Titanium dioxide

Long-term toxicity to fish

Description of key information

Results of several studies of the toxicity to early life-stages of different fish species indicate that dispersed microsized TiO2 is not chronically toxic to freshwater fish up to at least 160 mg/L (NOEC ≥ 160 - 1000 mg/L) whereas dispersed nanosized TiO2 is not chronically toxic up to 80 mg/L (NOEC ≥ 80 - < 1000 mg/L). Considering the low solubility of microsized TiO2, it is further concluded that microsized TiO2 is not chronically toxic up to its solubility limit.

Additional information

Microsized TiO2:

One reliable study demonstrates that microsized TiO2 materials are not chronically toxic to freshwater fish: In a study by Shaw et al. (2016), the titanium dioxide bulk material NM-100 (148.3 ± 109.7 nm) did not affect embryo survival, hatching rate and the time to hatch of Danio rerio embryos at the highest test concentration of 160 mg/L (as dispersion) in a fish early-life stage toxicity test (OECD TG 210) after 4 d and 6 d of exposure. Further, an effect on larvae length, muscle block width and yolk sac volume could not be observed after 6 d of exposure (NOEC ≥ 160 mg/L (nominal)). Due to the shorter exposure period of 6 d instead of 30 d as recommended in OECD 210, this study cannot be considered a key study and is instead applied in a weight-of evidence approach.

Further studies (rated Klimisch 3 due to individual shortcomings) investigating the toxicity of microsized TiO2 to early life-stages and adults of freshwater fish including Danio rerio, Phoxinus phoxinus, and Coregonus atumnalis migratorius demonstrate that microsized TiO2 is not chronically toxic to freshwater fish up to its solubility limit and dispersed TiO2 concentration of 500 mg/L. In a 115 d early-life stage toxicity test with embryos and in a 30 d juvenile fish growth test by Beim et al (1982) with Coregonus autumnalis migratorius, dispersed microsized TiO2 concentrations up to 3 mg/L did not negatively affect survival and growth of exposed fish (115 d NOEC and 30 d NOEC: ≥ 3 mg/L (nominal)). To the contrary, a stimulation of growth, weight and length, was observed (Beim et al. 1982). In a 28 d fish growth test by Beim et al (1982) with adult Phoxinus phoxinus, dispersed microsized TiO2 concentrations up to 1000 mg/L (nominal) did not affect survival and growth of exposed fish (28 d NOEC: ≥ 1000 mg/L (nominal)). Faria et al. (2014) also observed that dispersed microsized TiO2 did not significantly affect the survival and length of zebrafish embryos in an early life-stage toxicity test following OECD TG 212. Morphological effects were also not observed up to the highest test concentration and the determined 8 d NOEC amounts to ≥ 1000 mg TiO2/L (dispersed, nominal).

Beim et al. (1982) performed a 30 d fish growth test with juvenile Phoxinus phoxinus and microsized TiO2 concentrations of up to 1000 mg/L (dispersed, nominal) did not affect survival and growth (length) of Phoxinus phoxinus whereas the weight of fish exposed to microsized TiO2 concentrations of 1000 mg/L were significantly lower than the weight of control fish, resulting in a 30 d NOEC of ≥ 500- < 1000 mg/L (nominal).

However, all chronic fish experiments were performed at concentrations several magnitudes above the solubility limit of microsized TiO2. Transformation/dissolution data of different microsized TiO2 materials indicate a low solubility in environmental media as dissolved Ti concentrations after 28 d were below the respective LOD/LOQ (< 0.11 / < 0.34 µg Ti/L). In sum, it is concluded that microsized TiO2 is not chronically toxic to fish up to concentrations of ≥ 160 mg/L dispersed TiO2 and its solubility limit in freshwater.

Marine data do not exist.

Nanosized TiO2:

One reliable study demonstrates that nanosized TiO2 materials are not chronically toxic to freshwater fish: The titanium dioxide nanomaterial NM-105 (20.6 ± 3.9 nm) did not affect embryo survival and hatching rate of Danio rerio at the highest test concentration of 160 mg/L (as dispersion) in an early-life stage fish toxicity test (OECD TG 210) after 4 d and 6 d of exposure. No effect on larvae length, muscle block width and yolk sac volume could be observed after 6 d of exposure (NOEC ≥ 160 mg/L (nominal)). Although a delay of hatching could be observed at the highest tested nano-TiO2 concentration (6 d LOEC: 160 mg/L; 6 d NOEC: 80 mg/L (nominal)), effects occurred at nano-TiO2 concentrations exceeding the recommended limit test concentration of 10 mg/L according to OECD 210. Hence, nano-TiO2 is considered to be not chronically toxic to Danio rerio embryos. Due to the shorter exposure period of 6 d instead of 30 d, as recommended in OECD TG 210, this study cannot be considered key study and is instead applied in a weight-of-evidence approach.

Additionally, in a standard fish, juvenile growth test performed according to OECD TG 215 to evaluate chronic effects of wastewater-borne TiO2 NPs on rainbow trout (Oncorhynchus mykiss), no effects on growth rate, behavior and mortality were observed after exposure for 28 d at the highest test concentrations of wastewater-borne TiO2 NPs, effluent-supplemented TiO2 NPs and water-dispersed TiO2 NPs (Zeumer et al. 2020). Derived NOECs for effluent-supplemented NPs and water-dispersed TiO2 NPs were ≥ 0.08 mg TiO2/L and ≥ 0.07 mg TiO2/L, respectively, based on measured concentrations (time-weighted average). Since tested concentrations were selected based on environmental relevance, and testing was not performed at the limit concentration of 10 mg/L (as specified in OECD TG 210), results of this study cannot be considered as key results and are thus applied in a weight-of evidence approach.

In further studies (rated Klimisch 3 due to individual shortcomings), three different titanium dioxide nanomaterials P25, NM-103, and NM-104 did not significantly affect the survival of zebrafish embryos in an early life-stage toxicity test (OECD TG 212) and 8 d NOEC values of ≥ 1000 mg n-TiO2/L (dispersed, nominal) were derived by Faria et al (2014). Furthermore, morphological effects were not observed with the exception of marginal effects on body length at the highest test concentration of 1000 mg n-TiO2/L but not at 100 mg n-TiO2/L (dispersed). A NOEC should nevertheless not be derived from the study by Faria et al (2014) since the separation factor should not exceed 3.2 according to OECD TG 212. Further, according to OECD TG 212, concentrations higher than 100 mg/L should not to be tested. Three supporting studies, including one fish early-life stage toxicity test (Bar-Ilan et al. 2013), one short-term toxicity test on fish embryo and sac fry stages (Bar-Ilan et al. 2012), and one prolonged fish toxicity test (Clemente et al. 2015), additionally indicate that nanosized TiO2 is not chronically toxic to embryos and juveniles of Danio rerio and Piaracus mesopotamicus since unbounded NOEC values ranged from ≥ 100 to ≥ 10,000 mg/L dispersed n-TiO2. The latter studies are considered as supporting studies since validity criteria were not provided (Bar-Ilan et al. 2013, Clemente et al. 2015), validity criteria were not met (Bar-Ilan et al. 2013), and/or tests were performed with too much biomass (Clemente et al. 2015).

Finally, based on reliable chronic toxicity studies it is concluded that nanosized TiO2 dispersions are not chronically toxic to freshwater fish up to ≥ 80 mg/L.

Marine data do not exist.

Based on a weight of evidence approach, dispersed microsized TiO2 is not chronically toxic to freshwater fish up to ≥ 160 mg/L. Considering the low solubility of microsized TiO2, it is further concluded that microsized TiO2 is not toxic up to its solubility limit. Dispersions of nanosized TiO2 are not chronically toxic to freshwater fish up to ≥ 80 mg/L.