Nanoceria was found to be non toxic for Danio rerio embryos exposed up to 200 mg l−1 nanoceria during 72 h.Table S1† illustrates the diversity in the measured effect concentration of nanoceria. Even for a given species, the results varied widely between studies. For example, Lee et al. showed significant mortality of D. magna after 96 h of expo-sure to 1 mg l−1 of 15 and 30 nm nanoceria103 while no toxicity was measure in D. magma after the same duration at 10 mg l−1 or a 48 h exposure at 1000 mg l−1 nanoceria.Van Hoecke et al. exposed D. magna to higher concentrations of 14, 20, and 29 nm nanoceria for 21 days, and found an LC50 of approximately 40 mg l−1 for the two smaller particles and 71 mg l−1 for the 29 nm particles.When combining all aquatic toxicity data, including C. elegans , no trends were observed between the nanoparticle size and the toxicity. We observed one extreme value, which is a report of reduction in life span of C. elegans at a concentration of 0.172 μg L−1 . 92 Some have suggested that the toxicity at low concentration can be explained by differences in the aggregation state as a function of concentration. NPs may be less aggregated at lower concentration.105 However, the nanoceria used in this study were positively charged, coated with hexa-methyleneteramine . It is possible that this coating rendered nanoceria much more toxic. Another Fig. 1 depicts the median of the lowest observed effect concentration and the EC10 or LC10 toward different species. This figure illustrates the high variability of the observed LOEC/EC10 between studies for a same organism . Based on the LOEC/EC10, the more sensitive species is the cyanobacte-rium Anabaena, while the least sensitive is Daphnia magna. No toxicity was observed up to 5000 mg Ce/L for the zebrafish Danio rerio and Thamnocephalus platyurus Fig. 1. It is noteworthy that exposure models predict concentrations significantly lower than those for which ecotoxicity investigations have encountered toxic effects. Therefore, nanoceria might not have any impact at environmental concentrations,growing pot despite the fact that some results are more worrying. More-over, most of the nano-ecotoxicology performed on aquatic organisms used a single species or a short trophic links and do not take into account important parameters such as the colloidal destabilization of the nanoceria, their interactions with organic molecules/ particles naturally occurring or bio-excreted, or the flux between compartments of the ecosystems .

To work under more realistic scenario of exposure, few nano-ecotoxicological studies are now performed in aquatic mesocosms with low doses of nanoceria, chronic and long-term exposure. Such studies will allow obtaining reliable exposure and impact data and their integration into an environmental risk assessment model that is currently missing.Although the data on environmental effects are far from complete, it is useful to consider case studies in order to gain knowledge about key data gaps and to give a first impression of relative risks based on current knowledge. While this case study is useful to point out areas where research is most needed, it is critical to point out the limitations of this case study. First, nanoceria have not yet been detected or measured in environmental media, and the actual environmental concentrations are not known. Second, very little is known about the fate and transport of nanoceria in the environment. Third, the toxicity data base is still very limited. Only a select few ecological receptor species have been tested to date and few if any sub-chronic or chronic exposures have been performed in longer lived organisms or in environmentally realistic exposure scenarios. The following case study characterizes the likely exposure concentrations and compares them to toxicity values for soil and water based on emissions due to combustion of fuels containing nanoceria additives and for discharge of chemical mechanical planarization media into sanitary sewers.Based on Table S1,† with the exception of HMT coated nanoceria, which do not apply to this case study and for which coating controls are lacking, the lowest EC10 value measured so far is 8000 ng l−1 for luminescence inhibition in cyano-bacteria.Previous estimates have been made for nanoceria used as a fuel catalyst and arriving in soil and water following atmospheric discharge106 in the UK based on known market size for this product. Clearly there is a wide disparity between concentrations likely to occur due to fuel catalyst combustion106 and the lowest toxicity values observed so far . However, there remains concern that nanoceria may enter water courses through its uses in specialized industrial polishing or chemical/mechanical planarization.Without specialized local knowledge on where these industrial concerns are located, the quantities of nanoceria used, that are disposed of from the premises, and the capacity of the associated sewage treatment plant, the local receiving water concentrations cannot be predicted.

Unfortunately, knowing global or national consumption of nanoceria in the polishing industry would not allow us to predict water concentrations. This is because the use of the product would not be evenly geographically spaced, or linked directly to human population density. However, it is possible to ask: what discharge would be needed to exceed the 8000 ng L−1 toxicity threshold for aquatic exposures? The dilution factor for sewage effluent recommended by EU risk assessment is 10. So effluent would need to contain 80 μg L−1 nanoceria. However, it is estimated that on entering an WWTP 95% of the nanoceria would enter sludge and only 5% pass through into the effluent.In that case the influent concentration would need to be 1.6 mg l−1 nanoceria. WWTPs are designed around population equivalents which tend to be around 160–200 L per PE per day in the UK so a PE unit would need to discharge 256–320 mg Ce per day to receiving waters. Given the current uses of nanoceria, this only seems likely to occur if a large industrial facility is directly discharging wastewater containing high concentrations of nanoceria directly into a sanitary sewer. Note that a population equivalent is a unit describing a given biodegradable load as measured by its biological oxygen demand.We have comprehensively reviewed what is known for nano-ceria about the environmental releases, methods for detection and characterization, fate and transport, toxicity and likelihood of toxicity in soil and water from acute exposures. Initial estimates of releases suggest that the majority of nanoceria will ultimately end up in landfills, with lesser amounts emitted to air, soil and water in that order. Once nanoceria enters the environment, it has been shown that NOM will have a major impact on their fate, transport and toxicity. As with other nanomaterials, aggregation is a key consideration and this has been shown to be influenced by water chemistry and interactions with natural coatings such as NOM. An important feature of nanoceria with respect to its behavior and toxicity is its valence state. There are several techniques that can characterize this property in environmental and bio-logical media, such as XAS, but most require relatively high concentrations. While we didn’t identify studies that detected nanoceria in natural environments or environmental media, a suite of techniques have been used to detect and character-ize them in complex toxicity testing media and in controlled laboratory studies. Thus, a major data gap and area for future research is the prediction and measurement of actual nano-ceria concentrations in the environment, either from point sources or non-point sources.

As a whole nanoceria appears to exhibit similar aquatic toxicity values other commonly studied manufactured nano-materials. For example,square pot a recent review found that species average LC50 values for Ag nanoparticles ranged from 0.01 mg L−1 to 40 mg L−1 while species mean LC50 values for ZnO ranged from 0.1–500 mg L−1 . 116 The range of EC50 values reported for Ce are similar to those for ZnO. Although reported toxicity data here uses LC10 and LOEC values, the range of species means 0.05–25.9 mg L−1 and many of the reported LC50 values are within the range of 0.1–100 mg L−1 , suggesting similar acute toxicity to ZnO NPs in aquatic expo-sures. This is of course based on the available data, which are predominantly on the toxicity of nanoceria to aquatic organisms, with sediment and terrestrial organism data severely lacking. For example, few if any studies have investi-gated toxicity in sediment dwelling organisms, which are likely to be exposed to nanoceria in the aquatic environment due to aggregation, settling and accumulation of nanoceria in sediment. Given the persistence of nanoceria, chronic studies are lacking as we are aware of only the C. elegans study.Equally important, very few species from few taxonomic groups have been tested. Large taxonomic groups such as insects and gastropods have not been tested and only one non-mammalian vertebrate spe-cies has been tested . Another difficulty is that most of the studies were performed with different nano-particles, doses, duration, organisms, exposure media, and their results are not directly comparable. Perhaps due to these differences, there are no apparent patterns to suggest that, as a whole, particle size has a major impact on toxicity. A problem in conducting realistic toxicity studies is the likely transformation of the free particles into homo or hetero-aggregates or even organic complexes in the real environment. There have been few studies that investigated the impact of size across a wide range of systematically varied particle sizes within a single study. Such studies are needed to definitively establish weather size is important. On the other hand coating may be an important variable given the extreme sensitivity seen with HMT coated particles in C. elegans. Coating was demonstrated to be a major determinant of toxicity in a more well controlled study that systematically varied coating properties and used coating controls.2 Of all of the taxonomic groups, toxicity is most well studied in vascular terrestrial plants. Overt phytoxicity of nano-ceria seems minimal and, while root to shoot translocation of these particles is often measurable it is generally quite low. In summary, although the literature on nanoceria impacts on terrestrial plants is not extensive, it is clear that overt phytotoxicity is minimal, even at excessive exposure concentrations. The data do suggest accumulation of nano-ceria within plant tissues, although the precise form of the element that crosses into the plant and the mechanism driving that process remains unknown. The potential trans-generational effects noted in the literature,as well as the complete lack of information on trophic transfer, are areas of concern. In addition, studies investigating environmentally relevant concentrations, potentially secondary effects from nanoceria exposure, including impacts on symbiotic micro-organisms or on edible tissue nutritional quality, certainly warrant further investigation. As a whole, the aquatic and terrestrial toxicity testing data for animals and microorganisms spans multiple orders of magnitude for acute toxicity values . This large variation can be exhibited within a single species exposed to similar nanoceria. For example, toxicity values for D. magna range from around 1–100 mg l−1 for fairly similar particles. Based on the overall toxicity database, it appears that C. elegans is the most sensitive animal and Anabaena is the most sensitive microorganism tested to date, although an important caveat is that the same endpoints were not com-pared across all species and that exposure systems varied. Interestingly no toxicity was observed in the fish species that has been tested even at extremely high exposure concentrations . Unfortunately, only two fish studies have been reported in the literature. There is a complete lack of toxicity testing data for sediment dwelling organisms, and extremely limited data for soil invertebrates. As a whole the data suggest that acute toxicity is possible at low μg L−1 concentrations in the water column. Data are lacking on soils and sediments, but toxicity values are likely to be far lower. One study indicated toxicity at lower concentrations than these values when 8 nm nanoceria were coated with HMT. Since no coating controls were used, it is critical that the influence of this coating and other similar positively charged coatings be studied using a similar end-point and suitable controls. The use and disposal of any nanoceria containing products with this coating should also be evaluated. It is not clear whether the chronic nature of this exposure or the influence of the coating on uptake and toxicity explain why this toxicity threshold is so low.