Adsorption of phenanthrene, 2-naphthol, and 1-naphthylamine to colloidal oxidized multiwalled carbon nanotubes: Effects of humic acid and surfactant modification

Authors

  • Lei Hou,

    1. College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    Search for more papers by this author
  • Dongqiang Zhu,

    1. State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Jiangsu, People's Republic of China
    Search for more papers by this author
  • Ximeng Wang,

    1. College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    Search for more papers by this author
  • Lilin Wang,

    1. College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    Search for more papers by this author
  • Chengdong Zhang,

    1. College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    Search for more papers by this author
  • Wei Chen

    Corresponding author
    1. College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    • College of Environmental Science and Engineering, Ministry of Education Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, Tianjin, People's Republic of China
    Search for more papers by this author

Abstract

Carbon nanotubes (CNTs) can exist in the form of colloidal suspension in aquatic environments, particularly in the presence of natural organic matter or surfactants, and may significantly affect the fate and transport of organic contaminants. In the present study, the authors examined the adsorption of phenanthrene, 2-naphthol, and 1-naphthylamine to three colloidal CNTs, including a stable suspension of oxidized multiwalled carbon nanotubes (O-MWNT), a humic acid (HA)-modified colloidal O-MWNT, and a sodium dodecyl sulfate (SDS)-modified colloidal O-MWNT. All three colloidal O-MWNTs exhibit strong adsorption affinities to the three test compounds (with KOC values orders of magnitude greater than those of natural organic matter), likely resulting from strong nonhydrophobic interactions such as π–π electron donor–acceptor interactions and Lewis acid–base interactions. When thoroughly mixed, HA (at ∼310 mg HA/g CNT) and SDS (at ∼750 mg SDS/g CNT) significantly affected the aggregation properties of O-MWNT, causing individually dispersed tubes to form a loosely entangled network. The effects of HA or SDS modification on adsorption are twofold. Adsorption of HA/SDS significantly reduces surface areas of O-MWNT; however, the entangled network allows adsorbate molecules to interact simultaneously with multiple tubes. An important implication is that humic substances and surfactant-like materials not only facilitate the formation of colloidal carbon nanoparticles but also affect how these colloidal carbon nanoparticles adsorb organic contaminants. Environ. Toxicol. Chem. 2013;32:493–500. © 2012 SETAC

INTRODUCTION

Carbon nanotubes (CNTs) are produced and used in increasingly large quantities worldwide, and the release of these engineered carbon nanomaterials into the environment is inevitable 1–3. A number of studies have shown that CNTs can interact strongly with organic contaminants (e.g., polycyclic aromatic hydrocarbons, chlorinated benzenes, antibiotics), due to their strong hydrophobic nature and their unique graphitic structures that allow specific π-electron-related interactions 4–8. Thus, it is likely that CNTs released into the environment can significantly affect the partitioning and fate of organic contaminants.

Carbon nanotubes can exist in colloidal forms (i.e., stable suspensions of CNTs) in aqueous solutions, depending on their type, surface properties, and coating, as well as water chemistry. Because of their stable nature and relatively high mobility 9, 10, these colloidal CNTs are of the greatest environmental significance. Stable suspensions of CNTs can be prepared by introducing O-containing functional groups to the surfaces of CNTs (e.g., using strong oxidizing agents) 11 and by surface-coating CNTs with dispersants (e.g., surfactants and polymers) 12. Stable suspensions of CNTs can also form in natural waters, in particular, in the presence of natural organic matter (NOM) and/or surfactant-type materials 9. For example, Hyung et al. 9 reported that Suwannee River NOM can effectively stabilize CNTs in water, an observation made by vigorously agitating CNTs in actual Suwannee River water.

Compared with their noncolloidal counterparts, colloidal CNTs may exhibit different aggregation properties 9, 11, 12 and surface properties. For example, it has been reported that while CNTs are prone to aggregation, a portion of the tubes in the stable suspensions of CNTs are suspended as individual tubes 9, 11. Such unique properties associated with colloidal CNTs may have significant effects on CNT–contaminant interactions 13 because adsorption affinities of organic contaminants to CNTs is strongly affected by the surface area, pore structures, and surface chemistry of CNTs 4, 5, 8. Even though this has not been tested systematically, insights can be drawn from related work on C60 in that stable C60 aggregates behave very differently from C60 powder in terms of adsorption properties 14, 15. Furthermore, different organic molecules may respond to the morphological changes of CNTs differently in terms of adsorptive interactions.

In the present study, we examined adsorption of phenanthrene, 2-naphthol, and 1-naphthylamine to three types of colloidal CNTs, including a colloidal form of surface oxidized multiwalled carbon nanotubes (O-MWNT), a humic acid (HA)–modified colloidal O-MWNT, and a surfactant-modified colloidal O-MWNT. The noncolloidal form of O-MWNT was used as a comparative adsorbent. Physicochemical properties of the four adsorbents were characterized. The adsorption affinities of the test adsorbates to the colloidal CNTs were examined and compared with those to the noncolloidal form. The effects of HA and surfactant modification on the properties of colloidal CNTs and on colloidal CNT–contaminant interactions were analyzed.

MATERIALS AND METHODS

Materials

Pristine MWNTs were purchased from Nanotech Port. Based on the information provided by the manufacturer, the product was synthesized with a chemical vapor deposition method using cobalt, manganese, molybdenum, and nickel as the catalysts. It contains more than 95% carbon nanotubes and less than 5% impurities, mainly amorphous carbon (3%), as well as trace amounts of metals. The sizes of the outer diameter range from 10 to 30 nm and the length from 5 to 15 µm. Oxidized MWNT was obtained using an acid oxidation method 10. The pristine MWNT was first pretreated with concentrated HCl for 1 h to remove amorphous carbon and metal catalysts and then treated with mixed HNO3 and H2SO4 (1:3 by volume) to increase the amount of surface O-functional groups. The obtained product is referred to as O-MWNT.

Sodium dodecyl sulfate (SDS) was purchased from Sigma Chemical. Leonardite HA was obtained from the International Humic Substances Society and is reported to be composed of 63.81 weight % carbon, 3.70 weight % hydrogen, 31.27 weight % oxygen, and a small amount of nitrogen, sulfur, and phosphate.The distribution of functional groups was carboxylic (27%), aromatic (58%), aliphatic (14%), and heteroaliphatic (1%). Phenanthrene (>98%), 2-naphthol (>99%), and 1-naphthylamine (>99.5%) were purchased from Sigma-Aldrich.

Preparation of colloidal O-MWNTs

A stable suspension of O-MWNT was prepared using the methods reported by Han et al. 12. First, approximately 200 mg O-MWNT was added to 1 L 0.001 M NaCl and sonicated at 100 W (Vibra-Cell VCX800; Sonics & Material) for 30 min. Afterward, the suspension was centrifuged at 10,000 rpm for 15 min. Then, the upper 75 to 80% of the supernatant was carefully withdrawn and stored. This colloidal O-MWNT suspension is referred to as O-MWNT-colloid. Two more colloidal O-MWNT suspensions were prepared in a similar way (see Supplemental Data for detailed procedures), but the background electrolyte solution (0.001 M NaCl) was replaced with a 60 mg/L HA solution or a 300 mg/L SDS solution (the pH of the solution was adjusted to neutral using 0.1 M HCl, if needed). Additionally, for the HA- and SDS-modified suspensions, the collected supernatants were dialyzed using dialysis bags of 6,000 to 8,000 molecular weight cutoff to remove the freely dissolved HA or SDS from the colloidal O-MWNT suspensions, using the method reported by Li et al. 16. These HA- or SDS-modified suspensions are referred to as O-MWNT/HA-colloid and O-MWNT/SDS-colloid, respectively.

The concentrations of O-MWNT in the colloidal O-MWNTs (i.e., O-MWNT-colloid, O-MWNT/HA-colloid, and O-MWNT/SDS-colloid) were determined using the method of Hyung et al. 9, by examining absorbance at 800 nm, with an UV/vis spectrophotometer (UV-2401; Shimadzu Scientific Instruments). The amount of HA in O-MWNT/HA-colloid or the amount of SDS in O-MWNT/SDS-colloid was calculated by subtracting the concentration of O-MWNT from the total organic carbon concentration of the suspension, determined with a high-sensitivity total organic carbon analyzer (Shimadzu Scientific Instruments). The colloidal O-MWNTs were stable (the concentrations changed little) over the entire course of the study.

Characterization of adsorbents

Bulk elemental analysis of the four adsorbents (i.e., O-MWNT, O-MWNT-colloid, O-MWNT/HA-colloid, and O-MWNT/SDS-colloid) was performed using an Elementar elemental analyzer. The surface elemental compositions of the adsorbents were analyzed using X-ray photoelectron spectroscopy (Physical Electronics, see Supplemental Data for detailed procedures). Surface areas of the adsorbents were calculated from nitrogen adsorption and desorption isotherms by multipoint adsorption isotherms of N2 at 77 K in the region of 10−7 to 1 relative pressure using the Brunauer-Emmett-Teller (BET) method. Morphology of the adsorbents was examined with a JEOL-2010 transmission electron microscope (TEM), and the samples were prepared by air-drying a drop of suspension onto a copper TEM grid (Electron Microscopy Sciences). Hydrodynamic diameters of the three colloidal O-MWNTs were measured by dynamic light scattering 11, and ζ potentials were measured by electrophoretic mobility at 25°C using a ZetaPALS (Brookhaven Instruments).

Adsorption experiments

Adsorption isotherms of phenanthrene, 2-naphthol, and 1-naphthylamine to the noncolloidal O-MWNT were obtained using a previously developed method 4, 5. First, a series of 40-ml aluminum foil-wrapped amber glass vials each containing 5 mg O-MWNT and 40 ml electrolyte solution (0.001 M NaCl, pH 7.0 ± 0.2) were prepared. Then, different amounts of phenanthrene, 2-naphthol, or 1-naphthylamine stock solution (in methanol) were added to the vials using a microsyringe; and the volume percentage of methanol was kept below 0.1% to minimize cosolvent effects. The vials were filled with the electrolyte solution immediately to leave minimal headspace and then tumbled end-over-end at 1 rpm for 7 d to reach adsorption equilibrium 4, 5. Afterward, the vials were left undisturbed on a flat surface for more than 24 h to allow complete settlement of the carbon nanotubes, and the supernatant was withdrawn to analyze the concentrations of the adsorbates 4, 17. The adsorbed mass at each equilibrium concentration was calculated as the difference between the total mass and the mass in the dissolved phase.

The adsorption isotherms to the three colloidal O-MWNTs were obtained using the literature methods 18, 19. First, a series of dialysis bags (6,000–8,000 molecular weight cutoff) each containing 5 ml of 0.001 M NaCl solution were put in 40-ml amber glass vials containing approximately 35 ml of a colloidal O-MWNT suspension. The concentration of O-MWNT in the suspension was 67.1 mg/L for O-MWNT-colloid, 66.5 mg/L for O-MWNT/HA-colloid, and 63.7 mg/L for O-MWNT/SDS-colloid. Then, different amounts of phenanthrene, 2-naphthol, or 1-naphthylamine stock solution were spiked into the colloidal O-MWNT suspension, and the volume percentage of methanol was kept below 0.1% to minimize cosolvent effects. Afterward, the vials were filled with the respective O-MWNT suspension, sealed, and tumbled for 7 d to allow adsorption equilibrium. Then, the concentrations of the adsorbate inside of the dialysis bags were measured. These concentrations represent the freely dissolved concentrations of the adsorbate, and the mass in the adsorbed phase was calculated based on mass balance.

To take into account the loss of an adsorbate from processes other than adsorbent adsorption (e.g., adsorption to dialysis bag, septum, and glass vial glassware), calibration curves were obtained separately from controls receiving the same treatment as the adsorption samples but no adsorbent 20. Calibration curves included at least seven standards over the test concentration range. Based on the obtained calibration curves, the adsorbed mass of organic adsorbates was calculated by subtracting the mass in the aqueous phase from the mass spiked. Adsorption isotherm data points were run in duplicate. The blank control samples of phenanthrene, 2-naphthol, and 1-naphthylamine showed no degradation during the adsorption experiments, and pH remained constant during all of the experiments.

Analytical methods for adsorbates

The concentrations of phenanthrene, 2-naphthol, and 1-naphthylamine were determined using a Waters high-performance liquid chromatography system equipped with a symmetry reversed-phase C18 column (4.6 × 150 mm). Phenanthrene was detected with a Waters 2475 fluorescence detector at an excitation wavelength of 250 nm and an emission wavelength of 364 nm; the mobile phase was methyl cyanide–deionized water (80:20, v:v; 1.0 ml/min). Both 2-naphthol and 1-naphthylamine were detected with a Waters 2487 UV/visible detector at wavelengths of 328 and 305 nm, respectively; the mobile phase was methyl cyanide–deionized water (50:50, v:v, for 2-naphthol and 60:40, v:v, for 1-naphthylamine; 1.0 ml/min). No peaks were detected in the spectra for potential degraded/transformed products of the test compounds.

RESULTS AND DISCUSSION

Characterization of colloidal O-MWNTs

The elemental compositions of O-MWNT, O-MWNT-colloid, O-MWNT/HA-colloid, and O-MWNT/SDS-colloid are compared in Table 1. The X-ray photoelectron spectroscopic results indicate that O-MWNT contains 9.3% surface oxygen, and our previous studies indicate that a considerable amount of surface acidic groups (mainly phenolic and carboxyl groups) can be introduced to pristine CNTs using this mixed acid-oxidation treatment 20. The elemental compositions of O-MWNT-colloid are similar to those of O-MWNT, indicating that the preparation process, especially sonication, has no significant effects on the surface functionalities of O-MWNT. On HA or SDS modification, surface oxygen content increased considerably to 17.2% for O-MWNT/HA-colloid and 19.1% for O-MWNT/SDS-colloid. This is consistent with the measured concentrations of HA (20.8 mg/L) and SDS (48.0 mg/L) in these two colloidal suspensions (Table 1). Note that both O-MWNT/HA-colloid and O-MWNT/SDS-colloid were dialyzed to remove the freely dissolved HA or SDS. Thus, the residual HA or SDS in the respective suspension is likely adsorbed to O-MWNT.

Table 1. Characterization of O-MWNT, colloidal suspension of O-MWNT, and humic acid and sodium dodecyl sulfate–modified colloidal suspensions of O-MWNTs
Adsorbent/colloidal suspensionConc. of O-MWNT in suspension (mg/L)Conc. of HA/SDS in suspension (mg/L)Elemental composition (%)BET surface areac (m2/g)ζ potential (mV)
BulkaSurfaceb
CHNCNOS
  • a

    Analyzed with CHN elemental analyzer.

  • b

    Analyzed with X-ray photoelectron spectroscopy.

  • c

    Surface area determined by N2 adsorption using the Brunauer–Emmett–Teller (BET) method.

  • O-MWNT = oxidized multiwalled carbon nanotube; HA = humic acid; SDS = sodium dodecyl sulfate; nd = not detected.

O-MWNT88.811.030.2089.700.809.300.20114
O-MWNT-colloid67.187.412.43nd90.300.708.800.20251–35.50
O-MWNT/HA-colloid66.520.863.334.431.5880.801.5017.200.4059–37.14
O-MWNT/SDS-colloid63.748.061.154.45nd74.500.9019.104.6069–26.19

Figure 1 shows representative TEM images of the three colloidal O-MWNTs, and a TEM image of the noncolloidal O-MWNT is included as a comparison. Interestingly, the TEM image of O-MWNT-colloid shows individually dispersed tubes, an observation consistent with the literature 11. However, for O-MWNT/HA-colloid and O-MWNT/SDS-colloid, individual tubes are entangled, forming a complex, loosely coiled network (compared with that of the noncolloidal O-MWNT) of individual tubes. The morphological differences among the three colloidal O-MWNTs shown by the TEM images are consistent with the estimated average hydrodynamic diameters, which are 123 nm for O-MWNT-colloid, 184 nm for O-MWNT/HA-colloid, and 202 nm for O-MWNT/SDS-colloid. Note that even though HA or SDS modification had significant effects on the morphology of the colloidal O-MWNTs, it did not significantly enhance the concentration of O-MWNT in the colloidal O-MWNTs (Table 1). This was likely due to the strong suspension capability of O-MWNT itself.

Figure 1.

Transmission electron microscopic images of oxidized multiwalled carbon nanotube (O-MWNT) (A), O-MWNT-colloid (B), O-MWNT/humic acid [HA]-colloid (C), and O-MWNT/sodium dodecyl sulfate [SDS]-colloid (D).

In the TEM images of O-MWNT/HA-colloid, opaque and amorphously shaped patches can be seen among individual tubes. Based on the literature 21, these patches should be HA-coated or bonded on O-MWNT. It has been proposed that dissolved organic matter can interact with the graphitic surfaces of CNTs via van der Waals forces and π–π stacking, and with the surface functional groups of CNTs via polar interactions (mainly H-bonding) and electrostatic forces 17. The concentration of SDS (300 mg/L) involved in the formation of O-MWNT/SDS-colloid is only approximately one-eighth of the critical micelle concentration of SDS (2,382 mg/L), and thus, the interaction between SDS and O-MWNT is likely dominated by the hydrophobic interactions between the hydrophobic chains of SDS molecules and the surface of O-MWNT, due to strong van der Waals attraction 22. Furthermore, Schwyzer et al. 23 proposed that surfactants have the same dispersing effects on CNTs as NOM, which might explain the similarities in the TEM images between O-MWNT/HA-colloid and O-MWNT/SDS-colloid.

The BET surface area (Table 1) of O-MWNT-colloid is over twice as high as that of the noncolloidal form, further indicating that the colloidal O-MWNT is well dispersed. However, the surface areas of O-MWNT/HA-colloid and O-MWNT/SDS-colloid are markedly lower (only 24–27% of the surface area of O-MWNT-colloid) and even lower than the surface area of the noncolloidal O-MWNT. The much reduced surface areas from HA or SDS modification were possibly caused by the coating of HA molecules or coiled SDS molecules to the surfaces of individual tubes or pore blockage 17.

The three colloidal O-MWNTs were stable under the experimental conditions of the present study for over one month. This is likely attributable to the negative surface charge of these colloidal suspensions (see the ζ potential values in Table 1). The negative surface charges not only enhance the surface hydrophilicity of the nanoparticles but also prevent the aggregation of the nanoparticles via electrostatic repulsion.

Adsorption affinities of colloidal O-MWNTs

The adsorption isotherms of phenanthrene, 2-naphthol, and 1-naphthylamine to O-MWNT, O-MWNT-colloid, O-MWNT/HA-colloid, and O-MWNT/SDS-colloid are shown in Figures 2 and 3. The adsorption data were fitted with the Freundlich sorption model: q = KF × Cn, where q (mmol/kg) and C (mmol/L) are the equilibrium concentrations of an adsorbate to the adsorbent and in the solution, respectively; KF (mmol1–nLn/kg) is the Freundlich affinity coefficient; and n (unitless) is the Freundlich linearity index. The fitted Freundlich model parameters are summarized in Supplemental Data, Table S1, and the curve-fitting results are shown in Supplemental Data, Figure S1. In general, the Freundlich model fits the adsorption data reasonably, and most of the isotherms are highly nonlinear, indicating heterogeneous distributions of adsorption sites, but no apparent trends of linearity can be generalized based on either chemical properties or properties of the adsorbents.

Figure 2.

Adsorption isotherms of phenanthrene, 2-naphthol, and 1-naphthylamine to the noncolloidal oxidized multiwalled carbon nanotube (O-MWNT) and colloidal O-MWNTs (O-MWNT-colloid, O-MWNT/humic acid [HA]-colloid, and O-MWNT/sodium dodecyl sulfate [SDS]-colloid).

Figure 3.

Surface area–modified adsorption isotherms of phenanthrene, 2-naphthol, and 1-naphthylamine to the noncolloidal oxidized multiwalled carbon nanotube (O-MWNT) and colloidal O-MWNTs (O-MWNT-colloid, O-MWNT/humic acid [HA]-colloid, and O-MWNT/sodium dodecyl sulfate [SDS]-colloid). q' = q/(Brunauer-Emmett-Teller [BET] surface area).

The adsorption affinities of a given adsorbate to the four different forms of O-MWNT are compared in Figure 2. For O-MWNT/SDS-colloid and O-MWNT/HA-colloid, the q values were calculated based on the mass of O-MWNT, rather than the total mass of O-MWNT and HA/SDS, because the adsorption affinities of the test adsorbates to HA is negligible compared with those to O-MWNT 13, 24. For phenanthrene, adsorption affinities follow the order O-MWNT > O-MWNT-colloid > O-MWNT/SDS-colloid > O-MWNT/HA-colloid (as indicated by the relative positions of the isotherms in Fig. 2, i.e., by the differences in the q values at a given C value), but the difference between O-MWNT and O-MWNT-colloid is small. For 2-naphthol and 1-naphthylamine, adsorption affinities to O-MWNT and O-MWNT-colloid are similar, and adsorption affinities to O-MWNT/SDS-colloid and O-MWNT/HA-colloid are similar but weaker than those to O-MWNT or O-MWNT-colloid (the differences are approximately half orders of magnitude). Thus, for all three test adsorbates, a general trend can be established, that is, O-MWNT-colloid exhibits similar adsorption affinities to the noncolloidal form of O-MWNT, but SDS or HA modification decreases the adsorption affinities noticeably (see Fig. 2). It has been reported that HAs can inhibit the adsorption of organic adsorbates to CNTs via mechanisms such as competitive adsorption, pore blockage, and molecular sieving 17, 25. The observed effects in the literature (∼29–57% decrease in distribution coefficient [Kd] by Chen et al. 17 and ∼3.5–12% decrease in Kd by Wang et al. 25), however, are in general less significant than that shown in Figure 2. This is likely because, in the literature, studies adsorption of HAs to CNT aggregates was done by mild mixing, and HA molecules are excluded from a large fraction of the interstitial spaces formed by individual tubes, leaving a significant amount of surface area uncovered and available for the adsorption of small adsorbate molecules 17. However, in the present study HA or SDS was thoroughly mixed with O-MWNT by sonication, which likely resulted in more significant coverage of the surface areas of individual tubes by HA or SDS. This seems to be in line with the reduced BET surface area of O-MWNT/HA-colloid and O-MWNT/SDS-colloid (Table 1).

It is worth noting that the adsorption affinities of the three test adsorbates to the colloidal forms of O-MWNTs are much greater than those to natural sorbents (e.g., soils and NOM), and such high adsorption affinities have important environmental implications. For example, the log KOC (organic carbon-normalized distribution coefficient; see Table 2) values of phenanthrene to O-MWNT-colloid range from 5.38 to 6.68, which are over one to two orders of magnitude greater than the average log KOC value of 4.18 to NOM, estimated using the correlation of Schwarzenbach et al. 26. The difference between the adsorption affinities to the colloidal O-MWNTs and to NOM (e.g., HAs) is even more significant for the two polar aromatics (2-naphthol and 1-naphthylamine). Depending on the equilibrium concentration, the observed log KOC values of these two compounds to the colloidal O-MWNTs (Table 2) are as high as 4.53 for 2-naphthol and 7.15 for 1-naphthylamine, several orders of magnitude greater than the respective log KOC value to a peat HA (2.60 for 2-naphthol and 2.85 for 2-naphthylamine) reported by Novoszad et al. 24. In our previous studies 5, we found that adsorption of aromatics and hydroxyl-/amino-substituted aromatics to CNTs is strongly affected by the π-electron interactions between adsorbate molecules and the π-electron network of the CNT graphitic surfaces, rendering much greater adsorption affinities than to natural organic sorbents. The superior adsorption affinities of colloidal CNTs for aromatics and substituted aromatics, as demonstrated in the present study, indicate that colloidal CNTs, even at relatively low concentrations, likely will have a significant impact on the equilibrium partition, transport, and possibly bioavailability of aromatic compounds in the environment.

Table 2. Summary of adsorbate properties (water solubility [Csat] and n-octanol–water partition coefficient [KOW]) and distribution coefficients (Kd) and organic carbon-normalized distribution coefficients (KOC) obtained from adsorption results
AdsorbateCsat (mmol/L)Log KOWAdsorbentLog Kd (L/kg)Log KOCc (L/kg)
  • a

    From Schwarzenbach et al. 26.

  • b

    From the U.S. National Library of Medicine Hazardous Substance Data Bank (http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?HSDB).

  • c

    KOC = Kd/fOC, where fOC is the fraction organic carbon of the adsorbent (a value of 0.8881 was used for O-MWNT, and a value of 0.8741 was used for the three colloidal O-MWNTs).

  • O-MWNT = oxidized multiwalled carbon nanotube; HA = humic acid; SDS = sodium dodecyl sulfate.

Phenanthrene6.31 × 10–3a4.57aO-MWNT4.24–6.474.29–6.52
O-MWNT-colloid5.32–6.625.38–6.68
O-MWNT/HA-colloid4.60–5.354.66–5.41
O-MWNT/SDS-colloid4.87–5.744.93–5.80
2-Naphthol6.93 × 100b2.70bO-MWNT3.17–5.013.22–5.06
O-MWNT-colloid2.91–4.472.97–4.53
O-MWNT/HA-colloid3.04–3.663.10–3.72
O-MWNT/SDS-colloid3.10–3.573.16–3.63
1-Naphthylamine1.19 × 101a2.25aO-MWNT3.92–6.323.97–6.37
O-MWNT-colloid4.05–7.094.11–7.15
O-MWNT/HA-colloid3.54–4.263.60–4.32
O-MWNT/SDS-colloid3.46–4.673.52–4.73

Mechanisms controlling adsorption to colloidal O-MWNTs

To understand the mechanisms controlling the adsorption of the three adsorbates to different forms of colloidal O-MWNT, the BET surface area-modified adsorption isotherms of a given adsorbate to different adsorbents are compared in Figure 3. For phenanthrene, adsorption affinities follow the order O-MWNT > O-MWNT/SDS-colloid > O-MWNT/HA-colloid > O-MWNT-colloid (indicated by the relative positions of the isotherms), but the differences are small. The differences in adsorption affinity among these adsorbents might be attributable to their morphological differences. For O-MWNT, the bulk aggregates, significant cross-linking between the surface O-functionalities of individual tubes likely exist 27, possibly through the mechanism of H-bonding. For the well-dispersed O-MWNT-colloid, however, the cross-linking between surface O-functionalities is likely weakened (the TEM images in Fig. 2 indicate that individually dispersed tubes are the dominant form). The greater amount of unlinked surface O-functionalities of O-MWNT-colloid can increase the surface hydrophilicity of this adsorbent 11 and can inhibit the adsorption of phenanthrene because the hydrophobic effect is an important factor controlling the adsorption of phenanthrene 26. As opposed to the very well-dispersed O-MWNT-colloid, the HA- or SDS-modified O-MWNTs are characterized as a network of loosely entangled individual tubes as well as patches of HA or coiled SDS (Fig. 1). This allows adsorbate molecules to interact simultaneously with multiple O-MWNT tubes (sorption to HA or SDS is negligible), resulting in enhanced adsorption. The surface area-modified adsorption isotherms of 2-naphthol exhibit similar patterns to phenanthrene, and the mechanisms mentioned above are likely viable for 2-naphthol. This is understandable because for both 2-naphthol and phenanthrene adsorption is dominated by hydrophobic effect and π–π electron donor–acceptor interactions with the graphitic surfaces of O-MWNT (both chemicals are strong π-electron donors).

Interestingly, the difference in adsorption affinity between O-MWNT-colloid and O-MWNT is considerably smaller for 1-naphthylamine than for phenanthrene and 2-naphthol. In our previous studies 5, 20, we demonstrated that 1-naphthylamine can interact with oxidized CNTs via strong Lewis acid–base interactions, in which the –NH2 group of 1-naphthylamine serves as the Lewis base and the surface O-functional groups of CNTs (mainly carboxyl and phenolic groups) serve as the Lewis acids. Compared with its noncolloidal counterpart, O-MWNT-colloid has a greater amount of exposed (free) surface O-functionalities, which facilitates the adsorption of 1-naphthylamine via Lewis acid–base interactions and, thus, counterbalancing the adsorption inhibition effects (due to inhibited surface hydrophobicity) mentioned above.

In Figure 4, the adsorption affinities of the three adsorbates to a given adsorbent are compared. For all the adsorbents, adsorption of 1-naphthylamine is the greatest among the three adsorbates (as shown by the relative positions of the isotherms); adsorption of 2-naphthol is generally weaker than adsorption of phenanthrene on a given adsorbent, but the differences are very small for O-MWNT and O-MWNT-colloid. Note that phenanthrene is much more hydrophobic than the other two adsorbates (as indicated by the aqueous solubility and n-octanol–water partition coefficient in Table 2); thus, the relative adsorption affinities among the three adsorbates indicate that strong nonhydrophobic interactions control the adsorption of 2-naphthol and 1-naphthylamine, particularly the latter. Based on our previous studies 5, 20, π–π electron donor–acceptor interactions are an important adsorption-enhancement mechanism for all three adsorbates to O-MWNT, but for 1-naphthylamine adsorption can also be enhanced by Lewis acid–base interactions. The significantly greater adsorption of 1-naphthylamine than 2-naphthol and phenanthrene seems to indicate that for surface O-functionality-rich CNTs, Lewis acid–base interactions are a more significant adsorption-enhancement mechanism than π–π electron donor–acceptor interactions for amino-substituted aromatics. Figure 4 shows that the differences in adsorption affinity between 1-naphthylamine and phenanthrene/2-naphthol are considerably larger on O-MWNT-colloid than on O-MWNT. This further indicates that once O-MWNT is well dispersed, a large fraction of surface O-functionalities are freed up and, consequently, Lewis acid–base interactions are enhanced. An interesting observation in Figure 4 is that phenanthrene and 2-naphthol exhibited very similar adsorption affinities to O-MWNT-colloid but markedly different adsorption affinities to O-MWNT/HA-colloid and O-MWNT/SDS-colloid. This might indicate that 2-naphthol and phenanthrene favor different adsorption sites on O-MWNT, which are affected differently by the adsorption of HA or SDS. Note that in the present study the adsorptive interactions are complicated by many different (and often counterbalancing) factors—for example, the presence of HA or SDS not only directly affects contaminant O-MWNT interactions but also affects the morphology of the adsorbents. More studies are needed to better understand the relative contributions of different mechanisms and the effects of water chemistry parameters.

Figure 4.

Comparison of adsorption affinities of phenanthrene, 2-naphthol, and 1-naphthylamine to a given adsorbent. OWMT = oxidized multiwalled carbon nanotube; HA = humic acid; SDS = sodium dodecyl sulfate.

The most significant finding of the present study is that colloidal O-MWNTs exhibit strong adsorption affinities to both apolar and polar aromatics. Given the high mobility of colloidal CNTs, these materials likely will have very strong effects on the fate and transport of organic contaminants. Furthermore, HA and SDS can significantly affect adsorptive interactions between organic contaminants and colloidal O-MWNTs. Humic substances (e.g., HA, fulvic acid) and surfactant-like materials are ubiquitous in the environment. These materials not only can facilitate the formation of colloidal carbon nanoparticles 9, 12, which likely are of much greater environmental risk due to their relatively high stability and mobility, but also affect how these colloidal carbon nanoparticles adsorb organic contaminants, as shown by the findings of the present study. This is an important aspect to be considered in risk assessment of engineered carbon nanomaterials.

SUPPLEMENTAL DATA

Detailed procedures used to prepare HA- and SDS-modified O-MWNT suspensions, summary of fitted Freundlich model coefficients obtained from adsorption results, and comparison of adsorption data and fitted Freundlich isotherms (1.3 MB DOC).

Acknowledgements

This project was supported by the National Natural Science Foundation of China (grants 21237002, 21177063, and 20977050), Tianjin Municipal Science and Technology Commission (grants 09JCYBJC26900 and 10SYSYJC27200), and China–US Center for Environmental Remediation and Sustainable Development.

Ancillary