Registration Dossier

Diss Factsheets

Ecotoxicological information

Endpoint summary

Administrative data

Description of key information

Additional information

Since the original REACH dossier submission in 2010, four studies relating to the terrestrial chronic toxicity of silver to microbes, plants and soil invertebrates have been completed. As a full chronic terrestrial data set is available a soil PNEC can be derived based on the lowest EC10 with an assessment factor of 10. The most sensitive chronic effect is an EC10 of 0.13 mg Ag/kg dw for shoot dry weight in Lactuca sativa. The soil PNEC based on this data is therefore 0.013 mg Ag/kg dw.

The terrestrial PNEC of 0.013 mg Ag/kg is derived using a deterministic approach, by applying an assessment factor of 10 to the lowest NOEC from a base-set of chronic terrestrial data. The scientific validity and feasibility of implementing this PNEC in order to robustly assess potential risks from silver exposures in soils is questionable.

Silver is a naturally occurring element and therefore would be expected to be found in all soils, but at relatively low concentrations. The FOREGS dataset contains silver concentrations for 840 topsoils from relatively pristine locations from 26 European countries. These data give a mean silver concentration in topsoils of 0.34 mg/kg (standard deviation 0.23) and 5th percentile of 0.08 mg Ag/kg. These data are in line with others who have reported ambient background silver concentrations in topsoils (Purcell and Peters 1998; Kramer et al. 2002). The mean silver concentration for pristine soils is 26 times higher than the terrestrial PNEC. The PNEC is also below the 5th percentile measured concentration, suggesting that this effect concentration is probably not a robust metric by which to assess silver terrestrial risks as risks would be identified across nearly all of Europe even in relatively pristine locations.

The chronic toxicity testing used to derive the PNEC was conducted to comply with the relevant OECD Guidelines for Testing of Chemicals, yet some issues and concerns relating to the design and methodology of the experimental work have been identified. This context is provided below.

There are limitations in the interpretation of the results from these toxicity tests, as in all cases only one “soil type” was used for each test. The test media needs to be considered when interpreting toxicity results and the influence of the test media on the partitioning and toxicity of silver. This is particularly the case for the plant emergence, survival and growth test reported here as this was conducted in 100 % silica sand. The use of sand as a test media has little environmental relevance and will result in relatively high bioavailability and toxicity of silver and may considered as a very “worst-case scenario”.

Soil factors are known to influence the sorption-desorption behaviour of silver, with increasing pH, organic matter and clay content having all been noted to reduce desorption and reduce availability and so silver bioavailability (Hou et al. 2005; Jacobson et al. 2005; Janik et al., 2010; Smith and Carson 1977). While these affects have been observed, no targeted attempt has been made to quantify them or develop empirical relationships to explain them for silver. The chronic testing on which the PNEC is based have been performed on soils (sand) that would reasonably be expected to reflect conditions of relatively high bioavailability (low pH, low organic matter and low clay content). The mitigating influence of soil properties upon chronic ecotoxicity has long been identified for metals and is summarised in MERAG (ICMM 2007). The source of much of the material contained in this document was the research and development activities of commodity groups driven by the Existing Substances Regulations (EEC 793/93). The Risk Assessment Reports (RARs) for nickel and zinc (and their compounds)(EC 2008a; EC 2008b) and the Voluntary Risk Assessment Report (VRA) for copper and its compounds (ECI 2007) explicitly state that the sorption of substances in soils depends on soil characteristics such as organic matter content, clay content and pH, therefore bioavailability and toxicity to soil organisms will also depend on soil characteristics. Effectively this means that the summary statistic for a metal from an ecotoxicity test in a particular soil with certain characteristics can only be used to assess risk in that particular soil.

In order to overcome this challenge it is clear that the terrestrial ecotoxicity data for a specific metal need to be normalised to standard conditions (as had been done in the previous ESRs a range of bioavailability scenarios, different soils bracketing the 10th to 90th percentily of EU soil conditions). In order to do this there is a requirement for considerable investment of research time and resources to identify and quantify empirical relationships between soils properties and ecotoxicological effect of a specific metal. It also needs to be stressed that the relationships between soil properties and sorption/bioavailability of metals is dependent upon both abiotic and biotic factors. Importantly, it has been shown that the relationships are metal-specific, meaning that read across from the relationships established for one metal is not possible (Van Gestel et al. 1995).

In addition to the soil related factors there are also artifacts of the laboratory tests that can also have an impact on the bioavailability of metals and so the interpretation of the ecotoxicity data. There is considerable evidence in the literature showing that spiking soils with soluble metal salts (such as silver nitrate) to determine toxicity is inadequate in terms risk assessment. This is due to the immediate changes that result in the spiked soils, which make them considerably different from field contaminated soils. The addition of soluble metal salts to soils can lead to changes in various soil properties that can results in indirect and direct effects to exposed organisms. This is due to the added ions replacing the absorbed ions from the soil exchange sites. This process results in a sharp increase in the ionic strength of the soil solution and a decrease in the pH, which can in turn have a large influence on the bioavailability of the metals (Speir et al., 1999; Stevens et al., 2003; Smolders et al., 2009; McLaughlin et al, 2011). The approach to counteract this effect is to subject the spiked soils to a process of leaching and ageing. This process of leaching removes excess ions form the soil solution, which reduces the ionic strength. This process also results in an increase in pH, and makes soils more representative of field soils which naturally undergo a process of leaching. The process of metal ageing in soils will result in a decrease in the labile (effectively bioavailable) pool of metal ions as they become incorporated into the soil-solid phase. This ageing process of experimental soils is important as it allows the metals ions to equilibrate in the soil to reach a point that is more representative of field soils. By including processes of leaching and ageing in experiments that involve spiking metal salts, an influence of the resulting toxicity has been observed (Stevens et al., 2003; Smolders et al., 2009). The toxicity tests subject to review in this report did not account for, or include, leaching or ageing of the soils. Therefore the relevance of the results in terms of field effects is unknown.

As part of the EU RARs and under the research programs designed to provide material for REACH, considerable data has been generated on the behaviour and toxicity of metals such as cadmium, cobalt, copper, nickel, lead and zinc. For these metals, it was possible to develop models of metal bioavailability and ageing in soils spiked with soluble metal salts, taking note of artefacts in metal spiking procedures (Stevens et al. 2003) and ageing of metals in soils over time (which reduces bioavailability). This could be combined with species sensitivity distributions and models for bioavailability of metals across soils, to produce soil-specific metal predicted no effect concentrations (PNECs). A summary of this process for the EU regulatory process is described in Smolders et al. (2009) and best science procedures for developing soil quality standards and soil PNEC are described more generically in a recent book chapter by McLaughlin et al. (2010).

The EU RARs and VRA have been undertaken for metals that could be considered to be relatively data rich for terrestrial endpoints when compared to silver (e.g. there are over 150 chronic endpoints for zinc and 250 for copper). Therefore, it was clear from existing test data, prior to undertaking extensive testing programmes, that there was likely to be a dramatic effect of soil properties on mitigating ecotoxicity although the precise form of that effect was not established. From the testing programmes undertaken by the International Copper Association the mitigating effects of abiotic and biotic factors were observed to give an order of magnitude difference between the HC5s when the soil ecotoxicity database was normalised for soils that covered the calculated 10th to the 90th percentile of EU soils (ECI 2007). For nickel the difference between the HC5’s across the same soil range is over two orders of magnitude (EC 2008b).

Therefore, it can be hypothesised that, even in light of the current relatively limited data on the sorption of silver in soil, that soil properties (and leaching and aging) may influence the chronic ecotoxicity. Therefore, for silver, it was imperative to assess the likely magnitude of any mitigating effects prior to the development of a full testing proposal and research programme.

A ‘pilot’ exercise was undertaken to assess of the potential variation in silver toxicity to the most sensitive receptors so far tested (plants) in a range of soils with varying properties. The study was undertaken by scientists at CSIRO in Australia, led by Professor Mike McLaughlin. Professor McLaughlin is a Chief Research Scientist with the CSIRO Sustainable Agriculture Flagship and was a Chief Investigator on the Metals in Asia project (developing soil standards for Cu and Ni in China), the GEMAS metal partitioning project, and the EU risk assessment projects for Co, Cu, Mo, Ni, and Zn, as well as contributing to risk assessments of B, Cd, and Pb.

The report is summarised as a supporting study in the plant toxicity section and is attached in full to this endpoint summary. However, a brief summary of the findings are presented below: 

·        Six soils six soils were selected that represented a range of properties. The soil properties considered to have the greatest influence on sorption of silver were clay content (range from 1.4 - 42%), pH (range from 3.6 -8.0) and soil organic carbon (range from 1.5 -6.9%). Two of the soils were collected from Europe and the rest were from Australia. Broos et al. (2007) and Li et al. (2010) have demonstrated that the soil property-based relationships developed to predict the ecotoxicity of metals for Australian, European and Chinese soils are very similar.

·        The most sensitive endpoints from the chronic ecotoxicity data set available for silver were for plants. The pilot study was undertaken performing standard test methods to measure the acute effects of silver on shoot length and root elongation across a range of soils. The results showed that the toxicity of silver was found to vary considerably between the different soils. The concentrations corresponding to a 10% effect (EC10) ranged from 9.8 to 505 mg Ag/kg soil for the endpoint of root length and from 3.7 to 578 mg Ag/kg soil for the endpoint of shoot length. The soils that showed the lowest degree of toxicity contained high levels of clay and organic carbon and had high pH.

·        The spiking of the soils with AgNO3 (the most bioavailable silver salt) resulted in a decrease in soil pH and an increase in ionic strength of the soil solution. These artifacts of spiking with soluble metal salts may have an impact of the toxicity test results, therefore, for future tests leaching of the soils would be recommended.

Comparison of the soil PNEC derived from the base-set of chronic data tested in sand of 0.013 mg Ag/kg to monitored silver concentrations in European soils clearly demonstrate that the proposed PNEC is not a realistic metric to assess potential risks of silver in soils across Europe. Therefore, we propose a testing research programme, similar to those followed by other commodity groups, such as copper and nickel, be undertaken in order to deliver a method to account for experimental artefacts and biotic and abiotic factors that will influence silver bioavailability.

The programme will be under taken by Professor McLaughlin’s group at CSIRO in Australia and will have the following key modules that closely reflect the experiences of other commodity groups and the summary provided by Smolders et al. (2009):

·        Collection of 8 soils varying in physico-chemical characteristics and spiking with soluble Ag+ ion (silver nitrate), removal of artefacts due to metal spiking to develop “leaching factors”. 

·        Examination of changes in labile (potentially available) Ag over time in soils using isotope dilution procedures, to develop “ageing factors”.

·        Examination of the toxicity of Ag+ to soil microorganisms, terrestrial plants and soil invertebrates, to produce critical toxicity thresholds.

·        Examination of how soil factors affect toxicity of Ag+, to produce soil bioavailability models from which it will be possible to normalise all chronic silver ecotoxicity data to specific soil types (as undertaken for copper, nickel and zinc).

This testing proposal was submitted to ECHA as part of the 2010 silver and silver compound registrations.There were no objections to the testing proposal by ECHA or the Member State’s Competent Authorities and data will be added to this dossier once results are available.

Broos K. Warne MStJ. Heemsbergen DA. Stevens D. Barnes MB. Correll RL. McLaughlin MJ. 2007. Soil factors controlling the toxicity of copper and zinc to microbial processes in Australian soils. Environ Toxicol. Chem. 26:583–590.

European Commission (EC). 2008a. European Commission, 2008 European Union Risk Assessment Report: Zinc Metal. CAS-No. 7440-66-6, EINECS-No. 231-175-3. Part I Environment. Final report May 2008. European Commission Joint Research Centre, European Chemicals Bureau. Luxembourg: Office of Official Publications of the European Communities.

European Commission (EC). 2008b. European Union risk assessment report on nickel, nickel sulphate, nickel carbonate, nickel chloride, nickel dinitrate, Denmark, Final report May 2008. Prepared by Denmark, Danish Environmental Protection Agency on behalf of the European Union.

ECI, 2007. European Union risk assessment report on copper, copper(II) sulphate pentahydrate, copper(I) oxide, copper(II) oxide, dicopper chloride trihydroxide. Voluntary risk assessment, draft February 2007. European Copper Institute.

Hou H. Takamatsu T. Koshikawa MK. et al. 2005. Migration of silver, indium, tin, antimony, and bismuth and variations in the chemical fractions on addition to uncontaminated soils. Soil Sci. 170:624-639.

ICMM 2007. Metals environmental risk assessment guidance. ISBN: 978-0-9553591-2-5

Jacobson AR. McBride MB. Baveye P. et al. 2005. Environmental factors determining the trace-level sorption of silver and thallium in soils. Sci. Total Environ. 345, 191-205.

Janik L. Forrester S. Kirby JK. McLaughlin MJ. 2010. Silver: Predicting metals partitioning in soils of the GEMAS sampling program using isotopic dilution and infrared spectroscopic techniques. CSIRO Land and Water, Adelaide.

Kramer JR, Benoit G, Bowles KC, Di Toro DM, Herrin RT, Luther III GW, Manolopoulos H, Robillard KA, Shafer MM, Shaw JR. 2000 Environmental chemistry of silver. In: Andren AW, Bober TW, eds, Toxicity of Silver in the Environment: Transport, Fate, and Effects. SETAC Press, Pensacola, FL. 200pp.

Li B. Ma Y. McLaughlin MJ. Kirby J. Cozens G. Liu J. 2010. Influences of soil properties and leaching on copper toxicity to barley root elongation. Environ. Toxicol. Chem. 29:835–842.

McLaughlin MJ. Lofts S. Warne M. St.J. Amorim MJB. Fairbrother A. Lanno R. Hendershot W. Schelkat CE. Ma Y. Paton GJ. 2010. Derivation of ecologically based soil standards for trace elements. In Soil Quality Standards, Eds. G. Merrington and I. Schoeters, SETAC Press, Pensacola, Florida.

Purcell TW. Peters JP. 1998. Sources of silver in the environment. Environ. Toxicol. Chem. 17. 539-546.

Smith IC. Carson BL. 1977. Trace metals in the Environment. Vol 2. Silver. Ann Arbor Science. Ann Arbor MI, USA.

Speir, TW, Kettles, HA, Percival, HJ, Parshotam, A, 1999. Is soil acidification the cause of biochemical responses when soils are amended with heavy metal salts? Soil Biol. Biochem. 31, 1953-1961.

Smolders E. Oorts K. Van Sprang P. Schoeters I. Janssen CJ. McGrath SP. McLaughlin MJ. 2009. Toxicity of trace metals in soil as affected by soil type and aging after contamination: Using calibrated bioavailability models to set ecological soil standards. Environmental Toxicology and Chemistry 28, 1633-1642.

Stevens DP. McLaughlin MJ. Heinrich T. 2003. Determining toxicity of lead and zinc runoff in soils: salinity effects on metal partitioning and on phytotoxicity. Environmental Toxicology and Chemistry 22, 3017-3024.

Van Gestel CAM. Rademarker MCJ. Straalen NM. 1995. Capacity controlling parameters and their impact on metal toxicity. In: Salomons W. Stigliani WM. (eds), Biogeodynamics of pollutants in soils and sediments. Springer, Berlin. Pp 171-192.