Mill scale (ferrous metal)

Short-term toxicity to fish

Description of key information

A substantial number of acute toxicity studies with fish has been identified of iron, but due to one or more significant deficiencies in the testing procedure and/or limited reporting of relevant data, none of these studies can be considered as a key study on its own. Hence, a weight of evidence approach was used for assessing this endpoint.

 

 The main deficiency that is applicable to virtually all studies, is a limited characterization of exposure concentrations. All tests were conducted at nominal concentration levels that exceed the solubility of the ferrous/ferric free ion, and therefore exposure concentrations represent a mix of free iron ions, colloidal iron forms (low and high molecular weight), iron bound to dissolved organic matter, and precipitated Fe-species. None of the studies report a proper characterization of the Fe-speciation, but it is important to note that Fe-speciation is defined by a dynamic equilibrium that is governed by several processes with (short) transformation half-lives. Therefore, it is virtually impossible to properly qualify and quantify the time-dependent Fe-speciation profile during a standard 96h fish toxicity test.

 

Studies with LC50-values that are based on nominal values are therefore considered unreliable. The availability of measured data, however, does not ensure that the reported values are meaningful. Some studies (e.g. Debnath et al, 2015; Shuhaimi-Othman et al; 2013) report measured iron data but samples were acidified prior to analysis without any indication that a filtration step was included in the sample treatment.

 

The few studies that do refer to a measured dissolved fraction also report that adverse effects were apparently related to due to gill smothering by precipitated iron and a subsequent lack of oxygen (e.g. Decker and Menendez, 1974 ; Gonzalez et al, 1990). Therefore, the effects do not reflect intrinsic Fe-toxicity.

All studies that reported measured-based LC50-values were considered in the overall weight-of -evidence approach.

 

Taking all data into account (K1-K2-K3), dissolved LC50-concentrations of dissolved Fe2+/Fe3+ were situated  between 50 and 8650 µg/L for (fish species: O. mykiss, S. fontinalis, S.trutta) (Mattock, 2002; Dalzell and Macfarlane, 1999; Decker and Menendez, 1974). All other studies reported LC50-vaues well above 1 mg/L, but here it was often not clear which Fe-fraction was represented by the value (acidified, total, nominal,…).

 

In addition, three acute D.rerio studies, equivalent to OECD 203, were conducted with insoluble iron compounds (iron(II,III) oxides, iron hydroxide oxide) and no effects were observed at the highest test concentrations which ranged between 10-100 g/L test material (nominal values).

Additional information

Available data suggest that iron salts are relatively non-toxic and this was sufficient for the EU Classification and Labelling Committee to determine that there was no need for classification of iron salts. It was also concluded that iron massive and sparingly soluble forms of iron are highly insoluble and non-hazardous.

 

All of the identified studies used test solutions with iron concentrations above that of its solubility limit. Due to physical effects of precipitated material some of these studies are meaningless for the investigation of intrinsic toxicity. Iron ions released to surface waters quickly form insoluble iron hydroxides in mixing zones. These positively charged iron colloids will react with the negatively charged mucus that lines the fish gill. This accumulation of iron on the fish gill results in physical effects.

 

Iron has complex redox chemistry. In very special conditions transient iron species can be formed that cause toxicity. These conditions, however, are not typical of most ambient conditions and are more representative of specific mixing zones. In ambient conditions, the dissolved natural background concentrations of iron are, in most cases, at equilibrium; therefore addition of iron would lead to the precipitation of iron compounds from solution and are therefore not intrinsically toxic (Jackson, Versfeld & Adams 2010, Peters, Brown & Merrington 2010).