Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

The new ECHA CHEM database has been released by ECHA, and it now contains all REACH registration data. There are more details on the transition of ECHA's published data to ECHA CHEM here.

Diss Factsheets

Environmental fate & pathways

Adsorption / desorption

Currently viewing:

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
adsorption / desorption, other
Remarks:
adsorption
Type of information:
read-across from supporting substance (structural analogue or surrogate)
Adequacy of study:
key study
Justification for type of information:
Substance considered to fall within the scope of the read-across 'Silver metal: Justification of a read-across approach for environmental information requirements' (document attached in IUCLID section 13).
Reason / purpose for cross-reference:
read-across source
Key result
Type:
Kd
Value:
4 023 L/kg
Temp.:
20 °C
Remarks on result:
other: Median value of 497 soils. Range 159 -> 4700. Standard temperature assumed.
Type:
Kd
Value:
159 L/kg
Temp.:
20 °C
Remarks on result:
other: Minimum partitioning. Standard temperature assumed.
Type:
Kd
Value:
> 4 700 L/kg
Temp.:
20 °C
Remarks on result:
other: Maximum partioning. Standard temperature assumed.
Details on results (HPLC method):
No data reported
Adsorption and desorption constants:
No data reported
Recovery of test material:
No data reported
Concentration of test substance at end of adsorption equilibration period:
No data reported
Concentration of test substance at end of desorption equilibration period:
No data reported
Transformation products:
not specified
Details on results (Batch equilibrium method):
The range of Kd values was 159 - >4700 L/kg, with a median of 4023. The 10th percentile of values was 1200 L/Kg and the 90th percentile >4700 L/Kg.
Statistics:
Median values and percentiles were calculated.
Validity criteria fulfilled:
not specified
Conclusions:
Nearly 500 soils collected from across Europe were tested to determine the Kd of silver, assilver nitrate using isotope dilution. The range of Kd values was 159 - >4700 L/kg, with a median of 4023. The 10th percentile of values was 1200 L/Kg and the 90th percentile >4700 L/Kg.
Executive summary:

Janik et al. (2010) determined the Kd values of silver in nearly 500 soils collected from across Europe. Isotope dilution was used with 110 Silver Nitrate (12 KBq as Ag NO3). After 72 hours equilibration, post spike, the solid solution partitioning coefficient for each soil was determined on the soil solution filtrates using gamma spectrometry. The measured Kd values ranged from 159 -> 4700 L/Kg, with a median value of 4023 L/kg. It should be noted that 42% of the 497 soils tested gave Kd values that were greater than 4700 L/Kg. The median Kd value of 4023 L/Kg will be used through this assessment.

Description of key information

Read-across from the dissolved silver ion is also applied to fulfil information requirements for uncoated and coated nanosilver. Supporting information for this read-across is summarised in endpoint summaries and in further detail in the appended read-across summary/justification document.

Key value for chemical safety assessment

Other adsorption coefficients

Type:
log Kp (solids-water in suspended matter)
Value in L/kg:
5.28
at the temperature of:
20 °C

Other adsorption coefficients

Type:
log Kp (solids-water in sediment)
Value in L/kg:
4.05
at the temperature of:
20 °C

Other adsorption coefficients

Type:
log Kp (solids-water in soil)
Value in L/kg:
3.6
at the temperature of:
20 °C

Additional information

Summary of available data for uncoated and coated nanosilver

The available studies of the environmental fate of nanosilver particles fall into three categories:

·        Aggregation and dissolution behaviour of nanosilver in solution;

·        Partitioning of nanosilver in environmental media; and

·        Fate of nanosilver during sewage treatment.

None of the studies have followed standard methods and a variety of different nanosilver materials have been tested. Many of the studies have used nanosilver materials which were prepared in-house and a variety of different surface coatings have been employed. These issues make detailed comparisons of the observations made in the different studies difficult, although several consistent patterns of behaviour have been observed.

Aggregation and dissolution behaviour of nanosilver inenvironmental media

There have been numerous studies of the aggregation behaviour of nanosilver particles in aqueous media. These studies have included both studies in a variety of synthetic freshwaters as well as several natural waters. The influences of ionic strength, cation composition, pH and natural organic matter on the aggregation behaviour of nanosilver have all been studied over a range of conditions. Many of these studies have used nanosilver particles which were laboratory prepared, following a variety of different methods, although some studies have used commercially available nanosilver materials. Consequently, the materials studied exhibit a variety of different sizes and coating materials (including studies on uncoated materials), although considerable consistency between the behaviour of the different materials has been observed.

A potentially important distinction can be made between different types of nanoparticle aggregation. These may be defined as homoaggregation, where nanoparticles interact with each other, and heteroaggregation, where nanoparticles interact with other materials such as naturally occurring colloids or environmental particles. In many cases the studies have not been able to distinguish between these different types of aggregation, although homoaggregation is likely to dominate in relatively concentrated solutions and in the absence of appreciable concentrations of environmental particles, whereas heteroaggregation is likely to be dominant where nanosilver particles are present at low concentrations and environmental particles are present. Cumberland and Lead (2009) studied the effects of pH, ionic composition and the presence of natural organic macromolecules (humic substances) on laboratory prepared nanosilver particles using flow field-flow fractionation. The nanoparticles were prepared in citrate, although some of them were treated in order to remove excess citrate from the materials prior to use (referred to as cleaned particles). The size of the materials was determined both by transmission electron microscopy (after drying) as 13.7 and 13.6 nm for the uncleaned and cleaned particles, respectively and also by dynamic light scattering (in solution) as 25 and 22 nm for the uncleaned and cleaned particles, respectively.

pH was found to have a relatively limited effect on the size distribution of the particles, although increases in ionic strength (at the same ionic composition based on monovalent cations) resulted in increased aggregation of the particles. Experiments at the same ionic strength, but in a medium based on divalent cations (Ca rather than Na), resulted in very rapid aggregation and precipitation of the particles. The addition of humic substances to the media resulted in greater stability of the nanoparticles, which is likely to be as a result of the formation of surface films on the nanoparticles. The cleaning procedure to remove excess citrate after production of the particles did not have a significant effect on the behaviour of the particles.

Delay et al. (2011) studied the stability of nanosilver particles in solution, using particles which were prepared in-house, either in ultrapure water, humic rich lake water, or a mixture of the two. Nanoparticles prepared in the presence of natural organic matter (from lake water) were larger than those prepared in its absence and had a lower net charge. It is unlikely that these particles are representative of those supplied commercially within Europe, although their behaviour, in terms of the influence of ionic strength and organic matter on their aggregation properties, was comparable to that observed in other studies. Increased ionic strength and concentrations of divalent ions (e.g. Ca) both resulted in an increased tendency for aggregation of the particles and the presence of natural organic matter increased the solution stability of the particles.

El Badawy et al. (2010) also studied the effects of pH, ionic strength and ionic composition on the stability of nanosilver suspensions. The studies were performed on five different nanosilver materials, four of which were prepared in-house and one of which was commercially produced. Consequently, this is one of the few studies which compare the effect of different coating types on the behaviour of the nanomaterials. The materials prepared in-house were prepared using hydrogen reduction (25 nm), citrate reduction (10 nm), sodium borohydride reduction (11 nm), and one which was stabilised using branched polyethyleneimine (11 nm). The commercially produced material was stabilised with polyvinylpyrrolidone (PVP: 200 nm). These resulted in particles with a variety of different hydrodynamic diameters, as determined by dynamic light scattering.

Whilst an increase in ionic strength resulted in aggregation of some of the forms of nanoparticles, it did not have an effect on the commercial polyvinylpyrrolidone stabilised particles. This was considered to be due to the fact that the coating results in steric stabilisation for this material, whereas the other nanosilver materials were either not stabilised (hydrogen reduced), electrostatically stabilised (citrate and sodium borohydride stabilised), or electrosterically stabilised (branched polyethyleneimine stabilised). A large effect of both ionic strength and pH was observed for the branched polyethyleneimine stabilised material and it was not possible to isolate the effect of a single variable.

The type of anions in the electrolyte solution (Cl- or NO3-) also had a significant effect on both the stability of the nanosilver materials in solution and their surface charges. Solutions containing chloride resulted in the formation of stable silver chloride precipitates, which appeared to result in increased stability of some of the nanosilver materials. It also suggests that free silver ions are present in some of the nanosilver suspensions. As has been observed in other studies, solutions containing divalent cations (e.g. Ca) tended to result in increased aggregation of nanosilver materials. The PVP stabilised material, however, was not affected by these changes to the ionic composition of the solutions.

Gondikas et al. (2012) studied the effect of a thiol containing amino acid (cysteine) on nanosilver and found that it increased the rate of release of silver from the particles, although very high cysteine concentrations were used, which may not be environmentally relevant, with the possible exception of during sewage treatment. The addition of cysteine also had an effect on the surface properties and aggregation rates. Citrate coated nanosilver released less silver into solution than polyvinylpirrolidone coated nanosilver.

Numerous other studies (Chinnapongse et al. 2011, Huynh and Chen 2011, Jin et al. 2010, Li and Lenhart 2012) made similar findings to the studies summarised above. A more limited number of other studies resulted in alternative findings regarding the solution behaviour of nanosilver particles. Piccapitera et al. (2012) found no influence of humic substances on the solution behaviour of nanosilver at concentrations of between 5 and 20 mg/L, although carbonate coated particles were used for this study which may explain why these effects were observed. Prathna et al. (2011) compared nanosilver particles with chemical and biological coatings and, whilst the chemically treated materials behaved similarly to those in the other studies, the biologically coated materials (which may not be representative of commercially produced nanosilver materials) showed opposite effects. Scheckel et al. (2010) studied the effect of aging nanosilver particles in a kaolin suspension and found no significant effects after 12 months of aging.

There are several factors which affect the aggregation behaviour of nanosilver in solution, including the nature of any surface coatings, the concentrations of divalent cations (such as Ca2+ and Mg2+), the ionic strength, and the presence of natural organic matter (such as humic substances). Conditions which cause an increase in the aggregation of nanosilver particles, such as elevated concentrations of divalent cations, are likely to result in an increase in the removal of nanosilver from solution, whereas conditions which enhance or maintain the stability of nanosilver particles, such as some types of surface coatings including natural organic macromolecules, are likely to result in a decrease in the rate of removal of nanosilver from solution. Nanosilver materials can also release dissolved silver into solution, although the conditions which promote the dissolution of silver from nanosilver are not well understood. Dissolved silver can make up a significant proportion of the total silver concentration in some stable nanosilver dispersions.

Dobias and Berier-Latmani (2013) evaluate the release of dissolved silver from nanosilver particles of various sizes and surface coatings embedded in gels deployed in natural river and lake waters (in Switzerland). Deployments were up to four months in duration. A considerable difference in dissolution was observed between 5, 10 and 50 nm materials. Small particles dissolve rapidly and almost completely (>80%), while larger materials (50 nm) persisted for longer (<50% reduction in mass over the period of dissolution observed). However, the difference between materials was reduced when dissolution was normalised to surface area, highlighting the importance of this property to nanosilver fate in the environment. In addition, PVP and tannic acid coated materials released relatively greater amounts of silver ions compared with citrate coated nanosilver materials. Release of dissolved silver was characterised by an initial period of rapid dissolution, followed by a period of much slower dissolution. The initial peak of dissolution was considered to be associated with the release of chemisorbed Ag+ from the surface of particles. The subsequent sustained period of slower release was considered to be consistent with an oxidative dissolution mechanism. Rates of oxidative dissolution may be greater in waters with lower pH than those used in this study (pH of lake and river water in this study were 7.9 – 8.4).

As requested in the Silver Substance Evaluation Final Decision, T/D testing has been performed on a nanosilver powder (see section 5.6 in IUCLID). VITO NV (2017) determined the T/D behaviour of a silver powder with a median particle size of 8 nm (see section 4.5 of IUCLID) in modified Daphnia and algae media (loading of 1153 and 1230 µg Ag/l in Daphnia and algae medium, respectively) over a period of 28 days. Three silver fractions in solution were measured: total silver, conventional dissolved silver (< 0,45 µm; i.e. ionic silver + silver particles with diameter < 450 nm) and truly dissolved silver (< 1 kDa; i.e. ionic silver).

In the Daphnia medium, there is an initial loss of total silver (to 254 µg/l after 7 days). This loss is partially transient reaching a plateau level of 473 µg/l after 28 days. The same steep fall in concentration of conventional dissolved silver is seen during the first days (99 µg/l after 7 days), after which the concentration becomes stable at levels around 120 µg/l (127 µg/l after 28 days). The truly dissolved silver concentration is very low at the start of the test, then slightly increases to about the same level as the conventional dissolved silver concentration (88 µg/l after 7 days and 146 µg/l after 28 days).

In the algae medium, the T/D behaviour is very different from the one described above for Daphnia medium. The total and conventional dissolved silver concentrations in this medium decrease slightly over time but fluctuate around 1000 µg/l over the whole test period (total silver concentration is 1006 µg/l after 7 days and 957 µg/l after 28 days; conventional dissolved silver concentration is 974 µg/l after 7 days and 912 µg/l after 28 days). The truly dissolved silver concentration is very low at the start of the test, then slowly increases during the test (125 µg/l after 7 days and 214 µg/l after 28 days).

 

Partitioning of nanosilver in environmental media

Relatively few studies on the fate and partitioning of nanosilver materials in environmental media have been performed and the majority of those which have been performed have studied the retention of the nanosilver particles in soils.

Cornelis et al. (2010) studied the fate and partitioning of added silver to soil samples; a comparison was made between ionic silver, nanosilver and bulk silver (assumed to be metal powder). Retention coefficients (Kr) were calculated, which take account of the fact that metal dissolution from the nanomaterials may occur, in order to distinguish them from partition coefficients which are commonly calculated for chemicals interacting with solid phases. The nanosilver particles had sizes ranging from 20 to 100 nm, although the specified nominal size according to the manufacturer was 10 nm. No mention was made of the nature of any surface treatment of the nanosilver particles, which limits the usefulness of this study given the importance of the coating material in determining the aggregation and solution behaviour of nanosilver particles.

Some soluble silver was released from nanosilver additions but most of the silver was found to be present as nanosilver particles. Differences were observed between nanosilver, ionic silver and bulk silver partitioning in the soils. Kr values were calculated for nanosilver in five soils, which showed a considerable range of partitioning behaviours. The Kr values calculated were 77, 68, 76, 541, 2165 L/kg. The greatest degree of retention was observed in the soil with the highest pH (pH 6.4), although the lowest retention was observed in a soil with a pH which was only slightly lower (pH 6.1). The highest retention coefficients were observed for the soils with the highest clay contents and soil texture appears to be the dominant factor in controlling the mobility of nanosilver particles in soils.

Cornelis et al. (2012) studied the retention of polyvinylpyrrolidine coated nanosilver particles in 16 Australian soils. The authors observed dissolution of silver from nanoparticles in some cases, although in other cases this dissolution was probably not apparent due to the binding of free silver to the soils. The dissolution of silver from nanoparticles was found to be greater in soils than in solutions, probably due to the removal of dissolved silver by chloride complexation or adsorption to soil surfaces. Different retention coefficients were observed for different loading rates of nanosilver applied to the soils, although only minimum, median, and maximum values were reported. At low loading rates a median retention coefficient of 589 L/kg was observed (range 50 to 2511 L/kg), and at high loading rates a median retention coefficient of 9420 L/kg was observed (range 2459 to 2051761 L/kg). Clay content was found to be the key factor in controlling the retention of nanosilver particles in the soils.

Coutris et al. (2012b) studied the effect of aging of nanosilver particles in soils on their mobility. The study employed two types of nanosilver materials, one of which was citrate coated (4.7 nm) and the other was uncoated (19 nm), with ionic silver (silver nitrate) also used for comparative purposes. Treated soils were allowed to age for two hours, two days, five weeks and 10 weeks before they were subjected to a sequential extraction procedure. Silver added as silver nitrate was predominantly associated with the oxidisable (i.e. organic matter) and strong acid digestible fractions after aging, which are not considered to be bioaccessible. After a short aging period of two hours, approximately 50% of the added silver was found to be bioaccessible, although this fraction decreased to 35% after 2 days, 10% after 35 days and 4% after 70 days. A smaller fraction of silver added as citrate stabilised nanosilver showed a smaller bioaccessible fraction: 24% after two hours, 4.5% after two days and 1% after 35 days. Conversely, the bioaccessible fraction of uncoated nanosilver was initially very low, but increased over time to 7% after 70 days.

This study, and others which use similar extraction techniques to those developed for assessing metal availability (e.g. Tessier et al. 1979), may be problematic when addressing the fate of nanosilver particles. This is due to the possibility of oxidising metallic silver to Ag+, which could occur due to some of the extractants used for the oxidation of organic matter. Similarly, reduction of Ag+ to metallic silver could be caused by the reducing agents used to release metals from iron oxide precipitates. Given these concerns, the results of this study must be treated with caution. The association of silver between different soil fractions differed depending on both the form of silver added to the soil, and on the soil type. Two possibilities for the different behaviour observed for the uncoated nanosilver suggested by the authors were saturation of silver binding sites in the soil and also slower release of dissolved silver from larger aggregates of uncoated nanoparticles. Effects of the extractant solutions used on the form of silver were not considered.

Cleveland et al. (2012) studied the fate of ionic silver, two commercial nanosilver materials and three consumer products containing nanosilver in a tidal marine mesocosm study (natural seawater filtered and diluted to 20‰). The commercial nanosilver materials had sizes of 20 nm and 80 nm, although no mention was made of any surface coatings. The consumer products were an antimicrobial wound dressing (Acticoat #20601, Smith and Nephew Medical Limited, Hull, UK), a sock reported to be manufactured using silver-coated nylon fibres (Style 2115, Fox River Mills Inc, Osage, IA) and a stuffed toy made with nanosilver coated foam stuffing (Benny the Bear, Pure Plushy, Oak Brook, IL). All experiments employed only a single addition of silver and the consumer products were placed intact on the surface of the water in the mesocosm and left for the duration of the study.

The maximum total silver concentrations measured in the seawater of the mesocosms, and times at which they were observed, increased in the order: ionic silver (0.1 μg/L, 6 to 24 hours), 20 nm nanosilver particles (0.6 μg/L, 2 hours), 80 nm nanosilver particles (0.8 μg/L, 2 hours), sock (5.3 μg/L, 12 hours), dressing (5.7 μg/L, 12 hours), stuffed toy (14 μg/L, 24 hours). Leaching of nanosilver from the consumer products began after 2 to 12 hours and continued for about 30 days. Silver did not accumulate in the intertidal sediment of the mesocosms treated with ionic silver or commercial nanosilver particles but did in mesocosms treated with consumer products. This was considered to be due to the relatively short residence time of the ionic silver and commercial nanosilver materials in the experimental system, compared with the longer term leaching of silver from the consumer products containing nanosilver. The study did not specifically measure the leaching of nanosilver, only total silver.

Lin et al. (2011) studied the adsorption behaviour of uncoated silver nanoparticles (12 nm) in a porous medium of glass beads, either uncoated or coated with haematite. Increased adsorption was observed at both low pH and at high ionic strength. A difference between the degree of adsorption to the particle surfaces was similar for both the haematite coated and uncoated beads at pH 8.3 and at ionic strengths up to 20 mM but was considerably higher to haematite coated beads at pH 5.0. Sagee et al. (2012) studied the transport of citrate coated nanosilver particles in soils and found that their retention was greatest in soils with finer particles and lowest in coarser soils. This is consistent with other findings (Cornelis et al. 2010, Cornelis et al. 2012), and similar to the behaviour of many other substances, and was interpreted as being due to mechanical straining rather than complexation or adsorption.

There are some recent quantitative data published on the remobilisation of silver from silver nanoparticles and also silver sulphide nanoparticles that further the discussions in relation to the fate of nanosilver in soils (Navarro et al. 2014). Specifically, these authors investigated the retention and release of nanoparticulate silver in spiked soils as citrate, PVP, humic acid coated forms and silver sulphide nanoparticle cysteine suspension. These forms were compared to silver nitrate and bulk silver sulphide (> 200 nm) spikes. Batch release experiments were performed on the spiked soils, in potassium nitrate solution and potassium nitrate solutions plus an environmentally/ agriculturally relevant ligand (humic acids, citrate, thiosulphate, cysteine, mercaptopropionic acid). The measure of release of silver in soils was shown to vary between the two soils tested but to be less than 25% of that retained in the soils accept for the thiosulphate treatment. It is important to note that of the silver that was released in the thiosulphate extracts less than 25% was present in the < 1KDa size fraction. This study shows the importance of colloidal silver associations and highlights that while these nanosilver particles can be present in low concentrations in solutions they are not the original manufactured nanosilver particles.

 

Unrine et al. (2012) studied the behaviour of nanosilver particles with two different types of surface coatings (gum arabic and polyvinylpyrrolidone) in aquatic mesocosms with and without sediments and plants present. Organic exudates from the plants were found to have an effect on the particles, although the release of exudates by the plants appeared to be stimulated by the release of silver from the nanoparticles. The interactions with the plants appeared to stabilise the polyvinylpyrrolidone coated particles but appeared to destabilise the gum arabic coated particles either through deposition or dissolution. Silver in the particulate phase of the mesocosms was found to be predominantly (>85%) present as Ag0, although cysteine complexes were also present in the systems containing plants.

The retention of nanosilver in soils appears to be largely governed by physical processes, rather than chemical processes. Any dissolved silver released from the nanosilver materials would be expected to partition to soil organic matter or be precipitated as silver sulfides. It is possible, however, that the fate of nanosilver particles in soils is governed by chemical interactions such as silver sulfide formation or the release of Ag+ which would interact strongly with soil particles and organic matter. Current studies are able to observe the overall degree of retention of nanosilver by soils but have been unable to assess the fate of the retained nanoparticles in the soils. Soil texture appears to be the dominant factor in controlling the potential for nanosilver to leach from soils, with fine textured soils (i.e. those predominantly composed of relatively small particles) being most likely to retain nanosilver particles.

As requested in the Silver Substance Evaluation Final Decision, transformation/dissolution testing has been performed on a nanosilver powder in three different soils (see section 5.6 in IUCLID). The objective of the study was to determine at what rate and to what extent nanosilver dissolves in soils and to compare this with silver nitrate. Three agricultural soils, representative for the EU, were spiked with 50 mg Ag/kg, as silver nitrate and nanosilver and as a control the same soils were spiked with water. The pore waters of the three soils were sampled in the first 35 days after spiking. In soils spiked with silver nitrate, the concentrations of total silver in the pore water decreased with time illustrating ageing reactions. In contrast, in the nanosilver spiked soils, the total concentration of silver in the pore water increased during the first days after spiking, indicating dissolution of the nanoparticles, followed by a decrease due to the ageing reactions. In all soils and at all times, the total concentration of silver in the pore water was equal or higher for silver nitrate than for nanosilver treatment but differences became smaller at longer equilibration times. The concentrations of truly dissolved silver (<1 kDa) for both silver nitrate and nanosilver follows the same pattern as the total concentration of silver in the pore water but were always lower than the total concentration of silver in the pore water. In one soil (with highest organic content), the truly dissolved silver concentration in the nanosilver treatment decreased to that of the un-amended control soil in contrast with the silver nitrate treatment suggesting that almost all added silver in solution formed colloids >1 kDa.

Fate of nanosilver during sewage treatment

A number of studies of the fate of nanosilver during sewage treatment have been performed and there are also several studies concerned with the effects of nanosilver on sewage treatment bacteria which also include some information on the fate of nanosilver during sewage treatment. These studies have all been performed on simulated sewage treatment systems on a laboratory scale, using different experimental setups and different types of nanosilver materials. Despite the differences between the various studies, a considerable degree of consistency has been observed in terms of the fate of nanosilver during sewage treatment.

Hou et al. (2012) studied the fate of nanosilver particles during simulated wastewater treatment processes using citrate coated particles with an average size of 23 nm. Minimal removal was observed during primary settling, although virtually all (over 90%) of the nanosilver was removed to the sludge during the biological treatment phase.

Kaegi et al. (2011) studied the fate of polyoxyethylene fatty acid ester coated nanosilver particles during simulated sewage treatment. Nanosilver particles were only found in the effluent shortly after the initial spiking and effluent concentrations of silver were found to be three orders of magnitude lower than the concentrations in the reactor (99.9% removal). The silver present in the sludge was predominantly in the form of Ag2S and transformation of nanosilver to Ag2S appeared to take approximately two hours, although transformation to Ag2S was only 90% complete.

Tiede et al. (2010) studied the fate of ethylene glycol coated nanosilver in sewage sludge supernatant using batch sorption studies. The majority (90%) of the nanoparticles were found to partition to the sludge phase within six hours, although some of the remaining silver in the supernatant was found to be in the nanoparticle form.

In a study of the toxicity of nanosilver to nitrifying bacteria in sewage treatment (Choi et al. 2009), the majority of the nanoparticles added were found to react with sulfide to form silver sulfide complexes or precipitates. The silver sulfides formed were not found to be oxidised during 18 hours of aeration. This suggests that the silver sulfide formed is unlikely to be oxidised during sewage treatment processes. A further study (Choi and Hu 2009) confirmed that sulfide is an important complexant in the fate of nanosilver during sewage treatment.

Liang et al. (2010) also found that the majority of nanosilver added to a simulated sewage treatment system was associated with the sludge phase. This study also derived a partition coefficient for silver in the reactor of 7100 L/kg and identified that all of the silver in the effluent was in soluble forms using ultrafiltration. This finding differs from some of the other studies (e.g. Tiede et al. 2010) which found that a proportion of the silver in the effluent was present as nanoparticles.

Reinsch et al. (2012) studied the tendency of nanosilver particles to undergo sulfidation (reaction with HS-) and its effect on bacterial toxicity. Two types of nanosilver particles were examined: commercially available polyvinyl alcohol coated particles, which were polydisperse and had a tendency to aggregate, and laboratory prepared polyvinyl alcohol coated particles, which were monodisperse. Both types of particles had similar surface charges. The tendency of the nanoparticles to aggregate affected their tendency to undergo reactions with sulphur species, with the larger, aggregated, particles reacting more slowly with sulphur species than smaller, dispersed, particles. The tendency of silver nanoparticles to aggregate, which is affected by both solution conditions and surface coating, may therefore be an important factor affecting the fate of nanosilver during sewage treatment.

A study of the effect of nanosilver on nitrogen removal processes during simulated sewage treatment (Jeong et al. 2012) using laboratory prepared polyvinylpyrrolidine coated particles found a lower degree of removal of nanosilver than other studies. Up to 30% of the nanosilver was found to be present in the supernatant. The degree of removal of nanosilver was found to greatest (90%) at a low loading rate of 1 mg/L, decreased to 67.5% at 10 mg/L, and increased to 85.5% at a high loading rate of 100 mg/L. The greater removal at the highest loading rate was considered to be due to the settling of aggregates, although the loading rates used are unlikely to be representative of the exposure observed in real sewage treatment plants.

Lombi et al. (2013) used bench-scale anaerobic digesters to investigate the influence of nanoparticle surface functionality (i.e. coating material, nanoparticle core structure and the nature of post-processing) on the speciation of silver in sewage sludges and biosolids. Various types of nanosilver materials were investigated and neither surface coating (citrate [6.4 nm], polyvinyl sulphonate [10.7 nm] or mercaptosuccinic acid [8.9 nm]) nor core structure prevented the formation of silver sulphides during waste water treatment. Unlike some other metals, the silver sulfides present in sewage sludge were stable over a six month period of simulated composting/stockpiling.

Overall, the available studies indicate that typically 10% of added nanosilver may be present in the effluents of sewage treatment plants, although there is some uncertainty as to the form of the silver in the effluent, which may be ionic silver associated with organic ligands, dispersed particles of nanosilver (in zero-valent form), or dispersed nanoparticles of silver sulfide. Some studies have observed that under some conditions the proportion of silver released in the effluent may be greater than 10%, although the relevance of the conditions to real sewage treatment plants may be questionable. A number of studies have identified that the silver which is associated with the sludge is present as stable silver sulfide precipitates, rather than as intact particles of nanosilver.

Comparison between ionic silver and nanosilver

The behaviour of nanosilver in the environment appears to be largely controlled by physical processes, particularly the tendency for both homo- and hetero-aggregation of nanosilver particles, which may cause them to settle out of suspension, and their tendency to be retained in soils due to either physical straining or sorption processes. The possibility that nanosilver in soils undergoes partitioning reactions which are comparable to those of other chemicals cannot be discounted, due to the approaches which are used to assess this behaviour, although there is no direct evidence to support this. Nanosilver particles do, however, appear to undergo chemical reactions during sewage treatment to form silver sulfide precipitates and it is possible, although presently unconfirmed, that ionic forms of silver may also undergo similar processes during sewage treatment.

The fate and behaviour of ionic, or dissolved, forms of silver in the environment are often controlled to some extent by associations with sulphur containing ligands such as thiols, due to the very high affinity of silver for these groups. Sulphur containing ligands (such as cysteine) have also been shown to promote the dissolution of silver from nanosilver particles, although this does not necessarily imply a similarity between the behaviour of nanosilver and ionic silver. The ability of nanosilver particles to release dissolved (i.e. ionic) silver is the most likely explanation of these phenomena.

Despite the possible differences in the fate and behaviour of nanosilver when compared to ionic silver in the environment studies, the retention of nanosilver in soils which have reported retention coefficients (which are considered to be conceptually comparable to the partition coefficients derived for chemical adsorption interactions) have resulted in values which are broadly comparable to those determined for ionic silver in soils (median KD 4023 L/kg). The retention coefficients derived for nanosilver in soil showed a clear dependency on the loading rate, with greater retention observed at higher loading rates. There is a dependency on concentration for the partition coefficients derived for ionic silver in soil (Daniels and Rao 1983, Jones et al. 1986), although given that the partitioning of ionic silver in soils is controlled by the availability of specific binding sites and ligands, a reduction in the partition coefficient is observed at higher silver concentrations due to saturation of the available binding sites and ligands present in the soils. This suggests that, despite the apparent similarities in the partitioning and retention behaviour of nanosilver and ionic silver in soils, this may not result from the same processes.

There is also considerable similarity in the overall removal efficiencies observed for nanosilver in simulated sewage treatment systems when compared to the removal efficiencies observed for total silver in operating sewage treatment plants. The typical removal efficiency observed for total silver (assumed to be ionic) is 82.9% and the typical removal efficiency observed for nanosilver is 90%. Considerable variability was observed for the removal efficiencies of operating sewage treatment plants, with between 33 and 99% removal being observed. The differences in the removal efficiencies could be due to the operating conditions of the plants and the concentrations of dissolved silver binding ligands in the treated effluents, which could act to maintain silver in solution (although in a complexed form).

Several studies have identified that silver sulfides are formed from nanosilver in the sludge phase of simulated sewage treatment plants, although the form of silver in sewage sludge has been less well studied. If ionic silver is complexed by ligands during sewage treatment then this could reduce the formation of silver sulfides from ionic silver, although ionic silver is usually considered to be present as silver sulfides in sewage sludge (Lytle 1984). It is possible, therefore, that the behaviour of nanosilver and ionic silver is very similar under the conditions of a sewage treatment plant. This may be due to the presence of relatively high concentrations of sulphur containing ligands, which are able to increase the dissolution of ionic silver from the nanosilver particles (Gondikas et al. 2012). Although direct sulfidation of nanosilver has also been observed (Liu 2011) this is only expected to be a dominant process at high sulfide concentrations. The size of aggregates of nanosilver particles, and the nature of their surface coatings, may mean that some forms of nanosilver are converted into ionic silver more or less readily during sewage treatment than other forms.

It is likely that in practice the tendency for silver which is added to the environment as nanosilver to behave in the same way as ionic silver is dependent upon the rate at which dissolved silver is released from the nanosilver particles, although during sewage treatment the extent and rate of formation of silver sulfides may be more important. If the rate of release is slow then the material will tend to behave predominantly as nanosilver; if the rate of release is fast the material will tend to behave predominantly as ionic silver. Understanding the factors which control the rate of dissolution of nanosilver in the environment is, therefore, likely to be important in understanding the overall fate of nanosilver in the receiving environment.

The reaction of nanosilver to produce silver sulfides during sewage treatment is likely to mean that much of the total emission of nanosilver enters the environment in the form of silver sulfides in sewage sludge if the sludge is spread to land.

Literature cited

Chinnapongse S, MacCuspie R, Hackley V. 2011. Persistence of singly dispersed silver nanoparticles in natural freshwaters, synthetic seawater, and simulated estuarine waters. Science of the Total Environment 409: 2443-2450.

Cumberland S, Lead J. 2009. Particle size distributions of silver nanoparticles at environmentally relevant conditions. Journal of Chromatography A 1216: 9099-9105.

Delay M, Dolt T, Woellhaf A, Sembritzki R, Frimmel F. 2011. Interactions and stability of silver nanoparticles in the aqueous phase: Influence of natural organic matter (NOM) and ionic strength. Journal of Chromatography A 1218: 4206-4212.

Dobias J, Bernier-Latmani R. 2013. Silver release from silver nanoparticles in natural waters. Environmental Science and Technology. 47: 4140-4146.

El Badawy A, Luxton T, Silva R, Scheckel K, Suidan M, Tolaymat T. 2010. Impact of environmental conditions (pH, ionic strength, and electrolyte type) on the surface charge and aggregation of silver nanoparticles suspensions. Environmental Science and Technology 44: 1260-1266.

Gondikas AP, Morris A, Reinsch BC, Marinakos SM, Lowry GV, Hsu-Kim H. 2012. Cyseine-Induced Modifications of Zero-valent Silver Nanomaterials: Implications for Particle Surface Chemistry, Aggregation, Dissolution, and Silver Speciation. Journal of Environmental Science and Technology 46: 7037 – 7044.

Huynh K, Chen K. 2011. Aggregation kinetics of citrate and polyvinylpyrrolidine coated silver nanoparticles in monovalent and divalent electrolyte solutions. Environmental Science and Technology 45: 5564-5571.

Jin X, Li M, Wang J, Marambio-Jones C, Peng F, Huang X, Damoiseaux R, Hoeck E. 2010. High throughput screening of silver nanoparticle stability and bacterial inactivation in aquatic media: influence of specific ions. Environmental Science and Technology 44: 7321-7328.

Li X, Lenhart J. 2012. Aggregation and dissolution of silver nanoparticles in natural surface water. Environmental Science and Technology 46: 5378-5386.

Liu J, Pennell KG, Hurt RH. 2011. Kinetics and mechanics of nanosilver oxysulfidation. Environmental Science and Technology. 45: 7345-7353.

Piccapietra F, Sigg L, Behra R. 2012. Colloidal Stability of Carbonate-Coated Silver Nanoparticles in Synthetic and Natural Freshwater. Environmental Science and Technology, 46: 818 – 825.

Prathna TC, Chandrasekaran N, Mukherjee A. 2011. Studies on aggregation behaviour of silver nanoparticles in aqueous matrices: Effects of surface functionalisation and matrix composition. Colloids and Surfaces A: Phyiscochem. Eng. Aspects, 390: 216-224.

Sagee O, Dror I, Berkowitz B. 2012. Transport of silver nanoparticles (AgNPs) in soil. Chemosphere 88: 670-675.

Scheckel K, Luxton T, Badawy A, Impellitteri C, Tolaymat T. 2010. Synchrotron speciation of silver and zinc oxide nanoparticles aged in a kaolin suspension. Environmental Science and Technology 44: 1307-13121.