Registration Dossier

Ecotoxicological information

Ecotoxicological Summary

Currently viewing:

Administrative data

Hazard for aquatic organisms


Hazard assessment conclusion:
no hazard identified

Marine water

Hazard assessment conclusion:
no hazard identified


Hazard assessment conclusion:
no hazard identified

Sediment (freshwater)

Hazard assessment conclusion:
no hazard identified

Sediment (marine water)

Hazard assessment conclusion:
no hazard identified

Hazard for air


Hazard assessment conclusion:
no hazard identified

Hazard for terrestrial organisms


Hazard assessment conclusion:
insufficient hazard data available (further information necessary)

Hazard for predators

Secondary poisoning

Hazard assessment conclusion:
insufficient hazard data available (further information necessary)

Additional information

Approach for PNEC derivation for freshwater

The industry is attempting to provide a large dataset which can be used to determine PNEC using the Biotic Ligand model (BLM). In all cases the PNEC will be based on total aluminium concentrations and wil be considerably higher than the dissolved aluminium concentrations that are found in th eenvironment. While this work is considered important by th eindustry, there is currently no obligation to determine PNEC values as the substance is not classified under DSD and does not meet criteria for the classification of metals and metal compounds under GHS/CLP. While an argument may be made that the metal salt under review is soluble according to standard methods (OECD 105), this is not supported by a Transformation/Dissolution study on aluminium sulphate at pH6 and 8 in which a significant amount of the substance salted out within 24 hours and in a 28 d study the substance was demonstrated to be extrememly poorly soluble following TDp guidelines (CLP 2013). For this reason the following information on the determination of the PNEC has been included in this dossier for information purposes only.

The available ecotoxicity database for the effect of aluminum on freshwater organisms is relatively large. Therefore, the use of the statistical extrapolation method is preferred for HC5-50/PNEC derivation rather than the use of an assessment factor on the lowest NOEC, as specified by the REACH Guidance document on information requirements and chemical safety assessment Chapter R. The PNEC is further based on the 50% confidence value of the 5th percentile value of the chronic effect NOEC/EC10 data (HC5-50) and an additional assessment factor taking into account the uncertainty on the HC5-50 (thus PNEC = HC5-50/AF). The advantage of this statistical extrapolation method is that it uses the whole sensitivity distribution of species in a collection of laboratory test data to derive a PNEC instead of taking only the lowest long-term NOEC.

Biotic ligand model

Aluminum (Al) toxicity to aquatic organisms is strongly affected by water chemistry conditions and factors such as pH, DOC, and hardness that can change the bioavailability and therefore the toxicity of Al to aquatic organisms. The importance of water chemistry on the aquatic toxicity of Al suggests that interactions between Al and constituents of exposure waters influence bioavailability and toxicity of Al to aquatic organisms. These types of interactions have typically been well-described by the biotic ligand model (BLM) framework.

The specific methodology used to apply the BLM is further described in the attached document and CSR

Development of an Acute Ecotoxicity Reference Value (ERV)


There are two possible approaches for assessing the significance of T/D test results for classification: (1) the T/D test media at the end of the 7 or 28 day study can be used directly as the test media for the toxicity studies or (2) the amount of dissolved metal in solution can be compared with a toxicity test performed at the same pH using a soluble metal salt. For the first approach, if there is no toxicity, the substance would not classify and there is no need for an ERV as the test solutions are evaluated directly following the T/D test. For the second approach there is the need for an ecotox reference value. This value needs to be generated from the available (high quality) chronic ecotoxicity data for fish, invertebrates and algae and are compared against dissolved metal concentrations at the end of the T/D study. 

In order to provide suitable data for the so called "soluble aluminium salts", aluminium sulphate was tested in a recent long term (28d) Transformation Dissolution study followed by a filtration step just prior to then exposure of the solution to Ceriodaphnia in a 7d reprotoxicity test. Two chronic reprotoxicity studies usingCeriodaphniawereperformed further to a 28 d TDp studyon aluminium sulphate 14 hydrate.Although this study did not provide conclusive evidence of non-toxicity due to the fact that the validity criteria in the controls did not meet Guideline obligations, there was nonetheless no indication of dissolved aluminium toxicity in either of these 7 day reprotoxicity studies using solution obtained from the 28 day TDp study and this is considered strong supporting evidence that aluminium salts should not be classified as toxic or harmful for the environment.


In a further stepby the European Aluminium Association,extensive testing was recently performed with a series of aluminium salts for the purpose of developing an environmental quality standard (EQS). Unlike other cationic metals, total Al is a better indicator of toxicity than either soluble or monomeric Al. For purposes of determining the ERV from these data, only studies which report total aluminium and which were of sufficient quality to be used in the BLM modeling of aluminium were used for classification purposes. Theirreview of existing and recently generated data indicates that toxicity does not correlate well with dissolved or monomeric Al. This suggests that most, if not all, the acute effects being observed are due to physical effects on the gills or respiratory membranes due to coating/smothering with aluminium hydroxide and does not reflect intrinsic toxicity. This is true even for studies receiving a Klimish rating of 1, i.e., the study was performed well and followed all the guidelines, but solubility was exceeded and hydroxide polymers were formed. Recognizing the limitations of these data and the extreme difficulty of deciding which data were sufficiently reliable for use or not for classification, the EAA chose to evaluate the available acute data by developing BLM models for invertebrates, fish and algae using the best data available. The models were then used to generate LC/EC50 values under standard conditions, i.e., pH 6 and 8 with a hardness of 100 and a DOC of 2 mg/L. These water quality parameters are considered reasonably conservative values for classification purposes and mimic water quality in standard OECD test media. 


Table 1 "Ecotox reference values generated using acute BLMs for aluminium; hardness 100 mg/l and DOC 2 mg/l"

(below) lists the ERV developed for comparison against the CIMM T/D results for 7-day studies. The values generated for fish arefrom the BLM developed from the NIVA data (based upon Atlantic Salmon, most sensitive fish species).  The invertebrate and algae values were generated using multiple linear regression models (models developed based upon CIMM data). For the invertebrate model, pH 8 values were extrapolated since the model used pH 6 and pH 7 results because there was typically insufficient mortality in the pH 8 tests forC. dubiaeven though the test concentrations were in the high mg/L range (i.e., EC 50 values are reported as the highest test concentration).

The EAA did not include discussin on T/D data for 28 days as they stated that Aluminium is highly insoluble at pH 6 and so is rapidly lost from the water column at neutral and alkaline pH values. Hence, the rapid loss from the water column argues for no chronic classification as agreed to by the C&L Committee.

Table 1. Ecotox reference values generated using acute BLMs for aluminium; hardness  

100 mg/l and DOC 2 mg/l



EC 50 Value (mg/L): pH 6

EC 50 Value (mg/L): pH 8











(Selenastrum capricornutum)





Final ERV




Further to a TDp study on almuinium sulphate 14 hydrate a 7 day Ceriodaphnid reprotoxicity study was performed on the filtered TDp solution where no toxicity was observed. Furthemore, toxicity was estimated for both the unfiltered/unamended T/D solution and the toxicity test waters using an aluminum biotic ligand model (BLM). In both cases, aluminum concentrations predicted to be toxic by the BLM exceeded measured aluminum concentrations in the test solutions, thus suggesting that no toxicity would be expected.

It is concluded that no true PNEC can be determined for this substance within environmentally relevant boundaries of pH, hardness and DOC.



Blust, Ronny, Peter Campbell and Claude Fortin. 2010.Assessing the risks associated with metals that change speciation and/or form precipitates under natural environmental conditions – comparison of chemical equilibrium models, Report to Industry Ecotoxicity Technical Assessment Panel (ETAP).


Euras. 2007.  Development of a high quality aquatic ecotoxicity database for Al metal, Al oxide and Al hydroxide, ARCHE Company, February 2007.


NIVA. 1996. Summary of 1996 Ecotox Studies. Effects of aluminium oxide, hydroxide, powder and metallic shavings on fish, daphnia and algae. Sponsored by the European Aluminium Association; issued by the Norwegian Institute for Water Research.


Conclusion on classification

It is important to point out that the substances being registered in this dossier are, despite standardised study results from OECD 105 which suggest the contrary, sparingly soluble forms of aluminium and not soluble metal salts. This has been substantiated following the Transformation Dissolution protocol as recommended by CLP in which the dissolved concentration of Al was observed to decrease to the µg/L level within 24 h and decreased to approximately 12 µg/L over the remaining 27 days of the study. Information is provided below on the classification of the soluble form – aluminum chloride and on an evaluation of the transformation-dissolution of aluminum powders. An overall conclusion on classification is drawn and presented.

7.6.1 Non-classification of "Sparingly" Soluble Forms of Aluminium


Lines of Evidence for No Classification


  1. Transformation/dissolution (T/D) tests followed reprotoxicity testing of the T/D solutions on Ceriodaphnia indicated that the above forms of Al do not classify. Moreover, further T/D tests performed by other aluminium consortia on fish, invertebrates and algae did not indicate a need to classify their aluminium salts
  2. T/D tests performed at CIMM indicated insufficient solubility of the above forms of Al to classify.

These lines of evidence are examined in the sections below.


Review of the Existing Transformation – Dissolution and Toxicity Data on Al


Available data indicate that aluminium salts are relatively non toxic and this was sufficient for the EU Classification and Labelling Committee (1999) to determine that there was no need for classification of aluminium chloride. In fact there is considerable evidence from TD studies that the so called "soluble" aluminium salts become extremely sparingly soluble or even highly insoluble when placed in environmental media under non-extreme forms of pH, DOC and hardness and are thus non-hazardous. In principal, additional testing for classification is not needed. However, the aluminium industry decided to evaluate the dissolution of Al substances as a means of demonstrating proper stewardship of its products. 


Studies reported in the literature have been performed repeatedly with test solutions based on soluble salts with aluminium concentrations above that of its solubility limit. Due to physical effects of precipitated material most of these studies are meaningless for the investigation of intrinsic toxicity. Aluminium ions released to surface waters quickly form insoluble aluminium hydroxides in mixing zones. These colloids can sorb to fish gills resulting in asphyxiation and mortality in rare instances. In laboratory studies, reports of asphyxiation are common and true or intrinsic toxicity appears to be lacking. The rapid formation of the complex hydroxides in neutral and alkaline waters complicates the assessment for classification for several reasons. First, the observed mortality is does not appear to be due to intrinsic toxicity. Second, the LC and EC50 values are typically above 1 mg/L (there are a couple of values in the 0.5-1.0 mg/L range) and all of the effects values are above the solubility limit of aluminium hydroxide. A report on the solubility of Al in OECD test media is attached and was summarised at the January 26, 2010 meeting with ECHA (Blust et al 2010). And third, there is a lack of consistency in reporting the LC/EC50 values as total, dissolved or monomeric in the literature. Recent studies performed by the European Aluminium Association demonstrate that the only reliable predictor of toxic effects to aquatic organisms is total aluminium. This has been demonstrated across a broad range of pH, DOC and hardness values. Hence, for classification purposes we chose to use “total” Al (i.e., non filtered samples) in the test media as the most appropriate parameter for reporting LC and EC50 values.



Rodriguez, Patricio. 2007.Final report to the European Aluminium Association OECD Dissolution Transformation Tests for aluminium hydroxide, aluminium oxide and metallic aluminium powder and massive forms. Chilean Mining and Metallurgy Research Center (CIMM). March 2007.


Van Gestel CAM, Hogerwerf G. 2001. Influence of soil ph on the toxicity of aluminium for Eisenia andrei in an artificial soil substrate, Pedobiologia 45: 385-395 ,