Sulfidation Mechanism for Zinc Oxide Nanoparticles and the Effect ofSulfidation on Their SolubilityRui Ma,†,⊥,# Clement Levard,‡,⊥,# F. Marc Michel,‡,§,⊥,# Gordon E. Brown, Jr.,‡,§,∥,⊥,#

and Gregory V. Lowry*,†,⊥,#

†Department of Civil & Environmental Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States‡Department of Geological & Environmental Sciences, Stanford University, Stanford, California 94305, United States§Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States∥Department of Photon Science, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States⊥Center for the Environmental Implications of NanoTechnology (CEINT), Duke University, Durham, North Carolina 27708, UnitedStates

*S Supporting Information

ABSTRACT: Environmental transformations of nanoparticles (NPs)affect their properties and toxicity potential. Sulfidation is an importanttransformation process affecting the fate of NPs containing metalcations with an affinity for sulfide. Here, the extent and mechanism ofsulfidation of ZnO NPs were investigated, and the properties ofresulting products were carefully characterized. Synchrotron X-rayabsorption spectroscopy and X-ray diffraction analysis reveal thattransformation of ZnO to ZnS occurs readily at ambient temperature inthe presence of inorganic sulfide. The extent of sulfidation depends onsulfide concentration, and close to 100% conversion can be obtained in5 days given sufficient addition of sulfide. X-ray diffraction andtransmission electron microscopy showed formation of primarily ZnSNPs smaller than 5 nm, indicating that sulfidation of ZnO NPs occursby a dissolution and reprecipitation mechanism. The solubility of partially sulfidized ZnO NPs is controlled by the remainingZnO core and not quenched by a ZnS shell formed as was observed for partially sulfidized Ag NPs. Sulfidation also led to NPaggregation and a decrease of surface charge. These changes suggest that sulfidation of ZnO NPs alters the behavior, fate, andtoxicity of ZnO NPs in the environment. The reactivity and fate of the resulting <5 nm ZnS particles remains to be determined.

■ INTRODUCTION

Zinc oxide nanoparticles (ZnO NPs) are widely used innumerous applications and products such as catalysis, semi-conductors, solar cells, sensors, sunscreens, textiles, cosmetics,coatings, pigments, and optic, and electronic materials.1,2 Manyapplications lead to their release to wastewater or surfacewaters.3 Nanomaterials entering wastewater treatment plants(WWTPs) will ultimately reside in the biosolids, which areoften subsequently used on croplands as a fertilizer. Concernshave therefore been raised about the potential risks associatedwith exposures of sensitive aquatic and terrestrial organisms toZnO NPs and released Zn2+ ions. Recent studies have shownthat exposures to ZnO NPs are detrimental to bacteria,4,5

various aquatic species,6,7 and earthworms.8

The toxicity of ZnO NPs has been attributed to both the NPand to the released Zn2+.9−11 Most toxicity studies have usedpristine ZnO NPs rather than their transformation productsexpected in the environment, which can have very differentproperties than the pristine NPs.12 Important environmentaltransformation of ZnO NPs include dissolution,13 sulfidation,phosphorylation,14 and interaction with natural organic

matter.15 Understanding the environmental transformations ofZnO NPs and how they affect important physicochemicalproperties like release of Zn2+ in aqueous solutions is thusessential for properly assessing their risks.Zinc readily transforms to ZnS in the presence of sulfide.12,16

Sulfide produced from microbial sulfate reduction is ubiquitousin anoxic environments such as river and lake sediment andpresent at concentrations ranging from 1 to 100 μg/L.17 Inwastewater treatment plants, sulfide concentrations range froma few μg/L up to 10 mg/L.18 Sulfidation of ZnO NPs istherefore expected in these environments. ZnS has recentlybeen identified in fresh biosolids that were amended with ZnONPs19 and in fresh and dried biosolids without ZnO NPaddition.20 ZnS NPs have also been identified as a major Znspecies in suspended matter taken from the Seine River nearParis.21 Metal sulfides are generally less soluble in water than

Received: August 30, 2012Revised: February 19, 2013Accepted: February 20, 2013Published: February 20, 2013

Article

pubs.acs.org/est

© 2013 American Chemical Society 2527 dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−2534

their corresponding oxides or chlorides, for example, Ag2S (Ksp= 10−50)16 versus Ag2O (Ksp = 2 × 10−8)22 or HgS (Ksp =10−54) versus HgCl2 (Ksp = 10−2).23 Even small amounts ofsulfur dramatically decrease the solubility of Ag NPs by formingan Ag2S shell on the particles.16 Little is known about themechanisms of sulfidation of ZnO NPs in aqueous solution atambient temperatures. Some studies have demonstrated thatZnO NPs readily react at room temperature with either sulfideor organosulfur (e.g., thioacetamide);24−26 however, knowledgeof the mechanisms of the sulfidation processes of ZnO NPs isstill lacking. Also, the properties of sulfidized ZnO NPs,including chemical composition, size and morphology, surfacecharge, and aqueous solubility, have not been reported.The present study focuses on sulfidation of ZnO NPs by

reacting them with aqueous solutions with different S/Znratios, followed by careful characterization of the resultingsulfidized NPs. The objectives of this study were to determine(1) to what extent ZnO NPs are sulfidized at various S/Znratios, (2) the properties of the resulting ZnS particles or ZnS-coated ZnO NPs that form during sulfidation including theimpact of sulfidation on aqueous solubility, and (3) themechanism of sulfidation of ZnO NPs in aqueous solution. Toaddress these issues, commercially available ZnO NPs weresulfidized over a range of S/Zn ratios to obtain different extentsof transformation from ZnO to ZnS. Both synchrotron-basedX-ray diffraction (XRD) and extended X-ray absorption finestructure (EXAFS) spectroscopy were used to characterize thetransformation products of the ZnO NPs after their reactionwith Na2S in aqueous solution. Particle size and morphology,surface charge, and dissolution of the resulting ZnS or ZnO/ZnS core/shell NPs were also characterized using transmissionelectron microscopy (TEM), dynamic light scattering (DLS),electrophoretic mobility (EPM), and inductively coupledplasma (ICP) spectrometry.

■ MATERIALS AND METHODSZnO NPs. ZnO NPs (P99/30) were purchased from

Nanosun (Dandenong, Australia) (Figure S1 of the SupportingInformation). These NPs were selected because they are beingused in a large-scale, multi-institution, international effort(Transatlantic Initiative for Nanotechnology and the Environ-ment) to determine their fate in a wastewater treatment plantand in soils amended with the biosolids from the WWTP. Theparticles were used as received. The manufacturer claims thatthe ZnO NPs contain no organic capping agent.Sulfidation of ZnO NPs. The ZnO NPs were sulfidized in

50 mL plastic centrifuge tubes using sodium sulfide (Na2S).The S/Zn ratio was changed by fixing ZnO concentration at 15mM and varying the concentration of sulfide in a 10 mMNaNO3 electrolyte. Before adding sulfide, the suspensions weresonicated in an ice-bath (Branson Model 250) with a microtipat a power input of 10 W for 10 min to disperse the ZnO NPs.To minimize oxidation of sulfide by dissolved oxygen, allsolutions were prepared using N2-purged DI water and N2-purged headspace. Various amounts of Na2S were then addedinto the tubes to produce S/Zn ratios ranging from 0.038 to2.16 (Table S1 of the Supporting Information). In a glovebox,solution pH was adjusted to 12 ± 0.2 using NaOH and then thecentrifuge tubes were sealed. Two additional ratios (S/Zn 0.616and 0.864) were sulfidized at pH 7 (buffered with HEPES) toassess the effect of pH on sulfidation. After 5 days, the resultingsolutions were centrifuged at 6000 g for 20 min. Supernatantswere carefully decanted and DI water was added. The tubes

were shaken to suspend the particles in sulfide-free water, andthen the tubes were centrifuged again and the supernatantswere decanted. These steps were repeated twice to remove theextra sulfide in solution. Finally, the reacted ZnO NPs weresuspended in deoxygenated DI water. Aliquots were freeze-dried (Labconco, Freezone 4.5 freeze-dryer) at −40 °C and 0.1mbar overnight and kept in an anaerobic glovebox prior to X-ray characterization as described below. Stock solutions ofsulfidized particles were kept in N2 purged solution in ananaerobic glovebox prior to characterization. However, nospecial care was taken to prevent exposure to oxygen duringpreparation of samples for TEM, EXAFS, or XRD.

NP Characterization. The pristine and sulfidized ZnO NPswere characterized with multiple techniques includingthermogravimetric analysis (TGA) to determine the presenceof a capping agent, TEM, and DLS for size, EPM for charge,XRD, and EXAFS spectroscopy to determine Zn speciation ofthe NPs. Details of each of these characterization methods aredescribed next.Thermogravimetric analysis (SDT Q600, TA Instruments,

New Castle, DE) was used to confirm the manufacturer’s claimof limited or no organic coating on ZnO NPs. Approximately15 mg of ZnO NPs was placed in the TGA holder. Theparticles were heated at a rate of 5 °C/min from ambienttemperature up to 700 °C in air, and the weight change uponheating was determined. Details of the TGA results areprovided in Figure S2 of the Supporting Information.TEM analysis of the pristine ZnO NPs was performed as

follows. The ZnO NPs were dispersed in 10 mM NaNO3 toprovide 20 mg/L NP dispersion. A 1 μL aliquot of thesuspension was then deposited on a copper TEM grid coveredby an ultrathin carbon support film. The TEM grid was driedunder ambient air before TEM observations. Pristine ZnO NPswere imaged using a Hitachi H-7100 electron microscopeoperating at 70 kV. TEM images of the sulfidized ZnO NPswere acquired using a FEI Tecnai G2 F20 X-TWIN(accelerating voltage of 200 kV) equipped with a CCD camera.Fast Fourier Transform (FFT) patterns were calculated fromthe TEM images for phase identification using Image J.27 Thewidths of the rings in the FFT patterns were used to identifythe crystalline phase of the NPs.The EPMs of the pristine and sulfidized ZnO NPs were

measured as a function of pH. Aqueous dispersions of the ZnOand ZnO/ZnS NPs (20 mg/L) in 10 mM NaNO3 wereprepared at different pH values ranging from 4 to 12. HNO3and NaOH were used to adjust the pH to desired values. TheEPMs were measured using a Malvern Zetasizer, and Zetapotentials were estimated from Smoluchowski’s model usingexperimental EPM data.DLS measurements were made on both pristine and

sulfidized ZnO NPs to determine their hydrodynamicdiameters. The stock solutions of 1200 mg/L concentrationof different NPs were briefly sonicated at 10 W power input for2 min and then quickly diluted to 20 mg/L using a 10 mMNaNO3 solution with pH preadjusted to 7. Then the averageNP sizes of the samples were measured using an ALV/CGS-3Compact Goniometer System equipped with a 22 mW HeNeLaser (λ = 632.8 nm) at a scattering angle of 90°.Chemical and structural characterization of ZnO and ZnO/

ZnS NPs after reaction with different S/Zn ratios was carriedout using EXAFS spectroscopy and synchrotron-based XRD.Zinc K-edge EXAFS spectra were collected at the StanfordSynchrotron Radiation Lightsource (SSRL) on wiggler beam-

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342528

line 4−3 at room temperature in transmission mode. Sampleswere first pelleted after diluting them with glucose powder toachieve an optimized absorption jump (Δμ) of 1. An N2-cooledSi(220) double-crystal monochromator was detuned by 30%for harmonic rejection, and energy calibration was performedwith a Zn metal foil placed after the I1 transmitted beamdetector. EXAFS spectra of the following reference compoundswere also collected: bulk ZnS (sphalerite), ZnS NP modelcompound, ZnO, Zn5(CO3)2(OH)6, and ZnCO3. Linearcombination fitting (LCF) of the EXAFS spectra wasperformed using the SIXPack interface to the IFEFFIT XAFSanalysis package.28 Samples were characterized by X-raydiffraction on beamline 11-ID-B at the Advanced PhotonSource (APS, Argonne National Laboratory) using aPerkinElmer amorphous silicon detector. Samples were firstmounted in 1 mm Kapton capillaries and diffraction patternswere collected at λ = 0.2128 Å, (approximately 58 keV). Imageswere integrated into intensity vs 2θ plots using Fit2D.29

Crystallite sizes were determined using the Scherrer line-broadening equation on the (111) reflection.30 To apply theScherrer equation, the background was subtracted and thepeaks were deconvoluted using Fityk.31 Instrumental peakbroadening was estimated from a CeO2 model compound.Model Compounds. Model compounds were either

purchased or synthesized in the laboratory. Bulk ZnS(sphalerite) and (ZnO) (zincite) were obtained from AlfaAesar. Synthetic nanocrystalline ZnS were prepared in aglovebox under nitrogen atmosphere by mixing 1 mM Zn-acetate with 2 mM sodium sulfide (Na2S) in deoxygenated DI.After 3 days of reaction, the particles were centrifuged at 26000g, washed with DI water twice, and dried in air. Thecrystallite size of the nanocrystalline ZnS model compound wasestimated from the broadening of the XRD peaks and found toaverage 3 nm.Dissolution Measurements. To determine the dissolution

rate and solubility of the particles as a function of their degreeof sulfidation, an aliquot of the slurry of the fresh or sulfidizedZnO NPs was diluted to an initial NP concentration of 20 mg/L in an aqueous 10 mM NaNO3 solution of 500 mL at pH 7 ±0.3 without further pH adjustment. At specified time points, 7mL of solution was removed for analysis. Particles and ionswere separated by centrifugation (3000 g for 20 min) usingAmicon filters (NMWL 3K Da). The amount of released Zn2+

was measured by ICP spectrometry (TJA IRIS Advantage/1000Radical ICAP Spectrometer) after digestion in 5% nitric acid.The dissolved fraction (C/C0) was then determined andreported as the average and standard deviation of duplicatemeasurements. A control experiment on a sample consisting of2 mg/L of zinc ion adsorbed to the ultrafilter unit was alsoconducted. Solutions with a Zn2+ concentration of 2 mg/L wereflowed through the filter unit for 1 h, and filtrates weremeasured for [Zn]. The adsorptive loss of Zn2+ to the filter wasnegligible at a Zn concentration of 2 mg/L.

■ RESULTS AND DISCUSSIONCharacterization of Pristine ZnO NPs. The ZnO NPs are

a mixture of spherical and rod-shaped particles (Figure S1 ofthe Supporting Information). The average particle diameter ofthe ZnO NPs was 30 ± 10 nm based on TEM measurements of∼100 particles. XRD data (Figure 1) show that the initial ZnONPs are zincite. The N2−BET surface area of the ZnO NPsprovided by the manufacturer is 27.9 m2/g, which is slightlybelow the calculated geometric surface area of 35.7 m2/g,

assuming spherical particles of 30 nm diameter and the densityof bulk ZnO. The difference between the measured andcalculated values may be due to aggregation of ZnO NPs andthe presence of rod-shaped particles.The weight loss upon heating the ZnO NPs to 700 °C is

∼1.6% (Figure S2 of the Supporting Information). The lossbetween 0 and 100 °C can be attributed to water lossaccounting for 0.5%. Hence, the organic residual on the particleis estimated to be ∼1.1%. The presence of an organic coatingon the pristine ZnO NPs therefore appears to be limited andmight have been introduced during the synthesis to controlparticle size.

Sulfidation of ZnO NPs. Both EXAFS and synchrotron-based XRD were used to characterize the structures of thesulfidized ZnO NPs. X-ray diffraction patterns for the series ofsulfidized ZnO NPs are presented in Figure 1 and comparedwith those of 3 model compounds: bulk ZnO (zincite), bulkZnS (sphalerite), and nano ZnS (sphalerite). The XRD patternfor the nanocrystalline ZnS model compound exhibits muchbroader peaks than that of bulk sphalerite due to the smallercrystallite size (3 nm) of the former. The initial nanoparticlesexhibit the zincite structure. As sulfide is added to the system anadditional phase appears as the peaks characteristic to thezincite phase disappear, and the peak intensities of this newphase gradually increase as the S/Zn ratio increases. This newphase was identified as sphalerite (ZnS).The sphalerite XRD peaks for the most sulfidized samples are

relatively broad even though they are slightly sharper thanthose of the 3 nm nanocrystalline ZnS reference sample,indicating that the ZnS produced in our experiments arerelatively small. The crystallite size of newly formed ZnS wasestimated using the Scherrer line-broadening formula (Figure2). ZnS crystallite size increased from ∼2.5 nm to ∼5 nm withincreasing sulfidation. The small ZnS size even for the onesproduced at relatively high Zn/S ratios suggests that this size

Figure 1. XRD patterns of ZnO NPs sulfidized for 5 days with S/Znratios from 0 to 2.158. Model compounds are presented in red forcomparison (ZnO (zincite) and ZnS (sphalerite, nano, and bulk)).Note that the spectra for the initial ZnO NPs is identical to that ofbulk ZnO (zincite).

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342529

range is the most stable one for ZnS under these conditions.This size range is often observed for aqueous-based synthesis ofZnS NPs at ambient temperature.32,33 After reaching this size,rate of surface growth of the previously nucleated NPs may fallto levels that allow the parallel precipitation of new ZnS nuclei.Linear combination fitting (LCF) results of the EXAFS

spectra from a k of 3 to 15 Å−1 are shown in part a of Figure 3.The best fits were obtained using the initial ZnO NP and the 3nm nanocrystalline ZnS reference spectra. The quality of the fitis illustrated by small residuals (Figure S3 of the SupportingInformation). Adding a third phase did not improve the fits

significantly. Attempts were made to fit the EXAFS spectrausing the bulk sphalerite model compound instead of the 3 nmnanocrystalline ZnS but this yielded much higher residuals. TheEXAFS spectra of the two ZnS model compounds (bulk andnano) are slightly different as observed in part b of Figure 3.The EXAFS spectrum of the nanocrystalline ZnS exhibit lesspronounced features than that of the bulk ZnS modelcompound. This is mainly due to a less intense second shellcontribution (Zn−Zn distances) as observed in the FourierTransform (cf., the Fourier Transform in Figure S4 of theSupporting Information). This can be explained by the fact thatfor the small crystallite domain (about 3 nm estimated from theScherrer equation) a significant fraction of the Zn atoms arelocated at the surface. At a local scale, these surface atoms areunder-coordinated, which causes a decrease of the average Zn−Zn contribution in the EXAFS spectra. The difference betweenthe bulk and the nanocrystalline ZnS can also be due to thepresence of amorphous ZnS in addition to the nanocrystallineparticles which will also lead to the decrease of the Zn−Zn peakintensity. Similar behavior has been previously observed forZnS NPs.33 The relative abundances of ZnO and ZnS phaseswere obtained from the LCF and plotted in part c of Figure 3.Even though complete transformation of the ZnO NPs intoZnS was expected for the highest sulfide concentrations (S/Zn= 1.080 and 2.158), adding the ZnO NP model compoundimproved the fit significantly, and ∼3% residual ZnO isestimated for the most sulfidized ZnO NPs. The presence of fitcomponents at concentrations less than ∼10% of the total issomewhat speculative given the sensitivity of the LCF process.However, given the high quality of the EXAFS spectra and thefit of the background-subtracted spectra up to k of 15 Å−1, it islikely that some ZnO remains in the product despite the highS/Zn ratio used to sulfidize the NPs. The Fourier Transform(FT) for the most sulfidized sample (S/Zn = 2.158) wasplotted and compared to those of the nanocrystalline ZnSmodel compound and the initial ZnO NPs (part d of Figure 3).The main difference between the FT of the sample and the ZnSmodel compound can be observed on the left of the first mainZn−S pair correlation (dotted circle in part d of Figure 3). Thesmall shoulder present in the FT of the sample and not in thatof the ZnS model compound can be attributed to a Zn−O paircorrelation as observed for the initial ZnO NPs. Adding thisZnO component could explain why the fit was improved.However, if ZnO were present in this sample, we would expecta small contribution of the second shell (at R + ΔR = ∼2.9 Å)given its strong peak intensity (mainly attributed to Zn−Znpair correlations), which is not the case here. Attempts to fitthis second phase were made using other potential Zn−Obearing model compounds such as ZnCO3, Zn5(CO3)2(OH)6,and Zn-acetate. When using the EXAFS spectrum ofZn5(CO3)2(OH)6 as a second component, the resulting fitwas as good as that using the EXAFS spectrum of ZnO NPs,with the nanocrystalline ZnS accounting for 93% andZn5(CO3)2(OH)6 the remaining 7%. Using the EXAFSspectrum of ZnCO3 or Zn-acetate as a second fit componentgave higher residuals. It is very likely that a second Zn−Obearing phase is present accounting for less than 10% of thetotal Zn. However, we are at the limit of detection with thisLCF technique, and the exact nature of this Zn−O bearingphase cannot be clearly identified. A comparison betweensulfidation at pH 7 and pH 12 for S/Zn = 0.616 S/Zn = 0.864is summarized in Table S3 of the Supporting Information. Forboth pH values, the same amount of sulfidation was observed,

Figure 2. ZnS crystallite sizes obtained from the Scherrer line-broadening equation as a function of % ZnS determined from EXAFSlinear combination fitting.

Figure 3. a. EXAFS spectra (open dots) and linear combination fits(red lines) of the resulting ZnO/ZnS NPs with S/Zn ratios rangingfrom 0.038 to 2.158. b. EXAFS spectra of initial 30 nm ZnO NPs. ZnSmodel compounds that were used for the fit. c. Percentages of ZnOand ZnS as a function of S/Zn ratio. d. Fourier transforms of theEXAFS spectra of initial ZnO NPs, nanocrystalline ZnS modelcompound, and the sulfidized ZnO NP sample with HS−/Zn = 2.158.

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342530

and only ZnO and ZnS were formed. This indicates a minimaleffect of pH on the sulfidation, likely a result of the significantlyhigher affinity of Zn for S relative to OH.TEM analysis was performed on ZnO/ZnS particles

prepared using three S/Zn ratios (0.238, 0.616, and 1.08).Parts a−c of Figure 4 show lower resolution TEM images,

whereas parts d−f of Figure 4 show higher resolution TEMimages for identifying lattice fringe spacings for sulfidized ZnONPs. From TEM, both morphology and lattice spacing changesupon increases in S/Zn ratios can be observed. The initial ZnONPs were fairly well dispersed (Figure S1 of the SupportingInformation), but sulfide addition leads to significant particleaggregation (part a of Figure 4). Further increase of S/Zn ratioformed a core/shell structure with some smaller, newly formedparticles attached to the outside of the hollow structure (part bof Figure 4). At a high S/Zn ratio (= 1.08), which results inmostly ZnS from EXAFS analysis, less core/shell structures andlarger aggregates of smaller ZnS particles are apparent (part c ofFigure 4). There is little evidence of an amorphous ZnS phasein any of these TEM images.To obtain phase identity for the sulfidation products, fast

fourier transform (FFT) patterns were calculated from highresolution TEM images (parts d−f of Figure 4). At a S/Zn ratioof 0.238 (part d of Figure 4), the d-spacings estimated from theFFT image of the dotted area are 2.8 Å and 2.5 Å, which areconsistent with the d-spacings for zincite (010) and (011)planes, respectively. If we look more closely at the shell of the

NPs, one can see small NPs (∼2 nm in diameter) that have a d-spacing of 3.1 Å, which is consistent with the d-spacing of thesphalerite (111) plane. For a S/Zn ratio of 0.616 (part e ofFigure 4), the FFT image exhibits a broad diffuse ringcorresponding to d-spacings ranging from 2.5 to 3.1 Å, whichis consistent with a mixture of ZnO and ZnS that might beattached to ZnO NP cores. The ring observed at 1.9 Å isconsistent with the d-spacing for the sphalerite (220) plane. Fora S/Zn ratio of 1.08 (part f of Figure 4), two diffuse rings wereobserved at 1.9 and 3.1 Å, which are consistent with thesphalerite structure and the EXAFS analysis indicating thatthese particles are largely ZnS (sphalerite).TEM images not only confirmed close to 100% conversion of

ZnO NPs to ZnS NPs for S/Zn ratios larger than 1.08, but theyalso suggest a dissolution and reprecipitation mechanism forsulfidation of the ZnO NPs. It seems that aggregation of theZnO/ZnS NPs, and the shell of ZnS NPs formed did notprevent dissolution and release of Zn2+ as discussed next.

Rates of Zinc Ion Release/Solubility. Dissolution rates ofthe pristine ZnO NPs and the sulfidized ZnO/ZnS NPs withvarious S/Zn ratios were measured (Figure 5). The relatively

rapid dissolution observed for the pristine ZnO NPs isconsistent with previous studies.2,13,34 The concentration ofZn2+ released from the ZnO NPs after 80 h was ∼0.038 mM.This value is lower than previous reports of 0.059 mM Zn2+ for24 nm ZnO NPs13 and 0.086 to 0.09 mM Zn2+ for 40−70 nmZnO NPs.35 These differences suggest that equilibriumsolubility was not achieved after 80 h for the pristine ZnONPs. In general, sulfidation reduced the rate and extent of therelease of Zn2+ from the NPs; however, the effect was relativelysmall until the NPs were at least 50% sulfidized (a Zn/S ratio of0.616). For lower degrees of sulfidation, the rate and extent ofparticle solubility is close to that of the pristine ZnO NPs. Thisfinding is consistent with formation of a noncoherent shell ofZnS NPs around the ZnO NP cores (part b of Figure 4). Thesolubility of even the most sulfidized ZnO NPs was manyorders of magnitude higher than that expected for bulksphalerite (Ksp = 2 × 10−25 yields a solubility of 3.7 × 10−10

Figure 4. TEM images of ZnO/ZnS NPs after ZnO NPs were reactedwith Na2S-containing solutions with various S/Zn ratios: (a)(d) S/Zn= 0.238; (b)(e) S/Zn = 0.616; and (c)(f) S/Zn = 1.08.

Figure 5. Dissolution rate measurements for ZnO and ZnO/ZnS NPsresulting from various S/Zn ratios. The initial particle concentrationwas 0.25 mM. Error bars represent ± standard deviations of duplicates.Because many of the Zn2+ concentrations are small, most of them areblocked by the symbols.

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342531

mM). This could result from higher solubility of amorphousZnS formed, a higher solubility for nanosized ZnS compared tobulk ZnS,36 or of the presence of a more soluble second phase.There was little evidence of an amorphous phase in TEMimages, and the particle sizes observed in TEM (Figure 4) areconsistent with the crystallite size measured by XRD (Figure2). Thus, amorphous ZnS is not likely the cause of the highersolubility. The 8 orders of magnitude higher solubility observedhere compared to bulk ZnS is likely too large to be due to a sizeeffect alone.12 Using the Kelvin equation12 and solubility datafor 1 and 10 nm ZnS,36 we calculate an expected solubility of1.4 × 10−9 mM for 5 nm ZnS. This solubility is much lowerthan measured here suggesting that the higher solubility is likelythe result of some higher solubility residual Zn−O-containingphase in the sulfidized materials (e.g., ZnO or possiblyhydrozincite), as suggested by the EXAFS analysis. Indeed,about 3 to 7% of the Zn remained bound to O for S/Zn ratiosof 1.080 and 2.158. The amount of Zn released after 120 h for aS/Zn ratio of 1.080 is about 0.018 mM, which represents about7% of the initial Zn present in the solution at the beginning ofthe experiment.Surface Charge and Aggregation of Sulfidized

Particles. The surface charge of nanoparticles is an importantparameter controlling their fate in the environment. Electro-phoretic mobility (EPM) was measured as a function of pH(Figure S4 of the Supporting Information). The pHPZC of theinitial ZnO NPs was 9.1, consistent with a previously reportedvalue of 9.2.37 Any level of sulfidation decreased the pHPZCfrom 9.1 for the ZnO NPs to 7 ± 0.5 for the sulfidizedmaterials. This result is consistent with expectation for ZnS.38

Although interpretation of the EPM and zeta potential forparticles with such complex morphology is difficult, thisdecrease in pHPZC suggests that the incoherent shell of ZnSNPs formed on the ZnO core strongly influences the overallcharge on these particles, even at a low S/Zn ratio (S/Zn0.216).Sulfidation of the ZnO NPs increased their extent of

aggregation. This was obvious visually as the sulfidizedmaterials readily sedimented from suspension at pH 7, whereasthe pristine NPs were comparatively more stable. This behavioris consistent with a decrease in zeta potential and with anincrease in hydrodynamic diameter of the particles withincreasing degree of sulfidation (Figure S5 and Table S2 ofthe Supporting Information). The intensity-averaged DLS sizesincreased from 100 nm for the pristine ZnO NPs to 230 nm forthe most sulfidized NPs (Figure S5 of the SupportingInformation). There were no 2−5 nm peaks observed for thenumber-averaged DLS sizes, which would have been consistentwith the size of newly formed ZnS NPs. This finding isconsistent with the TEM images, which indicate that the ZnSNPs were predominantly attached to the ZnO NP cores.Implications for the Environment. Here, we demonstrate

that ZnO NPs readily transform to very small (2.5 to 5 nmcrystallite size) ZnS NPs when exposed to sulfide. This findingagrees with expectations from many environmentally relevantmedia (e.g., anoxic sediments) where the fate and solubility ofmetals such as Zn is controlled by available sulfide. Theproperties of the resulting ZnS/ZnO NPs have importantenvironmental implications. Sulfidation of the ZnO NPsdecreases their surface charge and promotes aggregationcompared to the pristine ZnO NPs. At higher degrees ofsulfidation, the solubility of the ZnO/ZnS NPs is decreased.These changes in the NP properties will affect the fate,

transport, and toxicity of ZnO NPs in natural systems.Aggregation can decrease the mobility of sulfidized ZnO NPs,and thus its bioavailability.38 The decreased dissolution mayalso decrease toxicity of ZnO NPs because less Zn2+ is released.The sulfidation of ZnO NPs observed here has also been

demonstrated for Ag NPs in the laboratory,16,18 the environ-ment,39 and WWTPs.40 However, the present study indicatesthat there are important differences in the mechanisms ofsulfidation and resulting sulfidized NPs for ZnO NPs comparedto Ag NPs. Ag NP sulfidation resulted in sulfidized Ag NPs ofroughly the same size and shape of the initial Ag NP oramorphous Ag2S at complete sulfidation.

16 The formation of aninsoluble Ag2S layer on the surface of the metallic Ag coredramatically decreases the rate of dissolution of Ag NPs, even atvery low S/Ag ratios. The lower solubility of the Ag2S-coatedAg NPs decreases their toxicity to sensitive organisms like E.coli or nitrifying bacteria.41 In contrast, sulfidation of ZnO NPsleads to the formation of a porous shell around the ZnO coreconsisting of 2.5 to 5 nm ZnS NPs. The different mechanismmay result from the higher rate of dissolution of ZnO NPscompared to Ag NPs, and because the Ag NPs need to beoxidized to Ag(I) species before they can dissolve. Thisoxidation step may in fact be the rate-limiting step for Ag NPsulfidation.18,42 As a consequence, sulfidation of ZnO NPs didnot inhibit dissolution nearly as much as for sulfidation of AgNPs (Figure 6). The ZnO/ZnS particles showed much higher

dissolution than Ag(0)/Ag2S NPs at same extent of sulfidation(% of metal sulfide obtained from LCF of the EXAFS spectra).Because the toxicity of ZnO NPs is strongly affected by therelease of Zn2+,6 the effect of sulfidation on ZnO toxicity toorganisms may be limited until high degrees of sulfidation areachieved.Sulfidation is very important transformation process for NPs

made from many types metal cations and needs to beconsidered in studies of the environmental fate of and effectson the NPs. We suggest that further investigation is needed tocompare the similarity and relevance of sulfidation of NPs madefrom metal cations with a high affinity for sulfide such as CuOand PbO. For ZnO sulfidation, the fate of the very small (2.5 to5 nm) ZnS NPs formed must be determined. Their very small

Figure 6. Comparison of dissolved percentage of sulfidized Ag NPsand ZnO NPs. Ag data are from Levard et al.16 Percent Zn dissolvedfor pristine ZnO NPs with S/Zn ratios of 0.432 and 1.08 werecalculated using dissolution data at a time point of 72 h. Dissolutionpercentages for S/Zn ratios of 0.616, 0.864, and 2.158 wereinterpolated between time points adjacent to 72 h, using either linearor polynomial fitting.

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342532

size suggests that they could be highly reactive in theenvironment. Stability diagrams for the Zn−S−P systemindicate that Zn3(PO4)2 should be a stable product underoxidizing conditions. The transformation of ZnS to Zn3(PO4)2was recently observed in biosolids upon successive wetting anddrying cycles at 37 C.19 While ZnS is a very stable form of Zn inreduced environments, a variety of other Zn species are stableand have been identified in biosolids including Zn3(PO4)2, Znbound to organic sulfur (e.g., thiols) and to minerals19 Hence,the environmental conditions (redox, pH, and concentration ofNOM/phosphate/sulfide/clay) that lead to predominance ofeach of these Zn species need further investigation.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional figures and details on experimental details ofsulfidation, TGA of ZnO NPs, DLS data on ZnO and sulfidizedparticles, zeta potential results, and linear combination fittingdetails are provided. This material is available free of charge viathe Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: glowry@cmu.edu.NotesThe authors declare no competing financial interest.#The Center for the Environmental Implications of Nano-Technology (CEINT) is headquartered at Duke University andhas collaborative efforts with researchers from other universitiesand agencies. The authors are not located at Duke University.

■ ACKNOWLEDGMENTSThis material is based on work supported by the US EPAScience to Achieve Results program (R834574), TransatlanticInitiative for Nanotechnology and the Environment (TINE),and the National Science Foundation (NSF and the Environ-mental Protection Agency (EPA) under NSF CooperativeAgreement EF-0830093, Center for the EnvironmentalImplications of Nanotechnology (CEINT). Any opinions,findings, conclusions or recommendations expressed in thismaterial are those of the author(s) and do not necessarilyreflect the views of the NSF or the EPA. This work has notbeen subjected to EPA review and no official endorsementshould be inferred. Portions of this research were carried out atthe Stanford Synchrotron Radiation Lightsource (SSRL), anational user facility operated by Stanford University on behalfof the U.S. Department of Energy, Office of Basic EnergySciences. Use of the Advanced Photon Source, a national userfacility operated for the U.S. Department of Energy (DOE)Office of Science by Argonne National Laboratory, wassupported by the U.S. DOE under Contract No. DE-AC02-06CH11357. We also thank the staff of APS (BL 11-ID-B) andSSRL (BL 4-3).

■ REFERENCES(1) Seow, Z. L. S.; Wong, A. S. W.; Thavasi, V.; Jose, R.;Ramakrishna, S.; Ho, G. W. Controlled synthesis and application ofZnO nanoparticles, nanorods and nanospheres in dye-sensitized solarcells. Nanotechnology.2009, 20 (4).(2) Bian, S. W.; Mudunkotuwa, I. A.; Rupasinghe, T.; Grassian, V. H.Aggregation and Dissolution of 4 nm ZnO Nanoparticles in AqueousEnvironments: Influence of pH, Ionic Strength, Size, and Adsorptionof Humic Acid. Langmuir. 2011, 27 (10), 6059−6068.

(3) Miller, R. J.; Lenihan, H. S.; Muller, E. B.; Tseng, N.; Hanna, S.K.; Keller, A. A. Impacts of metal oxide nanoparticles on marinephytoplankton. Environ. Sci. Technol. 2010, 44 (19), 7329−7334.(4) Brayner, R.; Ferrari-Iliou, R.; Brivois, N.; Djediat, S.; Benedetti,M. F.; Fievet, F. Toxicological impact studies based on Escherichia colibacteria in ultrafine ZnO nanoparticles colloidal medium. Nano Lett.2006, 6 (4), 866−870.(5) Jones, N.; Ray, B.; Ranjit, K. T.; Manna, A. C. Antibacterialactivity of ZnO nanoparticle suspensions on a broad spectrum ofmicroorganisms. Fems Microbiol Lett. 2008, 279 (1), 71−76.(6) Franklin, N. M.; Rogers, N. J.; Apte, S. C.; Batley, G. E.; Gadd, G.E.; Casey, P. S. Comparative toxicity of nanoparticulate ZnO, bulkZnO, and ZnCl2 to a freshwater microalga (Pseudokirchneriellasubcapitata): The importance of particle solubility. Environ. Sci.Technol. 2007, 41 (24), 8484−8490.(7) Poynton, H. C.; Lazorchak, J. M.; Impellitteri, C. A.; Smith, M.E.; Rogers, K.; Patra, M.; Hammer, K. A.; Allen, H. J.; Vulpe, C. D.Differential gene expression in Daphnia magna suggests distinct modesof action and bioavailability for ZnO nanoparticles and Zn ions.Environ. Sci. Technol. 2011, 45 (2), 762−768.(8) Canas, J. E.; Qi, B. B.; Li, S. B.; Maul, J. D.; Cox, S. B.; Das, S.;Green, M. J. Acute and reproductive toxicity of nano-sized metaloxides (ZnO and TiO2) to earthworms (Eisenia fetida). J EnvironMonitor. 2011, 13 (12), 3351−3357.(9) Muyssen, B. T. A.; Janssen, C. R. Importance of acclimation toenvironmentally relevant zinc concentrations on the sensitivity ofDaphnia Magna toward zinc. Environ. Toxicol. Chem. 2005, 24 (4),895−901.(10) Muyssen, B. T. A.; Janssen, C. R. Zinc acclimation and its effecton the zinc tolerance of Raphidocelis subcapitata and Chlorella vulgarisin laboratory experiments. Chemosphere 2001, 45 (4−5), 507−514.(11) Raghupathi, K. R.; Koodali, R. T.; Manna, A. C. Size-dependentbacterial growth inhibition and mechanism of antibacterial activity ofzinc oxide nanoparticles. Langmuir. 2011, 27 (7), 4020−4028.(12) Ma, R.; Levard, C.; Marinakos, S. M.; Cheng, Y. W.; Liu, J.;Michel, F. M.; Brown, G. E.; Lowry, G. V. Size-controlled dissolutionof organic-coated silver nanoparticles. Environ. Sci. Technol. 2012, 46(2), 752−759.(13) Mudunkotuwa, I. A.; Rupasinghe, T.; Wu, C. M.; Grassian, V. H.Dissolution of ZnO nanoparticles at circumneutral pH: A study of sizeeffects in the presence and absence of citric acid. Langmuir. 2012, 28(1), 396−403.(14) Lv, J. T.; Zhang, S. Z.; Luo, L.; Han, W.; Zhang, J.; Yang, K.;Christie, P. Dissolution and microstructural transformation of ZnOnanoparticles under the influence of phosphate. Environ. Sci. Technol.2012, 46 (13), 7215−7221.(15) Yang, K.; Lin, D. H.; Xing, B. S. Interactions of humic acid withnanosized inorganic oxides. Langmuir. 2009, 25 (6), 3571−3576.(16) Levard, C.; Reinsch, B. C.; Michel, F. M.; Oumahi, C.; Lowry,G. V.; Brown, G. E. Sulfidation processes of PVP-coated silvernanoparticles in aqueous solution: Impact on dissolution rate. Environ.Sci. Technol. 2011, 45 (12), 5260−5266.(17) Hydrogen Sulfide: Human Health Aspects; World HealthOrganization Geneva, 2003.(18) Liu, J. Y.; Pennell, K. G.; Hurt, R. H. Kinetics and mechanismsof nanosilver oxysulfidation. Environ. Sci. Technol. 2011, 45 (17),7345−7353.(19) Enzo Lombi, E. D.; Tavakkoli, Ehsan; Turney, Terence; Naidu,Ravi; Miller, Bradley W; Scheckel, Kirk G. Fate of zinc oxidenanoparticles during anaerobic digestion of wastewater and post-treatment processing of sewage sludge. Environ. Sci. Technol. 2012, 46(16), 9089−9096.(20) Donner, E.; Howard, D. L.; de Jonge, M. D.; Paterson, D.;Cheah, M. H.; Naidu, R.; Lombi, E. X-ray absorption and micro X-rayfluorescence spectroscopy investigation of copper and zinc speciationin biosolids. Environ. Sci. Technol. 2011, 45 (17), 7249−7257.(21) Priadi, C.; Le Pape, P.; Morin, G.; Ayrault, S.; Maillot, F.; Juillot,F.; Hochreutener, R.; Llorens, I.; Testemale, D.; Proux, O.; Brown, G.E. X-ray absorption fine structure evidence for amorphous zinc sulfide

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342533

as a major zinc species in suspended matter from the Seine riverdownstream of Paris, Ile-de-France, France. Environ. Sci. Technol. 2012,46 (7), 3712−3720.(22) Ghosh, S. K.; Kundu, S.; Mandal, M.; Nath, S.; Pal, T. Studieson the evolution of silver nanoparticles in micelle by UV-photo-activation. J Nanopart Res. 2003, 5 (5−6), 577−587.(23) Heaton, A. C. P.; Rugh, C. L.; Wang, N. J.; Meagher, R. B.Physiological responses of transgenic merA-tobacco (Nicotianatabacum) to foliar and root mercury exposure. Water Air Soil Poll.2005, 161 (1−4), 137−155.(24) Hu, Y.; Qian, H. H.; Liu, Y.; Du, G. H.; Zhang, F. M.; Wang, L.B.; Hu, X. A microwave-assisted rapid route to synthesize ZnO/ZnScore-shell nanostructures via controllable surface sulfidation of ZnOnanorods. Crystengcomm. 2011, 13 (10), 3438−3443.(25) She, Y. Y.; Yang, J.; Qiu, K. Q. Synthesis of ZnS nanoparticles bysolid-liquid chemical reaction with ZnO and Na2S under ultrasonic. TNonferr Metal Soc. 2010, 20, S211−S215.(26) Wu, D. P.; Xiao, B.; Liu, N.; Xiao, Y.; Jiang, K. One step fromZnO rod to ZnS porous tube. Mater Sci Eng B-Adv. 2010, 175 (3),195−200.(27) Woehrle, G. H.; Hutchison, J. E.; Ozkar, S.; Finke, R. G.Analysis of nanoparticle Transmission Electron Microscopy data usinga public-domain image-processing program, Image. Turk. J. Chem.2006, 30 (1), 1−13.(28) Webb, S. M. SIXpack: A graphical user interface for XASanalysis using IFEFFIT. Phys. Scr. 2005, T115, 1011−1014.(29) Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.;Hausermann, D. Two-dimensional detector software: From realdetector to idealised image or two-theta scan. High Pressure Res.1996, 14 (4−6), 235−248.(30) Birks, L. S.; Friedman, H. Particle size determination from X-rayline broadening. J. Appl. Phys. 1946, 17 (8), 687−691.(31) Wojdyr, M. Fityk: A general-purpose peak fitting program. J.Appl. Crystallogr. 2010, 43, 1126−1128.(32) Bahmani, B.; Moztarzadeh, F.; Rabiee, M. Synthesis of zincsulfide semiconductor nanoparticles by coprecipitation method forbiological diagnostics. J. Optoelectron Adv. M. 2007, 9 (11), 3336−3339.(33) Gilbert, B.; Huang, F.; Zhang, H. Z.; Waychunas, G. A.;Banfield, J. F. Nanoparticles: Strained and stiff. Science 2004, 305(5684), 651−654.(34) Milani, N.; McLaughlin, M. J.; Stacey, S. P.; Kirby, J. K.;Hettiarachchi, G. M.; Beak, D. G.; Cornelis, G. Dissolution kinetics ofmacronutrient fertilizers coated with manufactured zinc oxidenanoparticles. J. Agr. Food Chem. 2012, 60 (16), 3991−3998.(35) Reed, R. B.; Ladner, D. A.; Higgins, C. P.; Westerhoff, P.;Ranville, J. F. Solubility of nano-zinc oxide in environmentally andbiologically important matrices. Environ. Toxicol. Chem. 2012, 31 (1),93−99.(36) Moreau, J. W.; Webb, R. I.; Banfield, J. F. Ultrastructure,aggregation-state, and crystal growth of biogenic nanocrystallinesphalerite and wurtzite. Am. Mineral. 2004, 89 (7), 950−960.(37) Zhang, Y.; Chen, Y. S.; Westerhoff, P.; Hristovski, K.;Crittenden, J. C. Stability of commercial metal oxide nanoparticlesin water. Water Res. 2008, 42 (8−9), 2204−2212.(38) Deonarine, A.; Lau, B. L. T.; Aiken, G. R.; Ryan, J. N.; Hsu-Kim,H. Effects of Humic Substances on Precipitation and Aggregation ofZinc Sulfide Nanoparticles. Environ. Sci. Technol. 2011, 45 (8), 3217−3223.(39) Lowry, G. V.; Espinasse, B. P.; Badireddy, A. R.; Richardson, C.J.; Reinsch, B. C.; Bryant, L. D.; Bone, A. J.; Deonarine, A.; Chae, S.;Therezien, M.; Colman, B. P.; Hsu-Kim, H.; Bernhardt, E. S.; Matson,C. W.; Wiesner, M. R. Long-term transformation and fate ofmanufactured Ag nanoparticles in a simulated large scale freshwateremergent wetland. Environ. Sci. Technol. 2012, 46 (13), 7027−7036.(40) Kaegi, R.; Voegelin, A.; Sinnet, B.; Zuleeg, S.; Hagendorfer, H.;Burkhardt, M.; Siegrist, H. Behavior of metallic silver nanoparticles in apilot wastewater treatment plant. Environ. Sci. Technol. 2011, 45 (9),3902−3908.

(41) Reinsch, B. C.; Levard, C.; Li, Z.; Ma, R.; Wise, A.; Gregory, K.B.; Brown, G. E.; Lowry, G. V. Sulfidation of silver nanoparticlesdecreases Escherichia coli growth inhibition. Environ. Sci. Technol. 2012,46 (13), 6992−7000.(42) Liu, J. Y.; Hurt, R. H. Ion release kinetics and particlepersistence in aqueous nano-silver colloids. Environ. Sci. Technol. 2010,44 (6), 2169−2175.

Environmental Science & Technology Article

dx.doi.org/10.1021/es3035347 | Environ. Sci. Technol. 2013, 47, 2527−25342534