Aluminum cobalt oxide

Long-term toxicity to fish

Description of key information

NOEC (16d, survival) = 0.060 mg Co/L (Brachydanio rerio) (read-across from cobalt chloride hexahydrate)

Additional information

No data on long-term toxicity to fish are available for the test substance cobalt aluminium oxide. However, there are reliable data available for different structurally analogue substances.

The environmental fate pathways and ecotoxicity effects assessments for cobalt metal and cobalt compounds as well as for aluminium metal and aluminium compounds is based on the observation that adverse effects to aquatic, soil- and sediment-dwelling organisms are a consequence of exposure to the bioavailable ion, released by the parent compound. The result of this assumption is that the ecotoxicological behaviour will be similar for all soluble cobalt and aluminium substances used in the ecotoxicity tests.

As cobalt aluminium oxide has shown to be highly insoluble with regard to the results of the transformation/dissolution test protocol (pH 6, 28 d), it can be assumed that under environmental conditions in aqueous media, the components of the substance will be present in a bioavailable form only in minor amounts, if at all. Within this dossier all available data from cobalt and aluminium substances are pooled and used for the derivation of ecotoxicological and environmental fate endpoints, based on the cobalt ion and aluminium ion. For cobalt, only data from soluble substances were available and for aluminium, both soluble and insoluble substance data were available. All data were pooled and considered as a worst-case assumption for the environment. However, it should be noted that this represents an unrealistic worst-case scenario, as under environmental conditions the concentration of soluble Co2+ and Al3+ ions released is negligible.

 

Cobalt

Data on chronic single-species toxicity tests resulting in high quality NOEC/L(E)C10 values (expressed as Co) for freshwater fish (n= 21) are summarised in the WHO CICAD, 2006 (see attached table).

Chronic data for several different fish species were extracted and used in the effects assessment. The NOEC/L(E)C10 values for fish range from 60 μg Co/L for Brachydanio rerio to 2000 μg Co/L for Pimephales promelas (WHO CICAD, 2006). The most sensitive data for Brachydanio rerio are reported in detail as key study record (Dave and Xiu, 1991). In this study conducted to methods comparable to guidelines, a NOEC (16d) of 60 µg Co/L (tested as cobalt chloride hexahydrate) was obtained for survival. Further results from studies on other freshwater fish are comprised in the attached table (cicad_fish).

References: World Health Organization (2006). Concise International Chemical Assessment Document 69. COBALT AND INORGANIC COBALT COMPOUNDS.

Aluminium

Long Term Fish Toxicity Literature Review: Four long-term reliable chronic toxicity studies to two species of fish (Pimephales promelas and Salveninus fontinalis) were identified as acceptable from the published literature. NOECs and EC10s ranged from 0.088 to 2.3 mg Al/L and 0.078 to 5.19 mg Al/L, respectively. 

The following information is taken into account for long-term fish toxicity for the derivation of PNEC:

The Al BLM developed using gill accumulation data from S. salar was applied to the chronic Pimephales promelas data (Oregon State University Aquatic Toxicology Laboratory 2010; Figure 7.1.1.1.2.-1). Application of the model to new data requires development of a critical accumulation value appropriate for the exposure duration and toxicity endpoint. In addition, calibration of the model to these data benefited from two other changes in parameter values. First, since the chronic endpoints for this species and in these test conditions were at much higher aluminium concentrations and saturation of NOM binding sites included in the model was beginning to occur, resulting in a somewhat reduced predicted effect of NOM compared with the observed effect. The binding site density for NOM was increased by two fold to provide adequate binding sites at these high Al concentrations. In addition, although the effect of hardness on observed aluminium toxicity was consistent in acute and chronic exposures, the predicted effect of hardness could be improved by a small change in the binding strength of Ca (i.e. the log K for binding at the biotic ligand was increased from 4.2 to 4.8. 

After application of the Al BLM, the variability in the response curve between effects of aluminium on the biotic ligand was reduced compared with response curve based on total aluminium (Figure 7.1.1.1.2.-2). Values for critical accumulation were estimated

directly from the predicted response curve on the biotic ligand to establish the CA10, or the critical accumulation level that results in a chronic effect of 10% (in this case a reduction in growth). 

Figure 7.1.1.1.2.-3 provides an evaluation of the ability of the long-term fish BLM to predict EC10 values. In this case, most of the EC10 values are predicted within 2-fold of the reported EC10 values, and all of the predicted EC10 values are within 4-fold of the reported values. 

Conclusion
As the effect values derived from analogue cobalt compounds are considerably lower than those derived from analogue aluminium substances, it can be reasoned that the cobalt ion will mainly account for ecotoxicological effects of the substance. Hence, it was concluded to put forward the most sensitive and reliable results derived from analogue cobalt compounds for assessment purposes. Still, it should be noted that this represents an unrealistic worst-case scenario as under environmental conditions in aqueous media, the components of the highly insoluble substance will be present in a bioavailable form only in minor amounts, if at all, and hence, the concentration of soluble Co2+ and Al3+ ions released is negligible.