Article pubs.acs.org/est

Influence of Residual Polymer on Nanoparticle Deposition in Porous Media Yonggang Wang,† Matthew D. Becker,† Vicki L. Colvin,‡ Linda M. Abriola,† and Kurt D. Pennell*,† †

Department of Civil and Environmental Engineering, Tufts University, Medford, Massachusetts 02155, United States Department of Chemistry, Rice University, Houston, Texas 77005, United States

‡

S Supporting Information *

ABSTRACT: Although surface coatings and free polymers are known to affect the mobility of nanoparticles in water-saturated porous media, the influence of these compounds on nanoparticle deposition behavior has received limited attention. A series of column experiments was conducted to evaluate the transport and retention of quantum dots (QDs) coated with a synthetic polymer, polyacrylic acid-octylamine (PAA-OA). Initial column studies, conducted with three size fractions of Ottawa sand, resulted in unusual solid-phase retention profiles, characterized by low QD deposition near the column inlet and increasing solid-phase concentrations along the column until a plateau or limiting capacity was reached near the column midpoint. Mathematical modeling studies indicated that the observed retention behavior could not be reproduced using one-dimensional simulators based on either clean-bed filtration theory or a modified filtration theory (MFT) model that incorporated a maximum retention capacity. Additional column studies demonstrated that changes in the inlet end plate configuration designed to ensure uniform flow did not alter the observed effluent breakthrough curves (BTCs) or shape of the retention profile. Subsequent QD transport experiments, pretreated by flushing with a pulse of PAA-OA solution, resulted in almost complete QD breakthrough with minimal retention. It is postulated that free polymer was preferentially adsorbed onto the solid surface near the column inlet, thereby preventing QD attachment, whereas in the down-gradient portion of the column, QDs attached to the solid phase without competition from the polymer. These findings reveal the importance of accounting for the influence of coconstituents on nanoparticle deposition and demonstrate the need to simulate both transport and retention data when assessing nanoparticle mobility in porous media.

■

INTRODUCTION In their pure form most engineered nanomaterials are insoluble or sparingly soluble in aqueous suspensions. For example, the solubility of uncoated, nonaged fullerene (C60) nanoparticles in water is reported to be approximately 1 × 10−9 mg/L.1,2 Aqueous nanoparticle suspensions are also extremely sensitive to the presence of electrolytes, which serve to suppress the diffuse double layer and can lead to particle agglomeration or aggregation.3−5 To counteract these effects, nanomaterials are typically prepared in the presence of, or subsequently mixed with, electrostatic or polymeric stabilizing agents to create nanoparticle surface coatings and tune particle−particle and particle−solid interactions. For example, citrate-stabilized silver nanoparticles (nAg) have been prepared with silver nitrate as the precursor, sodium borohydride as the reducing agent, and trisodium citrate as the coating agent.6,7 Similarly, the core-shell of quantum dots (QDs) is hydrophobic and, thus, coating the external surface of the particles with amphiphilic polymers has been shown to greatly enhance their solubility in water.8 Such coatings often render QDs biocompatible, and when combined with their small size and tunable emission spectra, have led to numerous biomedical applications, including in vivo imaging and delivery of cancer therapies.9−11 © XXXX American Chemical Society

The coatings used to modify the surface properties of engineered nanomaterials also affect their mobility in natural and engineered porous media. For example, Wang et al.12 reported that when QDs were coated with polyacrylic acidoctylamine (PAA-OA) or linoleic acid (LA), more than 65 and 89%, respectively, of the introduced mass was transported through water-saturated columns packed with quartz sand. In contrast, complete retention (i.e., no recovery) was observed under identical conditions when QDs were coated with polyethylene glycol functionalized polymer (PEGP), which is widely used in medical applications. Nanoscale zerovalent iron (nZVI) particles have also been coated with polymers, including poly(styrenesulfonate) (PSS), to facilitate their stability and transport in soil columns and heterogeneous aquifer cells.13,14 In addition to surface coatings, the presence of synthetic polymers (e.g., surfactants) or natural organic matter (NOM) in solution has also been shown to alter nanoparticle stability Received: January 29, 2014 Revised: July 28, 2014 Accepted: August 18, 2014

A

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Table 1. Summary of Column Experiments Conducted To Measure QD Transport in Three Size Fractions of Ottawa Sand at Two Flow Rates, Effects of Endplate Conditions on Retention, and Transport Free Polymer (PAA-OA)a columnb

d50c (mm)

θwd

vpe (m/d)

PWf

Retg (%)

MBh (%)

katti (h−1)

Smaxj (pmol/g)

η0k

αl

CS MS-A MS-B FS-A FS-B MS-BR CS-GB FS-CNA FS-CNB CS-PAA

0.354 0.163 0.163 0.125 0.125 0.125 0.354 0.125 0.125 0.354

0.37 0.39 0.39 0.39 0.39 0.38 0.37 0.39 0.39 0.37

7.7 7.4 7.3 7.2 7.3 7.4 7.7 7.3 7.2 7.7

2.0 3.0 3.0 3.0 3.0 2.8 2.8 2.8 2.8 3.2

24.7 30.6 38.3 99.6 99.1 44.5 23.6 82.4 85.6 0.0

102.8 98.5 105.6 100.3 100.7 100.2 101.2 97.9 102.0 96.3

10.99 12.29 8.16 37.08 60.94 NAm NA NA NA NA

1.11 2.87 2.31 12.20 9.41 NA NA NA NA NA

0.098 0.146 0.146 0.174 0.174 NA NA NA NA NA

0.126 0.048 0.032 0.092 0.151 NA NA NA NA NA

a

All experiments were conducted with 3 mM NaCl as background electrolyte, adjusted to pH 7. For the QD transport experiments, effluent and retention profile data were fit simultaneously to the MFT-I model, which accounted for a first-order attachment kinetics and maximum retention capacity. bCS, coarse sand (40−50 mesh); MS, medium sand (80−100 mesh); FS, fine sand (100−140 mesh); GB, glass beads; CN, conical inlet end plate; BR, core boring; PAA, polyacrylic acid-octylamine copolymer; A and B indicate duplication. cMean grain diameter. dVolumetric water content. ePore water velocity. fPulse width. gRetention. hOverall mass balance. iQD attachment rate. jMaximum retention capacity. kSingle collector efficiency. lCollision efficiency factor. mNA, not applicable.

permeability at the column inlet. An additional column experiment was then performed to assess the transport and deposition behavior of QD-PAA-OA following pretreatment with a PAA-OA solution. Effluent concentration and retention profile data were simulated using a nanoparticle transport model that accounts for first-order attachment and detachment kinetics, as well as a maximum retention capacity.

and transport behavior in porous media. The addition of an ethoxylated nonionic surfactant greatly improved the stability of aqueous nC60 suspensions,15 and the presence of 1000 mg/L Tween 80, a food-grade nonionic surfactant, greatly enhanced nC60 transport through 40−50 mesh Ottawa sand.16 Wang et al.16 also reported that the addition of Suwannee River humic or fulvic acid (20 mg/L) to an aqueous suspension of fullerene nanoparticles (nC60) resulted in minimal particle attachment and effluent breakthrough curves that coincided with those of a nonreactive tracer. Similarly, Espinasse et al.17 found that the presence 1 mg/L tannic acid in an aqueous nC60 suspension decreased the attachment efficiency by a factor of 2−3. The observed enhancements in nanoparticle stability and mobility are generally attributed to carboxylic and phenolic moieties present in NOM, which can associate with suspended particles and solid surfaces (e.g., quartz), thereby increasing electrostatic repulsive forces. Although the ability of surface coatings and free polymers to enhance nanoparticle mobility in porous media is well documented, the influence of these compounds on nanoparticle deposition (attachment−detachment) behavior has received far less attention. Wang et al.16 reported that the presence of adsorbed-phase surfactant resulted in nC60 retention profiles that exhibited hyperexponential decay, with a corresponding effluent breakthrough curve that was consistent with filter ripening (i.e., increased attachment over time). Cationic surfactants, which strongly adsorb to negatively charged surfaces, have been shown to substantially increase nC60 retention in soil.18 Thus, the objective of this study was to investigate the influence of a surface coating polymer, polyacrylic acid-octylamine (PAA-OA), which may exist both in the free aqueous phase and as a surface-bound coating, on the deposition behavior of QDs in water-saturated quartz sand. Initial column studies, conducted using three size fractions of Ottawa sand, revealed that QDs prepared with PAA-OA as the coating agent exhibited unexpected retention profiles, where the solid-phase concentration was minimal near the column inlet and increased with travel distance until a plateau or limiting capacity was reached. To investigate potential physical (flow path) mechanisms contributing to this retention behavior, a second set of column experiments was conducted with modified end plate configurations and a layer of higher

■

MATERIALS AND METHODS Preparation of Nanoparticle Suspensions. Monodisperse CdSe/CdZnS nanocrystals with a diameter of 5.2 nm and a maximum emission wavelength of 600 nm were synthesized using the protocol of Zhu et al.8 Due to the hydrophobicity of the trioctylphosphine oxide (TOPO) layer on the outer shell,19 the QD surface was further modified with an amphiphilic polymer, polyacrylic acid-octylamine (PAA-OA), which was synthesized by combining polyacrylic acid (Mw = 1800 g/mol) with octylamine using 1-ethyl-3-(3-(dimethylamino)propyl) carbodiimide (EDC) as the cross-linker.20 To reduce the free polymer concentration, aqueous QD stock suspensions were concentrated three times by centrifugation (104667g for 4 h), followed by resuspension in deionized (DI) water. Column Transport Studies. Borosilicate glass columns (2.5 cm i.d. × 10 cm length) were packed with one of three size fractions quartz sand, 40−50, 80−100, or 100−140 mesh, which was sieved from bulk F-50 Ottawa sand (U.S. Silica, Berkeley Springs, WV, USA). Prior to use, the sand was cleaned using a sequential acid soaking, water rinse, sonication, and oven-drying procedure.21 Following dry packing in 1 cm lifts, each column was purged with CO2 gas for 15 min followed by flushing with at least 10 pore volumes (PVs) of degassed background electrolyte solution (3 mM NaCl at pH 7) to achieve complete water saturation.22 A pulse (2−3 PVs) of aqueous QD suspension (ca. 6 × 1015 particles/mL), prepared by diluting the QD stock 400-fold and conditioned to an ionic strength of 3 mM (NaCl) and pH 7, was introduced into the water-saturated column in an up-flow mode at a flow rate of 1.0 mL/min, which corresponds to a Darcy velocity of 2.8 m/d, respectively. The influent reservoir was then switched to the background electrolyte solution, which was injected at the same flow rate for 3 PVs to elute unattached QDs. Effluent samples were collected continuously, and measured aqueous-phase B

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

0.1−15 nM and 0.1−10 mg/L, were prepared for aqueous QDs and Cd by serial dilution of the QD stock solution and Cd standard (1000 mg/L, ULTRA Scientific, Kingstown, RI, USA), respectively. The method detection limits for aqueous- and solid-phase QDs were 0.06 nM and 0.1 pmol/g, respectively. The concentration of residual PAA-OA in the QD stock was determined using a total organic carbon analyzer (TOC-5000A, Shimadzu Scientific Instruments, Columbia, MD, USA) following separation by centrifugation at 5000 rpm for 30 min. The mean hydrodynamic diameter and zeta potential of QDs in aqueous suspension were characterized by dynamic light scattering (DLS) using a ZetaSizer Nano ZS analyzer (Malvern Instruments Ltd., Southborough, MA, USA) operated in noninvasive back scattering (NIBS) mode at an angle of 173°. Prior to use, the ZetaSizer was calibrated using monodispersed polystyrene spheres (Nanosphere Size Standards, Duke Scientific, Palo Alto, CA, USA) with a mean diameter of 97 ± 3 nm and a zeta potential transfer standard (Malvern Instruments Ltd.) with a mean zeta potential of −68 ± 6.8 mV. Approximately 1 mL of QD suspension (ca. 10 nM) was transferred into a disposable cuvette (Malvern Instruments Ltd.) and analyzed using a green laser at a wavelength of 532 nm. The mean hydrodynamic diameter and zeta potential of QDs in 3 mM NaCl at pH 7 were 30.8 ± 2.6 nm and −45.8 ± 1.8 mV, respectively. Mathematical Modeling. Assuming laminar flow conditions and the absence of particle−particle interactions, the transport and retention of nanoparticles in uniform watersaturated porous media can be described using a traditional mass balance equation that accounts for advection, hydrodynamic dispersion, and solid-phase deposition:

concentrations were plotted versus the number of dimensionless pore volumes introduced to construct effluent breakthrough curves (BTCs). At the conclusion of each experiment, the column was sectioned into eight increments and the corresponding solid-phase concentrations were plotted versus the distance from the column inlet to obtain a retention profile.21 In several experiments, solid-phase samples were collected in concentric rings to determine the radial distribution of nanoparticles at a particular location. To assess the potential effects of flow anomalies at the column inlet on QD retention behavior, four column experiments were conducted at the same flow rate (1 mL/ min) using various end plate configurations. In the first experiment (CS-GB), 1.5 cm of the column bed closest to the inlet was packed with 3 mm glass beads to achieve a highpermeability zone. In the following two experiments (FS-CNA and FS-CNB), a customized end plate was constructed using a milling machine to achieve a conical configuration (1.1 cm base radius × 1.0 cm height; Figure S1, Supporting Information). In the fourth experiment (MS-BR), each of the first three solid samples near the column inlet was split into two concentric samples consisting of a 1 cm diameter core and the surrounding material to directly assess the transverse distribution of retained QDs. One additional column experiment was performed to evaluate the influence of adsorbed polymer on QD transport and deposition (CS-PAA). The general experimental procedures described above were followed, with the exception that the column was flushed with approximately 3 PVs of PAA-OA solution (430 mg PAA-OA/L, 3 mM NaCl, and pH 7) prior to introducing the QD-PAA-OA suspension. Effluent concentrations of PAA were monitored throughout the experiments to examine PAA sorption−desorption in the absence and presence of QDs. In total, 10 separate column experiments were performed as summarized in Table 1. In selected experiments, a nonreactive tracer test was conducted prior to QD introduction to assess the hydrodynamic dispersion and flow properties of the packed column. For each tracer test, a 2−3 PV pulse of sodium bromide solution (3 mM NaBr) was introduced into the column at a flow rate of 1.0 mL/min. Immediately following the tracer injection, an additional 3 PVs of background solution were flushed through the column at the same flow rate to displace the tracer solution. The resulting effluent BTCs were fit to a one-dimensional (1-D) form of the advective−dispersive− reactive (ADR) transport equation using the CXTFIT model (ver. 2.1)23 to obtain hydrodynamic dispersivity values (see the Supporting Information). Analytical Methods. The concentration of QDs in the stock suspension was confirmed by UV absorbance at a wavelength of 532 nm.8,19 To measure solid-phase QD concentrations, samples were collected in 1.5 cm increments, dried at 95 °C, and digested in 10 mL of concentrated nitric acid (12.1 M) using a microwave-assisted digester (model Discover SP-D, CEM Corp., Matthews, NC, USA) operated at 190 °C and 230 psi. The amount of Cd in digested samples was quantified using inductively coupled plasma-optical emission spectrometry (ICP-OES; Optima7300DV, PerkinElmer, Inc., Shelton, CT, USA) operated at a RF power of 1500 W, nebulizer flow of 0.8 mL/min, and pump rate of 1.5 L/min at a wavelength of 214.44 nm. For aqueous samples, the concentration of QDs was determined by direction injection into the ICP-OES without pretreatment. Calibration curves, consisting of at least five points over concentration ranges of

ρ ∂S ∂C ∂C ∂ 2C + b = DH 2 − vp ∂t θw ∂t ∂x ∂x

(1) 3

Here, C and S are the aqueous (M/L ) and attached solidphase (M/M) concentrations, ρb is the solid-phase bulk density (M/L3), θw is the porosity of the solid phase (−), DH is the hydrodynamic dispersion coefficient (L2/t), and vp is the porewater velocity (L/t). The rate expression for QD deposition in quartz sand was modified from classical filtration theory24 to include a first-order irreversible attachment rate (katt) and a maximum solid-phase retention capacity (Smax):21 ρb ∂S θw ∂t Ψ=

= katt ΨC

(2)

Smax − S Smax

(3) 24

The expression for katt is given as katt =

3(1 − θw )vp 2d50

αη0

(4)

where η0 represents the theoretical single-collector efficiency frequency, which can be calculated using the equation developed by Tufenkji and Elimelech,25 d50 is the mean diameter of the porous medium, and α is the collision efficiency factor (i.e., fraction of collisions that result in particle attachment). The term Ψ represents the fraction of sites available for particle attachment. As the sites available for attachment are filled, fewer collisions occur and Ψ approaches zero, resulting in less attachment to the solid phase. When Smax C

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 1. Measured and simulated effluent breakthrough curves (A) and retention profiles (B) obtained for QD-PAA-OA transport in three size fractions (40−50, 80−100, and 100−140 mesh) of Ottawa sand at a flow rate of 1.0 mL/min. Column experiments were conducted at an ionic strength of 3 mM and pH 7. Effluent BTC and retention profile data were simultaneously fit to the MFT-1 model, which accounted for a first-order attachment kinetics and maximum retention capacity.

is much greater than S, Ψ approaches unity and eq 2 converges to the clean-bed filtration model.24 For the simulations presented herein, inlet and outlet boundary conditions were implemented as third and second types, respectively,26 and the initial QD concentrations in resident aqueous and solid phases were assumed to be zero. Equations 1−3 were approximated using an implicit-in-time and central-in-space finite difference scheme that was implemented in MATLAB R2010a (The MathWorks, Natick, MA, USA).27 Model fitting to experimental observations was accomplished through implementation of a nonlinear least-squares optimization routine provided by MATLAB R2010a.

decreased sharply without significant tailing, consistent with irreversible attachment (i.e., absence of particle detachment and re-entrainment), which has been observed for several other nanoparticle-porous media systems.21,22 At the conclusion of each transport experiment, the columns were destructively sectioned into discrete increments and the solid phase was analyzed to quantify the spatial distribution of QDs along the length of the column and to obtain the total mass balance (i.e., retained plus eluted mass versus the total mass injected). As shown in Table 1, the total measured QD mass balance for all experiments ranged from 98 to 106% of the introduced mass, demonstrating the ability of experimental procedures to achieve mass balance closure. Consistent with the general trends observed in QD effluent BTCs, the retention of QDs increased with decreasing sand grain size. The most interesting aspect of the measured QD deposition was the unusual shape of the retention profiles, which consistently exhibited lower QD retention near the column inlet (Figure 1B and S2B). Regardless of sand grain size, solidphase QD concentrations increased with increasing distance from the column inlet until a plateau was reached near the midpoint of the column. Such retention behavior is not consistent with typical colloid and nanoparticle retention profiles, which are characterized by elevated retention at the column inlet or injection point followed by a gradual or sharp decline with travel distance.16,29,30 However, the observed plateau in the measured QD solid-phase concentration profiles is consistent with a limiting or maximum retention capacity.21,22 Mathematical Modeling of QD Mobility. Transport models, based on clean-bed filtration theory (CFT), have been widely used to simulate nanoparticle mobility in water-saturated porous media.17,31,32 However, the observed asymmetric effluent BTCs and retention that increases with travel distance (Figure 1) are not reproducible with CFT-based models. For example, implementation of a CFT model to fit the CS data set yielded a symmetric BTC that underestimated the maximum effluent concentration and produced an exponentially decaying retention profile (see Figure S3). Therefore, the nanoparticle transport model, based on retention-capacity-modified filtration

■

RESULTS AND DISCUSSION QD Transport and Deposition. Effluent BTCs and retention profiles obtained for PAA-OA-coated QD transport in three size fractions of Ottawa sand at a flow rate of 1.0 mL/ min (pore-water velocity ≈ 7.4 m/day) are shown in Figure 1. The results of replicate experiments, conducted in 80−100 mesh (d50 = 163 μm) and 100−140 mesh (d50 = 125 μm) Ottawa sand under identical conditions are presented in Figure S2 (Supporting Information) to demonstrate the reproducibility of the experimental system. In these experiments, QD retention ranged from 25% of the injected mass for the coarsest sand (d50 = 354 μm) to >99% for the finest sand (d50 = 125 μm) (Table 1). Consistent with the retention data, the time of the initial appearance of QDs in the column effluent increased from approximately 1.1 to 3.5 PVs, and the maximum relative concentration (C/C0) decreased from 1.0 to nearly 0.1 as the mean diameter of the sand decreased from 354 to 125 μm. Under similar experimental conditions, Torkzaban et al.28 found that the initial breakthrough time of functionalized QDs increased from ca. 2.2 to 6.2 PVs when the mean diameter of the sand was decreased from 270 to 135 μm. In all experiments, the relative effluent QD concentration exhibited a gradual ascent to a maximum value, which was more pronounced in the finer size fractions of Ottawa sand. This finding suggests that QD deposition was governed by a ratelimited attachment process. The distal portion of QD BTCs D

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 2. Effluent breakthrough curve (A) and retention profile (B) obtained for QD-PAA-OA transport in 80−100 mesh Ottawa sand at an ionic strength of 3 mM, pH 7, and flow rate of 1.0 mL/min. Comparisons between retained QD concentrations in a 1 cm diameter core and the surrounding 1.5 cm annulus (C) are shown at three locations near the column inlet.

model was able to capture effluent BTC data for a range of media, but failed to simulate measured QD retention profiles. These findings clearly indicate that nanoparticle modeling efforts based solely on BTC data, a common practice in nanoparticle transport studies,28,31,33 can result in inaccurate representation and interpretation of deposition behavior in both space and time. Furthermore, the unusual shape of the QD retention profiles suggests that additional processes or mechanisms, not accounted for in the current version of the MFT model, influenced QD deposition near the column inlet. Interpretation of Measured Retention Profiles. A number of potential factors and processes could contribute to the observed reduction in QD retention near the column inlet, including rate-limited detachment, flow field variations, surface heterogeneity, and the presence of residual polymer. One possible explanation for the observed retention profiles is the detachment of deposited QDs (i.e., reversible particle attachment) during the 3 PV injection of QD-free background solution following QD pulse injection. To evaluate this possibility, the solid-phase expression (eq 2) in the MFT model was reformulated to include a first-order detachment rate term (kdet):

theory (MFT) (eqs 1−4), was used to simulate the measured QD transport and retention data.12,28,33 Initially, the MFT model was simultaneously fit to the measured effluent BTC and retention data, and the rate of detachment (kdet) was set to zero, consistent with irreversible attachment (MFT-I; Table 1 and Figure 1). The MFT model accurately captured the gradual ascent, concentration plateau, and sharp decrease in relative effluent QD concentrations at the end of pulse injection. The fitted attachment rates (katt) ranged from 8.16 to 60.94 1/h, and maximum retention capacities (Smax) ranged from 1.11 to 12.20 pmol/g. In general, the value of these parameters increased with decreasing sand grain size, consistent with results reported in a prior QD mobility study.28 Despite the ability of the model to capture the general effluent concentration behavior of QD nanoparticles, the retention profile trends were poorly represented in all cases; more specifically, the model was unable to capture the trend of increasing retention with distance near the column inlet. In addition, similar fitted parameter values (katt, Smax) were obtained when the MFT model (MFT-II) was only fit to the effluent BTCs (Table S1 and Figure S4), further demonstrating the inability of the MFT model to capture the observed retention profile data. The comparisons between model simulations and measured data demonstrate the importance of including retention profile data and performing a complete experimental mass balance when attempting to interpret nanoparticle fate and transport behavior in porous media. In this particular case, the MFT

ρb ∂S θw ∂t

= katt ΨC −

ρb θw

kdetS

(5)

Incorporation of nanoparticle detachment into the MFT model (MFT-DET) resulted in a slight improvement in simulated E

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 3. Effluent breakthrough curves (A) and retention profiles (B) obtained for QD-PAA-OA transport in 80−100 mesh Ottawa sand at an ionic strength of 3 mM, pH 7, and flow rate of 1.0 mL/min. The column end plates were modified to include a highly permeable zone of glass beads (circles) or a customized conical end plate (diamonds).

point where the 1/8-in. tubing connects with the 2.5 cm diameter column. To determine if QD deposition was nonuniform near the inlet, solid-phase QD concentrations were measured radially from the column center. This QD transport experiment (MS-BR) yielded an effluent BTC and retention profile (Figure 2) that were nearly identical to those obtained previously under the same conditions (80−100 mesh Ottawa sand at 1.0 mL/min). At three locations along the column axis, concentrations of retained QDs were nearly identical in the center of the column and within the surrounding annulus (Figure 2). To further explore potential effects of end plate configuration on QD retention, a set of duplicate column experiments (FS-CNA and FS-CNB) was conducted with a custom-milled conical end plate (Figure S1). An additional column experiment (CS-GB) was conducted with a packed zone of high-permeability glass beads (3 mm diameter) in the first 1.5 cm of the column, followed by 40−50 mesh Ottawa sand. As shown in Figure 3, regardless of these modifications to the end plate configuration, solid-phase concentrations of QDs increased with distance from the column inlet, consistent with the results obtained previously (Figure 1). As noted in the Introduction, QDs are typically prepared or coated with amphiphilic polymers to enhance their aqueous solubility and to improve their compatibility with biological systems. A re-examination of the QD-PAA-OA nanoparticles revealed that the copolymer PAA-OA was not grafted (i.e., covalent bond) onto the CdSe/CdZnS core-shell complex. Furthermore, although excess polymer remaining in the stock suspension was removed through repeated centrifugation steps, it is likely that some residual polymer remained in the stock suspension following the purification process. An equilibrium condition would then develop between the mass of PAA-OA in the free aqueous phase and that on QD surfaces. Thus, the influent suspension introduced to the packed columns would contain both PAA-OA-coated QDs and free PAA-OA. Polymers, existing in either the aqueous suspension or associated with the solid phase, have been shown to enhance or reduce nanoparticle mobility in porous media.12,14,16,37,38 Thus, we hypothesized that the anomalous retention profiles could be attributed to the adsorption of free PAA-OA onto the

trends in QD retention near the column inlet, in comparison with MFT-I model results, and also produced substantial tailing in BTCs (Figure S5). The fitted values of the QD detachment rate (kdet) for three size fractions of sand ranged from 0.01 to 0.58 1/h, whereas fitted values of the attachment rate (katt) and maximum retention capacity (Smax) were similar to those obtained with the MFT-I model (Table S1). As a consequence, the fitted katt values were at least 20 times greater than corresponding kdet values obtained with the MFT-DET model, indicating the relatively small role of detachment. Taken in concert with the absence of tailing in the observed BTCs, these findings indicate that QD detachment did not contribute to the observed reductions in QD retention near the column inlet. Additionally, this analysis demonstrates the value of using mathematical models to evaluate potential contributions of processes that influence nanoparticle transport and deposition behavior in porous media. A second possible explanation for the unusual retention profiles is the presence of variable flow conditions within the column, particularly near the column inlet. Effluent BTCs obtained for a nonreactive tracer (Br−) (Figure S6) were symmetric in shape, with no evidence of tailing. In addition, fitted values of the retardation factor (RF) and hydrodynamic dispersivity (α) were 1.02 and 0.068 cm, respectively, which are in close agreement with prior results.21 These findings, along with overall mass balances for all experiments (Table 1), indicate that QD transport and deposition in the watersaturated columns were not influenced by physical nonequilibrium processes, such as rate-limited mass transfer into regions of immobile water or preferential flow paths. However, several studies suggest that local flow velocities in 1-D column systems could be heterogeneous across the column diameter.34,35 Such nonuniform flow could affect nanoparticle deposition, which has been shown to depend upon flow velocity.22 In a recent study,36 faster movement of nanoparticles in the center of a column compared to the surrounding domain was reported, which was attributed to radially decreasing flow velocities from the center to the outer edge of the column. Thus, we hypothesized that the observed reduction in QD retention near the column inlet could be due to nonuniform particle deposition resulting from locally elevated flow near the F

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 4. Effluent breakthrough curves (A) and retention profiles (B) obtained for QD-PAA-OA transport through 40−50 mesh (circles) after preflushing the column with a solution containing the free polymer, PAA-OA.

quartz sand near the column inlet, effectively “blocking” sites that would have been available for nanoparticle attachment. To evaluate the effects of adsorbed PAA-OA on QD transport and retention, an additional column experiment was conducted in a column packed with 40−50 mesh (CS-PAA) Ottawa sand. In this experiment, the column was flushed with 3 PVs of PAA-OA solution (ca. 430 mg/L), followed by 3 PVs of background solution to remove free PAA-OA and, finally, a 3 PV pulse of QD-PAA-OA suspension. The resulting QD effluent BTC and retention profile (Figure 4) clearly demonstrate that adsorbed PAA-OA acts to block the solid-phase attachment sites, resulting in nearly complete QD breakthrough (96% mass recovery in effluent) and negligible QD attachment (Table 1). In a separate transport experiment in which PAA-OA was injected at an influent concentration of 50 mg/L, the average adsorption of PAA-OA on 40−50 mesh Ottawa sand was 0.40 μg/g. Because the levels of residual PAA-OA were much lower in the QD influent suspensions (0.9 mg/L), the observed blocking effect of the polymer in the original QD transport experiments (Figure 1) would be expected to be most prominent near the column inlet, as was observed. The experimental column results therefore support the hypothesis that residual PAA-OA polymer preferentially adsorbed onto the solid surface near the column inlet, thereby preventing QD attachment, whereas in the distal portion of the column QDs were able to attach to the solid phase without competition from the polymer. These findings demonstrate the importance of accounting for the influence of residual stabilizing agents on nanoparticle fate and transport in the environment.

■

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. Huiguang Zhu for preparing PAA-OA-coated quantum dots. This work was supported by the Advanced Energy Consortium (http://www.beg.utexas.edu/aec) under Projects BEG08-11 and BEG13-01. Member companies include BP America Inc., BG Group, Petrobras, Repsol, Schlumberger, Statoil, Shell, and Total.

■

(1) Andrievsky, G. V.; Kosevich, M. V.; Vovk, M.; Shelkovsky, V. S.; Vashchenko, L. A. On the production of an aqueous colloidal solution of fullerenes. J. Chem. Soc., Chem. Commun. 1995, 12, 1281−1282. (2) Deguchi, S.; Alargova, R. G.; Tsujii, K. Stable dispersions of fullerenes, C60 and C70, in water. Preparation and characterization. Langmuir 2001, 17, 6013−6017. (3) Chen, K. L.; Elimelech, M. Aggregation and deposition kinetics of fullerene (C-60) nanoparticles. Langmuir 2006, 22, 10994−11001. (4) Fortner, J.; Lyon, D.; Sayes, C.; Boyd, A.; Falkner, J.; Hotze, E.; Alemany, L.; Tao, Y.; Guo, W.; Ausman, K. C60 in water: nanocrystal formation and microbial response. Environ. Sci. Technol. 2005, 39, 4307−4316. (5) Min, Y. J.; Akbulut, M.; Kristiansen, K.; Golan, Y.; Israelachvili, J. The role of interparticle and external forces in nanoparticle assembly. Nat. Mater. 2008, 7, 527−538. (6) Liu, J.; Hurt, R. H. Ion release kinetics and particle persistence in aqueous nano-silver colloids. Environ. Sci. Technol. 2010, 44, 2169− 2175. (7) Mittelman, A. M.; Taghavy, A.; Wang, Y.; Abriola, L. M.; Pennell, K. D. Influence of dissolved oxygen on silver nanoparticle mobility and dissolution in water-saturated quartz sand. J. Nanoparticle Res. 2013, 15, 1−13. (8) Zhu, H. G.; Prakash, A.; Benoit, D. N.; Jones, C. J.; Colvin, V. L. Low temperature synthesis of ZnS and CdZnS shells on CdSe quantum dots. Nanotechnology 2010, 21, 255604−255614. (9) Gao, X.; Cui, Y.; Levenson, R. M.; Chung, L. W. K.; Nie, S. In vivo cancer targeting and imaging with semiconductor quantum dots. Nat. Biotechnol. 2004, 22, 969−976. (10) Huang, H.-C.; Barua, S.; Sharma, G.; Dey, S. K.; Rege, K. Inorganic nanoparticles for cancer imaging and therapy. J. Controlled Release 2011, 155, 344−357.

ASSOCIATED CONTENT

* Supporting Information S

Additional data and descriptions of nonreactive tracer test. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*(K.D.P.) Phone: (617) 627-3099; fax: (617) 627-3994; email: [email protected]. G

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(11) Xue, B.; Deng, D. W.; Cao, J.; Liu, F.; Li, X.; Akers, W.; Achilefu, S.; Gu, Y. Q. Synthesis of NAC capped near infrared-emitting CdTeS alloyed quantum dots and application for in vivo early tumor imaging. Dalton Trans. 2012, 41, 4935−4947. (12) Wang, Y.; Zhu, H.; Becker, M. D.; Englehart, J.; Abriola, L. M.; Colvin, V. L.; Pennell, K. D. Effect of surface coating composition on quantum dot mobility in porous media. J. Nanoparticle Res. 2013, 15, 1805−1821. (13) Phenrat, T.; Kim, H.-J.; Fagerlund, F.; Illangasekare, T.; Tilton, R. D.; Lowry, G. V. Particle size distribution, concentration, and magnetic attraction affect transport of polymer-modified Fe(0) nanoparticles in sand columns. Environ. Sci. Technol. 2009, 43, 5079−5085. (14) Phenrat, T.; Song, J. E.; Cisneros, C. M.; Schoenfelder, D. P.; Tilton, R. D.; Lowry, G. V. Estimating attachment of nano- and submicrometer-particles coated with organic macromolecules in porous media: development of an empirical model. Environ. Sci. Technol. 2010, 44, 4531−4538. (15) Bensasson, R. V.; Bienvenue, E.; Dellinger, M.; Leach, S.; Seta, P. C60 in model biological-systems − a visible-UV absorption study of solvent-dependent parameters and solute aggregation. J. Phys. Chem. 1994, 98, 3492−3500. (16) Wang, Y.; Li, Y.; Costanza, J.; Abriola, L. M.; Pennell, K. D. Enhanced mobility of fullerene (C(60)) nanoparticles in the presence of stabilizing agents. Environ. Sci. Technol. 2012, 46, 11761−11769. (17) Espinasse, B.; Hotze, E. M.; Wiesner, M. R. Transport and retention of colloidal aggregates of C60 in porous media: effects of organic macromolecules, ionic composition, and preparation method. Environ. Sci. Technol. 2007, 41, 7396−7402. (18) Wang, L. L.; Huang, Y.; Kan, A. T.; Tomson, M. B.; Chen, W. Enhanced transport of 2,2′,5,5′-polychlorinated biphenyl by natural organic matter (NOM) and surfactant-modified fullerene nanoparticles (nC(60)). Environ. Sci. Technol. 2012, 46, 5422−5429. (19) Yu, W. W.; Chang, E.; Falkner, J. C.; Zhang, J. Y.; Al-Somali, A. M.; Sayes, C. M.; Johns, J.; Drezek, R.; Colvin, V. L. Forming biocompatible and nonaggregated nanocrystals in water using amphiphilic polymers. J. Am. Chem. Soc. 2007, 129, 2871−2879. (20) Wu, X. Y.; Liu, H. J.; Liu, J. Q.; Haley, K. N.; Treadway, J. A.; Larson, J. P.; Ge, N. F.; Peale, F.; Bruchez, M. P. Immunofluorescent labeling of cancer marker Her2 and other cellular targets with semiconductor quantum dots. Nat. Biotechnol. 2003, 21, 41−46. (21) Wang, Y. G.; Li, Y. S.; Fortner, J. D.; Hughes, J. B.; Abriola, L. M.; Pennell, K. D. Transport and retention of nanoscale C-60 aggregates in water-saturated porous media. Environ. Sci. Technol. 2008, 42, 3588−3594. (22) Li, Y. S.; Wang, Y. G.; Pennell, K. D.; Abriola, L. M. Investigation of the transport and deposition of fullerene (C60) nanoparticles in quartz sands under varying flow conditions. Environ. Sci. Technol. 2008, 42, 7174−7180. (23) Toride, N.; Leij, F. J.; van Genuchten, M. T. The CXTFIT Code for Estimating Transport Parameters from Laboratory or Field Tracer Experiments, version 2.1; Research Report 137; U.S. Department of Agriculture: Riverside, CA, USA, 1999. (24) Yao, K. M.; Habibian, M. M.; Omelia, C. R. Water and waste water filtration − concepts and applications. Environ. Sci. Technol. 1971, 5, 1105−1112. (25) Tufenkji, N.; Elimelech, M. Correlation equation for predicting single-collector efficiency in physicochemical filtration in saturated porous media. Environ. Sci. Technol. 2004, 38, 529−536. (26) Simunek, J.; Genuchten, M. T. V.; Sejna, M. The Hydrus-1d Software Package for Simulating the One-Dimensional Movement of Water, Heat, and Multiple Solutes in Variably-Saturated Media; U.S. Department of Agriculture: Riverside, CA, USA, 2005. (27) Li, Y.; Wang, Y.; Pennell, K. D.; Abriola, L. M. Investigation of the transport and deposition of fullerene (C60) nanoparticles in quartz sands under varying flow conditions. Environ. Sci. Technol. 2008, 42, 7174−7180.

(28) Torkzaban, S.; Kim, Y.; Mulvihill, M.; Wan, J. M.; Tokunaga, T. K. Transport and deposition of functionalized CdTe nanoparticles in saturated porous media. J. Contam. Hydrol. 2010, 118, 208−217. (29) Liang, Y.; Bradford, S. A.; Simunek, J.; Vereecken, H.; Klumpp, E. Sensitivity of the transport and retention of stabilized silver nanoparticles to physicochemical factors. Water Res. 2013, 47, 2572− 2582. (30) Bradford, S. A.; Torkzaban, S.; Kim, H.; Simunek, J. Modeling colloid and microorganism transport and release with transients in solution ionic strength. Water Resour. Res. 2012, 48, W09509, DOI: 10.1029/2012WR012468. (31) Quevedo, I. R.; Tufenkji, N. Mobility of functionalized quantum dots and a model polystyrene nanoparticle in saturated quartz sand and loamy sand. Environ. Sci. Technol. 2012, 46, 4449−4457. (32) Lecoanet, H. F.; Bottero, J.-Y.; Wiesner, M. R. Laboratory assessment of the mobility of nanomaterials in porous media. Environ. Sci. Technol. 2004, 38, 5164−5169. (33) Torkzaban, S.; Wan, J.; Tokunaga, T. K.; Bradfor, S. A. Impacts of bridging complexation on the transport of surface-modified nanoparticles in saturated sand. J. Contam. Hydrol. 2012, 136−137, 86−95. (34) Greiner, A.; Schreiber, W.; Brix, G.; Kinzelbach, W. Magnetic resonance imaging of paramagnetic tracers in porous media: quantification of flow and transport parameters. Water Resour. Res. 1997, 33, 1461−1473. (35) Paulsen, J. L.; Donaldson, M. H.; Betancourt, S. S.; Song, Y. Q. Quantitative measurements of injections into porous media with contrast based MRI. J. Magn. Reson. 2011, 212, 133−138. (36) Shang, J. Y.; Liu, C. X.; Wang, Z. M.; Wu, H.; Zhu, K. K.; Li, J. A.; Liu, J. In-situ measurements of engineered nanoporous particle transport in saturated porous media. Environ. Sci. Technol. 2010, 44, 8190−8195. (37) Song, J. E.; Phenrat, T.; Marinakos, S.; Xiao, Y.; Liu, J.; Wiesner, M. R.; Tilton, R. D.; Lowry, G. V. Hydrophobic interactions increase attachment of gum arabic- and PVP-coated Ag nanoparticles to hydrophobic surfaces. Environ. Sci. Technol. 2011, 45, 5988−5995. (38) Lerner, R. N.; Lu, Q. Y.; Zeng, H. B.; Liu, Y. The effects of biofilm on the transport of stabilized zerovalent iron nanoparticles in saturated porous media. Water Res. 2012, 46, 975−985.

H

dx.doi.org/10.1021/es500523p | Environ. Sci. Technol. XXXX, XXX, XXX−XXX