Transport and Retention of Colloidal Aggregates of C60 in Porous Media

Oct 6, 2007 - The physical−chemical behavior of the fullerene C60 in environmental and physiological media is of interest for understanding the pote...
0 downloads 10 Views 162KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 7396-7402

Transport and Retention of Colloidal Aggregates of C60 in Porous Media: Effects of Organic Macromolecules, Ionic Composition, and Preparation Method BENJAMIN ESPINASSE, ERNEST M. HOTZE, AND MARK R. WIESNER* Duke University, Department of Civil and Environmental Engineering, PO Box 90287, Durham, North Carolina 27708-0287

The physical-chemical behavior of the fullerene C60 in environmental and physiological media is of interest for understanding the potential transport, exposure, and impacts of these materials on organisms and ecosystems. We consider the role of electrolyte composition and concentration, the effect of organic macromolecules, and the mode of preparation of colloidal aggregates of C60 (nC60) on the deposition of these colloids in a porous medium such as a groundwater aquifer or a water treatment filter. Results for nC60 deposition are qualitatively consistent with trends anticipated by theory. Deposition was found to increase with increasing ionic strength, the presence of polysaccharide-type organic matter, and lower Darcy velocities. Factors that will tend to decrease the retention of these materials in porous media include a low ionic strength and the presence of humic-like substances, while the ionic strengths typical of many natural waters and the presence polysaccharide-based natural organic matter, as may be produced by algae or bacteria, will tend to favor deposition and reduced potential for exposure. Variability in the method of preparing colloidal aggregates of fullerenes was observed to yield significant differences in nC60 properties and transport behavior.

Introduction The behavior of the fullerene C60 in environmental and physiological fluids is of interest because of the wide range of projected uses and processing conditions that may involve the introduction of these materials into aqueous solutions. The physical-chemical properties of C60 that affect fullerene transport and interactions with organisms are also of interest in understanding therapeutic applications and potential risks to ecosystems and human health. Although the solubility of the C60 molecule in water is negligible, there are a number of known pathways that may produce stable colloidal dispersions of C60, each generally yielding aggregates that differ with respect to size, surface chemistry, and structure (1). Colloidal aggregates of C60 (referred to as nC60, where n is typically observed to be in a range of 102-106 molecules of C60 per aggregate) may be formed by first dissolving C60 in an organic solvent, mixing the resultant solution with water, * Corresponding author e-mail: [email protected]. 7396

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007

and then evaporating the organic solvent to yield an aqueous nC60 suspension (2-4). Alternatively, colloidal suspensions can be formed by long-duration stirring of C60 in water or by sonication of C60 in water (5). Some studies have suggested that nC60 derived using the organic solvent tetrahydrofuran (THF) may be toxic (6-8). The extent and mechanism of toxicity, as well as the role of the organic solvent in the production of this toxicity, remain unclear. Aggregation and deposition of nC60 may reduce the effective persistence of and exposure to C60 in aquatic environments (9, 10). The ability to predict nanomaterial exposure and toxicity as a function of the physicochemical properties of the nanomaterial would be a powerful tool in risk assessment for these materials. However, such predictions are likely to be complicated by interactions between nanomaterials and naturally occurring materials such as salts, proteins, and macromolecules that may be present in environmental and physiological systems. We have previously considered the transport of one variety of nC60 and two other fullerenes in a laboratory porous medium using a background ionic strength of 10-2 M NaCl (10, 11). In this paper, we consider the role of electrolyte composition and concentration, the effect of organic macromolecules, and the mode of nC60 preparation on the deposition of these colloids in a model porous medium. Transport in such a system is relevant to understanding the potential mobility of this material in a natural system such as a groundwater aquifer or an engineered system such as a water treatment filter. We obtain results for nC60 deposition that are qualitatively consistent with trends anticipated by classical DLVO theory and macromolecular stabilization and destabilization.

Materials and Methods Column experiments were performed to evaluate the retention of suspensions of fullerenes in porous media. Reproducibility of the method was evaluated, followed by systematic variation in the solution chemistry of the fullerene suspensions, to evaluate changes in ionic strength and composition and the presence of two varieties of organic macromolecules. Because there are no standard procedures for the production of colloidal suspensions of fullerenes, one element of this effort involved a comparison of the transport properties of different fullerene suspensions produced in this work and with those used in previous studies. Column Preparation and Procedures. Particle mobility was assessed using a well-defined porous medium in a packed column. Spherical silicate glass beads (Potters Industries Inc. Berwyn, PA), with a mean diameter of 360 µm were used as the porous medium. As in earlier studies (9-11), flow was introduced to a 10 cm long cylindrical column, 2.65 cm in diameter. The compacted medium had an effective porosity of approximately 0.36, and the total volume of the system between the point of injection and the effluent detector was 35 mL. In contrast with previous studies, the flow was introduced in an upflow configuration. A procedure for cleaning glass beads (see Supporting Information) was adopted that was found to reduce the presence of extraneous materials initially present on the glass bead surface, relative to that in previous methods (10, 12-14), as viewed by scanning electron microscopy. Aqueous suspensions of fullerenes were pumped through the packed bed of silicate beads at a velocity of either 12 × 10-4 (fast) or 4 × 10-4 m s-1 (slow) using a magnetic-drive gear pump (Ismatec variable-speed pump drive and micropump head N-07002-27, Cole-Parmer Instrument Company, 10.1021/es0708767 CCC: $37.00

 2007 American Chemical Society Published on Web 10/06/2007

Chicago, IL). These values of Darcy or approach velocity were selected to allow comparison with data from a previous study of fullerene deposition in porous media11. A 0.2 µm filter was placed after the pump to remove any impurities from the bulk of the suspension and from the pump elements. Flow rates were measured by rotameter and verified by weighing the mass of the permeate over time on a digital balance. Suspensions of fullerene aggregates were injected into a carrier-feed solution that included any solutes introduced to the column (salts and organic macromolecules) at a ratio of injection flow to carrier-feed flow of 0.18. Concentrations of the carrier solution were adjusted to achieve the desired final concentrations of ionic strength or organic macromolecules at the inlet to the column. Column tests were conducted without buffering pH, which was continuously monitored during the experiments and ranged from 6.5 to 7.5, except in experiments where organic macromolecules were added and NaHCO3 was used to buffer the feed solution to a value of 7.2. The concentration of fullerenes in the column effluent was monitored continuously by UV absorbance at a wavelength of 269 nm (Hitachi 2810 UV-vis spectrophotometer equipped with a temperature controller and autosipper), following calibration in each case to a standard curve as determined by measurements of total organic carbon (Shimadzu 5050A). Variations in each chemical condition were reproduced at two different flow rates. The effect of ionic strength on THF/nC60 and TTA/nC60 retention was investigated over a range of 0.01-0.6 M using NaCl, while the effect of composition was explored for the TTA/nC60 suspensions using a matrix of electrolyte solutions, where the cation or anion was substituted (e.g., CaCl2, NaNO3, etc.) at a given ionic strength (0.01 or 0.1 M). The role of organic matter on fullerol and THF/nC60 retention was evaluated using two classes of naturally occurring macromolecules, tannic acid (average molecular weight 1700 g mol-1, Sigma Aldrich, St. Louis, MO), and a polysaccharide mixture (molecular weight (15) ranging from 10 000 to 250 000 g mol-1 with an average molecular weight of 49 000 g mol-1) of algal origin (Sigma Aldrich, St. Louis, MO). The effect of these organic macromolecules on the fullerene aggregates was tested at three different ionic strengths using NaCl as the electrolyte. The organic matter was reacted with the fullerene suspensions in the carrier fluid just prior to their introduction into the column. Alginate and tannic acid were tested at concentrations of 2 and 1 ppm, respectively. These concentrations of alginate and tannic acid were selected because they were observed to be thresholds, beyond which the electrophoretic mobility of the fullerene aggregates remained nearly constant with increasing concentration of organic matter in the solution. The deposition of particles in a porous medium, such as a sand filter or aquifer, was modeled as a sequence of particle transport to the immobile surface or “collector”, such as a sand grain, followed by attachment (16) where transport is described by a collector efficiency, η0, that is multiplied by an attachment efficiency, R. The attachment or stickiness coefficient is often interpreted as encompassing very shortrange interactions, largely chemical in nature in contrast with the collector efficiency that embodies physical effects such as porous media characteristics, flow velocity, and particle size. A mass balance of particles over a differential volume of porous medium can be integrated over distance within a homogeneous medium to yield an expression for the attachment efficiency factor, R, as a function of the observed removals, the characteristics of the porous medium, and the flow

4rC R)ln(C/C0) 3(1 - )η0L

(1)

where rC is the radius of a collector (assumed to be spherical) in the porous medium,  is the porosity of the porous medium, L is the length of the porous medium, and C and C0 are the particle concentrations in the column effluent at distance L and the influent at L ) 0, respectively. The single-collector efficiency, η0, is calculated as function of the Darcy velocity, porous medium grain size, porosity, and temperature among other variables (17). With the experimental C/C0 values (fraction of influent particles remaining following passage through the porous medium) obtained from the plateau of breakthrough curves and theoretical ηo values, values of R were calculated for fullerene suspensions in each column experiment. The plateau portion of the curve of C/C0 was averaged over at least one non-dimensional time unit. Reproducibility of the value of the plateau portion of the breakthrough curves was within (2%. Fullerene Suspensions. Dispersions of C60 clusters (nC60) and nC60(OH)18 were prepared from powdered C60 and powdered C60(OH)18 (MER, Tucson, AZ). Two varieties of nC60 were prepared and are referred to as TTA/nC60 and THF/ nC60. All of the suspensions were characterized by dynamic light scattering (DLS), measurements of electrophoretic mobility (Nanosizer ZS, Malvern Instruments, Worcestershire, U.K.) and electron microscopy at concentrations and solution chemistries produced at the inlet to the column. TTA/nC60. A procedure similar to that reported by Scrivens et al. (18), was used involving three different organic solvents, toluene, tetrahydrofuran (THF), and acetone (Aldrich Chemical Co., Inc. Milwaukee, WI). The C60 was first dissolved in toluene and then added to a volume of THF at room temperature. The resulting magenta solutions were added dropwise to rapidly stirred acetone. Doubly deionized water (DDW) was slowly added to the resulting acetone/THF/ toluene/ C60 solution. Following the first 100 mL of DDW, the TTA/nC60 appeared as a fine brownish-yellow suspension. No additional treatment was needed to initiate colloid formation. Upon complete addition of water, the organic solvents and some water were removed by slow boiling to a final volume of 200 mL yielding brownish-yellow aqueous suspensions of TTA/nC60. The evaporation step was necessary to remove the residual solvents; however residual organic solvent remains within the nC60. We have previously shown that these residual solvents affect the physical chemical characteristics of the nC60, such as charge and hydrophobicity (1, 19). The concentration of TTA/nC60 in the resulting suspension as determined by TOC analysis (including residual solvent in the aggregate) was 7 ppm as organic carbon. The TTA/nC60 colloids produced by this method ranged from 50 to 200 nm in diameter with an average diameter of 92 nm (standard deviation 18 nm) as determined by dynamic light scattering (DLS). The DLS instrument (Zetasizer nano ZS, Malver Instrument, Bedford, MA) employs a He-Ne laser (633 nm) and collects time-variable scattering at a fixed angle of 173°. The mean diameter, as measured by DLS, of the TTA/nC60 produced for these experiments was significantly smaller than the number mean diameter of 168 nm for the TTA/nC60 that we have previously produced. Consistent with previous descriptions of these materials (1, 11, 19), the colloids typically displayed somewhat facetted edges and were often hexagonal in a 2-D projection as observed by TEM. Samples were imaged with a JEOL 2100 at 100 kV TEM after wetting carbon-coated grids with the suspension and then drying at room temperature. Aggregates of these colloids were frequently observed, but it was not possible to distinguish specifically which aggregates may have existed in suspension or those that may have been created in drying the sample for TEM imagery. THF/nC60. THF/nC60 was prepared following the method outlined by Deguchi (3). Powdered C60 (MER Corporation, Tucson, AZ) was first dispersed in THF (Fisher Scientific, VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7397

FIGURE 1. Breakthrough curves for fullerol suspensions (13 mg TOC/L, 0.03 M NaCl, Darcy velocity of 12 × 10-4 m.s-1) after injection at θ ) 0 were followed by a plateau of roughly constant concentration (left). When injection of the fullerol suspension was stopped at θ ≈ 3.5, concentrations in the column effluent decreased monotonically. Attachment efficiencies calculated from plateau values of C/C0 increased with ionic strength (right). Houston, TX) at a concentration of 25 mg/L. The mixture was then purged with nitrogen to remove any dissolved oxygen and stored overnight in the dark while being continuously stirred to allow the solution to become saturated with C60. The C60 solution was then filtered through a 0.22 µm nylon filter to remove excess solid material. An equal amount of DDW was then added to the C60 in THF solution, while being continuously stirred. THF was subsequently removed from the solution using a rotary evaporator (Buchi Rotovap, Flawil, Switzerland), where the more volatile THF was evaporated off at a temperature of approximately 80 °C. This solution was filtered through a 0.22 µm nylon filter and stored in the dark. The resulting THF/nC60 had a mean particle diameter of 111 (standard deviation 38 nm), as determined by dynamic light scattering (DLS). The concentration of these THF/nC60 was 13 ppm as TOC, with residual THF known to remain within the nC60 (1) aggregate. As was observed with the TTA/nC60, THF/nC60 colloids frequently exhibited faceted edges. Colloids often appeared in groups with an underlying, more or less round primary structure approximately 20 nm in diameter, as well as longer more-rectangular objects as projected in 2-D. Fullerol Aggregates. Fullerol (C60(OH)18, MER, Tucson, AZ) was introduced directly to water by mechanical stirring and sonication with a direct probe sonicator. This produced a colloidal suspension of nC60(OH)18 with an average diameter of 120 nm (standard deviation 20 nm). In contrast with the TTA and THF nC60, fullerol aggregates were more rounded in appearance. The concentration of carbon in the suspension measured by TOC was 13 mg/L corresponding to a concentration of 18 mg/L of fullerol in suspension, accounting for the average of 18 hydroxyl groups per fullerene.

Results and Discussion The effect of ionic strength on the retention of each of the fullerene suspensions is summarized in terms of the observed concentrations in the column effluents and the corresponding attachment efficiencies calculated from the plateau portions of the break through curves. The effect of organic macromolecules on the retention of these suspensions is then presented. Effect of Ionic Strength and Composition. Fullerol Suspensions. Results from a series of column experiments are shown in Figure 1 for the case of fullerol (nC60(OH)18) when the background ionic strength was fixed at 3 × 10-2 M NaCl (Figure 1). In response to a step function increase and then decrease of fullerol aggregates in the feed, the con7398

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007

centration of fullerol in the column effluent increased monotonically, showing little retention in the column and reaching a value of C/C0 of 0.5 near one pore volume (θ ) 1) similar to an ideal tracer. The earlier reported “affinity transition” at the higher Darcy velocity in such experiments (11) was not observed. The reproducibility of three experiments is illustrated in the left panel of Figure 1. On the basis of nine consecutive experiments, the average plateau value of passage of fullerol aggregates through the column at this ionic strength was calculated to be 95 ( 2%. The descending portion of the curve was also monotonic. Experiments repeated at three different ionic strengths yielded a trend of increasing attachment efficiency with increasing ionic strength (Figure 1, right). At the highest ionic strength the attachment efficiency was calculated to be unity corresponding to a plateau value of C/C0 ) 0.33 at a Darcy velocity of 12 × 10-4 m s-1. As retention increased, variability in the plateau values and associated collision efficiencies also increased. TTA/nC60. Deposition of TTA/nC60 in the porous medium also increased with increasing ionic strength (Figure 2), again consistent with classic DLVO theory (20, 21), which predicts a decrease in electrostatic repulsion between particles as ionic strength increases. This trend was observed at both the faster and slower flow velocities applied in these experiments. Plateau values obtained (using NaCl as the electrolyte) are plotted in Figure 2 for the two velocities studied. Retention of TTA/nC60 was less in the experiments conducted at the higher Darcy velocity in nearly every case but one (presented below), where the electrolyte was sodium sulfate at an ionic strength of 0.01 M. At the highest ionic strength, plateau values approach the theoretical limit for a value of R approaching unity where the plateau values are calculated to be approximately 0.10 and 0.37 for the lower and higher Darcy velocities, respectively. In contrast with our earlier reported observations (11) of TTA/nC60 removal in a similar porous medium, these results show very little removal of the nC60 at an ionic strength of 10-2 M, no evidence of an “affinity transition”, and a classic dependence on velocity at higher ionic strengths. Differences in experimental conditions between these two studies include a TTA/nC60 of a smaller mean size (80 vs 168 nm), differences in glass bead preparation, and an upflow column configuration. But above all, these results underscore the divergence in experimental results that may be obtained for different preparations of nC60, as is illustrated later in this paper using two types of nC60.

FIGURE 2. Evolution of plateau values of C/C0 for the TTA/nC60 as a function of ionic strength at two Darcy velocities (left) and the associated average attachment efficiencies calculated from these data (right).

TABLE 1. Ratio of TTA/nC60 Outlet Concentration over Inlet Concentration, C/C0 (Plateau Value within (2%), for Different Concentrations and Types of Saltsa C/ C0

salt

ionic strength (M)

12 × 10-4 m s-1

4 × 10-4 m s-1

attachment efficiency (r ( SD)

NaCl NaCl NaCl NaCl NaCl Na2SO4 Na2SO4 (NH4)2SO4 NaNO3 NaNO3 CaCl2 MgCl2

0.6 0.3 0.1 0.06 0.01 0.01 0.1 0.1 0.01 0.1 0.01 0.01

0.32 0.47 0.64 0.83 0.96 0.96 0.66 0.54 1 0.42 0.79 0.82

0.15 0.22 0.4 0.67 0.94 1 0.43 0.45 0.96 0.21 0.59 0.53

0.73 ( 0.14 0.52 ( 0.04 0.31 ( 0.05 0.13 ( 0.00 0.03 ( 0.01 ∼0 0.29 ( 0.02 0.36 ( 0.13 0.01 ( 0.03 0.57 ( 0.08 0.17 ( 0.00 0.18 ( 0.05

a Values of R represent an average between values calculated in experiments performed at the two Darcy velocities.

Variations in TTA/nC60 retention with ionic strength and composition are summarized in Table 1, where the affinity of the TTA/nC60 for the porous medium in these experiments is expressed by the stickiness or attachment efficiency, R. The values of R reported in Table 1 are mean values calculated from experiments conducted at the high and low velocities, assuming the physicochemical properties of the particles and column do not change with velocity. This assumption is supported by the small standard deviation in values for R calculated from data obtained at the two Darcy velocities. Increasing ionic strength leads to an increase in attachment efficiency that is attributed to reduced electrostatic repulsion resulting from charge screening. At equal ionic strengths, both divalent cations (Ca2+ and Mg2+) increased the retention of the TTA/nC60 in the column compared to the monovalent Na+, consistent with the ability of these ions to specifically associated with charged functional groups, as well as Schultz-Hardy effect. The plateau value was slightly lower for the magnesium compared to the calcium-based salt. In experiments where Na+ was the counterion, TTA/ nC60 stability decreased in the order of NO3- > Cl- ≈ SO42-. When the anion, Cl-, was held constant stability decreased in the order of Mg2+ ≈ Ca2+ > Na+. THF/nC60. The trends in THF/nC60 deposition as a function of ionic strength were similar to those observed for the TTA/

nC60. Fullerene aggregates were deposited in the column to a greater degree as the ionic strength increased (Figure 3). Consistent with theory, the higher Darcy velocity resulted in less deposition. The same trend of increasing deposition (and increasing R) with ionic strength was observed at both velocities and similar values for the attachment efficiencies were calculated at a given ionic strength for experiments performed at the two Darcy velocities investigated. In comparison with the TTA/nC60, the THF/nC60 exhibited less mobility through the column (higher attachment efficiencies) at all of the ionic strengths evaluated (Figure 4). At both velocities, the TTA/nC60 exhibited a higher plateau value of C/C0 compared to that measured for the THF/nC60. However, as the ionic strength increased, the differences in mobility for these two varieties of nC60 decreased, with the differences being smallest in experiments performed at the higher Darcy velocity (12 × 10-4 m s-1). The differences in attachment efficiency are consistent with theoretical expectations based on differences in particle surface charge, as estimated from measurements of electrophoretic mobility for these suspensions. The electrophoretic mobility of the TTA/nC60 was greater (more negative) than that of the THF/nC60 at every ionic strength evaluated (see Supporting Information). This is in contrast with measurements we have previously reported for similar materials prepared for an earlier study (1), where the THF/nC60 in those suspensions was observed to have a more negative electrophoretic mobility than the TTA/nC60, once again highlighting the variability that may come into play in the preparation of colloidal suspensions of fullerenes for subsequent transport or toxicity studies, as well as the possible impact of residual solvents in altering the surface chemistry of nC60 in aqueous suspensions. These observations underscore the modifications to colloidal aggregates of fullerenes that may occur as a consequence of interaction with ionic components present in environmental and physiological systems. Because the factors affecting particle deposition in porous media tend to parallel those controlling particle aggregation (22, 23), it is reasonable to conclude that, other factors being equal, the stability of nC60 is likely to decrease substantially as these aggregates encounter solutions with ionic strengths typical of those in fresh waters (near 10-2 M). In more saline environments, such as seawater or fluids present within the lung, the propensity for fullerene deposition, and very likely that of aggregation as well (9), are predicted to increase. VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7399

FIGURE 3. Evolution of C/C0 for THF/nC60 as a function of ionic strength at two Darcy velocities (left) and the associated average attachment efficiencies calculated from these data (right).

FIGURE 4. Comparison of attachment efficiencies calculated for the THF/nC60 and TTA/nC60. Effect of Organic Macromolecules. Organic macromolecules are also widely present in natural and physiological solutions. Interactions between these macromolecules and nanomaterials may also have a profound effect on the surface chemistry and transport of nanomaterials in aquatic systems. Adsorption of organic solutes on colloids has been observed to both increase and decrease particle stability, depending on the nature of the macromolecule. For example, proteins and humic, fulvic, and tannic acids (24) tend to reduce the potential for particle aggregation and deposition through steric or charge stabilization, while polysaccharides, such as alginate, tend to promote particle aggregation and deposition by a bridging process (15, 25). Indeed, in experiments performed using colloidal aggregates of C60, we observed that the nature of the organic macromolecule, as well as the initial functionality of the fullerene, influenced the particle mobility in porous media. In comparison with the attachment efficiencies calculated from experiments where organic material was absent, (NaCl only) the presence of tannic acid in the feed solution stabilized the THF/nC60, resulting in a greater mobility of the fullerene aggregates through the column as expressed by a lower attachment coefficient, R (Figure 5). The characteristics of tannic acid with respect to aggregation, deposition, and membrane fouling are very similar to those of humic and fulvic acids, stemming in large part from the combination of hydrophobic portions of these materials and ionizable functional groups (26-28). The presence of natural organic 7400

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007

FIGURE 5. Attachment efficiencies (average of values from experiments at fast and slow velocities) determined for THF/nC60 as a function of ionic strength of the solution for 3 solutions: NaCl alone, NaCl + tannic acid, and NaCl + alginate. matter of this variety with fullerenes would be expected to enhance fullerene mobility in aquatic environments. Adsorption of the relatively small tannic acid molecules (e2 nm) on the fullerol and THF/nC60 aggregates (typically greater than 100 nm) may increase the negative charge of these fullerene colloids which would in turn increase electrostatic repulsion and colloid stability (29). At lower ionic strengths, reduced charge shielding may extend the reach of the adsorbed tannic acid molecules such that some steric stabilization is produced. In contrast, as in other colloidal systems, the presence of polysaccharide-like natural organic matter increased the deposition of the fullerene aggregates because of an increase in the attachment efficiency (Figure 5) and reduced the impact of ionic strength. The presence of alginate in the feed solution resulted in attachment efficiencies that were relatively high and relatively insensitive to the concentration of NaCl. The large molecular weight of the more linear polysaccharide molecules corresponds to molecular diameters on the order of several hundred nanometers. The association between large polysaccharides and smaller colloids is therefore more appropriately described as a collection of colloids along the length of the polysaccharide

(30), leading to a relatively large assembly. Although experiments were not performed using both Ca and alginate, the ability of calcium to initiate polysaccharide gel formation is well-known (e.g (31)). It is reasonable to conclude that the destabilizing effect of polysaccharides on colloidal aggregates of fullerenes would be even more pronounced in the presence of calcium ions. Measurements of electrophoretic mobility as a function of tannic acid or alginate concentration show a concentration-dependent relationship, with the electrophoretic mobility of the nC60 becoming progressively more negative at higher concentrations of these macromolecules (see Supporting Information). While it is likely that tannic acid and alginate also interacted directly with the silica beads, the electrophoretic mobility data further illustrate the role of natural organic matter in altering aggregate surface chemistry and ultimately determining the mobility of fullerenes in aquatic environments. Indeed, in some cases, it appears that association of macromolecules with fullerene aggregates and porous media may dominate the transport properties of these aggregates. However, it appears that the initial surface chemistry of the fullerenes may play a role in the determination of whether or not macromolecules modify the aggregate surface. When fullerol aggregates were exposed to tannic acid or alginate solutions before entering the porous media, little modification in deposition behavior was produced. While the retention of the aggregates of the hydroxylated C60 was much less than that of aggregates of underivatized C60 (THF/nC60), the addition of alginate had very little impact on removal across the column and the associated attachment efficiency for the fullerol (see Supporting Information). In summary, the observed trends in deposition of these fullerene nanoparticles are consistent with those observed in other colloidal systems. As with other colloidal systems, the physicochemical properties of aggregates of fullerene nanomaterials are likely to be influenced by the concentrations and composition of solutes in the environmental and physiological systems. Factors that will tend to decrease the retention of these materials in porous media and other environmental systems include a low ionic strength and the presence of hydrophobic, charged macomolecules, such as humic, fulvic, and tannic acids. We note that the high attachment efficiencies observed at higher ionic strengths are typically associated with higher rates of aggregation as well as deposition. Studies of nanomaterial toxicity are typically conducted in physiological media characterized by a high ionic strength (∼10-1 M). These results suggest that special care is required in performing and interpreting such studies to account for increased aggregation and affinity for both biotic and nonbiotic surfaces. In environmental systems, the mobility of fullerenes will likely be reduced at ionic strengths typical of many surface waters and estuaries and in the presence of polysaccharide-based natural organic matter as may be produced by algae or bacteria.

Acknowledgments We gratefully acknowledge the contributions Miguel Maldonado and Chennan Li to portions of the experimental work and comments from Dr. Jonathan Brant during the preparation of this manuscript. This work was supported in part by NSF grant BES 0653659 and EPA STAR awards 83241301 and 91650901.

Supporting Information Available Background information on the procedure for cleaning glass beads and data on electrophoretic mobility (Tables S1 and S2) and mobility of the fullerol material in the presence of

organic macromolecules (Figure S1). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Brant, J. A.; Labille, J.; Bottero, J.-Y.; Wiesener, M. R. Characterizing the impact of preparation method on fullerene cluster structure and chemistry. Langmuir 2006, 22, 3878-3885. (2) Andrievsky, G. V.; Klochkov, V. K.; Karyankina, E. L.; MchedlovPetrossyan, N. O. Studies of aqueous colloidal solutions of fullerene C60 by electron microscopy. Chem. Phys. Lett. 1999, 300, 392-396. (3) Deguchi, S.; Rossitza, G. A.; Tsujii, K. Stable dispersions of fullerenes, C60 and C70 in water. Preparation and characterization. Langmuir 2001, 17, 6013-6017. (4) Scrivens, W. A.; Tour, J. M. Synthesis of 14C-labeled C60, its suspension in water, and its uptake by human keratinocytes. J. Am. Chem. Soc. 1994, 116, 4517-4518. (5) 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. (6) Zhu, S. Q.; Oberdorster, E.; Haasch, M. L. Toxicity of an engineered nanoparticle (fullerene, C-60) in two aquatic species, daphnia and fathead minnow. Mar. Environ. Res. 2006, 62, S5S9. (7) Jia, G.; Wang, H.; Yan, L.; Wang, X.; Pei, R.; Yan, T.; Zhao, Y.; Guo, X. Cytotoxicity of carbon nanomaterials: Single-wall nanotube, multi-wall nanotube, and fullerene. Environ. Sci. Technol. 2005, 39 (5), 1378-1383. (8) Yamago, S.; Tokuyama, H.; Nakamura, E.; Kikuchi, K.; Kananishi, S.; Sueki, K.; Nakahara, H.; Enomoto, S.; Ambe, F. In-vivo biological behavior of a water-miscible fullerenesC-14 labeling, absorption, distribution, excretion and acute toxicity. Chem. Biol. 1995, 2 (6), 385-389. (9) Brant, J. A.; Lecoanet, H.; Wiesner, M. R. Aggregation and deposition characteristics of fullerene nanoparticles in aqueous systems. J. Nanopart. Res. 2005, 7, 545-553. (10) 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. (11) Lecoanet, H. F.; Wiesner, M. R. Velocity effects on fullerene and oxide nanoparticle deposition in porous media. Environ. Sci. Technol. 2004, 38 (16), 4377-4382. (12) Tobiason, J. E. Chemical effects on the deposition of nonBrownian particles. Colloids Surf. 1989, 39, 53-77. (13) Veerapaneni, S.; Wiesner, M. R. Role of suspension polydispersivity in granular media filtration. J. Environ. Eng., ASCE 1993, 119 (1), 172-190. (14) Veerapaneni, S. Formation and morphology of colloidal deposits in porous media. Ph.D. dissertation, Rice University, Houston, TX, 1996. (15) Singh, R. P.; Tripathy, T.; Karmakar, G. P.; Rath, S. K.; Karmakar, N. C.; Pandey, S. R.; Kannan, K.; Jain, S. K.; Lan, N. T. Novel biodegradable flocculants based on polysaccharides. Curr. Sci. 2000, 78 (7), 798-803. (16) O’Melia, C. R.; Stumm, W. Theory of water filtration. J. Am. Water Works Assoc. 1967, 59, 1393-1412. (17) Tufenkji, N.; Elimelech, M. Correlation equation for predicting single-collector efficiency in physiochemical filtration in saturated porous media. Environ. Sci. Technol. 2004, 38, 529-536. (18) Scrivens, W. A.; Tour, J. M.; Creek, K. E.; Pirisis, L., Synthesis of 14C-labeled C60, its suspension in water, and its uptake by human keratinocytes. J. Am. Chem. Soc. 1994, 116, 4517-4518. (19) Brant, J. A.; Lecoanet, H.; Hotze, M.; Wiesner, M. R. Comparison of electrokinetic properties of colloidal fullerenes (n-C60) formed using two procedures. Environ. Sci. Technol. 2007, in press. (20) Derjaguin, B. V.; Landau, L. D. Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes. Acta Physicochim. URSS 1941, 14, 733-762. (21) Hunter, R. J. Zeta Potential in Colloid Science; Academic Press Inc.: New York, 1981; p 386. (22) O’Melia, C. R. Aquasols: The behavior of small particles in aquatic systems. Env. Sci. Technol. 1980, 14 (9), 1052-1060. (23) O’Melia, C. R. Particles, pretreatment, and performance in water filtration. J. Environ. Eng. 1985, 111 (6), 874-890. (24) Tsai, C. Y.; Lee, D. S.; Tsai, Y. H.; Chan, B.; Luh, T. Y.; Chen, P. J.; Chen, P. H. Shrinking gold nanoparticles: Dramatic effect of a cryogenic process on tannic acid/sodium citrate-generated gold nanoparticles. Mater. Lett. 2004, 58 (14), 2023-2026. VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7401

(25) Nguyen, T. P.; Hankins, N. P.; Hilal, N. Effect of chemical composition on the flocculation dynamics of latex-based synthetic activated sludge. J. Hazard. Mater. 2007, 139 (2), 265274. (26) Inskeep, W.; Silvertooth, J. Inhibition of hydroxyapatite precipitation in the presence of fulvic, humic, and tannic acids. Soil Sci. Soc. Am. J. 1988, 52 (4), 941-946. (27) Lahoussine-Turcaud, V.; Wiesner, M. R.; Bottero, J.-Y. Fouling in tangential-flow ultrafiltration: The effect of colloid size and coagulation pretreatment. J. Membr. Sci. 1990, 52 (1), 173-190. (28) Wiesner, M. R.; Laıˆne´, J. M. Coagulation and membrane separation. In Membrane Processes in Water Treatment; Mallevialle, J., Odendaal, P., Wiesner, M. R., Eds.; McGraw-Hill, 1996.

7402

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007

(29) Narkis, N.; Rebhun, M. Stoichiometric relationship between humic and fulvic acids and flocculants. J. Am. Water Works Assoc. 1977, 69 (6), 325. (30) Ferretti, R.; Stoll, S.; Zhang, J.; Buffle, J. Flocculation of hematite particles by a comparatively large rigid polysaccharide: The schizophyllan. J. Colloid Interface Sci. 2003, 266, 328-338. (31) Jermann, D.; Pronka, W.; S., M.; Boller, M. Interplay of different nom fouling mechanisms during ultrafiltration for drinking water production. Water Res. 2007, 41 (8), 1713-1722.

Received for review April 13, 2007. Revised manuscript received August 31, 2007. Accepted September 5, 2007. ES0708767