Science of the Total Environment 420 (2012) 327–333

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Aging and soil organic matter content affect the fate of silver nanoparticles in soil

Claire Coutris a,⁎, Erik Jautris Joner b,1, Deborah Helen Oughton a,2

a Department of Plant and Environmental Sciences, Norwegian University of Life Sciences, P.O. Box 5003, N-1432 Aas, Norwayb Bioforsk Soil and Environment, Fredrik Dahls vei 20, N-1432 Aas, Norway

⁎ Corresponding author. Tel.: +47 6496 6217; fax: +E-mail addresses: claire.coutris@umb.no (C. Coutris)

(E.J. Joner), deborah.oughton@umb.no (D.H. Oughton).1 Tel.: +47 9283 3168; fax: +47 6300 9410.2 Tel.: +47 6496 5544; fax: +47 6496 6007.

0048-9697/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.scitotenv.2012.01.027

a b s t r a c t

a r t i c l e i n f o

Article history:Received 16 June 2011Received in revised form 11 January 2012Accepted 11 January 2012Available online 10 February 2012

Keywords:Metal bioaccessibilitySequential extractionEngineered nanomaterialsBioavailability

Sewage sludge application on soils represents an important potential source of silver nanoparticles (Ag NPs)to terrestrial ecosystems, and it is thus important to understand the fate of Ag NPs once in contact with soilcomponents. Our aim was to compare the behavior of three different forms of silver, namely silver nitrate,citrate stabilized Ag NPs (5 nm) and uncoated Ag NPs (19 nm), in two soils with contrasting organic mattercontent, and to follow changes in binding strength over time. Soil samples were spiked with silver and left toage for 2 h, 2 days, 5 weeks or 10 weeks before they were submitted to sequential extraction. The ionic silversolution and the two Ag NP types were radiolabeled so that silver could be quantified by gamma spectrom-etry by measuring the 110mAg tracer in the different sequential extraction fractions. Different patterns of par-titioning of silver were observed for the three forms of silver. All types of silver were more mobile in themineral soil than in the soil rich in organic matter, although the fractionation patterns were very differentfor the three silver forms in both cases. Over 20% of citrate stabilized Ag NPs was extractible with water inboth soils the first two days after spiking (compared to 1–3% for AgNO3 and uncoated Ag NPs), but the frac-tion decreased to trace levels thereafter. Regarding the 19 nm uncoated Ag NPs, 80% was not extractible at all,but contrary to AgNO3 and citrate stabilized Ag NPs, the bioaccessible fraction increased over time, and by day70 was between 8 and 9 times greater than that seen in the other two treatments. This new and unexpectedfinding demonstrates that some Ag NPs can act as a continuous source of bioaccessible Ag, while AgNO3 israpidly immobilized in soil.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Sewage sludge application on soils represents an important poten-tial route of engineered nanomaterials (ENMs) into terrestrial ecosys-tems (Blaser et al., 2008; Gottschalk et al., 2009; Mueller and Nowack,2008). Soils are the basis for a large part of human food productionand there are concerns for the possible negative impacts of ENMs en-tering terrestrial ecosystems. Soils contain a wide variety of solutesand colloidal materials including dissolved organic carbon, whichmay coat ENMs, affect their interactions with soil surfaces, and ulti-mately interfere with the uptake and various aspects of organisms'exposure to ENMs (Lowry et al., 2010). Thus determination of thetotal amount of ENMs in a soil may not provide the full picture withrespect to exposure and estimation of their bioavailability. In orderto assess risks of posed by ENMs to terrestrial fauna, microorganismsand plants, a fair amount of data on toxicity has been published (seePeralta-Videa et al., 2011 for a review), but comparatively little isknown on exposure and even less on bioavailability (for studies on

47 6496 6007., erik.joner@bioforsk.no

rights reserved.

earthworms see Coutris et al., 2011; Shoults-Wilson et al., 2011;Unrine et al., 2010a, 2010b).

Bioavailability can be defined as the fraction freely available to crossan organism's cellular membrane from the medium the organism in-habits at a given time (Semple et al., 2004). In soils, bioavailability is acomplex result of interactions between contaminants and soil constitu-ents, and can be strongly influenced by the organisms' activity such asfeeding or burrowing behavior (Ahlf et al., 2009). Bioaccessibility, onthe other hand, can be defined as the fraction available to cross an orga-nism's cellular membrane from the environment, if the organism hasaccess to the chemical (Semple et al., 2004). Bioaccessibility encom-passes both bioavailability and potential availability of a compoundwhich is physically or temporally constrained. Although it is unlikelythat chemical fractionation techniques can imitate the process ofmetal absorption or uptake in organisms, such techniques have givengood estimates of the bioaccessibility of metals in the environment(see Peijnenburg et al., 2007 for a review). Sequential extraction proce-dures, for instance, use a series of chemical extractants to displace anddissolve metals from the soil matrix, and have been used extensivelyto assess the partitioning of trace elements in environmental samples(first formulated by Tessier et al., 1979). Using the experience andknowledge gained from studies on bioaccessibility of trace elements insoils, and applying similar techniques to estimate bioaccessibility ofme-tallic or oxide ENMs, show promise. By comparing ENM extractability

328 C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

with that of corresponding metal salts, it may also be possible to de-scribe some of the assumed unique properties of ENMs with respect totheir behavior in contact with environmental matrices. Aging of envi-ronmental pollutants in soil is a well known phenomenon that de-scribes temporal aspects of their bioaccessibility. This is also aparameter for which ENMs may have unique properties, and whereknowledge is scarce. Scheckel et al. (2010), usingX-ray absorption spec-troscopy, studied the time-dependent evolution of Ag NPs suspended inkaolin. Uncoated Ag NPs (mean particle diameter 100 nm) remainedunchanged over 18 months, even in the presence of NaCl, whereas or-ganically coated Ag NPs (mean particle diameter 148 nm) became coat-ed with chloridewhen NaCl was present. No destabilization/dissolutionof these particles was observed during 18 months. Shoults-Wilson et al.(2011), also usingX-ray absorption spectroscopy, studied the speciationof PVP coated AgNPs (meanparticle diameter 10 and 30 to 50 nm) aged28 days in soil.Most silver remained asAg0,while 10 to 17%waspresentas Ag2O, indicating that Ag NPs had undergone partial oxidation in soil.Other studies on silver behavior in sewage sludge suggested that Ag2Swas the dominant species originating either from Ag ions or Ag NPs,and that the very low solubility of Ag2S would limit the bioavailabilityand adverse effects of silver in the environment (Kim et al., 2010;Nowack, 2010).

One of the major reasons why bioaccessibility measurementsthrough sequential extraction have not yet been attempted for ENMsis the difficulty in measuring their amounts in soils or aqueous soil ex-tracts. A solution to this problem is the use of radioactively labeledENMs and their quantification by gamma spectrometry, which allowsfor rapid, sensitive and simple determination in a large number of sam-ples (Coutris et al., 2011; Oughton et al., 2008). The aim of the presentstudy was to compare the extractability of two contrasting Ag nanopar-ticles (AgNPs) and their corresponding ionic species (Ag+ ions added asAgNO3) using standard protocols for sequential extraction of metals insoil. Further, wewanted to examine the temporal differences in extract-ability of the selected Ag species as well as differences in extractabilityin two soils with contrasting organic matter content.

2. Materials and methods

An experiment was set up with a full factorial design, featuringthree forms of silver (ionic Ag, citrate stabilized Ag NPs and

Table 1Physico-chemical properties of the study soils. n/a: not available; MU: measurement uncer

Parameter Organic soil Mineral so

Coordinates 59°39′47″N, 10°45′3″ E 59°39′34″pH (CaCl2) 4.8 5.8Total organic carbon 14.2 1.49Carbonate 0.4 0.3Nitrogen Kjeldahl 0.5 n/aChloride 11.2 n/aSulfate 11 n/aSulfur 770 n/aWater holding capacity 48.0 n/aSand 50 40

Coarse sand (0.63–2 mm) 14 11Medium sand (0.2–0.63 mm) 8 17Fine sand (0.063–0.2 mm) 28 12

Silt 30 43Coarse silt (20–63 μm) 6 11Medium silt (6.3–20 μm) 11 15Fine silt (2–6.3 μm) 13 17

Clay (b2 μm) 20 17Cation exchange capacity 24.3 9.1Calcium 19.6 4.1Magnesium 3.0 0.9Potassium 0.4 b0.1Sodium b0.1 b0.1Base saturation 95 55Exchangeable acidity 0.18 0.22

uncoated Ag NPs), two soil types (organic and mineral), four con-tact times (2 h, 2 d, 35 d and 70 d), and 4 replicates for each treat-ment combination. In addition, there were two control (non-spiked) soil samples per contact time. All chemicals were of ana-lytical grade and purchased from Merck, Darmstadt, Germany(unless otherwise stated). Ultrapure MilliQ water (18 MΩ cm)was used throughout the experiment.

2.1. Test soils

Two agricultural loam soils of similar geological origin in southernNorway (Ås, Akerhus County) were used in this study. Soils weresieved at 2 mm, air-dried and characterized by an accredited labora-tory (Eurofins Moss, Norway) (Table 1). The two soils had similar par-ticle size distribution, particularly for clay content, but the organiccarbon content was ten times higher in one soil (referred to as organ-ic soil) compared to the other (referred to as mineral soil).

2.2. Silver materials

2.2.1. Silver nitrateThe ionic test solution was made from AgNO3 amended with

110mAg tracer. The tracer and the stable Ag were mixed in a glassvial using 1 mL tracer, 3.8 mL 2.6 mmol L−1 AgNO3, 15.2 mL MilliQwater.

2.2.2. Citrate stabilized silver nanoparticlesSilver nanoparticles were formed by chemical reduction of AgNO3

(stock solution containing the 110mAg tracer) with sodium borohydride(NaBH4) in the presence of sodium citrate (Na3C6H5O7·2H2O) to stabi-lize the nanoparticles (Doty et al., 2005). Briefly, a freshly preparedaqueous solution of NaBH4 (6 mL, 10 mmol L−1) was added to a200 mL solution of AgNO3 (0.25 mmol L−1) and sodium citrate(0.25 mmol L−1) under vigorous stirring. The nanoparticle suspensionwas kept at 4 °C in the dark until use. Specimens for microscopy wereprepared by evaporating 10 μL on a 400-mesh carbon coated coppergrid (Chemi-Teknik, Oslo). Images of nanoparticles were acquired lessthan 2 h after specimen preparation, on a FEI Morgagni 268 transmis-sion electron microscope (TEM) operating at 80 keV (Fig. 1a). Threegrids were prepared and five representative pictures were analyzed

tainty in % (except for pH, absolute value).

il Unit Method MU

N, 10°45′12″ ENone DIN ISO 10390 0.2g/100 g dw DIN EN 13137 5g/100 g dw DIN ISO 10693 5g/100 g dw EN 13654-1 5mg/kg dw EN ISO 10304-2m 20mg/kg dw EN ISO 10304-2m 20mg/kg dw EN ISO 11885 20vol.% DIN ISO 11274 n/a% DIN ISO 11277% 10% 10% 10% DIN ISO 11277% 10% 10% 10% DIN ISO 11277 10cmol(+)/kg DIN ISO 11260 5cmol(+)/kg DIN ISO 11260 3cmol(+)/kg DIN ISO 11260 4cmol(+)/kg DIN ISO 11260 6cmol(+)/kg DIN ISO 11260 6% 5cmol(+)/kg DIN ISO 14254 0.2

size (nm)

Par

ticle

num

ber

0 5 10 15 20 25 30

020

4060

8010

0

size (nm)

Par

ticle

num

ber

0 10 20 30 40

010

2030

40

a

b

c

Fig. 1. Citrate stabilized silver nanoparticles (a) and uncoated silver nanoparticles (b, c)observed in a transmission electron microscope (FEI Morgagni 268, operating at80 keV). Insets show size distribution histograms.

329C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

per grid. The particles were spherical and the mean particle diameterwas 4.7±3.7 nm (mean±SD throughout the text, n=500).

2.2.3. Uncoated silver nanoparticlesAg NP powder (QSI-nano silver) purchased from Quantum

Sphere (Santa Ana, CA, USA) was submitted to neutron activationat a flux of 1012 neutrons cm−2 s−1 for 48 h at the reactor of theInstitute of Energy Technology (Kjeller, Norway). One gram of soilwas added to 21 mL MilliQ water, shaken and centrifuged 2 min at2000×g, and 20 mL of the supernatant was collected. The nano-particle suspension was prepared by adding 17 mg of the neutronactivated Ag NP powder to the collected supernatant, and placed5 min in an ultrasound bath. These particles had previously beencharacterized with respect to particle size, shape, crystallographicstructure, and specific surface area (Oughton et al., 2008). The par-ticleswere spherical and could present a networked aggregate structure(Fig. 1c). They had a face-centered cubic structure and themeanparticlediameter was 19.2±6.8 nm (n=300). Images of nanoparticles wereacquired as above (Fig. 1b and c).

2.3. Spiking procedure

Soil samples (730 mg dry soil) in glass vials were spiked with ei-ther of the following: 1 mL of AgNO3 solution, 2 mL of citrate stabi-lized Ag NP suspension, 1 mL of uncoated Ag NP suspension,resulting in soil slurries in all cases. Actual concentrations of addedAg in soil were determined by gamma spectrometry. In mineral soil,Ag concentrations were 56±3 mg kg−1 dry soil, 65±2 mg kg−1

dry soil, and 1010±14 mg kg−1 dry soil, for silver nitrate, citrate sta-bilized Ag NPs and uncoated Ag NPs, respectively. In organic soil, Agconcentrations were 61±3 mg kg−1 dry soil, 69±10 mg kg−1 drysoil, and 1300±170 mg kg−1 dry soil, for silver nitrate, citrate stabi-lized Ag NPs and uncoated Ag NPs, respectively. Preliminary studiesshowed that the water soluble Ag fraction was very low for uncoatedAgNPs. Thereforewe chose a higher Ag concentration for these particlesin order to ensure their detection in the water fraction. The Ag concen-trations used in this study are several orders of magnitude higher thanbackground concentrations of Ag in soils (0.07–0.29 mg kg−1, Hou etal., 2005), but close to that in biosolids (1.94–856 mg kg−1, USEPA,2009). Samples were stored at room temperature (20 °C) in thedark. There were 4 replicates for each silver type and for eachtime point (2 h, 2 d, 35 d, 70 d), plus 2 non-spiked soil samplesper time point.

2.4. Sequential extraction procedure

Soil samples spikedwith silver were submitted to sequential extrac-tion after 2 h, 2 d, 35 d and 70 d after adding silver to the soil. The ex-traction procedure is presented in Table 2 with information onextractants, solid to liquid ratios, contact time between soil samplesand extractants, and temperature. The extractions were carried out in50 mL centrifuge tubes (Nalgene), and supernatants were separatedfrom the solid fraction by centrifugation (15 min at 10,000×g). Steps1, 2, and 3 (inert electrolytes) involve displacement processes, i.e. re-versible processeswithout rupture of chemical bonds. Steps 4 (weak re-ducing agent), 5 (oxidizing agent), and 6 (strong mineral acid) involveirreversible dissolution processes through changing the physico-chemical forms of sorbed species or by attacking the structure withwhich the element is associated. The overall recovery of silver at theend of this multi-step procedure was 91±3.6%, with losses probablydue to adsorption to centrifuge tubes, funnels and pipettes.

2.5. Gamma spectrometry

All samples were measured for radioactivity on a NaI detector(1480 automatic gamma counter WIZARD 3″, Perkin Elmer). Targeted

Table 2Summary of the sequential extraction procedure. Volumes of reagents are given for one gram of soil. Each fraction was collected by centrifugation (15 min at 10,000×g) and filteredthrough a 2 μm blue ribbon filter (Whatman 589/3). The residue in steps 4 to 6 was washed once with 10 mL of MilliQ water, centrifuged and filtered the same way.

Fraction Operational definition Extractant Condition

1 Water soluble 20 mL of deionized water 1 h at 20 °C2 Exchangeable 20 mL of 1 M CH3COONH4 (pH 7) 2 h at 20 °C3 Weak acid dissolution (e.g. carbonate bound) 20 mL of 1 M CH3COONH4 (adjusted to pH 5 with CH3COOH) 2 h at 20 °C4 Easily reducible (e.g. Fe/Mn oxides) 20 mL of 0.04 M NH2OH·HCl in 25% v/v CH3COOH (pH 3) 6 h at 80 °C5 Oxidizable (e.g. organic matter) 15 mL of 30% H2O2 (adjusted to pH 2 with HNO3) 5.5 h at 80 °C

then 5 mL of 3.2 M CH3COOH in 20% v/v HNO3 30 min at 20 °C6 Acid digestible 20 mL of 7 M HNO3 6 h at 80 °C7 Residual Total activity−(F1+F2+F3+F4+F4+F5+F6)

330 C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

energies were from 884 keV to 1332 keV. Each sample, consisting ofca. 20 mL of the extractant or wash after centrifugation or solid resi-due for the residual fraction, was counted up to 30 min to obtain acounting error b10%, and in most cases below 1%. All reagents werealso counted to determine background radiation and verify the ab-sence of contamination during the experiment. To determine thetotal activity in a sample prior to extraction, all samples were countedon the NaI detector just before the sequential extraction was made.For time-course comparisons, all activities were corrected for radioac-tive decay, with the half-life of 110mAg being 249.9 days.

2.6. Statistical analyses

Changes in fractionation over time were tested using one-wayANOVA. In a few cases, the normality and homoscedasticity assumptionswere not reached, which does not preclude the use of the ANOVA, giventhe robustness of this test (Norman, 2010). Differences due to agingwereidentified with a Tukey's Honestly Significant Difference (HSD) test. Sta-tistical analyses were performed using the R package (version 2.11.1)(The R Foundation for Statistical Computing, Vienna, Austria, 2010).

3. Results and discussion

The water soluble and ion exchangeable fractions in soils are themost easily accessible to living organisms and will thereafter be re-ferred to as the bioaccessible fraction.

Partitioning patterns for silver differed significantly (1) over time,(2) among the silver initial forms, and (3) between the organic andmineral soils (Fig. 2).

3.1. Silver nitrate

In the organic soil, Ag added as AgNO3 was recovered mainly usingstringent extraction conditions, like hydrogen peroxide and concentratednitric acid, which accounted for a recovery of 13–23% and 58–68% ofadded Ag, respectively (Fig. 2a). The bioaccessible fraction represented6.4±0.7% 2 h after spiking and thereafter dropped to 0.9±0.1% (Fig. 3).Approximately 12% of added Ag remained bound in the residual fraction.In the mineral soil, 49.5±0.6% of Ag was bioaccessible 2 h after spiking,but this value decreased over time (35±1.8% on day 2, 10±2.6% onday 35, and 4±1.2% on day 70). This indicates that Ag fixation inmineralsoil is slower than in organic soil. Unlikewhatwas seen in the organic soil,silver extracted by hydrogen peroxide and nitric acid increased over timeto reach 18.7±1.2% and 40.8±0.4% of total Ag after 70 days, respectively(Fig. 2b).

The results of our study are in line with those published previouslyon silver nitrate spiked soils. Hou et al. (2005) contaminated four natu-ral soils (Andosol, Cambisol, Fluvisol and Regosol) with AgNO3 and ex-posed them toprecipitation in a grass-coveredfield for 18 months. Theyobserved a rapid fixation of Ag to the soils and found thatmost Ag (88%,total Ag concentration 3 mg kg−1) remained in the uppermost layer(0–2 cm). They also calculated transfer rates (by measuring the totalAg concentration in the uppermost and lower soil layers) between 3%

(Andosol) and 18% (Cambisol). The key soil parameters for Ag retentionwere organicmatter, phyllosilicate content, and Al concentrations, all ofwhich had higher levels in Andosol. Silver distribution was assessed bysequential extraction and indicated that Ag was present mainly in theresidual fraction (45±15%), the H2O2 extractible, organically boundfraction (27±15%), the metal–organic complex-bound fraction(18±7%), and the amorphous metal oxide-bound fraction (5.4±4.5%).The reasons reported for the rapid immobilization of Ag ions in soilswere: (1) incorporation into minerals by substitution of K+ with Ag+,due to the similar ionic radius of these two elements; (2) stabilizationby intercalation between silicate layers of clays; (3) precipitation of me-tallic Ag through reduction by organic matter such as humic acids, fulvicacids, and lignite (Akaighe et al., 2011; Hou et al., 2005).

In another study, Hou et al. (2006) investigated the background con-centration of Ag and its fractionation in various soil types, using sequen-tial extraction. Soils contained 0.17±0.08 mg Ag kg−1, and Agdistribution was consistent with that found earlier with added Ag: re-sidual (60%), H2O2 extractible, organically bound (18%), metal–organiccomplex-bound (10%), amorphous metal oxide-bound (8.3%).

Recent studies on silver behavior in sewage sludge suggest thatAg2S is the dominant species originating either from Ag ions or AgNPs (Kim et al., 2010; Nowack, 2010). While the sequential extractionprocedure cannot determine the exact silver species in our soils, theresults do show that AgNO3 is rapidly fixed by soils, thus, ultimately,Ag ion sources would expect to show low bioaccessibility.

3.2. Citrate stabilized silver nanoparticles

In organic soil, 21% of Ag added as citrate stabilized silver nanoparti-cles was extracted by water 2 h after spiking, but this value dropped to3.1±0.3% after 2 days and to 0.2±0.01% by day 35 (Fig. 2c). The wholebioaccessible fraction followed the same trend, showing a steep expo-nential decrease (from 24±1.4% 2 h after spiking, to 4.5±0.4% on day2, and 0.9±0.1% by day 35) (Fig. 3). The considerable drop in thewater soluble fraction between 2 h and 2 d was counterbalanced by arise in the nitric acid fraction (from 50 to 65%) and, to a lesser extent,in the residual fraction. The partitioning of Ag from AgNO3 and citratestabilized Ag NPs was very similar in organic soil, except from themuch larger water soluble fraction 2 h after spiking in the case of citratestabilized Ag NPs (Fig. 2c). In mineral soil, 37±4.8% of the silver addedas citrate stabilized AgNPswas bioaccessible 2 h after spiking, mainly ina water extractible form (25±5.5) % (Figs. 2d and 3). The bioaccessiblefraction decreased over time, more slowly than observed in the organicsoil, but still reaching very lowvalues by day70 (1.8±0.1%) (Fig. 3). Thehydrogen peroxide extractible fraction remained low, and the nitric acidextractible fraction around one third of total Ag throughout the agingprocess (Fig. 2d). On the other hand, the fraction extractedwith hydrox-ylamine (easily reducible material, e.g. Mn/Fe oxides) increased overtime and represented the largest fraction (Fig. 2d). The hydroxylaminefraction was also larger in mineral soil than in organic soil, whichcould be explained by the fact that Ag, in organic soil, is more likely tobind to organic material than to Mn/Fe oxides.

0

20

40

60

80

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

H2O NH4Ac pH7 NH4Ac pH5 HONH3Cl H2O2 HNO3 Residual

%

AgNO3 in organic soil2 h2 d35 d70 d

2 h2 d35 d70 d

2 h2 d35 d70 d

2 h2 d35 d70 d

2 h2 d35 d70 d

2 h2 d35 d70 d

a b c c

a a b c

a b c ca b c c

a b b b

a b c a

a b b b

0

20

40

60

80

%

AgNO3 in mineral soil

d

d

d

c

b

ba b c

a b c

a b c

a b c

a a b

a b b

a b b c

0

20

40

60

80

%

Citrate stabilized Ag NPs in organic soil

a b c c

a b c c a b c c

a b b c

a b c ca a b c

a a a a

0

20

40

60

80

%

Citrate stabilized Ag NPs in mineral soil

ca b c d a b c a a b b

a a b b

a ac b bc

a a b c

a a b b

0

20

40

60

80

%

Uncoated 20 nm Ag NPs in organic soil

abc

abab b a b c ca a a a a a a a

a a a a

a a ab b

a a ac

0

20

40

60

80

%

Uncoated 20 nm Ag NPs in mineral soil

aa b a b a b c d a b a a a b a aa a a

a ab ab b

a ab ab b

a b

c d

e f

Fig. 2. Relative distribution of 110mAg in sequential extraction fractions. Organic (a, c, e) or mineral soil samples (b, d, f) were spiked with either radiolabeled AgNO3 (a, b), citratestabilized Ag NPs (c, d), or uncoated Ag NPs (e, f) and submitted to sequential extraction 2 h, 2 d, 35 d or 70 d after spiking. Results are means±SD (n=4). Identical letters indicatetreatments that do not differ significantly (P>0.05).

331C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

The comparison between citrate stabilized Ag NPs and AgNO3 indi-cates that soil properties influence Ag speciation to a larger extent thanthe initial form of Ag (citrate stabilized Ag NPs vs. AgNO3). These resultssupport the hypothesis that some Ag NPs do not behave very differentlyfrom AgNO3 in soils, since a rapid reduction of Ag ions occurs upon con-tact with soil organic matter, transforming Ag ions into NPs and colloids.

Some differences were still found between citrate stabilized AgNPs and AgNO3. First, the water extractible fraction was larger for cit-rate stabilized Ag NPs than for AgNO3 during the first days following

spiking. This could be explained by Ag–citrate complexes reactingmore slowly than Ag ions, or by the likely presence of small NPs inthe water soluble fraction. Ultrafiltration and cloud point extraction(Chao et al., 2011) are examples of techniques that could be used toseparate and quantify ionic from particulate Ag, although this differ-entiation was not attempted in the present study. After 35 days, theamount of Ag extracted by water was similar for citrate stabilizedAg NPs and AgNO3 in both soils. Secondly, an interesting resultcame from the comparison between the H2O2 extracted fractions. In

0

10

20

30

40

50

0 10 20 30 40 50 60 70

%

Days

AgNO3 Organic

Citrate AgNPs Organic

Uncoated AgNPs Organic

AgNO3 Mineral

Citrate AgNPs Mineral

Uncoated AgNPs Mineral

a

cb

b

d

c

b bb

cc

c

a

c

a

b

a

a

b

ba

a

c

a

Fig. 3. Bioaccessible 110mAg in soil spiked with either radiolabeled AgNO3, citrate stabi-lized Ag NPs, or uncoated Ag NPs and submitted to sequential extraction 2 h, 2 d, 35 dor 70 d after spiking. Results are means±SD (n=4). Different letters indicate signifi-cant changes along time within a treatment (Pb0.05). For each of the four time points,gray boxes group points that are not significantly different (P>0.05).

332 C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

mineral soil, because of the very low organic matter content, Agextracted by H2O2 would reflect particle and colloid dissolution ratherthan organic matter oxidation. Therefore, the 4-fold higher H2O2

extractible fraction obtained in mineral soil spiked with AgNO3

could indicate that Ag colloids formed from AgNO3 are more easilyoxidizable than citrate stabilized Ag NPs.

3.3. Uncoated silver nanoparticles

Uncoated Ag NPs exhibited a strikingly different partitioning insoil compared to AgNO3 and citrate stabilized Ag NPs. In the or-ganic soil, the large majority of silver was not extractable(~85%), even under stringent conditions (Fig. 2e). Two hoursafter spiking, the bioaccessible fraction was the lowest of all treat-ments, but this fraction increased over time and reached 7.2±0.9% of the total Ag on day 70, which was higher than any ofthe other treatments at this time (Fig. 3). The partitioning of Agin mineral soil was very close to that in organic soil, with alarge majority not extractable and a bioaccessible fraction that in-creased over time (8.6±0.2% on day 70) (Figs. 2f and 3). In thecase of these uncoated Ag NPs, soil properties had a very limitedimpact on Ag speciation.

On day 70, bioaccessible Ag from uncoated Ag NPs in mineralsoil was twice and nearly 5-fold as high as in AgNO3 and citratestabilized Ag NP treatments, respectively. In organic soil the dif-ferences were larger, with bioaccessible Ag from uncoated AgNPs being 9-fold and 8-fold as high as in AgNO3 and citrate sta-bilized Ag NP treatments, respectively. Contrary to Ag(I), Ag0 isnot likely to be bound to cation exchange binding sites. Regard-ing sulfur binding sites, reaction of sulfur species (H2S, HS−)with Ag0 has been demonstrated in anaerobic environments(Kim et al., 2010; Liu et al., 2011). However, Liu et al. (2011) ob-served no reaction between Ag0 and sulfate, which is the mainform of sulfur in aerobic environments (i.e. the conditions inthe present study). It is therefore unlikely that the higher totalconcentration of Ag in the uncoated Ag NP treatment and bind-ing site saturation are the cause of the increased bioaccessibility.The initial Ag ionic/water soluble content of the Ag powder wasmuch lower than that added as AgNO3, thus the higher Ag con-centrations used for Ag NPs should not impair the validity ofthe comparison.

Nevertheless, since the AgNO3 showed rapid binding in soil, it isinteresting to consider why Ag released from the uncoated NPs did

not also become bound to soils during this time. This might be drivenby the slow dissolution rate of the uncoated Ag NPs compared toAgNO3 and citrate stabilized Ag NPs. This explanation is supportedby the larger size of individual uncoated Ag NPs and their tendencyto form aggregates (Fig. 1c), inasmuch as this reduces the specific sur-face area and, in turn, the dissolution rate. In addition, Ag could be re-leased from the uncoated Ag NPs in form of smaller particles (Gloveret al., 2011) or relatively stable complexes, and the soil sorption ki-netics for these species would probably be slower than for AgNO3

complexed ions, resulting in a greater “equilibrium” proportion inthe bioaccessible fraction.

4. Conclusions

One of the main issues in risk assessment of Ag NPs is whetherthey are suitably covered under the framework for metal risk as-sessment. Since Ag ions also are known to become stronglybound to soil components, there have been suggestions that theframeworks should also be suitable for Ag NPs (Nowack, 2010).However, our study has shown for the first time that someforms of Ag NPs can be more bioaccessible than Ag ions, sincethey can act as a constant source of relatively stable and bioacces-sible Ag. This increase in bioaccessible Ag over time in soil spikedwith uncoated Ag NPs, being between 8 and 9 times greater thanthe bioaccessible fraction of AgNO3 after 70 days contact time, isan important and unexpected finding that calls for caution in NPimpact assessment.

Acknowledgments

Financial support from the Research Council of Norway (projectsNanoEnvironment NFR 182069/S10 and NanoTrace NFR 183758/S30)is gratefully acknowledged. The authors thank Hilde Kolstad and ElinØrmen for help at the Microscopy Lab.

References

Ahlf W, Drost W, Heise S. Incorporation of metal bioavailability into regulatoryframeworks—metal exposure in water and sediment. J Soil Sediment 2009;9:411–9.

Akaighe N, MacCuspie RI, Navarro DA, Aga DS, Banerjee S, Sohn M, et al. Humic acid-induced silver nanoparticle formation under environmentally relevant conditions.Environ Sci Technol 2011;45:3895–901.

Blaser SA, Scheringer M, MacLeod M, Hungerbuhler K. Estimation of cumulative aquaticexposure and risk due to silver: contribution of nano-functionalized plastics andtextiles. Sci Total Environ 2008;390:396–409.

Chao JB, Liu JF, Yu SJ, Feng YD, Tan ZQ, Liu R, et al. Speciation analysis of silver nanopar-ticles and silver ions in antibacterial products and environmental waters via cloudpoint extraction-based separation. Anal Chem 2011;83:6875–82.

Coutris C, Hertel-Aas T, Lapied E, Joner EJ, Oughton DH. Bioavailability of cobalt andsilver nanoparticles to the earthworm Eisenia fetida. Nanotoxicology 2011.doi:10.3109/17435390.2011.569094. (online Apr 2011).

Doty RC, Tshikhudo TR, Brust M, Fernig DG. Extremely stable water-soluble Ag nano-particles. Chem Mater 2005;17:4630–5.

Glover RD, Miller JM, Hutchison JE. Generation of metal nanoparticles from silver andcopper objects: nanoparticle dynamics on surfaces and potential sources of nano-particles in the environment. ACS Nano 2011;5:8950–7.

Gottschalk F, Sonderer T, Scholz RW, Nowack B. Modeled environmental concentra-tions of engineered nanomaterials (TiO2, ZnO, Ag, CNT, fullerenes) for different re-gions. Environ Sci Technol 2009;43:9216–22.

Hou H, Takamatsu T, Koshikawa MK, Hosomi M. Migration of silver, indium, tin, anti-mony, and bismuth and variations in their chemical fractions on addition touncontaminated soils. Soil Sci 2005;170:624–39.

Hou H, Takamatsu T, Koshikawa MK, Hosomi M. Concentrations of Ag, In, Sn, Sb and Bi,and their chemical fractionation in typical soils in Japan. Eur J Soil Sci 2006;57:214–27.

Kim B, Park C-S, Murayama M, Hochella MF. Discovery and characterization of silversulfide nanoparticles in final sewage sludge products. Environ Sci Technol2010;44:7509–14.

Liu JY, Pennell KG, Hurt RH. Kinetics and mechanisms of nanosilver oxysulfidation. En-viron Sci Technol 2011;45:7345–53.

Lowry GV, Hotze EM, Bernhardt ES, Dionysiou DD, Pedersen JA, Wiesner MR, et al. En-vironmental occurrences, behavior, fate, and ecological effects of nanomaterials:an introduction to the special series. J Environ Qual 2010;39:1867–74.

333C. Coutris et al. / Science of the Total Environment 420 (2012) 327–333

Mueller NC, Nowack B. Exposure modeling of engineered nanoparticles in the environ-ment. Environ Sci Technol 2008;42:4447–53.

Norman G. Likert scales, levels of measurement and the “laws” of statistics. Adv HealthSci Educ 2010;15:625–32.

Nowack B. Nanosilver revisited downstream. Science 2010;330:1054–5.Oughton DH, Hertel-Aas T, Pellicer E, Mendoza E, Joner EJ. Neutron activation of engi-

neered nanoparticles as a tool in studies on their environmental fate and uptake inorganisms. Environ Toxicol Chem 2008;27:1883–7.

Peijnenburg W, Zablotskaja M, Vijver MG. Monitoring metals in terrestrial environ-ments within a bioavailability framework and a focus on soil extraction. EcotoxicolEnviron Saf 2007;67:163–79.

Peralta-Videa JR, Zhao LJ, Lopez-Moreno ML, de la Rosa G, Hong J, Gardea-Torresdey JL.Nanomaterials and the environment: a review for the biennium 2008–2010. J Haz-ard Mater 2011;186:1-15.

Scheckel KG, Luxton TP, El Badawy AM, Impellitteri CA, Tolaymat TM. Synchrotron spe-ciation of silver and zinc oxide nanoparticles aged in a kaolin suspension. EnvironSci Technol 2010;44:1307–12.

Semple KT, Doick KJ, Jones KC, Burauel P, Craven A, Harms H. Defining bioavailabilityand bioaccessibility of contaminated soil and sediment is complicated. EnvironSci Technol 2004;38:228A–31A.

Shoults-Wilson WA, Reinsch BC, Tsyusko OV, Bertsch PM, Lowry GV, Unrine JM. Role ofparticle size and soil type in toxicity of silver nanoparticles to earthworms. Soil SciSoc Am J 2011;75:365–77.

Tessier A, Campbell PGC, Bisson M. Sequential extraction procedure for the speciationof particulate trace metals. Anal Chem 1979;51:844–51.

Unrine JM, Hunyadi SE, Tsyusko OV, Rao W, Shoults-Wilson WA, Bertsch PM. Evidencefor bioavailability of Au nanoparticles from soil and biodistribution within earth-worms (Eisenia fetida). Environ Sci Technol 2010a;44:8308–13.

Unrine JM, Tsyusko OV, Hunyadi SE, Judy JD, Bertsch PM. Effects of particle size onchemical speciation and bioavailability of copper to earthworms (Eisenia fetida)exposed to copper nanoparticles. J Environ Qual 2010b;39:1942–53.

US EPA. Targeted National Sewage Sludge Survey Statistical Analysis Report 2009. [Accessed2011 Jun 7]. Available from http://water.epa.gov/scitech/wastetech/biosolids/upload/TNSSS_stat_appendix-A-Apr-2009.pdf.