Mill scale (ferrous metal)

Endpoint summary

Description of key information

 

Reviews on the toxicity iron to aquatic organisms have pointed out that it is difficult to assess clearly whether the observed effects from exposure to iron are due to chemical toxicity or physical effects (See Chapter 4 of the Chemical Safety Report: Environmental Fate Properties). Many of the observed effects appear to be due to particulate iron hydroxides, and no clear evidence of chemical toxicity has been identified (Vangheluwe and Versonnen, 2004).Indeed, the majority of both historic and recent ecotoxicity tests, have probably tested the physical effects of precipitated iron (oxy)hydroxides, and reliable information about actual exposures is often limited. In many cases the exposure concentrations have been reported as nominal concentrations only, and analysis of both total and dissolved (i.e. 0.45 µm filtered) iron are relatively uncommon. As a result of this the mechanisms by which iron acts to cause adverse effects on aquatic organisms are unclear. Whilst it is theoretically possible that chemical toxicity of some iron species is the cause of the observed effects, it is highly likely that the effects reported in laboratory experiments are due to physical effects of precipitated iron minerals clogging gills and other sensitive biological membranes, or reducing the efficiency of food collection by filter feeding organisms. This is supported by the fact that these studies are always performed with oxic water, most often at pH values above 7 and well above the solubility of ferric hydroxide.

Acute laboratory toxicity studies with aquatic species (fish, invertebrates, algae) indicate that the range of adverse nominal concentration levels (total added) is situated between 1 and 1000 mg/L, with the majority between 10-100 mg/L. The long-term adverse concentration levels are observed in a comparable range, i.e. 0.3 - 1000 mg/L (nominal levels, total added), with the vast majority exceeding 1 mg/L, and most values higher than 10 mg/L.

The majority of data that were generated with soluble iron salts showed the accumulation of iron on gills (smothering effects), fouling of feeding appendages and covering of algal cell walls (blocking of photosynthesis), thus causing a number of physical effects that no not represent an intrinsic property of iron. Data from these studies cannot be used for the derivation of a PNEC:

• The vast majority of available ecotoxicity data that were generated under laboratory conditions were generated in tests were adverse effects of soluble iron salts (FeCl3, Fe2(SO4)3) were investigated rather than using less soluble iron compounds (iron metal/powder, iron oxides). Effects data are often expressed as nominal concentrations, despite the fact that concentration levels were well above the solubility limit. Hence, nominal concentration level do not properly reflect the dissolved fraction under the test conditions. 


• As only a very small fraction of iron ions will be present in test media as ‘true dissolved’ iron, it is likely that secondary effects arising from the precipitation of ferric hydroxide and/or interaction of iron colloids with biological membranes are the main contributor to the adverse effects that were observed in ecotoxicity tests.


• All of the identified studies fail to report the adverse effects as a function of Fe(II) or Fe(III). Effect levels are reported as dissolved or total iron and are therefore not meaningful for assessing the intrinsic toxicity as such values predominantly represent the colloidal/particulate fraction.


• Many studies fail to provide adequate measurements of key physicochemical parameters (pH, DOC) that affect the iron speciation and bioavailability (pH, DOC) before/during/at the end of the exposure period. 


• Most of the historic ecotoxicity tests on iron are likely to have tested the effects of a suspension of precipitated iron (oxy)hydroxide: exposure details are generally relatively limited, leading to significant uncertainties surrounding the relevance and reliability of many laboratory toxicity tests.

 

 

Intrinsic toxic effect such as disruption of ionic transport systems and oxidative stress have been noted for iron in aqueous solutions of low pH and high DOC content, but such conditions are atypical of the natural environment: at ambient conditions, the natural background concentrations of dissolved iron are expected to be at equilibrium. Therefore any addition of iron would lead to precipitation from solution, and no intrinsic toxicity would be expected under ambient conditions. Some studies have suggested that ferrous ions possess a higher intrinsic toxicity, but a reliable quantification of this intrinsic toxicity cannot be achieved. Based on the half-life for oxidation and precipitation of ferrous iron, it is anticipated that during the typical timescale of standard test protocols, a significant proportion of added ferrous salts would have been converted to ferric hydroxide when added to standard artificial test medium. Precipitates of ferric ions contribute to physical adverse effects through mechanisms like ‘smothering’ of respiratory surfaces, fouling of feeding appendages and covering of photosynthetic surfaces. Consequently, reported adverse effects are most likely due to physical effects. This is further supported by the observation that the identified L(E)C50 -, EC10- and NOEC-values exceed the equilibrium concentration of dissolved ferric iron by several orders of magnitude (solubility of 6.16 * 10^-5 µg Fe(III)/L and 6.16 * 10^-10 µg Fe(III)/L at pH 6 and 8, respectively).

It is thus concluded that standard toxicity tests cannot provide ecotoxicity reference values that represent the intrinsic toxicity of iron (Fe(2 +) and/or Fe (3 +)) and which could serve as starting point for determination of a PNEC (SSD-approach or Assessment Factor (AF) approach).

Additional information

Information on field studies

 

Reviews of the toxicity of iron to aquatic organisms have indicated that invertebrates are likely to be the most sensitive group of organisms, suggesting that an assessment of the effects of iron on invertebrate communities in the field may provide a reliable indication of the effects of iron on aquatic ecosystems. A review by the Environment Agency (2007), for instance, identified Daphnia magna as the most sensitive species in chronic tests, with a chronic NOEC for reproduction of 0.16 mg/L total iron (nominal) (Dave 1984). This study however, likely assessed exposure to a suspension of iron precipitates, rather than truly dissolved iron, as the pH of the test was in the range 7 to 8. The formation of precipitates was noted, and the exposure concentrations were expressed as nominal concentrations only. This study used solutions which had been aged for less than 30 minutes, and showed greater toxicity than similar tests where longer ageing periods were used (e.g. Biesinger and Christensen 1972).

Analyses of field data are not able to identify the specific cause of any effects observed, in terms of whether any adverse effects due to iron are caused by either chemical toxicity or as a result of smothering by precipitates. Whilst a deterioration in ecological quality with increasing iron concentrations may be identified in some cases, and this deterioration may be linked to iron exposures, the specific cause of the effects cannot be identified.

A field study of the effects of iron on benthic invertebrates (Rasmussen and Lindegaard 1988) provides some evidence of effects on aquatic ecosystems. Water samples were taken between 1979 and 1980 from the River Vidaa, Denmark. The samples were analysed for total iron, dissolved iron and pH. Concentrations of annual average dissolved iron from 28 sites ranged from 0 to 32 mg Fe(II)/L. Dissolved iron was reported to consist almost entirely of Fe(II), and consequently dissolved (i.e. <0.45 μm filtered) iron concentrations were assumed to be Fe(II). Samples of benthic invertebrates were taken at the same sites. pH was between 6.7 and 8.8 at the different sites. Numbers of taxa were correlated to the concentrations of different iron components.

The number of taxa was negatively correlated to iron concentrations expressed as annual average Fe(II), maximum recorded Fe(II), annual average total iron and winter average Fe(II). At concentrations below 0.2 mg Fe(II)/L 67 taxa were collected, and between 0.2 and 0.3 mg Fe(II)/L 53 taxa were recorded. Taxa that were eliminated were primarily grazers that feed on biofilm. Up to concentrations of 10 mg Fe(II)/L taxa continued to be eliminated, with 10 taxa left at this concentration. It is not clear from this study whether dissolved iron is genuinely a better metric of iron exposure; other potential pressures on the benthic macroinvertebrate communities were not considered in this study. A further limitation is that no established thresholds or guidelines for the quality of benthic macroinvertebrate communities were used to derive the critical concentrations of iron reported in the study.

 

More recently the development of ecological assessment tools for various components of aquatic ecosystems, such as benthic macroinvertebrates, diatoms, macrophytes, and fish, have been developed for assessing the ecological quality of surface waters under the Water Framework Directive in Europe. Similar monitoring programmes have also enabled analyses to be made for North America. Through combining the results of these ecological analyses with chemical monitoring data for iron exposures it has been possible to undertake further studies into the effects of iron on ecological communities (e.g. Crane et al. 2007, Linton et al. 2007, Peters and Crane 2009, Environment Agency 2010a).

In order to avoid complications in the analysis, due to other potentially confounding factors on the observed responses, these studies have considered limiting functions which describe the maximum achievable ecological quality as a function of increasing iron exposure. One advantage of some of these more recent studies is that it has been possible to express the ecological quality at each individual site relative to a predicted reference condition, which significantly reduces uncertainties related to whether or not particular taxa would be expected to be present if the site were unimpacted.