less toxicity compared to As alone in solution. Daily survival data indicated that the inhibition of the mobility of D. magna increased from 13.33 to 100 % when As concentration in solution increased from 0.01 to 0.5 mg/L. In fact, a micrograph of D. magna exposed to As alone without NPs (Figure SI-2B) shows the dramatic impact of As on this invertebrate. Finally, the presence of As made all tested NPs more toxic. Most important is the impact of As on Fe0-NP toxicity. In the presence of As, metallic Fe-NPs displayed a highly significant toxicity effect as compared to the biological responses recorded from D. magna exposed to metallic Fe-NPs and no As. It appears that As enhances the toxicity of metallic Fe-NPs, and it is possible that a Trojan horse-like effect is responsible of the observed adverse biological effects.

Conclusion

This study focused on the potential toxicity of PVP-coated iron-based NPs and formed PVP-coated iron-based NP and arsenic complexes, which can form if iron-based NPs were to be used in water treatment processes. Based on the bioassays used in this study, the results suggest that the biological responses of exposed organisms would vary with both the chemical composition of NPs and types of aqueous suspensions (surfactant stabilized or not). Differences were obvious when comparing the biological impacts of Fe-oxide NPs to that of Fe0-NPs, regardless of the tested model organisms. Overall, Fe-oxide NPs were more toxic than Fe0-NPs, in either the presence or absence of arsenic adsorbed onto the NPs. However, it is important to note that the toxicological implications assessed in this study relied on short-term exposure experiments. While this approach takes into account the effect of water chemistry (e.g., pH, types of ions present, ionic strength, dissolved organic matter) on the recorded toxicity of tested NPs, it leaves out the potential role of NP aging and associated physicochemical transformations that can be accounted for only through long-term exposure studies or by use of NPs aged in tested media. For instance, when Fe0-NPs are injected into aquatic systems, they would certainly undergo corrosion which results in the formation of an encapsulating layer of surface oxide (Scott et al. 2010; Crane et al. 2011). In fact, the bulk of Fe0-NPs injected into groundwater during a remediation process do form iron oxide particulates (Adeleye et al. 2013), and both the chemical composition and degree of crystallinity of such iron oxide minerals are not necessarily similar to Fe3O4 and γ-Fe2O3 tested in this study. On the other hand, over time, the buildup of exudates secreted by microorganisms could impact the stability of NPs, and therefore, their interactions with living organisms (Adeleye et al. 2016). Since toxicity results obtained from short-term exposure studies give a near-term insight into the potential biological impacts, further studies should emphasize long-term exposure experiments coupled with the characterization of the NPs as a function of time. Accordingly, whenever one looks at nZVI as treatment option for polluted water bodies, a long-term monitoring should be part of the strategy (Grieger et al. 2010).

Overall, findings from this study suggest that efforts to develop NP-based technologies for water treatment should take into account the fate of generated NP-toxicant wastes to avoid downstream negative impacts on ecosystem functions, and ultimately, human health.