Environ. Sci. Technol. 2010, 44, 1386–1391
Impact of Source Water Quality on Multiwall Carbon Nanotube Coagulation R . D A V I D H O L B R O O K , * ,† C A R L Y N . K L I N E , ‡,§ A N D JAMES J. FILLIBEN| Surface and Microanalysis Science Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899; Department of Chemistry, Kutztown University, Kutztown, Pennsylvania 19530-0730; and Statistical Engineering Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899
Received September 28, 2009. Revised manuscript received December 4, 2009. Accepted December 17, 2009.
Potable water treatment facilities may become an important barrier in limiting human exposure to engineered nanoparticles (ENPs) as ENPs begin to contaminate natural aquatic systems. Coagulation of ENPs will likely be a major process that controls the ENP fate and the subsequent removal in the aqueous phase. The influence that source water quality has on ENP coagulation is still relatively unknown. The current study uses a 23 × 24-1 fractional factorial design to identify seven key surface water constituents that affect multiwall carbon nanotube (MWCNT) coagulation. These seven factors include: influent concentrations of kaolin, organic matter (OM), alginate, and MWCNTs; type and dosage of coagulant; and method of MWCNT stabilization. MWCNT removal was most affected by coagulant type and dosage, with alum outperforming ferric chloride at circumneutral pH. None of the other factors were universally significant but instead depended on coagulant type, dose, and method of stabilization. In all cases where factors were found to have a significant impact on MWCNT removal, however, the relationship was consistent: higher influent concentrations of kaolin and alginate improved MWCNT removal while higher influent concentrations of OM hindered MWCNT coagulation. Once MWCNTs are released into the natural environment, their coagulation behavior will be determined by the type and quantity of pollutants (i.e., factors) present in the aquatic environment and are governed by the same mechanisms that influence the colloidal stability of “natural” nanoparticles.
Introduction The discussion regarding unintentional release of engineered nanoparticles (ENP) into the natural environment is usually conducted with a tone of inevitability; the questions never really begin with “if” but instead with “when, how much, * Corresponding author phone: (301) 975-5202; fax: (301) 4171321; e-mail:
[email protected]. † Surface and Microanalysis Science Division, National Institute of Standards and Technology. ‡ Present Address: PMRS, Inc., Horsham, Pennsylvania 19044. § Department of Chemistry, Kutztown University. | Statistical Engineering Division, National Institute of Standards and Technology. 1386
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 4, 2010
and what environments would be most affected”? Recent evidence suggests that this tone is warranted; early estimates of ENP concentrations in natural systems (1) are being substantiated through actual measurements (2, 3). ENPs in aquatic systems are of particular concern since the ENPs’ small size, high surface reactivity, and composition can have a detrimental impact on resident fauna, flora, and microbial populations (4-8). ENPs present in aquatic systems may also represent a pathway for human exposure, especially if these contaminated aquatic systems are used as sources of potable water. One type of ENP that has received considerable attention is carbon nanotubes (CNTs), caused in part by their structural and morphological similarity to a naturally occurring nanomaterial, asbestos fibers (9). Mueller and Nowack (1) estimate that CNT concentrations in Swiss waters may range between 0.5 ng/L to 0.8 ng/L, which equates to a number density of ≈ 108 CNTs/L (assuming a CNT mass on the order of tens of attograms per CNT (10)). This number density is similar to that of asbestos fibers in both source and treated drinking waters (11, 12). Although some studies have suggested that populations heavily exposed to asbestos in drinking water may possess elevated risks of certain cancers (13), the total body of evidence supporting this link is neither consistent nor convincing (14, 15). Such epidemiological or exposure studies have not been conducted for CNTs, but the precautionary principle suggests that a more thorough understanding of environmental factors that affect CNT behavior in potable water treatment facilities are therefore deemed critical. Yet considering their potential importance, potable water treatment unit processes (specifically coagulation) as related to ENP removal have, to date, received only scant attention. Westerhoff et al. (16) examined the relationship between ENP characteristics and behavior in aquatic systems, suggesting aggregation potential was determined by surface charge and potential for ligand complexation (16). Zhang et al. (17) reported that coagulation removal efficiencies of select metal oxide nanoparticles ranged between 20% and 60% and Holbrook et al. (18) confirmed that MWCNTs could be removed from the aqueous phase via coagulation using either ferric chloride or aluminum sulfate (alum). Hyung and Kim (19) concluded that removal of nC60 depended on coagulant dose, the presence of natural organic matter (OM), and ENP surface properties. Potable source waters can vary widely in the nature and concentration of OM, turbidity, biopolymer concentration, pH, temperature, and ionic strength. ENP aggregation and coagulation, processes that will ultimately determine ENP fate and behavior once released into the environment, are likely influenced by these same parameters (16, 20). ENP fate is further complicated by different methods of stabilization. The previous studies, therefore, are somewhat limited in their utility. Conducted under constant ambient conditions (e.g., constant OM concentrations) or in the absence of other pollutants (e.g., clay), they failed to capture the inherent complexity in water quality so prevalent in natural aquatic systems. Consequently, extrapolating the ENP removal efficiencies obtained in these studies to other source waters could prove challenging. The present study attempts to fill this knowledge gap by considering the effect of seven factors on multiwall carbon nanotubes (MWCNT) removal during coagulation, including: (1) concentrations of influent kaolin, OM, alginate and MWCNTs; 10.1021/es902946j
2010 American Chemical Society
Published on Web 01/21/2010
TABLE 1. Sample 24-1 Fractional Factorial Design (n = 8)a condition 1 2 3 4 5 6 7 8
kaolin conc. OM conc. alginate conc. MWCNT conc. (factor 1) (factor 2) (factor 3) (factor 4) + + + +
+ + + +
+ + + +
+ + + +
optimal coagulant dosage 0.27 0.27 0.07 0.29 suboptimal coagulant dosage ferric chloride 0.04 0.04 alum 0.02 0.04
TABLE 2. Description of Experimental Factors and Levels factor
5 6 7
kaolin organic matter alginate MWCNT concentration surfactant-stabilized OM-stabilized MWCNT stabilization coagulant type coagulant dosage
low (-)
surfactantOMstabilized MWCNTs stabilized MWCNTs ferric chloride alum
a This design was repeated for each of the eight conditions in the 23 factorial design and both duplicated and randomized for statistical purposes.
1 2 3 4
TABLE 3. Summary of Coagulant Dosages
high (+)
units
10 2 2
30 10 10
mg/L mg/L mg/L
2.2 0.21 surfactant ferric chloride suboptimal
4.4 0.45 OM alum optimal
mg/L mg/L
(2) type and dosage of coagulant (ferric chloride vs alum); and (3) method of MWCNT stabilization (surfactant vs OM). The main objectives of this work are as follows: (1) to determine whether or not the type of coagulant (ferric chloride vs alum) was a critical factor in MWCNT removal; (2) to determine what role typical surface water pollutants play in MWCNT coagulation, and; (3) to ascertain whether or not mechanisms of coagulation would be identical for MWCNTs and “natural” nanoparticles (colloids).
Methods and Materials Experimental Overview and Design. Full and fractional factorial designs allow the determination of how different parameters and their interactions affect a response variable (in this case, MWCNT removal from the aqueous phase via coagulation, hereinafter referred to as MWCNT removal). A two-level design was chosen to address the first objective. A 27 full factorial design requires n ) 27 ) 128 conditions, which was considered prohibitive. A more efficient design is to run a 23 × 24-1 fractional factorial design, which would require only n ) 64 conditions. This design is formed by running a 23 full factorial design on the three “controllable” factors (coagulant type, coagulant dosage, and MWCNT stabilization method) in concert with a 24-1 fractional factorial design on the remaining four “uncontrollable” factors that represent surface water pollutants and contaminants (kaolin, NOM, alginate and MWCNT concentrations, respectively), where “controllable” factor refers to that can be chosen by the operator/engineer and “uncontrollable” factor refers to what would be found in the source water. Each 24-1 experimental design was run a total of eight times, once for each of the eight different combinations of coagulant type, coagulant dosage, and MWCNT stabilization method. A sample of the 24-1 experimental design (n ) 8) is provided in Table 1. Concentrations for all seven factors are detailed in Tables 2 and 3. A more thorough discussion of the experimental design approach is provided in the Supporting Information (SI). Since there were no significant 2-term interactions involving
units µM Fe µM Al µM Fe µM Al
MWCNTs, the ensuing analysis and discussion will focus only on main effects. Note that due to the intrinsic variability of environmental samples, each of the 23 × 24-1 ) 64 experimental conditions was run in duplicate, thus yielding n ) 128 runs. Raw Synthetic Water and Test Reactors. Daily stock solutions consisting of kaolin (Fisher Scientific, Pittsburgh, PA) and Na2CO3 (Fisher Scientific) were mixed in 3 L of room temperature (22 °C ( 1 °C) Montgomery County, MD tap water for 18 h. More detailed information about the tap water quality can be found at http://tinyurl.com/yb4rhq6. Following mixing, a known volume of this stock solution was transferred into one of two test reactors. Two different reactors were necessary due to the disparity in MWCNT stock concentrations; surfactant-stabilized MWCNTs can be produced at much higher concentrations compared to their OMstabilized counterparts, thereby affording a much larger reactor volume in the experimental design without compromising MWCNT detection limits. Subsequently, 2 L B-KER2 square jar test beakers (10.8 cm × 10.8 cm, Phipps and Bird, Richmond, VA) were used in the surfactantstabilized MWCNT experiments, whereas 50 mL conical Falcon tubes (115 × 30 mm diameter) were used in the OMstabilized MWCNT experiments. An appropriate amount of either synthetic humic acid (HA, Fisher Scientific and used in the 2 L jar test beakers) or Suwannee River natural organic matter (SRNOM, International Humic Substance Society, St. Paul, MN and used in the 50 mL tubes) and alginate (Fisher Scientific) was taken from stock solutions prepared in Milli-Q water (Millipore, Bedford, MA) and added to the reactors. Comparative jar tests using either HA or SRNOM revealed no statistical difference in MWCNT removal (SI) and are therefore collectively referred to as OM throughout this work. Finally, tap water was used to bring the working reactor volume to 2 L and 15 mL, respectively, the pH was adjusted to 7.3 ((0.3) with sulfuric acid, and MWCNTs were added and allowed to mix for 20 min prior to initiating coagulation. Literature values, experience, and laboratory observations were used to determine the factor levels (low (-) and high (+)) and the coagulant dosing concentrations for the specific conditions tested (Tables 2 and 3). Initial kaolin concentrations were chosen to reflect turbidity measurements in natural waters (21), and both OM and alginate concentrations were consistent with literature values of NOM and biopolymer, respectively (21-23). Coagulant dosages were determined during preliminary experiments; the optimal coagulant dose refers to the lowest coagulant concentration that could achieve at least 50% MWCNT removal while the suboptimal dose reflects both under-dosed coagulant conditions in an engineered system as well as free metal concentrations found in natural systems (24, 25). Although the MWCNT concentrations used in these experiments are several orders of magnitude higher than what may be expected in the natural environment (1), such concentrations allow for improved detection and therefore a better analysis of MWCNTs behavior. Carbon nanotube suspension, coagulation simulation and analytical descriptions are included in the SI. VOL. 44, NO. 4, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1387
FIGURE 1. Main effects plots for coagulant type analysis. Open squares represent statistically significant relationships (p < 0.05) and the symbols “-” and “+” represent the low and high levels of each factor (Table 2) or coagulant type (“-” ) ferric chloride, “+” ) alum).
Results and Discussion Analysis of Factorial Designs. The primary graphical tool used to assess the relative importance of a specific factor is the main effects plot (27). Figure 1 shows four such plotssone for each of the four combinations of coagulant dosage (optimal vs suboptimal) and MWCNT stabilization method (surfactant vs OM). Each main effects plot in Figure 1 illustrates the relative importance of five factors: the four influent concentration factors (kaolin, OM, alginate and MWCNT concentrations) plus a factor of primary interest (coagulant type). Each plot contains 10 plot pointsstwo points for each of the five factors. For a given factor, the left-most plot point is the average of n ) 16 responses in MWCNT removal when that factor was in the “low” (-) setting, while the right-most plot point is the same with the factor in the “high” (+) setting (Tables 1 and 2). The dotted line is the overall average response (n ) 32). When such a factor “has an effect” on the response variable (i.e., MWCNT removal), these two averages will be considerably different and hence the line connecting the two plot points will be steep and long. Conversely, a factor having no effect will typically yield a shallow sloped, short connecting line. Statistical significance of each factor is quantified by the Student’s t test and statistically significant factors in Figures 1 and 2 have been denoted using open square symbols. Coagulant Type Analysis. The four main effects plots in Figure 1 are grouped according to method of stabilization and coagulant dose. Focusing on the surfactant-stabilized MWCNTs (panels A and B in Figure 1), the main effects plot for the optimal coagulant dosage (panel A, Figure 1) illustrates 1388
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 4, 2010
that the only significant factor on MWCNT removal is influent kaolin concentration, with greater MWCNT removal being observed at higher kaolin levels (i.e., positive slope). OM, alginate, coagulant type, and MWCNT concentrations have no statistically significant impact on MWCNT removal, and the average MWCNT removal is 49% (dotted line, Figure 1). Average MWCNT removal at the suboptimal coagulant dose (panel B, Figure 1) is approximately 10% and both influent OM and alginate concentrations have a significant albeit opposing impact on MWCNT removal; lower MWCNT removal was observed at higher influent OM concentrations while higher MWCNT removal was observed at higher influent alginate concentrations. When OM is used as the stabilizing agent (panels C and D in Figure 1), the only significant factor at the optimal coagulant dose is the coagulant type (panel C, Figure 1), while influent kaolin, OM and MWCNT concentrations are either significant or suggestively significant (p < 0.08) at the suboptimal dose (panel D, Figure 1). Coagulant type and dosage are both critical factors in MWCNT removal. Both coagulants can, in fact, effectively remove MWCNTs given sufficient coagulant is present. In the case of the surfactant-stabilized MWCNTs, the required iron concentration is approximately 4× higher compared to the required aluminum concentration for a given level of treatment (Table 3). When iron and aluminum concentrations are nearly equivalent (OM-stabilized MWCNTs, Table 3), coagulant type is a significant factor (panel C, Figure 1) with alum providing greater MWCNT removal compared to ferric chloride. As discussed below, much of these differences are due to coagulant behavior at circumneutral pH. But these
FIGURE 2. Main effects plots when coagulant dosage is treated independently. Top row of plot points in each plot depicts behavior at optimal coagulant dosage while bottom row is for suboptimal dosage. Open squares represent statistically significant relationships (p < 0.05) and the symbols “-” and “+” represent the low and high levels of each factor (Table 2). differences also suggest that coagulant type should be treated distinctly from the other factorssthere is a confounding issue of coagulant type vs dosage. Separation from the other factors is afforded by the chosen experimental design; influent conditions in the 24-1 factorial design (Table 2) are identical for both coagulant types and therefore can be analyzed separately but compared directly for each condition. Coagulant Dosage Analysis. In Figure 2, the results from the optimal and suboptimal dosages are placed on the same main effects plot, resulting in a total of eight scenarios (i.e., two types of coagulant × two coagulant dosages × two MWCNT stabilization methods). The distance between the two dotted lines, which represent the average MWCNT removal at the optimal (top line) and suboptimal dosages (bottom line), respectively, illustrates the large influence that coagulant dosage has on MWCNT removal. Figure 2 illustrates three main points. First, of all the factors considered in this investigation, MWCNTs removal is most strongly dependent on coagulation dose. The distance between the dotted lines in a given plot (between 40% and 70% points) is much greater than the vertical distance between the high and low concentrations of any individual factor. Second, the individual factors that contribute to changes in raw water quality do impact MWCNT removal, but their effect depends on both coagulant type and coagulant dose. In other words, no factor is universally significant. Kaolin has a significant impact on MWCNT removal in six of the eight scenarios, both OM and alginate in four (OM is also suggestively significant in the surfactant-stabilized, suboptimal alum dosage), and MWCNT in one (Figure 2). The same factor, however, is significant for a given coagulant at both dosages and MWCNT-stabilization method only four
times and is never the same for both coagulants with the same reactor configuration. Third, and most critically, when these factors are significant, the direction of the main effects slope is consistent: a positive slope for kaolin and alginate; a negative slope for OM. So although the significant individual factors may differ among coagulant type and dose, their influence on MWCNT removal is nevertheless robustly reliable. The potential MWCNT removal mechanisms for each of the factors are discussed below. Significant Factors and MWCNT RemovalsA Physical and Chemical Interpretation. MWCNT coagulation is consistent with well-known coagulant behavior and/or behavior of naturally produced nanomaterials (colloids) in both engineered and natural systems. Correspondingly, the same mechanisms that govern colloidal stability will also dictate MWCNT behavior during coagulation. In this sense, the results obtained in this study using a complex environmental matrix confirm what other investigators have either hypothesized or demonstrated using simple aqueous systems (16, 19, 20). Furthermore, once MWCNTs are released into the natural environment, their coagulation behavior will be determined by the type and quantity of pollutants (i.e., factors) present in the aquatic environment. Perhaps such behavior is to be expected. Rapid sorption of OM and alginate to the MWCNT surface was observed prior to any coagulant addition (SI). Rather than providing a uniform coating, the sorbed OM and alginate appear at rather discrete intervals along the length of the MWCNT. Initial sorption of these organic macromolecules may bind to surface defects sites. Smith et al. (28) reported that even pristine MWCNTs have oxygen functional groups and Hyung et al. (26) hypothesized that NOMs’ charged functional groups VOL. 44, NO. 4, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1389
may contribute to CNT complexation, suggesting electrostatic attraction between the MWCNTs and OM and alginate may be occurring. Such complexation would alleviate any unique surface charge contribution of the MWCNTs, creating a homogeneous surface that would be more similar to the sorbed macromolecules (i.e., factors) than to the pristine MWCNT itself. Each factor, along with some proposed coagulation mechanisms, is discussed below with relevance to their impact on colloidal stability (in general) and MWCNT removal. Coagulant Type. Alum is much more efficient at MWCNT removal than ferric chloride at circumneutral pH levels used in these experiments, an observation that is consistent with DOC removal (29, 30). At this pH range, alum has higher concentrations of positively charged species compared to ferric chloride and therefore both adsorption and sweep floc are responsible for contaminant removal (31, 32). Such cationic species can assist in the complexation, neutralization, and subsequent flocculation of the negatively charged materials, including OM and MWCNTs. The presence of MWCNTs in the raw water, therefore, does not change wellestablished metal salt behavior in the coagulation process. Kaolin. MWCNT removal improved with higher influent kaolin concentrations (Figure 2). Purposeful addition of supplementary particles is often used to improve coagulation efficiency in certain applications as a way to increase opportunities for particle-particle contact (33). Higher kaolin concentrations would not only increase opportunities for kaolin-MWCNT contact but also translate to a greater available surface area for subsequent MWCNT removal. Such MWCNT removal is potentially achieved by two distinct mechanisms: (1) direct MWCNT surface adsorption on the kaolin particle followed by removal via particle destabilization; or (2) MWCNT enmeshment in a floc particle and removed via sweep flocculation. Both attachment mechanisms are observed for surfactant-stabilized MWCNTs (18, 34) and remain plausible mechanisms for removal of OMstabilized MWCNTs, especially since both kaolin and MWCNTs preferentially adsorb OM with similar characteristics (i.e., enriched in aromatic carbon, high molecular weight) (35, 36). OM. Poorer MWCNTs removal at higher influent OM concentrations (Figure 2) can be explained primarily by three mechanisms. First, coagulant dose and influent OM concentrations are stoichiometrically related for desired level of treatment; additional coagulant is necessary to react with OM that adsorbs to any surfaces as well as any OM left in solution. With constant coagulant dosage, a reduction in removal efficiencies would be expected at higher OM concentrations since there would be less available coagulant for colloid/MWCNT destabilization. Second, higher OM concentrations may saturate sorption sites on the kaolin surface, thereby rendering them unavailable for MWCNT sorption and subsequent removal via kaolin destabilization. Third, additional OM may help to further stabilize the MWCNTs through surface adsorption (36). Alginate. Improved MWCNT removal was observed at higher influent concentrations of alginate (Figure 2). Alginate, which is an anionic polyelectrolyte at circumneutral pH, has been proposed as a feasible alternative to synthetic coagulants for water treatment (37). Bernhartdt et al. (38) observed improved flocculation of quartz particles in the presence of alginate and proposed interparticle bridging as the destabilization mechanism provided that: (1) the alginate surface coverage of the particle is intermittent, and; (2) the alginate molecules possess a high enough charge density and sufficient molecular weight (>2 kDa) to facilitate the formation of bridging surfaces. Both conditions appear to have been fulfilled as intermittent alginate surface coverage on the 1390
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 4, 2010
MWCNTs is observed (SI) and the molecular weight of the alginate ranges between 20 and 80 kDa (39). While improving MWCNT removal, the presence of alginate also appears to alter the relation between turbidity removal and MWCNT removal by preferentially increasing turbidity removal relative to MWCNT removal (SI). Therefore, one universal relationship between MWCNT removal and turbidity removal does not appear to exist but must be developed for different source waters. MWCNTs. Initial MWCNT concentration is a significant factor in only one scenario (Figure 2) and illustrates that poorer MWCNT removal is concomitant with higher influent concentrations. We had previously hypothesized about the existence of a stoichiometric relationship between coagulant dose and influent MWCNT concentration. This hypothesis was tested using surfactant-stabilized MWCNTs (in an effort to achieve the largest possible MWCNT concentration range), both types of OM and a constant coagulant dose. As hypothesized, an inverse trend between influent MWCNT concentrations and MWCNT removal was observed for both coagulants (SI). Variations in removal efficiency, however, were sufficiently large that a statistically significant relationship could be not developed. Consequently, a sufficient mechanistic explanation of how influent MWCNT concentrations impact MWCNT removal via coagulation is still lacking. Environmental Significance. MWCNT coagulation in the natural environment would be expected to be, in general, rather limited due to the relatively low concentrations of free inorganic metals and the necessary change in environmental conditions that would facilitate metal-hydroxide formation (e.g., redox potential). Potable water treatment facilities, however, can achieve significant amounts of MWCNT removal but will require optimization on an individual basis. Additionally, as influent concentration of clay (kaolin), OM and biopolymer (alginate) can be highly variable, it can be expected that MWCNT removal in the coagulation process will also change, making downstream processes such as filtration important in reducing MWCNT concentrations in the final effluent. Other controllable parameters, such as pH, reactor configuration, mixing intensity, etc. will require future research to determine what role, if any, they have in optimizing MWCNT removal.
Acknowledgments C.N.K. was financially supported by a 2007 NIST Summer Undergraduate Research Fellowship (SURF). Bastian Pellegrin (TGA analysis) and Matt Staymates (impeller fabrication), both of NIST, and Charles Bott (jar tester) of VMI are gratefully acknowledged. We also thank the three anonymous reviewers for improving this manuscript with their constructive comments. Certain commercial equipment, instruments, or materials are identified in this work to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.
Supporting Information Available A thorough detail of the experimental design, including statistical analysis, and carbon nanotubes suspension is included in an expanded Methods section. Figure SI.1 shows the TGA analysis of the surfactant-stabilized MWCNTs. Figure SI.2 shows the TGA analysis and UV-vis spectrum of the NOM-stabilized MWCNTs. Figure SI.3 compares MWCNT removal using both HA and SRNOM as OM surrogates. Figure SI.4 shows ESEM images of rapid humic acid and alginate sorption to the MWCNT surface. Figure SI.5 illustrates the impact alginate has on the relationship between turbidity
removal and MWCNT removal during coagulation. Figure SI.6 shows the relationship between influent MWCNT concentration and MWCNT removal. This information is available free of charge via the Internet at http://pubs.acs. org/.
Literature Cited (1) Mueller, N. C.; Nowack, B. Exposure modeling of engineered nanoparticles in the environment. Environ. Sci. Technol. 2008, 42 (12), 4447–4453. (2) Kiser, M. A.; Westerhoff, P.; Benn, T.; Wang, T.; Perez-Rivera, J.; Hristovski, K. Titanium nanomaterial removal and release from wastewater treatment plants. Environ. Sci. Technol. 2009, 43 (17), 6757–6763. (3) Kaegi, R.; Ulrich, A.; Sinnet, B.; Vonbank, R.; Wichser, A.; Zuleeg, S.; Simmler, H.; Brunner, S.; Vonmont, H.; Burkhardt, M.; Boller, M. Synthetic TiO2 nanoparticle emission from exterior facades into the aquatic environment. Environ. Pollut. 2008, 156 (2), 233–239. (4) Bradford, A.; Handy, R. D.; Readman, J. W.; Atfield, A.; Muhling, M. Impact of silver nanoparticle contamination on the genetic diversity of natural bacterial assemblages in estuarine sediments. Environ. Sci. Technol. 2009, 43 (12), 4530–4536. (5) Kang, S.; Mauter, M. S.; Elimelech, M. Microbial cytotoxicity of carbon-based nanomaterials: implications for river water and wastewater effluent. Environ. Sci. Technol. 2009, 43 (7), 2648– 2653. (6) Griffitt, R. J.; Luo, J.; Gao, J.; Bonzongo, J. C.; Barber, D. S. Effects of particle composition and species on toxicity of metallic nanomaterials in aquatic organisms. Environ. Toxicol. Chem./ SETAC 2008, 27 (9), 1972–1978. (7) Federici, G.; Shaw, B. J.; Handy, R. D. Toxicity of titanium dioxide nanoparticles to rainbow trout (Oncorhynchus mykiss): gill injury, oxidative stress, and other physiological effects. Aquat. Toxicol. 2007, 84 (4), 415–430. (8) Sayes, C. M.; Gobin, A. M.; Ausman, K. D.; Mendez, J.; West, J. L.; Colvin, V. L. Nano-C60 cytotoxicity is due to lipid peroxidation. Biomaterials 2005, 26 (36), 7587–7595. (9) Poland, C. A.; Duffin, R.; Kinloch, I.; Maynard, A.; Wallace, W. A. H.; Seaton, A.; Stone, V.; Brown, S.; Macnee, W.; Donaldson, K. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos-like pathogenicity in a pilot study. Nat. Nanotechnol. 2008, 3 (7), 423–428. (10) Wu, W.; Palaniapan, M.; Wong, W. K. Multiwall carbon nanotube resonator for ultra-sensitive mass detection. Electron. Lett. 2008, 44 (18), 1060–U20. (11) Browne, M. L.; Varadarajulu, D.; Lewis-Michl, E. L.; Fitzgerald, E. F. Cancer incidence and asbestos in drinking water, Town of Woodstock, New York, 1980-1998. Environ. Res. 2005, 98 (2), 224–232. (12) Bales, R. C.; Newkirk, D. D.; Hayward, S. B. Chrysotile asbestos in California surface waterssfrom upstream rivers through water-treatment. JAWWA. 1984, 76 (5), 66–74. (13) Cantor, K. P. Drinking Water and Cancer. Cancer Causes Control 1997, 8 (3), 292–308. (14) Blumberg, I.; Starosvetsky, J.; Bilanovic, D.; Armon, R. TiO2 P-25 anatase rapid precipitation from water by use of struvite formation. J. Colloid Interface Sci. 2009, 336 (1), 107–110. (15) World Health Organization, Guidelines for Drinking-Water Quality, 3rd ed.; WHO: Geneva, 2008. (16) Westerhoff, P.; Zhang, Y.; Crittenden, C.; Chen, C. Properties of commercial nanoparticles that affect their removal during water treatment. In: Nanoscience and Nanotechnology; John Wiley & Sons, Inc.: Hoboken, NJ, 2008. (17) Zhang, Y.; Chen, Y. S.; Westerhoff, P.; Hristovski, K.; Crittenden, J. C. Stability of commercial metal oxide nanoparticles in water. Water Res. 2008, 42 (8-9), 2204–2212. (18) Holbrook, R. D., Mansfeld, C. B. Bott, C. B., Behavior and removal of multiwall carbon nanotubes during simulated drinking water treatment process. J. Environ. Eng., Submitted. (19) Hyung, H.; Kim, J. H. Dispersion of C60 in natural water and removal by conventional drinking water treatment processes. Water Res. 2009, 43 (9), 2463–2470.
(20) Klaine, S. J.; Alvarez, P. J. J.; Batley, G. E.; Fernandes, T. F.; Handy, R. D.; Lyon, D. Y.; Mahendra, S.; Mclaughlin, M. J.; Lead, J. R. Nanomaterials in the environment: behavior, fate, bioavailability, and effects. Environ. Toxicol. Chem./SETAC 2008, 27 (9), 1825–1851. (21) Holbrook, R. D.; Yen, J. H.; Grizzard, T. J. Characterizing natural organic material from the Occoquan watershed (northern Virginia, U.S.) using fluorescence spectroscopy and PARAFAC. Sci. Total Environ. 2006, 361 (1-3), 249–266. (22) Wilkinson, K. J.; Joz-Roland, A.; Buffle, J. Different roles of pedogenic fulvic acids and aquagenic biopolymers on colloid aggregation and stability in freshwaters. Limnol. Oceanogr. 1997, 42 (8), 1714–1724. (23) Tipping, E.; Marker, A. F. H.; Butterwick, C.; Collett, G. D.; Cranwell, P. A.; Ingram, J. K. G.; Leach, D. V.; Lishman, J. P.; Pinder, A. C.; Rigg, E.; Simon, B. M. Organic carbon in the Humber rivers. Sci. Total Environ. 1997, 194345–355. (24) Kopacek, J.; Hejzlar, J.; Borovec, J.; Porcal, P.; Kotorova, I. Phosphorus inactivation by aluminum in the water column and sediments: lowering of in-lake phosphorus availability in an acidified watershed-lake ecosystem. Limnol. Oceanogr. 2000, 45 (1), 212–225. (25) Urban, N. R.; Gorham, E.; Underwood, J. K.; Martin, F. B.; Ogden, J. G. Geochemical processes controlling concentrations of Al, Fe, and Mn in Nova-Scotia lakes. Limnol. Oceanogr. 1990, 35 (7), 1516–1534. (26) Hyung, H.; Fortner, J. D.; Hughes, J. B.; Kim, J. H. Natural Organic matter stabilizes carbon nanotubes in the aqueous phase. Environ. Sci. Technol. 2007, 41 (1), 179–184. (27) Filliben, J. J.; et. al. An EDA Approach to Experiment Design. In: NIST/SEMATECH e-Handbook of Statistical Methods. 2003. Available at http://www.itl.nist.gov/div898/handbook/pri/ section5/pri59.htm. (28) Smith, B.; Wepasnick, K.; Schrote, K. E.; Bertele, A. H.; Ball, W. P.; O’Melia, C.; Fairbrother, D. H. Colloidal properties of aqueous suspensions of acid-treated, multi-walled carbon nanotubes. Environ. Sci. Technol. 2009, 43 (3), 819–825. (29) Volk, C.; Bell, K.; Ibrahim, E.; Verges, D.; Amy, G.; Lechevallier, M. Impact of enhanced and optimized coagulation on removal of organic matter and its biodegradable fraction in drinking water. Water Res. 2000, 34 (12), 3247–3257. (30) Chadik, P. A.; Amy, G. L. Removing trihalomethane precursors from various natural-waters by metal coagulants. JAWWA 1983, 75 (10), 532–536. (31) Duan, J. M.; Gregory, J. Coagulation by hydrolysing metal salts. Adv. Coll. Inter. Sci. 2003, 100-102, 475–502. (32) Chowdhury, Z. K.; Amy, G. L.; Bales, R. C. Coagulation of submicron colloids in water-treatment by incorporation into aluminum hydroxide floc. Environ. Sci. Technol. 1991, 25 (10), 1766–1772. (33) Stumm, W.; O’Melia, C. R. Stoichiometry of coagulation. JAWWA 1968, 60514–539. (34) Han, Z. T.; Zhang, F. W.; Lin, D. H.; Xing, B. S. Clay minerals affect the stability of surfactant-facilitated carbon nanotube suspensions. Environ. Sci. Technol. 2008, 42 (18), 6869–6875. (35) Balcke, G. U.; Kulikova, N. A.; Hesse, S.; Kopinke, F. D.; Perminova, I. V.; Frimmel, F. H. Adsorption of humic substances onto kaolin clay related to their structural features. Soil Sci. Soc. Amer. J. 2002, 66 (6), 1805–1812. (36) Hyung, H.; Kim, J. H. Natural organic matter (Nnm) adsorption to multi-walled carbon nanotubes: effect of nom characteristics and water quality parameters. Environ. Sci. Technol. 2008, 42 (12), 4416–4421. (37) Kawamura, S. Effectiveness of natural polyelectrolytes in watertreatment. JAWWA 1991, 83 (10), 88–91. (38) Bernhardt, H.; Hoyer, O.; Schell, H.; Lusse, B. ReactionMechanisms Involved in the Influence of algogenic organicmatter on flocculation. Z. Was. Abwas. Forsch. 1985, 18 (1), 18–30. (39) Chen, K. L.; Mylon, S. E.; Elimelech, M. Aggregation kinetics of alginate-coated hematite nanoparticles in monovalent and divalent electrolytes. Environ. Sci. Technol. 2006, 40 (5), 1516–1523.
ES902946J
VOL. 44, NO. 4, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1391
