Grantee Research Project Results

2007 Progress Report: Fate and Transformation of C₆₀ Nanoparticles in Water Treatment Processes

EPA Grant Number: R832526
Title: Fate and Transformation of C₆₀ Nanoparticles in Water Treatment Processes
Investigators: Kim, Jae Hyung , Hughes, Joseph
Institution: Georgia Institute of Technology
EPA Project Officer: Packard, Benjamin H
Project Period: September 1, 2005 through December 31, 2008
Project Period Covered by this Report: September 1, 2006 through December 31, 2007
Project Amount: $375,000
RFA: Exploratory Research: Nanotechnology Research Grants Investigating Environmental and Human Health Effects of Manufactured Nanomaterials: A Joint Research Solicitation - EPA, NSF, NIOSH (2005) RFA Text |  Recipients Lists
Research Category: Nanotechnology , Human Health , Safer Chemicals

Objective:

B. WORK STATUS/PROGRESS

B.1. Overview

In Year 2 of the project, we continued to evaluate the interaction of natural organic matter (NOM) with multi-walled carbon nanotubes (MWNTs) and the mechanism of C₆₀ photochemical reactivity in the aqueous phase. These studies were based on our findings made during the first year of the project. Three manuscripts were submitted (currently under review) during this project year, all to ACS Environmental Science & Technology. The findings from these studies are presented in Chapters B.2, B.3, and B.4 (each chapter representing each manuscript submitted).

First, we investigated the effect of NOM characteristics and water quality parameters on NOM adsorption to MWNT (Chapter B.2). The isotherm experimental results fitted well with a modified Freundlich model that took into account of the heterogeneous nature of NOM. Accordingly, the preferential adsorption of the higher molecular weight fraction of NOM was observed by a size exclusion chromatographic analysis. Experiments performed with various NOM samples suggested that the degree of NOM adsorption varied greatly depending on the type of NOM and was proportional to the aromatic carbon content of NOM. The adsorption of NOM to MWNT was also dependent on water quality parameters: adsorption increased as pH decreased and ionic strength increased. As a result of NOM adsorption to MWNT, a fraction of MWNTs formed a stable suspension in water, the concentration of which depended on the amount of NOM adsorbed per unit mass of MWNT. The amount of MWNT suspended in water was also affected by ionic strength and pH. The findings in this study suggested that the fate and transport of MWNT in the natural system would be largely influenced by NOM characteristics and water quality parameters.

We also continue to explore the mechanism of photochemical reactivity of C₆₀ in the aqueous phase based on our findings made in the first year of the project. First, in Chapter B.3., we demonstrate that the degree of C₆₀ clustering in the aqueous phase is strongly dependent on the type and concentration of encapsulating agents such as surfactant, polymer and natural organic matter that interact with C₆₀. Degree of C₆₀ clustering was quantitatively analyzed using UV-Vis spectral characteristics. The dispersion status played a critical role in determining C₆₀’s photochemical reactivity, in particular, its ability to mediate energy transfer and produce singlet oxygen in the presence of oxygen. Consistent with findings in the organic phase, C₆₀ in the aqueous phase lost its intrinsic photochemical reactivity when they formed aggregates. Experiments performed using a laser flash photolysis suggested that the loss of reactivity resulted from drastic decrease in lifetime of key reaction intermediate, i.e. triplet state C₆₀. This study suggests that the photochemical reactivity of C₆₀ in the aqueous phase, which has been linked to oxidative damage in biological systems in earlier studies, is strongly dependent on the media environment surrounding C₆₀.

In work presented in Chapter B.4., The mechanism involved with energy and electron transfer by C₆₀ in the aqueous phase during UV irradiation and subsequent production of reactive oxygen species (ROS) such as singlet oxygen and superoxide radical anion was investigated. Electron paramagnetic resonance (EPR) study showed that C₆₀ embedded in micelles of nonionic surfactant (Triton X 100) or anionic surfactant (sodium dodecylbenzenesulfonate) produced ROS, but aggregated C₆₀ did not, consistent with our earlier findings made using indicator chemicals. Nanosecond and femtosecond laser flash photolysis showed that the aggregation of C₆₀ significantly accelerates the decay of excited triplet state C₆₀, which is a key intermediate for energy and electron transfer, thus blocking the pathway for ROS production. This finding suggests that C₆₀ clusters will not contribute to oxidative damage or redox reactions in natural environment and biological systems in the same way molecular C₆₀ in organic phase reportedly does. In contrast, C₆₀ embedded in surfactant micelles produces ROS and the evidence is presented for the formation of C₆₀ radical anion as an intermediate.

Progress Summary:

B.2. Natural Organic Matter (NOM) Adsorption to Multi-Walled Carbon Nanotubes: Effect of NOM Characteristics and Water Quality Parameters

B.2.1. Acknowledgements

This work was submitted to ACS Environmental Science & Technology and currently under review. The authors thank Dr. Jim Millette and Whitney Hill at MVA Scientific Consultants (Duluth, Georgia) for their assistance on TEM analysis. Dr. Vernon Snoeyink at University of Illinois at Urbana-Champaign is also thanked for reviewing the manuscript and providing valuable comments.

B.2.2. Introduction

As the evidence for toxicological effects of carbon nanotubes (CNT) is rapidly accumulating [1-3], understanding the fate and transport characteristics of CNT in the natural environment during unintended discharge is becoming an important issue. In particular, an exposure route involving natural waterways, which has traditionally not been considered as these molecules are extremely hydrophobic, has been receiving a widespread interest. Our recent study [4] showed that pristine multi-walled carbon nanotubes (MWNT) could be stabilized (suspended) in the aqueous phase by natural organic matter (NOM) which might provide sterically and electrostatically stable surfaces to MWNT after adsorption to the MWNT surface. This finding suggested that the dispersal of carbon based nanomaterials, CNT in particular, in the natural aquatic environment might occur to a higher extent than predicted based only on the hydrophobicity of these materials. In order to accurately predict the behaviors of MWNT in the environment, the mechanism of interaction between NOM and CNT and the effect of water quality on this interaction need to be elucidated.

NOM is a mixture of chemically complex polyelectrolytes with varying molecular weights, produced mainly from the decomposition of plant and animal residues [5]. Due to the carboxylic and phenolic moieties distributed throughout the entire molecule, NOM generally carries a negative charge in the natural environment [6]. These physical and chemical characteristics of NOM are likely to be closely related to the mechanism of NOM interaction with CNT. Compared to NOM adsorption onto CNT, the mechanism of NOM adsorption onto activated carbon is relatively well known due to the rich history of application to water treatment. A few characteristics of NOM interaction with activated carbon are noteworthy and might be helpful for interpretation of CNT-NOM interaction.

First, the adsorption capacity and strength strongly depend on the type of NOM and the type of activated carbon. Factors affecting adsorption have been reported to include size and chemical characteristics of NOM as well as pore structure and surface chemistry of activated carbon [5-8, 11]. Second, due to the polydisperse nature of NOM, different fractions of NOM tend to have a different degrees of adsorptive interactions with the adsorbent [7]. This preferential adsorption is reflected by the occurrence of dose dependent isotherm relationship. For example, the strongly adsorbable fraction of NOM exhibits a more favorable adsorption at lower activated carbon dose. Finally, NOM adsorption is affected by water quality parameters such as ionic strength and pH which influence the charge and configuration of NOM [6]. Specifically, the adsorption of negatively charged NOM to the activated carbon surface generally increases as ionic strength increases and pH decreases [5, 8-11].

Differences between activated carbon and CNTs need to be also recognized for the proper interpretation of CNT-NOM adsorption phenomena. First, the activated carbon consists of pores of different sizes which provide sites for NOM adsorption. CNTs in contrast provide adsorption sites only along the surface of a cylindrical structure [12]. Second, the chemical structure of activated carbon, which contains carbons of varying degree of saturation and oxidation state as well as functional groups formed during activation process [7], is fundamentally different from that of CNT, which consists only of globally conjugated unsaturated carbons in two dimensional arrays.

In this study, the effect of NOM characteristics on its adsorption to MWNT and the stability of MWNT in water was studied using NOMs obtained from different sources and extracted by different methods. The effect of water quality parameters such as ionic strength and pH was also assessed. Batch isotherm experiments were performed and the results were analyzed using the Freundlich adsorption isotherm equation which has been frequently used to describe aqueous phase adsorption. Preferential adsorption of different molecular size fractions of NOM was examined using size exclusion chromatography (SEC). Finally, the amount of stable MWNT suspension formed in water as a result of NOM adsorption under varying conditions was quantitatively analyzed.

B.2.3. Experimental

Materials. MWNT (10-20 nm diameter × 10-30 μm length) with over 95% purity was obtained from the Cheap Tubes Inc. (Brattleboro, VT). Suwannee River natural organic matter (SRNOM), Suwannee River humic acid standard II (SRHA), Suwannee River fulvic acid standard II (SRFA), Leonardite humic acid standard (LHA), Elliott soil humic acid standard (ESHA), Nordic lake humic acid reference (NLHA), Nordic lake fulvic acid reference (NLFA), Waskish peat humic acid standard (WPHA), and Waskish peat fulvic acid standard (WPFA) were purchased from the International Humic Substances Society (IHSS) (St. Paul, MN). Elemental and carbon compositions of the NOMs are provided in Table B.2.1. NOM stock solution was prepared by mixing a known amount of NOM with ultrapure water for 24 hours. Dissolution of NOM was facilitated by adding NaOH to increase the solution pH to 7. After measuring the total organic carbon (TOC) content of the stock solution by a TOC-Vw analyzer (Shimadzu, Columbia, MD), the target NOM concentration was prepared by diluting the stock solution. Ultrapure water produced by a Milli-Q water filtration system (Millipore, Billerica, MA) was used for the preparation of all the solutions.

Isotherm test. Isotherm relationships for NOM adsorption to MWNT were evaluated by a modified bottle-point technique [5], where each data point of the isotherm was determined by an individual batch experiment. Both constant adsorbent method and constant adsorbate method were adopted. For all the isotherm experiments, NOM solution was buffered with 1 mM phosphate (NaH₂PO₄) and ionic strength was adjusted with NaCl. The solution pH was adjusted using NaOH and HCl. The mixture of MWNT powder and NOM solution in a 40 mL vial was agitated using a magnetic stirrer for 6 days. Independent kinetic study suggested that the adsorption of NOM reached saturation within 2 days of mixing. After 2 days of settling, each suspension was filtered using a Whatman 541 filter (20-25 μm nominal pore size, Florham Park, NJ).

Analysis. The concentration of MWNT suspended in the filtrate was determined by visible light absorbance at 800 nm (VIS₈₀₀) (8453 UV-Vis Spectroscopy System, Agilent, Palo Alto, CA). Our past research verified that the absorbance at this wavelength (i.e., transmittance decrease due to light scattering by suspended MWNT) was linearly correlated with MWNT concentrations measured by a Thermal Optical Transmittance (TOT) analyzer [4]. That study also verified that the TOT method provided an accurate measure of the concentration of MWNT in a solution containing both nanotubes and NOM [4]. In this study, VIS₈₀₀ were also calibrated with those obtained using a TOT analyzer (Sunset Laboratory, Tigard, OR). The equilibrium concentration of NOM in the liquid phase, C_e (mg C/L), was measured by UV absorbance at 254 nm (UV₂₅₄) (8453 UV-Vis Spectroscopy System, Agilent, Palo Alto, CA) after removing MWNT with a 0.2 μm Acrodisc nylon membrane syringe filter (Pall Corporation, Ann Arbor, MI). UV₂₅₄ measurement was calibrated with TOC measurement (TOC-Vw, Shimadzu, Columbia, MD) for each NOM type. Once C_e was obtained, the equilibrium concentration of NOM adsorbed on the unit mass of MWNT, q_e (mg C/g MWNT), was obtained by the following equation:

(1)

where C₀ (mg C/L) = the initial concentration of NOM and D (g MWNT/L) = dosage of MWNT.

The molecular weight distribution of NOM was analyzed by a Hewlett-Packard 1100 high performance liquid chromatography (HPLC) system (Wilmington, DE) equipped with a Waters Protein-Pak™ 125 SEC column (Milford, MA), a commonly used column for NOM fractionation [13-15] (mobile phase = 0.1 M NaCl solution buffered with 1 mM phosphate at pH 6.8, flow rate = 1 mL/min at 40°C). UV absorbance at 254 nm (UV₂₅₄) was monitored by a diode-array detector (DAD). Electron microscopic images were analyzed by a Philips 120 transmission electron microscope (TEM) (New York, NY). A TEM specimen was prepared by placing a droplet of fullerene suspension on a copper carbon grid (Electron Microscopy Science, Hatfield, PA) and drying overnight at room temperature.

B.2.4. Results and Discussion

Adsorption Isotherm. The results obtained from isotherm experiments performed under varying initial concentrations of SRNOM and MWNT are shown in Figure B.2.1a. The isotherms were adsorbent-dose dependent, i.e., different linear isotherms were obtained at different MWNT doses. Similar adsorbent-dose dependent phenomenon has been observed for adsorption of SRNOM onto activated carbons in past studies [5, 10]. Each isotherm at the same MWNT dose fitted well with the following Freundlich adsorption model which has been commonly used to represent aqueous phase adsorption phenomena:

(2)

where K_F (L/mg) and n (dimensionless) represent Freundlich constant and Freundlich exponent, respectively. Generally, K_F increases as the adsorption capacity of the adsorbent increases and n increases as the adsorption strength increases. As the initial MWNT dose decreased, more SRNOM adsorbed per unit mass of MWNT and consequently q_e increased. This increase was less pronounced when equilibrium SRNOM concentration (C_e) increased, and the isotherms merged at the highest C_e.

Deviation from a single solute isotherm (i.e. a unique isotherm is obtained regardless of the initial adsorbate concentration or adsorbent dose) is attributed to the heterogeneity of NOM and consequential occurrence of preferential adsorption [16, 17]. In such a case, it is known that a unique isotherm is obtained by normalizing the equilibrium adsorbate concentration (C_e) by adsorbent dose (D) as follows [5, 13, 18, 19]:

(3)

A normalized Freundlich adsorption model fitted well with experimental data (Figure B.2.1b) regardless of MWNT dose. The experimental results obtained with the other NOMs used in this study also matched well with the normalized model (Figure B.2.2). The fitted model parameters for all the NOMs are summarized in Table B.2.2.

Effect of NOM Type. Results summarized in Table B.2.2. suggest that the adsorptive interaction between NOM and MWNT was strongly dependent on the type of NOM. For example, less soluble, higher molecular weight humic acids had generally higher adsorption capacity than fulvic acids. Various functionalities of the NOMs identified in the previous study [20] using ¹³C NMR were compared with adsorption characteristics. Among various carbon functionalities present in NOM (e.g., carbonyl, carboxyl, aromatic, acetal, heteroaliphatic, and aliphatic carbons), aromatic carbon content showed the most strong linear relationship with K_F (Figure B.2.3) regardless of source (i.e., lake, soil, or river) and type (i.e., bulk, fulvic or humic) of NOMs. Some deviation from the linearity, which might originate from the existence of different elemental composition and functional groups in NOMs, was also observed. Nevertheless, the finding that the adsorption capacity is closely related to the aromatic functional group content in NOM is consistent with past studies which reported that the attractive interaction between chemical compounds containing the aromatic moiety and CNT was largely driven by π-π interactions [21, 22]. Specifically, a previous study [21] suggested that a benzene ring present in a surfactant that shielded the CNT surface would stack upon benzene ring present in the CNT. Gotovac et al. [23] reported that the adsorption of the tetracene (4 benzene rings) was 6 times greater than that with phenanthrene (3 benzene rings). The strong correlation between adsorption capacity and aromatic content of NOM implies that aromatic fractions of NOM, which can range from ca. 10-40 % (C/C) depending on the source and age [20] could be a useful measure to evaluate the level of NOM adsorption onto MWNT and consequently the dispersion of MWNT in natural waters.

From the isotherm parameters calculated for each NOM (Table B.2.2.), it is also notable that K_F was inversely proportional to 1/n for all the NOMs tested in this study (Figure B.2.4). This suggests that the capacity (K_F) and adsorption strength (n) of NOM adsorption are proportional to each other, which is not always the case for NOM adsorption to activated carbons. In case of activated carbon, adsorption capacity is mainly determined by surface area of carbon’s porous structure available for adsorption, and many of the smaller pores are not accessible to the larger, more strongly adsorbing NOM molecules. Even for the same activated carbon, the area available for adsorption varies with molecular size distribution and characteristics of NOM. Such variation is not necessarily related to chemical composition of NOM that governs adsorption strength between the NOM and the surface. In contrast, MWNT do not have pores available for adsorption and, therefore, NOM adsorption is less influenced by the physical structure of the adsorbent. Consequently, NOMs with greater adsorption strength are likely to have greater adsorption capacity to MWNT as experimental results suggested.

Preferential Adsorption. Preferential adsorption of the higher molecular weight faction of SRNOM to MWNT was evident when size exclusion chromatograms of the non-adsorbed portion of NOM at different MWNT doses were examined (Figure B.2.5a). As MWNT dose was increased, higher molecular weight faction (i.e., fractions appearing at lower retention time) was removed to a greater extent. Similar phenomena were observed from the SEC analysis performed with other types of NOM such as SRHA and SRFA (Figures B.2.5b and B.2.5c). This observation is in accordance with previous studies performed with the polydispersed polymers and non-porous adsorbents where preferential adsorption of high molecular weight fractions was observed [16, 17]. However, the opposite molecular weight dependence was reported when adsorbent was porous (e.g., activated carbon). From the SEC study on the Laurentian humic acid adsorbed to activated carbon, Kilduff et al. [10] suggested that lower molecular weight fraction of the humic acid more favorably adsorbed to activated carbon than the higher molecular weight fraction. Summers and Roberts [6] examined activated carbon adsorption of Aldrich humic acid fractionated by molecular weight and observed greater adsorption of the low molecular weight fraction. The apparent preferential adsorption of the low molecular weight fraction in the case of activated carbon would result from the size exclusion effect of pores within the activated carbon structure (i.e., small adsorbates can access both small and large pores, but large adsorbates can not access small pores) [6]. However, when the adsorption was evaluated on the basis of the accessible surface area, the larger molecules were found to have greater adsorption capacity to activated carbon [5], which is consistent with findings from this study.

Effect of Water Quality Parameters. The adsorption of NOM to MWNT was greatly influenced by water quality parameters such as ionic strength and pH. At the same equilibrium SRNOM concentration in the liquid phase (C_e), the solid phase SRNOM concentration (q_e) was the highest in the solution containing 0.1 M NaCl and the lowest without NaCl (Figure B.2.6a). The experimental data fitted well with the Freundlich isotherm model at each ionic strength. Adsorption capacity, expressed in terms of K_F, increased as ionic strength increased (K_F = 6.72 at NaCl = 0 M, 7.05 at NaCl = 0.01 M and 7.55 at I = 0.1 M) presumably due to change in molecular configuration of NOM. Gosh and Schnitzer [24] suggested that humic substances would become increasingly coiled and form more compact structures as ionic strength increased. Adsorption capacity would consequently increase as more molecules could occupy the same surface area [10]. As the NOM molecule becomes more compact, the area for NOM-MWNT interaction would be reduced and the attractive force per individual NOM molecule to MWNT surface would decrease. Consequently, the adsorption strength expressed in terms of 1/n, decreased as ionic strength was increased (1/n = 0.39 at NaCl = 0 M, 0.48 at NaCl = 0.01 M and 0.60 at NaCl = 0.1 M). This is also consistent with the observation made in Figure B.2.5 that NOM with larger molecular size adsorbed more effectively to MWNT due to greater level of adsorptive interaction possible per NOM molecule with MWNT surface. At higher ionic strength, enhanced double layer compression in SRNOM-MWNT agglomerates would also enhance SRNOM adsorption onto MWNT.

Figure B.2.6b shows the effect of pH on SRNOM adsorption to MWNT. The adsorption capacity decreased as pH increased (i.e., K_F = 8.31 at pH = 5.0, 5.10 at pH = 7.0 and 2.96 at pH = 9.0). As pH increases, the NOM molecules will become less coiled and less compact due to greater charge repulsion, and the adsorption capacity might consequently decrease as discussed above. In addition, as pH increases, weakly acidic SRNOM with carboxylic and phenolic moieties becomes more negatively charged [25]. Thus, in higher pH, repulsion between SRNOM and MWNT surface coated with SRNOM would increase, hindering further adsorption of SRNOM. However, the strength of adsorption did not show appreciable change (i.e. 1/n = 0.39 at pH = 5.0, 0.40 at pH = 7.0 and 0.47 at pH = 9.0).

Stability of MWNT in the Aqueous Phase. As a result of NOM adsorption to MWNT, a fraction of MWNTs are dispersed in the aqueous phase and form a stable suspension. TEM images of MWNT in SRNOM solutions (Figure B.2.7) indicate that most of the MWNT were individually dispersed. The amount of stable MWNT suspension in aqueous phase (C_MWNT) was mainly determined by the amount of SRNOM adsorbed per MWNT (q_e) (Figure B.2.8). The MWNT suspension is facilitated by the shielding of extremely hydrophobic MWNT surface with NOM which provides thermodynamically more favorable surface. Adsorbed NOM is also expected to contribute to steric and electrostatic stabilization. Therefore, the amount of the NOM adsorbed onto MWNT would determine the extent of MWNT stability in water, i.e., C_MWNT increased as q_e increased. C_MWNT seemed to be also influenced by the type of NOM. For example, at the same q_e, C_MWNT with SRHA was higher than that with SRNOM, implying that SRHA might have greater MWNT stabilization capacity than SRNOM (Figure B.2.8). However, it was difficult to find a obvious relationship between C_MWNT and properties of NOM. For the same q_e, more MWNTs were suspended when more MWNTs were initially added to the solution (i.e., higher D).

The amount of MWNT suspended in water also strongly depended on solution ionic strength. The amount of MWNT suspended in water (C_MWNT) was plotted versus q_e at different ionic strengths in Figure B.2.9. Experimental results scattered at lower C_MWNT due to analytical limitations. At higher ionic strength, much less MWNT was suspended despite the similar amount of SRNOM was adsorbed. For instance, even at high q_e (> 10 mg C/g MWNT), MWNTs were only slightly dispersed (i.e., less than 1 mg/L) in the SRNOM solution with 0.1 M NaCl. This might be due to greater double layer compression in the high ionic strength solution. However, as MWNT used in this study has a very high aspect ratio (ca. diameter ranges from 10 to 20 nm and length ranges from 10 to 30 μm) it is difficult to quantitatively analyze the zeta potential of MWNT. The effect of pH was less obvious than that of ionic strength, although electrostatic stabilization of NOM-MWNT agglomerates would be more efficient due to the deprotonation of NOM at higher pH and experimental data showed slight increase in C_MWNT at higher pH for the same q_e.

Environmental Significance. The results of this study suggested that the environmental fate of MWNT will be largely influenced by the amount and the type of NOM as well as solution chemistry such as ionic strength and pH. Other water quality parameters such as divalent ions and inorganic composition, which were not examined in this study, might play a critical role in determining the degree of NOM-MWNT interaction. For natural waters from different sources with different characteristics, the aromatic content and molecular weight distribution of NOM might be useful parameters to predict the extent of NOM adsorption and level of MWNT dispersion. Even though a wide range of NOM concentrations (5 to 50 mg C/L) was investigated, it is noted that NOM concentrations in some natural surface and ground waters are lower than concentration levels used in this study. Therefore, further study with actual surface and ground waters with relatively low NOM contents might be necessary for a comprehensive understanding of interaction between NOM and MWNT.

In some aspects, NOM adsorption to MWNT was similar to that to activated carbon, i.e., occurrence of preferential adsorption and fitting to the Freundlich adsorption isotherm model. However, difference in physical structure of MWNT and activated carbon led to a few key differences in adsorption behavior such as preferential adsorption of higher molecular weight NOM onto MWNT. Understanding both similarities and differences between well-characterized activated carbon adsorption and MWNT adsorption should be helpful to understand not only the fate of MWNT in natural waters but also that of other carbon based nanomaterials such as single-walled carbon nanotubes and C₆₀.

B.2.5. Literature Cited

1. Kang, S.; Pinault, M.; Pfefferle, L. D.; Elimelech, M. Single-walled carbon nanotubes exhibit strong antimicrobial activity. Langmuir 2007, 23, 8670-8673.
2. Lam, C. W.; James, J. T.; McCluskey, R.; Hunter, R. L. Pulmonary toxicity of single-wall carbon nanotubes in mice 7 and 90 days after intratracheal instillation. Toxicol. Sci. 2004, 77, 126-134.
3. Warheit, D. B.; Laurence, B. R.; Reed, K. L.; Roach, D. H.; Reynolds, G. A.; Webb, T. R. Comparative pulmonary toxicity assessment of single-wall carbon nanotubes in rats. Toxicol. Sci. 2004, 77, 117-25.
4. 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, 179-184.
5. Summers, R. S.; Roberts, P. V. Activated carbon adsorption of humic substances.1. Heterodisperse mixtures and desorption. J. Colloid Interface Sci. 1988, 122, 367-381.
6. Summers, R. S.; Roberts, P. V. Activated carbon adsorption of humic substances. 2. Size exclusion and electrostatic interactions. J. Colloid Interface Sci. 1988, 122, 382-397.
7. Sontheimer, H.; Crittenden, J. C.; Summers, R. S., Activated Carbon for Water Treatment. DVGW-Forschungsstelle: Karlsruhe, Germany, 1988.
8. McCreary, J. J.; Snoeyink, V. L. Characterization and activated carbon adsorption of several humic substances. Wat. Res. 1980, 14, 151-160.
9. Randtke, S. J.; Jepsen, C. P. Effects of salts on activated carbon adsorption of fulvic acids. J. Am. Water Works Assoc. 1982, 74, 84-93.
10. Kilduff, J. E.; Karanfil, T.; Weber, W. J. Competitive interactions among components of humic acids in granular activated carbon adsorption systems: Effects of solution chemistry. Environ. Sci. Technol. 1996, 30, 1344-1351.
11. Letterman, L. D. Water Quality and Treatment. McGraw Hill: New York, 1999.
12. Yang, K.; Xing, B. S. Desorption of polycyclic aromatic hydrocarbons from carbon nanomaterials in water. Environ. Pollut. 2007, 145, 529-537.
13. Kilduff, J. E.; Karanfil, T.; Chin, Y. P.; Weber, W. J. Adsorption of natural organic polyelectrolytes by activated carbon: A size-exclusion chromatography study. Environ. Sci. Technol. 1996, 30, 1336-1343.
14. Li, Q. L.; Snoeyink, V. L.; Mariaas, B. J.; Campos, C. Elucidating competitive adsorption mechanisms of atrazine and NOM using model compounds. Wat. Res. 2003, 37, 773-784.
15. Her, N.; Amy, G.; Foss, D.; Cho, J.; Yoon, Y.; Kosenka, P. Optimization of method for detecting and characterizing NOM by HPLC-size exclusion chromatography with UV and on-line DOC detection. Environ. Sci. Technol.2002, 36, 1069-1076.
16. Stuart, M. A. C.; Scheutjens, J.; Fleer, G. J. Polydispersity effects and the interpretation of polymer adsorption-isotherms. J. Polym. Sci. B 1980, 18, 559-573.
17. Koopal, L. K. The effect of polymer polydispersity on the adsorption-isotherm. J. of Colloid Interface Sci. 1981, 83, 116-129.
18. Karanfil, T.; Kitis, M.; Kilduff, J. E.; Wigton, A. Role of granular activated carbon surface chemistry on the adsorption of organic compounds. 2. Natural organic matter. Environ. Sci. & Technol. 1999, 33, 3225-3233.
19. Harrington, G. W.; Digiano, F. A. Adsorption equilibria of natural organic-matter after ozonation. J. Am. Water Works Assoc. 1989, 81, 93-101.
20. Thorn, T. A.; Folan, D. W.; MacCarthy, P. Characterization of the International Humic Substances Society Standard and Reference Fulvic Acids by Solid State Carbon-13 and Hydrogen-1 Nuclear Magnetic Resonance Spectrometry; Water-Resources Investigation Report 89-4196; U.S. Geological Survey: Denver, CO, 1989.
21. Islam, M. F.; Rojas, E.; Bergey, D. M.; Johnson, A. T.; Yodh, A. G. High weight fraction surfactant solubilization of single-wall carbon nanotubes in water. Nano Lett. 2003, 3, 269-273.
22. Tan, Y.; Resasco, D. E. Dispersion of Single-Walled Carbon Nanotubes of Narrow Diameter Distribution. J. Phys. Chem. B 2005, 109, 14454-14460.
23. Gotovac, S.; Honda, H.; Hattori, Y.; Takahashi, K.; Kanoh, H.; Kaneko, K. Effect of nanoscale curvature of single-walled carbon nanotubes on adsorption of polycyclic aromatic hydrocarbons. Nano Lett. 2007, 7, 583-587.
24. Ghosh, K.; Schnitzer, M. Macromolecular structures of humic substances. Soil Sci. 1980, 129, 266-276.
25. Ritchie, J. D.; Perdue, E. M. Proton-binding study of standard and reference fulvic acids, humic acids, and natural organic matter. Geochim. Cosmochim.Acta 2003, 67, 85-96.

Figure B.2.1. Adsorption of SRNOM to MWNT. (a) Isotherm experimental data and Freundlich adsorption isotherm model fit. (b) Normalized data and normalized Freundlich adsorption isotherms model fit. (both at 22 °C, pH=7.0, [NaCl] = 5 mM, [NaH₂PO₄] = 1 mM)

Figure B.2.2. Isotherm experimental result and Freundlich adsorption model fit for (a) SRHA, (b) SRFA, (c) ESHA, (d) LHA, (e) NLHA, (f) NLFA, (g) WPHA, and (h) WPFA. (Isotherm experimental condition: pH 7, 5 mM NaCl, and 1 mM NaH₂PO₄)

Figure B.2.3. Relationship between the aromatic group content of NOM and NOM-MWNT adsorption capacity (K_F).

Figure B.2.4. Correlation between K_F and 1/n for various NOMs.

Figure B.2.5. Size exclusion chromatograms of (a) SRNOM (b) SRHA and (c) SRFA that were not adsorbed after adding MWNT at varying doses.

Figure B.2.6. Effect of (a) ionic strength (at 22 °C, pH 7.0, [NaH₂PO₄] = 1 mM) and (b) pH (at 22 °C, [NaCl] = 5 mM, [NaH₂PO₄] = 1 mM) on SRNOM adsorption to MWNT. Lines denote Freundlich adsorption isotherm model fit.

(a) (b)

Figure B.2.7. TEM images of MWNT stabilized in SR NOM shown at different magnifications. The bars represent 2 μm and 25 nm for (a) and (b), respectively. Evolution of bubble-like structures adjacent to MWNT surfaces, which formed as organic matter sublimed by high energy electron beam irradiation, was visible during the first a few seconds of TEM analysis (Figure 5b).

Figure B.2.8. Dependency of the amount of MWNT suspended in water (C_MWNT) on the amount of various NOMs adsorbed onto MWNT (q_e).

Figure B.2.9. Effect of (a) ionic strength (at 22 °C, pH 7.0, [NaH₂PO₄] = 1 mM) and (b) pH (at 22 °C, [NaCl] = 5 mM, [NaH₂PO₄] = 1 mM) on the amount of MWNT suspended in water.

Table B.2.1. Carbon Distribution and Elemental Composition of NOMs Investigated in This Study. Data Are Excerpted from International Humic Substances Society (IHSS) webpage (http://www.ihss.gatech.edu Exit ).

	Carbon Distribution							Elemental Composition
	Carbonyl	Carboxyl	Aromatic	Acetal	Heteroaliphatic	Aliphatic		H2O	Ash	C	H	O	N	S
SRHA	6	15	31	7	13	29		20.4	1.04	52.63	4.28	42.04	1.17	0.54
ESHA	6	18	50	4	6	16		8.2	0.88	58.13	3.68	34.08	4.14	0.44
LHA	8	15	58	4	1	14		7.2	2.58	63.81	3.7	31.27	1.23	0.76
SRFA	5	17	22	6	16	35		16.9	0.58	52.34	4.36	42.98	0.67	0.46
NLHA	10	19	38	7	11	15		9.1	0.31	53.33	3.97	43.09	1.16	0.58
WPHA	8	18	42	6	8	18		6.93	1.6	54.72	4.04	38.54	1.47	0.36
NLFA	10	24	31	7	12	18		9.2	0.45	52.31	3.98	45.12	0.68	0.46
WPFA	7	19	36	6	12	20		8.29	0.16	53.63	4.24	41.81	1.07	0.29
SRNOM	8	20	23	7	15	27		8.15	7	52.47	4.19	42.69	1.1	0.65

Table B.2.2. Freundlich Adsorption Model Parameters for Various NOMs.

	SRNOM	SRHA	SRFA	ESHA	LHA	NLHA	NLFA	WPHA	WPFA
1/n	0.3840	0.1998	0.3476	0.2247	0.2122	0.2223	0.3237	0.2247	0.2777
KF	5.47	9.82	5.45	11.34	13.08	11.63	6.94	12.42	8.44
r2	0.9702	0.8395	0.9692	0.9085	0.9004	0.9026	0.9652	0.9341	0.9029

B.3. Effect of Encapsulating Agents on Dispersion Status and Photochemical Reactivity of C₆₀ in the Aqueous Phase

B.3.1. Acknowledgements

This work was submitted to ACS Environmental Science & Technology and currently under review. The authors would like to thank Dr. Prashant Kamat, Dr. Anusorn Kongkanand, and Dr. Yoichiro Matsunaga for their assistance during laser flash photolysis, which was performed at Notre Dame Radiation Laboratory, South Bend, IN.

B.3.2. Introduction

With unique physical and chemical properties, C₆₀ has been widely recognized as a promising nanomaterial with various potential applications (1). One of notable characteristics is its strong photochemical reactivity: i.e. provided with photon energy above 2.3 eV (2), C₆₀ is readily excited with very high quantum yield (nearly 1.0) to a triplet state (³C₆₀*) which subsequently produces reactive oxygen species (ROS) such as singlet oxygen (¹O₂) and superoxide radical anion (O₂·^-) in the presence of oxygen (1, 3-9). C₆₀’s capability to generate ROS has been instrumental in biomedical applications such as enzyme inhibition, antiviral activity, and photodynamic therapy (9-12). Accordingly, various methods have been developed to effectively disperse extremely hydrophobic C₆₀ in the aqueous media and biological systems. These include: 1) generation of water-stable fullerene colloidal aggregates (often termed as nano-C₆₀ or nC₆₀) via ultrasonication (13) or stepwise solvent exchange (14, 15), 2) incorporation of fullerene into micelle/polymer/vesicle structures (4, 6, 11, 16-18), and 3) surface modification by addition of hydrophilic functional moieties (10, 19-21).

Strong photosensitizing activity and aqueous availability combined have been at the center of concerns in environmental and ecological impacts of this material. The exactly same property mentioned above (i.e. production of ROS) has been claimed responsible for oxidative damages and cell death in biological receptors (15, 22-24). In addition, following the similar mechanisms discussed above, dispersion of C₆₀ in the natural aqueous environment might be facilitated by: 1) formation of colloidal aggregates, which has been demonstrated possible after prolonged exposure to water (15, 25), 2) interaction with natural encapsulant analogues such as natural organic matter (NOM) and other anthropogenic contaminants such as surfactant that strongly interact with C₆₀, and 3) functionalization of C₆₀ by chemical reactions in natural environment or during water and wastewater treatment processes (e.g., chemical treatment). Therefore, it is possible that C₆₀ can be widely dispersed in the natural waterways, potentially interacting with biological systems, due to unintended events of mass discharge to the environment.

However, accurate assessment of environmental impact of this material should consider how these two properties are related to each other. Our recent study suggested that C₆₀’s photochemical reactivity (i.e. related to toxicological effect) would be dependent on its dispersion status in the aqueous phase (i.e. related to exposure route). Although C₆₀ associated with encapsulating agents such as surfactants or polymers retained the above intrinsic photochemical property (size of C₆₀ colloids < 5 nm), this property was lost when C₆₀ formed aggregates of size ca. 100 nm (7). Alternatively, varying levels of photochemical reactivity and ROS production might be possible with C₆₀’s that are released to the aqueous phase under different disposal scenarios, since dispersion status (i.e. degree of aggregation/clustering) of C₆₀ varies as it interacts with different encapsulating molecules present in the aqueous system (14, 15, 30). This study therefore examines the effect of a wide spectrum of both anthropogenic and natural encapsulants on the C₆₀’s dispersion status and photochemical reactivity at the same time. Degree of aggregation (clustering) was quantified based on UV-Vis spectral characteristics and compared to ¹O₂ production rate. A laser flash photolysis was performed to trace the decay kinetics of ³C₆₀* which is a critical transient intermediate for energy transfer and subsequent ¹O₂ production. Our study shows that the pathway for ¹O₂ production is prohibited when C₆₀ is present as an aggregate in the aqueous phase, since the lifetime of ³C₆₀* is drastically reduced, consiste nt with observations made in organic phase (1).

B.3.3. Experimental

Materials. C₆₀ (99.9%, sublimed) was purchased from MER Corp. Other chemicals used were of the highest purity available and used without the further treatment. Chemicals used as received in this study include: furfuryl alcohol (FFA, Aldrich), Na₂HPO₄·7H₂O (Aldrich), NaH₂PO₄·H₂O (Aldrich), sodium dodecyl sulfate (SDS, Sigma-Aldrich), sodium decyl sulfate (SDES, Fluka), sodium octyl sulfate (SOS, Fluka), sodium dodecylbezene sulfonate (SDBS, Aldirch), cetyltrimethylammonium bromide (CTAB, Sigma), dodecyltrimethylammonium bromide (DTAB, Sigma), octyltrimethylammonium bromide (OTAB, Sigma), Triton X100 (TX 100, Sigma-Aldrich), Triton X100 reduced (TX 100-R, Sigma-Aldrich), Triton X405 (TX 405, Sigma-Aldrich), Brij 35 (Sigma), Brij 78 (Sigma), Tween 65 (Sigma), sodium polyacrylate (PA, Fluka), polyvinylpyrrolidone (PVP, Sigma-Aldrich), polyethylene glycol (PEG, Fluka), γ-cyclodextrin (Sigma), tetrahydrofuran (THF, where), and toluene (Fisher Scientific). Suwannee River humic acid (SRHA) and fulvic acid (SRFA) isolated by RO were purchased from International Humic Substances Society (IHSS, St. Paul, MN). Ultrapure water (>18.2 MΩ) produced by a Milli-Q water purification system (Millipore, Billerica, MA) was used for the preparation of all solutions.

Preparation of aqueous stable C₆₀. Ultrasound (50/60 Hz, 125 W) was applied to a heterogeneous mixture of 10 mL toluene containing 5 mg C₆₀ and 90 mL ultrapure water in a sealed bottle for 24 hrs using an ultrasonicator (Model 8845-40, Cole-Parmer, US). The water phase gradually assumed brownish-orange hue as C₆₀ was dispersed. Ultrasound was further applied to the mixture open to atmosphere for additional 24 hrs at 60 °C in order to evaporate toluene and the solution was further filtered through a 0.45-μm PTFE filter (Millipore Corp.). Aqueous suspension of aggregate form of C₆₀ prepared according to this specific method is herein referred to as son/C₆₀. An alternative method based on solvent exchange (15) was also used to prepare aqueous suspension of C₆₀ aggregate (termed separately as nC₆₀) for comparison purpose. C₆₀ associated with target encapsulating agent, referred to as C₆₀/encapsulant, was prepared following the same method, except that varying concentration of encapsulating agent was dissolved in the aqueous phase prior to sonication. Note that the previously established preparation methods (4, 18) were not applicable in our study, since some target compounds (e.g., ionic surfactants) were negligibly soluble in organic solvents such as toluene and chloroform. However, C₆₀/TX 100 (above c.m.c.) and C₆₀/ γ-cyclodextrin prepared by our method was comparable to one prepared according to the methods described in (4, 7, 30) based on UV spectral analysis. Preliminary experiments also suggested that UV spectra of C₆₀ suspensions prepared by this method (C₆₀/SDS, C₆₀/SDBS, C₆₀/TX 100, and C₆₀/CTAB) were reproducible (see Table B.3.1. for abbreviations of chemical names).

Photochemical Experiments. Photochemical experiments were performed following the method previously described by Lee et al. (2006). Under air-equilibrated condition at ambient temperature (22 °C) in a 60 mL cylindrical quartz reactor which was surrounded by six 4-W black light blue (BLB, Philips TL4W) lamps. The incident light intensity in the active wavelength region (350 – 400 nm) was measured at 3.33 x 10^-4 Einstein×min^-1L^-1 by ferrioxalate actinometry. Production of ¹O₂ by aqueous colloidal C₆₀ during UV irradiation was monitored using FFA as a probe compound [k(FFA + ¹O₂) = 1.2 x 10⁸ M^-1s^-1 (31)]. An aliquot of 10 mM FFA stock solution was added to a reaction mixture to achieve target concentration of 1 mM. Initial concentration of aqueous C₆₀ suspension was fixed at 5 mg/L and reaction solutions were buffered at pH 7 using 10 mM phosphate. As the photochemical reaction proceeded, sample aliquots of 1 mL were withdrawn from the reactor using a syringe, filtered through a 0.45-μm PTFE filter (Millipore), and transferred to a 2-mL amber glass vial for further analyses. The FFA concentration was measured using an Agilent 1100 HPLC equipped with a C-18 column (Agilent Zorbax RX-C18) and a diode-array detector. Each photochemical experiment was duplicated.

Laser Flash Photolysis. Prior to experiment, a solution (3 mL) containing 5 mg/L C₆₀ was placed in a rectangular quartz reactor, purged with argon gas for 30 min and sealed to inhibit the transfer of absorbed energy from ³C₆₀* to oxygen. A laser pulse at 355 nm (5 mJ, pulse width = 6 ns) generated from a Quanta Ray Nd:YAG laser system was used as an excitation source for C₆₀. A monochromatic laser at 740 nm, produced by filtering the light from a xenon lamp using a monochromator and aligned perpendicular to the excitation laser, was used to monitor the concentration of ³C₆₀* at 740 nm. Each laser flash photolytic experiment was duplicated.

B.3.4. Results and Discussion

UV-Vis Spectral Characteristics of C₆₀ associated with Surfactants. UV-Vis analysis of C₆₀ associated with selected surfactant suggests that spectral characteristics that reflect dispersion status of C₆₀ are strongly affected by the presence of surfactants (30) (Figure B.3.1). Spectra of son/C₆₀ and nC₆₀ both showed a specific absorption peak at around 350 nm and broad band absorption in the wavelength region from 400 to 500 nm, consistent with the previous literature (14, 15). However, C₆₀ associated with TX 100, a non-ionic surfactant widely used as an artificial cell membrane (4, 6, 30), showed a sharp absorption peak at 330 nm (i.e. blue-shifted compared to that in son/C₆₀ and nC₆₀) and negligible absorption in the wavelength ranges from 400 to 500 nm. This spectrum is fairly similar with that of C₆₀ molecularly dissolved in toluene or hexane (30). The spectrum of C₆₀ further depended on the type and amount of surf

Future Activities:

C.5. Next Phase Plan

In year 3, we will continue to study the response of aqueous stable C₆₀ to UV irradiation, especially at lower wavelength to verify the potential for direct photolysis of C₆₀. Experiments are currently performed to evaluate the fate of C₆₀ colloids in conventional water treatment processes.

Journal Articles on this Report : 3 Displayed | Download in RIS Format

Publications Views

Other project views:	All 20 publications	10 publications in selected types	All 10 journal articles

Publications

Type	Citation	Project	Document Sources
Journal Article	Fortner JD, Kim DI, Boyd AM, Falkner JC, Moran S, Colvin VL, Hughes JB, Kim J-H. Reaction of water stable C60 aggregates with ozone. Environmental Science & Technology 2007;41(21):7497-7502.	R832526 (2007) R832526 (Final)	Abstract from PubMed
Journal Article	Hyung H, Fortner JD, Hughes JB, Kim J-H. Natural organic matter stabilizes carbon nanotubes in the aqueous phase. Environmental Science & Technology 2007;41(1):179-184.	R832526 (2006) R832526 (2007) R832526 (Final)	Abstract from PubMed
Journal Article	Lee J, Fortner JD, Hughes JB, Kim J-H. Photochemical production of reactive oxygen species by C60 in the aqueous phase during UV irradiation. Environmental Science & Technology 2007;41(7):2529-2535.	R832526 (2007) R832526 (Final)	Abstract from PubMed

Supplemental Keywords:

Health, RFA, Scientific Discipline, PHYSICAL ASPECTS, Water, Health Risk Assessment, Physical Processes, Risk Assessments, Environmental Chemistry, Engineering, Chemistry, & Physics, Biochemistry, Drinking Water, epidemelogy, community water system, health effects, toxicity, toxicokinetics, nanomaterials, water quality, human exposure, engineered nanomaterials, ambient particle health effects, exposure, nanotechnology, drinking water system, drinking water contaminants, fate and transport, other - risk assessment, cellular responses, human health risk, human health effects, particle exposure, biochemical research

Progress and Final Reports:

Original Abstract

2006 Progress Report

2008

Final Report