Registration Dossier

Ecotoxicological information

Endpoint summary

Currently viewing:

Administrative data

Description of key information

Additional information

In line with the risk assessment/classification approach adopted for other metals and inorganic metal compounds (ECHA 2012), terrestrial toxicity information requirements are “read across” from the properties of the soluble silver salts. Almost exclusively these are studies that used silver nitrate (AgNO3) as the test substance. Silver nitrate is considered to be the form of silver with the greatest toxicity. Where applicable, further justification of the validity of this read-across approach is made in endpoint summaries.

The derivation of the PNECsoil is based on the findings of a research programme undertaken by the Commonwealth Scientific and Industrial Research Organisation (CSIRO) into the toxicity and bioavailability of silver in soils (funded by the European Precious Metals Federation). The PNECsoil accounts for the bioavailability of silver in six soil types typical of ecoregions in Europe. This approach is consistent with previous risk assessments under the former Existing Substances Regulations (e. g. EC 2008) and most recently for copper, nickel and lead under REACH. Further information on the PNECsoilderivation is provided in the appended PNEC summary document.

The PNECsoil for silver has been derived from chronic, reliable and relevant ecotoxicity data. There are no formal criteria for minimum species and taxonomic diversity before a terrestrial dataset is considered sufficiently comprehensive for use in PNEC derivation via a species sensitivity distribution (SSD). However, the requirements of the aquatic compartment (REACH Guidance Chapter 10, Section R.10.3.1.3) offers a useful benchmark against which to assess the appropriateness of the available terrestrial dataset for an SSD approach to PNEC derivation. The aquatic criteria outline a minimum requirement of NOECs/EC10s for more than 10 species (preferably 15) from across eight different taxa. The normalised dataset for silver contains 11 species and a single microbial functional process from across six taxa (oligochaeta, crustaceans, insects, monocotyledons, dicotyledons and nitrifying bacteria). The diversity requirement of at least eight different taxa from the aquatic compartment is therefore not met. However, the diversity of organisms represented in the silver database is consistent with other metals that have adopted an SSD approach to PNEC derivation. An SSD approach is therefore also considered appropriate for silver.

The normalisation process adopted enables the ecotoxicity database to be corrected to a specific soil scenario, which is effectively the same as having all of the tests in the database undertaken in that same soil. However, the requirement for REACH is to assess the potential risks across soil types representing conditions between the 10th and 90th percentile of all those in Europe (EC 2008). We have followed the same eco-region approach to that used in previous risk assessments undertaken to fulfil the requirements under the Existing Substances Regulation and REACH for other metals.

This approach uses six soil types from different regions across the EU that are indicative of typical conditions and which also cover a wide range of physico-chemical conditions (pH between 3.0 and 7.5, clay between 7 and 46 %). These conditions would be expected to provide a broad range of silver bioavailability. The bioavailability scenarios are shown in Table 6.3. However, the German Forest soil with the pH of 3, while used in previous risk assessments for trace elements, falls considerably below the 10th percentile of pH values in the FOREGS dataset and is less than the 2nd percentile of pH values from the more recent GEMAS European soils dataset. Therefore, it has not been taken further in this assessment as, while it may be of site-specific relevance, it is not of regional relevance.

The soil physico-chemical parameters from the eco-regions (excluding the German forest soil) have similar pH values to the soils used in the CSIRO study and the 10th to 90th percentiles from FOREGS. Further, the soils shown in Table 6.3 have a lower organic carbon minimum of 0.3% compared to 0.9% for the CSIRO soils and a higher maximum of 25%, compared to 12% for CSIRO soils. The range of clay contents is greater for the CSIRO soils compared to the eco-region soils. Unsurprisingly, the physico-chemical ranges for soils from laboratory ecotoxicity testing of silver tend to be narrower than those from the eco-regions and evaluation undertaken by CSIRO.

For each soil eco-region scenario, the HC5-50 was determined by full inter-species normalisation approach using ETX (RIVM ETX 2.0 software). A single species/function geometric mean was calculated from normalised EC10/NOEC data for each species and endpoint. The lowest (most sensitive) resulting value for each species was then applied in the SSD. This resulted in HC5-50 values which range over two orders of magnitude from 0.64 mg Ag/kg dry weight in the Loamy Spanish soil, with the lowest organic carbon content, to 10.56 mg Ag/kg dry weight in clay-rich soil from Greece. Goodness of fit criteria (in both the body and the tails of log-normal distribution) are met for all the SSDs.