Transformations of Ag and ZnO Ag NPs in the environment.
ENMs will transform in the environment. Transformations (e.g. sulfidation or ligand-promoted dissolution) affect their persistence, fate, transport, and toxicity. The objectives of the current work are to 1) determine the effect of environmental ligands other than sulfur (halogens) on Ag NP oxidation and dissolution, 2) determine the speciation and distribution of transformed Ag NPs in CEINT mesocosms and in WWTP biosolids, 3) determine the effect of sulfidation on Ag NP toxicity to a broad range of organisms, and 4) determine if sulfidation is a common transformation for other NPs made from class B metal cations (ZnO, CuO, etc.). We hypothesize that sulfidation will dominate the transformation process of such NPs in most environments, and will decrease toxicity broadly across many organism types regardless of exposure route or media type.
Sulfidation of Ag NPs is rapid (Levard et al., 2011), but halogens including iodide, bromide, and chloride also can promote the oxidative dissolution of 40 nm PVP coated Ag NPs. Ag NPs transformation ranged from 22%-84% after 6 weeks aging with each ligand present at 5mM. The following ligand strength was determined; Sulfide > Iodide > Bromide > Chloride. The transformed Ag NPs have a lower solubility than pristine Ag NPs and solubility reduction followed the same order as ligand strength. This suggests that ligand promoted oxidation of Ag NPs is expected over relatively short time scales, even in oxic environments without sulfide.
We determined the distribution and speciation of Ag NPs in CEINT mesocosms and in WWTP biosolids. About 18 months after addition to the mesocosms (either in the water column or to the terrestrial soils), Ag NPs resided largely in the compartment where they were added, but some transfer between compartments (~20% of the material) occurs. EXAFS analysis of mesocosm sediment samples confirmed that the chemical composition of Ag NPs added to the mesocosms was largely Ag2S, but that some Ag(0) character remained. The extent of transformation was greater in sediments compared to the terrestrial soils. Despite this partial transformation to Ag2S, Ag remained bioavailable to aquatic organisms and plants as determined by theme 2 (Matson) and theme 3 researchers (Richardson). The mechanisms of Ag uptake in organisms are still being investigated. EXAFS investigations determined that the Ag0 from Ag NPs was completely oxidized/transformed to Ag2S after two weeks of exposure to biosolids. Post-sulfidation, the Ag is associated with the solid fraction of the BS slurry; < 0.6% of the total Ag remains in the aqueous fraction. When Ag NP amended biosolids are mixed with soil (as would occur from use on croplands as fertilizer), the Ag associates with, and becomes enriched within, the water-dispersible clay fraction of soils, implying an exposure risk from runoff away from croplands and into streams and rivers.
The toxicity of sulfidized Ag NPs was compared to the toxicity of pristine Ag NPs for five organisms (embryonic Zebrafish, Killifish, C. Elegans, Duckweed, and Daphnia) for the toxicity endpoints of either mortality or a sub-lethal effect. These experiments were conducted in five different labs across CEINT using one batch of well characterized pristine and sulfidized Ag NPs to ensure consistency across the experiments. Each organism was tested according to individual lab protocols. Table 1 shows the results of these experiments and demonstrates that across all organisms and media types sulfidation of Ag NPs greatly reduces toxicity responses. We are currently testing the hypothesis that the decrease in the amount of ionic Ag due to sulfidation is decreasing toxicity to all the organisms tested. The general decrease in toxicity with sulfidation across multiple plant, fish, and terrestrial organisms suggests that sulfidation will have broad impacts on Ag NP toxicity. This fact, combined with the mesocosm study indicating that Ag NPs are readily sulfidized in the environment, suggest that the toxicity potential of Ag NPs may be overestimated from tests using the pristine material.
A 30±5 nm ZnO nanoparticle was sulfidized with different S/Zn ratios to obtain different extents of transformation from ZnO to ZnS. Sulfidation was rapid (< 96 hours). Synchrotron based X-ray Diffraction (XRD) was used to characterize the chemical speciation of the sulfidized particles. ZnO NP sulfidation resulted in 4nm ZnS particles as determined using Rietveld fitting of XRD data. This indicated a dissolution re-precipitation mechanism for ZnO sulfidation, different than for Ag NPs which resulted in fused Ag2S NPs with roughly the same particle size as the initial Ag(0) particles. Additional characterization of the resulting ZnS is underway. These studies suggest that sulfidation will be a common process to NPs made from class B metal cations and that those materials my ultimately be classified by their environmental reactivity rather than their chemical composition.
Uncertainty analysis and model validation for characterization of NP macromolecular coatings using soft particle electrokinetic theory.
The coatings are an essential part of the “surface” of NPs and therefore characterization of the coating layers is essential to explain attachment and aggregation behavior. A soft particle electrokinetic model may be used to determine the coating properties if uncertainty in the fitted parameters is reasonably small. Identifiability of the fitted layer thickness for theoretical cases was compared for both an approximate analytical model and an exact numerical model (which includes polarization/relaxation effects and relaxes assumptions on particle size and layer thickness). The uncertainty analysis indicated that the model will be most useful for characterizing thin, low-charge coatings. Model validation against relatively monodisperse BSA-coated gold (provided by Vikesland) and polyacrylic acid coated hematite nanoparticles was conducted. The numerical model provides layer thicknesses comparable to those determined using DLS. Independent measurement of some parameters, such as charge density by titration, is expected to further improve the model’s ability to provide accurate values of the macromolecular properties needed for attachment models.