Cerium trichloride

Toxicity to aquatic algae and cyanobacteria

Description of key information

The key study (Hefner, 2013) yielded a 72-h NOEC and EC50 value (growth rate-based) of 0.46 and 0.63 mg Ce/L, respectively, for the unicellular green alga Pseudokirchneriella subcapitata (i.e., 0.81 and 1.1 mg CeCl3/L, respectively). However, these values will not be taken forward to PNEC derivation because the effects on growth were observed to be concurrent with phosphate depletion in the test medium due to complexation with cerium, suggesting that the observed effect on growth inhibition is due to phosphate deprivation rather than direct toxicity of the rare earth. This is confirmed by modelling calculations using Visual MINTEQ v3.0. Further testing is not considered useful because the technical issue of phosphate depletion cannot be overcome (phosphate dosing during the test would result in 100% cerium depletion from the test medium). Finally, the results will not be used for classification purposes of CeCl3 either.

Additional information

Three studies on the toxicity of cerium to algae were included in this dossier.

The first study of Bringmann and Kühn (1969) investigated the effects of CeCl3 on growth of Scenedesmus algae. The lowest concentration at which cerium exerted an adverse effect on biomass (growth) was 0.15-0.20 mg Ce/L. Results can however not be considered reliable because cerium concentrations in test media were not analytically verified, and because it is not clear to what extent the observed effect was an indirect effect such as a phosphate deprivation effect.

In a second study (Tai et al., 2010), the toxicity of cerium (added as Ce(NO3)3) to the marine unicellular alga Skeletonema costatum was investigated in a 72-h growth inhibition study. The 72-h EC50 of cerium was reported to be 4.16 mg Ce/L. However, since no analytical verification of test substance or element concentrations was performed, results can not be considered reliable. Further, it is not clear to what extent the observed effects were due to phosphate deprivation.

The third study is the study of Hefner (2013). This study was a 72-h algal growth inhibition test with Pseudokirchneriella subcapitata in which CeCl3 was used as test item. The growth rate-based 72-h EC50 and NOEC were 0.63 and 0.46 mg Ce/L (i.e., 1.1 and 0.81 mg CeCl3/L), respectively.

Due to the known issue with phosphate complexing by rare earth elements in algal growth inhibition tests, phosphate concentrations were monitored during this study. It was observed that at the lower test concentrations, all cerium was precipitated shortly after addition to the test medium, because phosphate was in excess. However, at the higher test concentrations, cerium was in excess, complexing all phosphate in the test medium. Algal growth was completely impeded at those test concentrations where all phosphate had disappeared from the test medium from the start of the test already. The ErC50 value (0.63 mg Ce/L) was therefore somewhat lower than the lowest test concentration at which complete immediate phosphate depletion occurred. Therefore, the observed effects are considered to be mostly due to phosphate deprivation instead of direct cerium toxicity. Since all algal growth inhibition tests need to be performed in test media containing a phosphate source, testing is considered technically not feasible if reliable results on cerium toxicity to algae are to be obtained. Further, the phosphate depletion effect is not considered an environmentally relevant effect, since it would only occur very locally where point source emissions or accidental releases occur, and will never affect an entire ecosystem. Therefore, the effects on algal growth are not taken into account for PNEC derivation and classification. As the study was well conducted and the "phosphate issue" was investigated, it is totally justified to attribute a Klimisch score of 1 to the study.

To further argument the conclusion that the observed toxicity is due to disappearance of phosphate from the test medium as a result of complexing with the rare earth, modelling calculations have been performed using Visual MINTEQ v3.0 using data from the algal growth inhibition study of Hefner (2013). Modelling of cerium speciation was performed as follows:

- All components of the test medium (nominal concentrations) were added to the modelling solution.

- The pH was set to 7.0 (although pH of test solutions was adjusted to 6.5, at the start of testing pH was around 7.0 in most treatments).

- Ionic strength was not set to a fixed value, but the model was allowed to calculate it (default).

- Temperature was set to 23°C (average temperature during the test).

- Five 'modelling problems' were added, using set total cerium levels (i.e., the measured dissolved Ce levels at test initiation: 0.198, 0.46, 0.943, 2.01 and 4.48 mg Ce/L). For each 'modelling problem' the program should then calculate speciation in the test medium.

- The following aqueous Ce species were modelled by Visual MINTEQ: Ce3+, Ce(CO3)2-, CeCO3+, CeHCO3+2, Ce(SO4)2-, CeSO4+, CeCl+2, CeEDTA-, CeHEDTA, CeOH+2, CePO4, and CeH2PO4+2.

- Two possible solid phases were added for Ce: Ce(OH)3 and CePO4. When solubility products are exceeded in the aqueous solution, the model allows precipitation of these phases. The model does not contain (by default) other solid phases for Ce, although especially at higher pH levels some carbonate precipitation may be expected too. For the modelling exercise presented here, the absence of possible cerium carbonate solid phases does not affect the outcome of the calculations. Note that the nominally added total phosphate (PO4 3-, total) concentration is 1.18E-05 M, hence no more CePO4 than that can be formed.

- Under the abovementioned conditions, the model calculations for dissolved versus precipitated Ce and phosphate can be summarised as follows (measured dissolved Ce at the end of testing was added for comparison):

Ce total (initial measured Ce dissolved) (mg/L)	Ce dissolved (model calculation) (mg/L)	% Ce dissolved (model calculation)	Ce dissolved (measured after 72 h) (mg/L)	% PO4 3- precipitated (model calculation)
0.198	0.00	0.00	0.0017	11.95
0.46	0.00	0.00	< LOQ	27.80
0.943	0.00	0.00	0.022	57.03
2.01	0.35	17.43	0.738	100
4.48	2.83	63.18	2.96	100

- Ce(OH)3 was not calculated to precipitate under the set conditions of testing, only CePO4 precipitation occurred. Further, when checking phosphate speciation, it became clear that phosphate precipitation was practically entirely due to precipitation with cerium.

- Based on this modelling exercise it is confirmed that under the conditions of the test all cerium is precipitated as CePO4 whenever phosphate is in excess and vice versa. The good agreement between measured dissolved Ce concentrations at the end of testing and the modelled dissolved Ce concentrations further increase the credibility of the model calculations.