Adsorption of Monoaromatic Compounds and Pharmaceutical

Jul 27, 2010 - Advanced functional materials in solid phase extraction for ICP-MS determination of trace elements and their species - A review. Man He...
1 downloads 8 Views 493KB Size
Environ. Sci. Technol. 2010, 44, 6429–6436

Adsorption of Monoaromatic Compounds and Pharmaceutical Antibiotics on Carbon Nanotubes Activated by KOH Etching LIANGLIANG JI, YUN SHAO, ZHAOYI XU, SHOURONG ZHENG, AND DONGQIANG ZHU* State Key Laboratory of Pollution Control and Resource Reuse, and School of the Environment, Nanjing University, Jiangsu 210093, P.R. China

Received May 2, 2010. Revised manuscript received June 25, 2010. Accepted July 15, 2010.

The relatively low surface area and micropore volume of carbon nanotubes limit their potential application as effective adsorbents for hydrophobic organic contaminants. In this study, KOH dry etching was explored to prepare activated singlewalled carbon nanotubes (SWNT) and multiwalled carbon nanotubes (MWNT) for adsorption of model monoaromatic compounds (phenol and nitrobenzene) and pharmaceutical antibiotics (sulfamethoxazole, tetracycline, and tylosin) in aqueous solutions. With activation, the specific surface area was increased from 410.7 m2/g to 652.8 m2/g for SWNT and from 157.3 m2/g to 422.6 m2/g for MWNT, and substantial pore volumes were created for the activated samples. Consistently, adsorption of the test solutes was enhanced 2-3 times on SWNT and 3-8 times on MWNT. Moreover, the activated carbon nanotubes showed improved adsorption reversibility for the selected monoaromatics, as compared with the pristine counterparts, which was attributed to the more interconnected pore structure and less pore deformation of the activated adsorbents. This is the first study on the adsorption/desorption of aqueous organic contaminants by KOH-activated carbon nanotubes. The findings indicate that KOH etching is a useful activation method to improve the adsorption affinity and adsorption reversibility of organic contaminants on carbon nanotubes.

Introduction Engineered carbon nanotubes have shown great promise in many biomedical and environmental applications (1-3). For example, a number of studies have indicated that carbon nanotubes are potentially efficient drug delivery agents (2, 4). The possibility of using carbon nanotubes as special adsorbents for the removal of organic contaminants from water has also received much attention (3, 5). Both applications require a deep understanding of the mechanisms and factors that control the organic-carbon nanotube interactions. Carbon nanotubes are composed of cylindrical graphite sheets having superhydrophobicity and very large van der Waals index (6), leading to remarkably strong hydrophobic effect in adsorption of hydrophobic organic chemicals. Additionally, the benzenoid rings of graphite sheets contain * Corresponding author phone/fax: +86 025-8359-6496; e-mail: [email protected]. 10.1021/es1014828

 2010 American Chemical Society

Published on Web 07/27/2010

sp2-hybridized carbon atoms with high electronic polarizability and hence may interact strongly with aromatic compounds via π-π coupling/stacking (7-10). Long and Yang (7) have attributed strong dioxin adsorption on carbon nanotubes to the π-π stacking between the two benzene rings of dioxin and the graphite sheets of carbon nanotubes. Recently, a mechanism of π-π electron-donor-acceptor (EDA) interaction has also been proposed to account for the extraordinarily strong adsorption of π-electron-donor compounds (amino- and hydroxyl-substituted aromatics) and π-electron-acceptor compounds (nitroaromatics and sulfonamide and tetracycline antibiotics) on carbon nanotubes (11-16). Carbon nanotubes aggregate in aqueous solutions through hydrophobic effect to form coagulated bundles that are dominated by flexible mesoporous interstices and grooves (17, 18). Thus, the size-exclusion effect that occurs to microporous activated carbon when adsorbing bulky organic chemicals can be largely prevented by exploring carbon nanotube adsorbents. Activated carbon is a conventional adsorbent and commonly used in water treatment for the removal of low-sized, undesirable organic chemicals such as phenolics. However, because activated carbon consists primarily of micropores (97%, Aldrich) and nitrobenzene (>99.5%, Fluka). The three antibiotic adsorbate compounds are sulfamethoxazole (99%, Sigma), tetracycline (hydrate, 99%, International Laboratory), and tylosin (95%, DeBioChem). Adsorbents. Single-walled carbon nanotubes (SWNT) and multiwalled carbon nanotubes (MWNT) were purchased from Nanotech Port Co. (Shenzhen, Guangdong Province, China). On the basis of the information provided by the manufacturer, the carbon nanotubes were synthesized by a chemical vapor deposition method using cobalt, magnesium, or nickel as catalysts; SWNT contained >90% (by volume) of carbon nanotubes, and the content with outer diameter less than 2 nm was >50%; MWNT contained >95% of carbon nanotubes, and the size of outer diameter ranged from 10 to 30 nm; the length of SWNT and MWNT ranged from 5 to15 µm. The SWNT and MWNT samples were treated to remove amorphous carbon and trace heavy metals using a previously developed method (27). The obtained samples of carbon nanotubes were activated by KOH dry etching according to previously developed methods (22, 23). Briefly, a mixture of carbon nanotubes and KOH powder at a mass ratio of 1/5 was heated to 800 °C with a ramp rate of 10 °C/min and held for 2 h under a N2 stream (40 mL/min). After cooling to room temperature, the resulted material was repeatedly washed with distilled water, followed by filtration and drying at 100 °C for 4 h. The KOH-activated SWNT and MWNT are referred to as K-SWNT and K-MWNT, respectively. The yield was 62% for K-SWNT and was 57% for K-MWNT. Other portions of SWNT and MWNT were treated the same but without KOH and were used as pristine (control) counterparts of the activated samples, referred to as P-SWNT and P-MWNT, respectively. Characterization of Adsorbents. Surface elemental compositions of the carbon nanotubes were measured by X-ray photoelectron spectroscopy (XPS) (Perkin-Elmer PHI 550 ESCA/SAM, USA). Transmission electron microscopy (TEM) images were performed on a JEM-200 CX (JEOL, Japan). The carbon nanotubes were also characterized by X-ray diffraction (XRD) analysis, N2 adsorption/desorption, and Raman spectra in the same manner as for our previous studies (16, 17). The average hydrodynamic diameter of carbon naotube aggregates in 10 mM NaCl solution at a pH of 6.0 was determined with a scattering angle of 0-135 degree using a laser particle size analyzer (Mastersizer 2000) (Malvern, UK). Batch Adsorption/Desorption. Adsorption/desorption experiments were conducted using a batch approach developed in our previous studies (15-17). All adsorption/ desorption experiments were run in duplicate. To initiate the experiments, about 10-25 mg carbon nanotubes were transferred to a 40-mL amber U.S. Environmental Protection Agency vial equipped with polytetrafluoroethylene-lined screw cap, followed by a background electrolyte solution containing 0.02 M NaCl. Afterward, stock solution of an adsorbate (prepared in pure water for tetracycline, and in methanol for other solutes) was added to the vial, and the volume percentage of methanol, if used, was kept below 0.1% to minimize cosolvent effects. The vial was then filled with 6430

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010

FIGURE 1. Pore size distributions of different adsorbents: pristine single-walled carbon nanotubes (P-SWNT), KOH-activated singlewalled carbon nanotubes (K-SWNT), pristine multiwalled carbon nanotubes (P-MWNT), and KOH-activated multiwalled carbon nanotubes (K-MWNT). the electrolyte solution to leave minimal headspace, covered with aluminum foil, and tumbled at room temperature for 21 d for tylosin and 3 d for other solutes. These times were used to reach apparent adsorption equilibrium based on the adsorption kinetics determined in this study for phenol and in our previous studies (17, 28) for other solutes. After centrifugation, solute was analyzed directly by highperformance liquid chromatography (HPLC) with a ultraviolet detector using a 4.6 × 150 mm SB-C18 column (Agilent). Isocratic elution was performed under the following conditions: acetonitrile-water (35:65, v:v) with a wavelength of 270 nm for phenol; methanol-water (60:40, v:v) with a wavelength of 266 nm for nitrobenzene; 0.05 M phosphoric acid-acetonitrile (75:25, v:v) with a wavelength of 265 nm for sulfamethoxazole; 0.01 M oxalic acid-acetonitrilemethanol (80:16:4, v:v:v) with a wavelength of 360 nm for tetracycline; 20 mM ammonium acetate-acetonitrile (65: 35, v:v) with a wavelength of 290 nm for tylosin. No peaks were detected in the HPLC spectra for potential degraded/ transformed products of the test adsorbates. To take account for possible solute loss from processes other than adsorbent sorption (sorption to glassware and septum and volatilization), calibration curves were obtained separately from controls receiving the same treatment as the adsorption samples but no adsorbent. Calibration curves included at least 14 standards over the test concentration range. Based on the obtained calibration curves, the adsorbed mass of solute was calculated by subtracting mass in the aqueous phase from mass added. Consecutive desorption experiments were performed using a single-step, centrifuge-withdraw-refill method for phenol and nitrobenzene after the adsorption experiments for 3 d to reach apparent desorption equilibrium based on the desorption kinetics of phenol. About 90% of supernatant was replaced with fresh background electrolyte solution. The adsorption ratios of the selected adsorbate/adsorbent combinations covered a favorable range (27-85%) for assuring the quality of desorption data. Adsorption reversibility was not examined for other adsorbates due to the high adsorption ratios. The equilibrium pH as measured at the end of the adsorption/desorption experiments was 6.0 ( 0.2.

Results and Discussion Characterization of Carbon Nanotubes. All test adsorbents were predominated by graphitized C (>88.5%) on the surfaces; however, a fairly high content of O-containing groups was

FIGURE 2. Adsorption isotherms plotted as adsorbed concentration (q, on unit mass basis) vs aqueous-phase concentration (CW) at equilibrium for different adsorbents: pristine single-walled carbon nanotubes (P-SWNT), KOH-activated single-walled carbon nanotubes (K-SWNT), pristine multiwalled carbon nanotubes (P-MWNT), and KOH-activated multiwalled carbon nanotubes (K-MWNT): (a) phenol, (b) nitrobenzene, (c) sulfamethoxazole, (d) tetracycline, and (e) tylosin. introduced to the carbon nanotubes (11.48% for SWNT and 7.89% for MWNT) by activation. The specific surface areas of the carbon nanotubes were increased remarkably by activation - from 410.7 m2/g to 652.8 m2/g for SWNT and from 157.3 m2/g to 422.6 m2/g for MWNT. The TEM observations were consistent with previous studies (22, 23) that large concentrations of defects were generated by the KOH activation, while the nanotubular morphology was preserved. The pore size distribution profiles of the four carbon nanotubes are compared in Figure 1. It is evident that substantial pore volumes were created for the activated carbon nanotubes. Clearly, the increases in surface area and pore volume were more prominent for MWNT than for SWNT. The interlayer spacing of the coaxial tubes of MWNT is close to that of the bulk graphite (0.335 nm) (29) and impenetrable to N2 molecules; however, the activation treatment generated more defects in the thicker wall of MWNT and hence higher abundance of adsorption sites, as compared with SWNT. The XRD results demonstrated that the graphitic structures were preserved in the activated carbon nanotubes but in less order. Raman spectra indicated that with activation the size of graphite sheets was lowered for SWNT but was nearly unchanged for MWNT. The measured hydrodynamic diameter was 542 ( 2 nm for P-MWNT (standard deviation

calculated from three replicate measurements) and was 460 ( 10 nm for K-MWNT (equivalently calculated to the diameter of a spherical particle that has the same translational diffusion speed of the rod-like carbon nanotubes), and the polydisperisty indexes were 0.312 and 0.286 for P-MWNT and K-MWNT, respectively, indicting less effective coagulation of K-MWNT. However, no consistent results could be measured for P-SWNT or K-SWNT due to the low particle suspendability. Adsorption Isotherms. Adsorption isotherms plotted as adsorbed concentration (q, mmol/kg) versus aqueous-phase concentration (CW, mmol/L) at equilibrium are presented in Figure 2. The adsorption data were fitted to the Freundlich model, q ) KFCWn, by nonlinear regression weighed on 1/q, where KF (mmol1-n Ln/kg) is the Freundlich affinity coefficient and n (unitless) is the Freundlich linearity index. The fitting parameters are summarized in Table 1. The Freundlich model fits most of the adsorption data very well (R2 > 0.98). The Freundlich n values are much smaller than 1, particularly for tetracycline and tylosin (n generally not exceeding 0.2), reflecting the high adsorption nonlinearity. Additionally, for a given adsorbate adsorption on the activated carbon nanotubes was consistently more nonlinear than adsorption on their pristine counterparts. In general, more nonlinear VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6431

TABLE 1. Freundlich Model Parameters, KF and n ± Standard Deviation, Distribution Coefficient (Kd), and Thermodynamic Index of Irreversibility (TII) for Adsorption/Desorption on Pristine Single-Walled Carbon Nanotubes (P-SWNT), KOH-Activated Single-Walled Carbon Nanotubes (K-SWNT), Pristine Multiwalled Carbon Nanotubes (P-MWNT), and KOH-Activated Multiwalled Carbon Nanotubes (K-MWNT) compound

phenol

nitrobenzene

sulfamethoxazole

tetracycline

tylosin

adsorbent

KF (mmol1-nLn/kg)

n

R2

Kda (L/kg)

TII b

P-SWNT K-SWNT P-MWNT K-MWNT P-SWNT K-SWNT P-MWNT K-MWNT P-SWNT K-SWNT P-MWNT K-MWNT P-SWNT K-SWNT P-MWNT K-MWNT P-SWNT K-SWNT P-MWNT K-MWNT

303 ( 5 800 ( 20 64 ( 1 457 ( 6 790 ( 10 1840 ( 20 175 ( 8 1100 ( 10 2200 ( 100 5200 ( 400 490 ( 30 2300 ( 100 690 ( 20 1400 ( 300 320 ( 10 800 ( 20 740 ( 20 910 ( 10 350 ( 10 650 ( 10

0.421 ( 0.005 0.403 ( 0.007 0.337 ( 0.004 0.328 ( 0.005 0.389 ( 0.005 0.300 ( 0.005 0.34 ( 0.01 0.282 ( 0.006 0.395 ( 0.009 0.40 ( 0.01 0.394 ( 0.009 0.347 ( 0.007 0.105 ( 0.006 0.095 ( 0.004 0.171 ( 0.007 0.104 ( 0.005 0.151 ( 0.006 0.097 ( 0.004 0.219 ( 0.008 0.127 ( 0.004

0.998 0.996 0.998 0.998 0.998 0.996 0.990 0.994 0.994 0.992 0.994 0.996 0.962 0.976 0.981 0.977 0.986 0.983 0.984 0.989

4390 12600 1370 10200 16100 54100 4370 34800 16900 38600 3780 24100 38900 83400 12500 45300 34100 55500 11300 33900

0.23 ( 0.02 0.24 ( 0.04 0.42 ( 0.02 0.15 ( 0.02 -0.03 ( 0.03 -0.05 ( 0.01 0.19 ( 0.04 -0.06 ( 0.02 NDc NDc NDc NDc NDc NDc NDc NDc NDc NDc NDc NDc

a Calculated from the Freundlich model (CW ) 10-5 · SW for phenol and CW ) 10-3 · SW for other adsorbates; CW is the aqueous phase concentration at adsorption equilibrium; and SW is the adsorbate aqueous solubility. The selected CW was located approximately in the middle of the examined concentration ranges.). b Averaged value ( standard deviation from all desorption points. c Not determined.

adsorption isotherms imply more heterogeneous adsorption sites due to specific adsorptive interactions and/or wider pore size distribution of adsorbent (see more details below). All test adsorbate/adsorbent combinations showed very strong adsorption affinity, as reflected by the high distribution coefficients (Kd, L/kg) (calculated from the Freundlich model at selected CW, results presented in Table 1). Recently, the adsorptive interactions of nitro-, amino-, and hydroxylsubstituted aromatic compounds, sulfonamides, and tetracycline with carbon nanotubes and graphite have been investigated in depth (11-16). An important point raised in these studies is that depending on the electron-giving/ withdrawing ability, the substituted groups cause the aromatic/unsaturated moieties of the adsorbate molecules to be electron-rich or electron-depleted and hence function as π-electron-donors (hydroxyl- and amino-substituted aromatics) or π-electron-acceptors (nitroaromatics, sulfonamides, and tetracycline). Therefore, the adsorbate molecules may interact with the polarized electron-depleted or electronrich regions on the graphitized carbon surfaces via the mechanism of π-π EDA interactions. Moreover, the protonated amino group in the tetracycline molecule and the tylosin molecule may facilitate strong cation-π bonding with π-electrons on the graphitized carbon surfaces (15, 28). The multiple functional groups/structures in the tetracycline and tylosin molecules led to a variety of specific adsorptive interactions on the surface of carbon nanotubes, and consequently higher degrees of adsorption nonlinearity were induced (see the Freundlich n values in Table 1). It is clearly shown in Figure 2 that adsorption of all test solutes was markedly stronger on the activated carbon nanotubes than on the respective pristine carbon nanotubes. With activation adsorption of SWNT was increased by 2-3 times, and adsorption of MWNT was increased by 3-8 times, as justified by the ratios of Kd at selected CW (see Table 1). For all adsorbates, the Kd of the activated carbon nanotubes was as high as 104 L/kg order of magnitude. The observed trends of adsorption can be assigned to the enhanced surface areas of carbon nanotubes from the activation treatment. Similarly, previous studies found that the KOH activation 6432

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010

improved hydrogen adsorption on carbon nanotubes greatly, mainly attributed to the enhanced surface area of adsorbent (24). In addition to the effect on adsorption affinity, the KOH activation led to a wider pore size distribution of the activated carbon nanotubes (see Figure 1), giving rise to more heterogeneity of adsorption sites and hence higher degrees of adsorption nonlinearity. Due to the extraordinarily strong adsorption affinity of the graphite surfaces, carbon nanotubes are effective adsorbents for many organic contaminants, including polycyclic aromatic hydrocarbons (PAHs), dioxin, and chlorinated benzenes (7-10). However, the moderate surface areas (normally in the range of 290 ( 170 m2/g, ref 30) of carbon nanotubes may cause restriction on their practical use in water treatment. Notably, despite the usually much lower manufacturing price, MWNT often exhibits prominently weaker adsorption per unit mass than SWNT, mainly due to the inaccessibility of the interlayer spacing of MWNT. However, compared with SWNT the activation treatment produced higher-abundance defects and hence more microporosity in the thicker wall of MWNT. Accordingly, more enhanced adsorption was shown on MWNT than on SWNT. Adsorption of K-MWNT was only slightly lower than that of K-SWNT but higher or similar to that of P-SWNT. Conclusively, with the activation treatment MWNT can be turned into a much more efficient and cost-effective adsorbent. It is evident from Figure 2 that the disparity in adsorption affinity between the activated carbon nanotubes and their pristine counterparts is smaller for the large-sized adsorbates (tylosin and tetracycline) than for the low-sized adsorbates (phenol, nitrobenzene, and sulfamethoxazole), suggesting the occurrence of adsorbate size-dependent adsorption regulated by adsorbent porosity. To better understand the impact of adsorbent porosity on adsorption, the surface areanormalized adsorption isotherms are compared between the four carbon nanotubes for the different adsorbates in Figure 3. For phenol, nitrobenzene, and sulfamethoxazole, adsorption was slightly stronger on KOH-activated carbon nanotubes than on the pristine carbon nanotubes. This was likely caused by the micropore-filling effect due to the closeness

FIGURE 3. Adsorption isotherms plotted as adsorbed concentration (q, on unit surface area basis) vs aqueous-phase concentration (CW) at equilibrium for different adsorbents: pristine single-walled carbon nanotubes (P-SWNT), KOH-activated single-walled carbon nanotubes (K-SWNT), pristine multiwalled carbon nanotubes (P-MWNT), and KOH-activated multiwalled carbon nanotubes (K-MWNT): (a) phenol, (b) nitrobenzene, (c) sulfamethoxazole, (d) tetracycline, and (e) tylosin. of the molecular sizes of the adsorbates and the pore sizes of a small portion of micropores created in the activation process. Previous studies have proposed the same mechanism to explain enhanced adsorption of low-sized adsorbates to highly microporous wood-made charcoal and activated carbon (15, 31, 32). However, the micropore-filling effect was expected to be negligible in adsorption of tetracycline and tylosin because of the bulky molecular sizes of the adsorbates. The normalized adsorption data were nearly overlapping between the activated and pristine carbon nanotubes for tetracycline and tylosin, implying that most micropores generated by activation were available for adsorption sites. This is understandable because the micropores were created on the surface of carbon nanotubes and could be readily accessed by adsorbate molecules. Comparison of the pore size distribution profiles between the adsorbents with and without the presence of the pharmaceutical antibiotics also lends a strong support to this view. Adsorption Reversibility. The adsorption and desorption data of phenol and nitrobenzene on the four carbon nanotubes are presented in Figure 4. Adsorption hysteresis was quantified by the thermodynamic index of irreversibility (TII) according to a previously developed approach (33, 34).

Theoretically, the value of TII varies between 0 and 1, corresponding to the two boundary states of completely reversible adsorption and completely irreversible adsorption, respectively. The calculated TII values averaged on all desorption points for the selected adsorbate/adsorbent combinations are also listed in Table 1. The examined adsorption-desorption data exhibited varied degrees of hysteresis, with the TII ranging from approximately 0 to 0.42. Several trends can be generalized: 1) with the exception of phenol/SWNT hysteresis was significantly lowered by the KOH activation; 2) for a given adsorbate P-MWNT showed higher hysteresis than P-SWNT, while no clear trend emerged for the KOH-activated samples; and 3) for a given adsorbent higherhysteresiswasobservedonphenolthanonnitrobenzene. Condensation and presence of high energy adsorption sites are the two main causes proposed for adsorptiondesorption hysteresis for carbon nanotubes (5). The available adsorption sites of carbon nanotube bundles include the external surface, the interstitial spaces and grooves formed between individual carbon nanotubes, and the inner tube surface. Multilayer adsorption could lead to surface condensation of organic chemicals on the carbon nanotubes (10). Moreover, the carbon nanotubes contained large portions of mesopores and thus were susceptible to capillary VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6433

FIGURE 4. Adsorption and desorption isotherms plotted together as adsorbed concentration (q) vs aqueous-phase concentration (CW) at equilibrium for pristine and KOH-activated single-walled carbon nanotubes (SWNT) and multiwalled carbon nanotubes (MWNT): (a) phenol on SWNT, (b) phenol on MWNT, (c) nitrobenzene on SWNT, and (d) nitrobenzene on MWNT. condensation effect causative to hysteresis (5). The higher hysteresis observed on P-MWNT than on P-SWNT could be attributed to the larger mesoporosity of P-MWNT. On the other hand, the KOH activation generated substantial micropore volumes for additional adsorption sites; therefore, the effect of capillary condensation was attenuated because less adsorbate molecules entered the mesopores, resulting in less hysteresis. Additionally, during the activation process the one-dimensional mesoporous groove channels between individual carbon nanotubes and the inner tube pores could be networked with the new pores created in the wall, leading to the formation of somewhat three-dimensional, interconnected pore structures to facilitate desorption. As a result, the hysteresis effect induced by the mesopore-condensation mechanism was less on the KOH-activated carbon nanotubes than on their pristine counterparts. However, the higher hysteresis observed on phenol than on nitrobenzene could not be explained by the mechanism of condensation. At room temperature nitrobenzene is a liquid, while phenol is a solid. Thus, nitrobenzene would have shown higher hysteresis if condensation were the only cause. Compound-specific hysteresis effect on carbon nanotubes suspended in aqueous solutions has also been reported previously. Strong hysteresis was observed for 17R-ethinyl estradiol, bisphenol A, and oxytetracycline (14, 21), while negligible hysteresis was observed for PAHs and atrazine (18, 35). A plausible mechanism for the significant hysteresis effect has been proposed that the adsorptive interactions on intrananoparticulate surfaces are very strong (due to additive adsorptive energies as similar to the pore-filling effect) and thus enable deformation of the interstitial pores by rearrangement of the nanoaggregates, thereby entrapping the 6434

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010

solute to cause hysteresis (14, 17, 21). It is interesting to note that the above-mentioned adsorbates capable of inducing prominent hysteresis effect all contain aromatic/unsaturated moieties substituted with strong electron-giving group (hydroxyl) or electron-withdrawing group (ketone) and act as effective π-electron-donors or π-electron-acceptors. These adsorbates are thus expected to invoke strong π- π EDA interactions with the intrananoparticulate surfaces of carbon nanotubes, leading to pore deformation by physical rearrangement of the nanoaggregates and in turn adsorptiondesorption hysteresis. A deeper insight into the compound-specific hysteresis effect can be obtained by comparing the adsorption data between phenol and nitrobenzene with normalization of solute hydrophobicity. Because n-hexadecane is an inert solvent dictated by the methylene structure which invokes van der Waals forces predominantly, the n-hexadecane-water partition coefficient (KHW) is an ideal index for justifying solute hydrophobicity (36, 37). Consistent with the observed trends of hysteresis effects, phenol displayed stronger hydrophobicity-normalized adsorption than nitrobenzene for all carbon nanotubes, wherein the disparities were larger for the pristine carbon nanotubes than for the activated carbon nanotubes. This is probably because phenol is a relatively strong π-electron-donor, while nitrobenzene is a relatively weak π-electron-acceptor; therefore, phenol invoked strong π-π EDA interactions than nitrobenzene. Nonetheless, the physical rearrangement dependent hysteresis effect induced by adsorptive interactions with the intrananoparticulate surfaces of carbon nanotubes of phenol was stronger than that of nitrobenzene, in particular for the pristine carbon nanotubes. Due to the presence of substantial micropores

and defects in the wall of activated carbon nanotubes, the formation of intrananoparticulate spaces through nanoaggregation would be less effective when compared with the pristine carbon nanotubes. To be consistent with the hypothesis, the measured hydrodynamic diameter of KMWNT was lower than that of P-MWNT. As a result, the adsorptive interactions between phenol and the intrananoparticulate surfaces of the activated carbon nanotubes and the associated hysteresis effect were suppressed. Undoubtedly, more studies are needed to further understand the impact of nanoaggregation on the adsorption reversibility of carbon nanotubes. Implication. The moderate surface area and onedimensional pore structure of carbon nanotubes limit their potential use as environmental adsorbents, in spite of the high binding affinity of the graphite surface. This study indicate that adsorption affinity and reversibility of carbon nanotubes, especially multiwalled carbon nanotubes, could be markedly improved by KOH dry etching due to the enhanced specific surface area and the formation of more interconnected pore structure. In fact, one advantage of carbon nanotubes over conventional carbonaceous adsorbents such as activated carbon is that they can be easily functionalized or derivatized on purpose to enhance the adsorption efficiency of target compounds. Moreover, it has been well recognized that carbon nanotubes can be deposited on biological tissues and cause severe toxic effects (38). The functionalization or derivatization of carbon nanotubes could heavily affect their exposure and toxicity. Further studies are needed to better understand this aspect to advance both potential applications and environmental implications of the carbon nanomaterials.

Acknowledgments This study was supported by the National Science Foundation of China (Grants 20637030, 20677026, and 20777031), the Ministry of Education of China (Grant NCET-06-0453), the Ministry of Science and Technology of China (Grant 2010DFA91910), and the Science Foundation of Jiangsu Province, China (BK2009248).

Supporting Information Available Table S1 summarizes selected physicochemical properties of adsorbates. Table S2 presents surface elemental compositions and surface areas of the carbon nanotubes. Table S3 presents ratio of pore volume of the carbon nanotubes occupied by adsorbed pharmaceutical adsorbates. Figure S1 presents chemical structures and molecular sizes of adsorbates. Figure S2 presents adsorption/desorption kinetics for selected adsorbates on the carbon nanotubes. Figure S3 presents TEM observations of the carbon nanotubes. Figure S4 compares XRD patterns of the carbon nanotubes. Figure S5 presents Raman spectra of the carbon nanotubes and graphite. Figure S6 compares porosities of the carbon nanotubes with and without the presence of pharmaceutical adsorbates. Figure S7 presents adsorption data of phenol and nitrobenzene upon normalization of solute hydrophobicity using n-hexadecane as reference solvent. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chemistry of carbon nanotubes. Chem. Rev. 2006, 106, 1105–1136. (2) Bianco, A.; Kostarelos, K.; Partidos, C. D.; Prato, M. Biomedical applications of functionalised carbon nanotubes. Chem. Commun. 2005, 5, 571–577. (3) Mauter, M. S.; Elimelech, M. Environmental applications of carbon-based nanomaterials. Environ. Sci. Technol. 2008, 42, 5843–5859.

(4) Kam, N. W. S.; O’Connell, M.; Wisdom, J. A.; Dai, H. Carbon nanotubes as multifunctional biological transporters and nearinfrared agents for selective cancer cell destruction. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11600–11605. (5) Pan, B.; Xing, B. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 2008, 42, 9005– 9013. (6) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry, 2nd ed.; Wiley-Interscience: New York, 2003. (7) Long, R. Q.; Yang, R. T. Carbon nanotubes as superior sorbent for dioxin removal. J. Am. Chem. Soc. 2001, 123, 2058–2059. (8) Peng, X.; Li, Y.; Luan, Z.; Di, Z.; Wang, H.; Tian, B.; Jia, Z. Adsorption of 1,2-dichlorobenzene from water to carbon nanotubes. Chem. Phys. Lett. 2003, 376, 154–158. (9) Fagan, S. B.; Filho, A. G. S.; Lima, J. O. G.; Filho, J. M.; Ferreira, O. P.; Mazali, I. O.; Alves, O. L.; Dresselhaus, M. S. 1,2Dichlorobenzene interacting with carbon nanotubes. Nano Lett. 2004, 4, 1285–1288. (10) Gotovac, S.; Hattori, Y.; Noguchi, D.; Miyamoto, J.; Kanamaru, M.; Utsumi, S.; Kanoh, H.; Kaneko, K. Phenanthrene adsorption from solution on single wall carbon nanotubes. J. Phys. Chem. B 2006, 110, 16219–16224. (11) Chen, W.; Duan, L.; Zhu, D. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 2007, 41, 8295–8300. (12) Chen, W.; Duan, L.; Wang, L.; Zhu, D. Adsorption of hydroxyland amino-substituted aromatics to carbon nanotubes. Environ. Sci. Technol. 2008, 42, 6862–6868. (13) Chen, J.; Chen, W.; Zhu, D. Adsorption of nonionic aromatic compounds to single-walled carbon nanotubes: Effects of aqueous solution chemistry. Environ. Sci. Technol. 2008, 42, 7225–7230. (14) Pan, B.; Lin, D.; Mashayekhi, H.; Xing, B. Adsorption and hysteresis of bisphenol A and 17R-ethinyl estradiol on carbon nanomaterials. Environ. Sci. Technol. 2008, 42, 5480–5485. (15) Ji, L.; Chen, W.; Duan, L.; Zhu, D. Mechanisms for strong adsorption of tetracycline to carbon nanotubes: A comparative study using activated carbon and graphite as adsorbents. Environ. Sci. Technol. 2009, 43, 2322–2327. (16) Ji, L.; Chen, W.; Zheng, S.; Xu, Z.; Zhu, D. Adsorption of sulfonamide antibiotics to multi-walled carbon nanotubes. Langmuir 2009, 25, 11608–11613. (17) Ji, L.; Liu, F.; Xu, Z.; Zheng, S.; Zhu, D. Zeolite-templated microporous carbon as a superior adsorbent for removal of monoaromatic compounds from aqueous solution. Environ. Sci. Technol. 2009, 43, 7870–7876. (18) Yang, K.; Xing, B. Desorption of polycyclic aromatic hydrocarbons from carbon nanomaterials in water. Environ. Pollut. 2007, 145, 529–537. (19) Kilduff, J. E.; Karanfil, T.; Chin, Y.-P.; Weber, W. J., Jr. Adsorption of natural organic polyelectrolytes by activated carbon: A size exclusion chromatography study. Environ. Sci. Technol. 1996, 30, 1336–1343. (20) Liu, G.; Zheng, S.; Yin, D.; Xu, Z.; Fan, J.; Jiang, F. Adsorption of aqueous alkylphenol ethoxylate surfactants by mesoporous carbon CMK-3. J. Colloid Interface Sci. 2006, 302, 47–53. (21) Oleszczuk, P.; Pan, B.; Xing, B. Adsorption and desorption of oxytetracycline and carbamazepine by multiwalled carbon nanotubes. Environ. Sci. Technol. 2009, 43, 9167–9173. (22) Raymundo-Pin ˜ ero, E.; Cazorla-Amoro´s, D.; Linares-Solano, A.; Delpeux, S.; Frackowiak, E.; Szostak, K.; Be´guin, F. High surface area carbon nanotubes prepared by chemical activation. Carbon 2002, 40, 1614–1617. (23) Jiang, Q.; Qu, M.; Zhang, B.; Yu, Z. Preparation of activated carbon nanotubes. Carbon 2002, 40, 2743–2745. (24) Chen, C.; Huang, C. Hydrogen storage by KOH-modified multiwalled carbon nanotubes. Int. J. Hydrogen Energy 2007, 32, 237– 246. (25) Jurewicz, K.; Babel, K.; Pietrzak, R.; Delpeux, S.; Wachowska, H. Capacitance properties of multi-walled carbon nanotubes modified by activation and ammoxidation. Carbon 2006, 44, 2368–2375. (26) Frackowiak, E.; Delpeux, S.; Jurewicz, K.; Szostak, K.; CazorlaAmoro´s, D.; Be´guin, F. Enhanced capacitance of carbon nanotubes through chemical activation. Chem. Phys. Lett. 2002, 361, 35–41. (27) Lu, C.; Chiu, H. Adsorption of zinc (II) from water with purified carbon nanotubes. Chem. Eng. Sci. 2006, 61, 1138–1145. (28) Ji, L.; Liu, F.; Xu, Z.; Zheng, S.; Zhu, D. Adsorption of pharmaceutical antibiotics on template-synthesized ordered VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6435

(29) (30)

(31)

(32)

(33)

6436

micro- and mesoporous carbons. Environ. Sci. Technol. 2010, 44, 3116–3122. Maddox, M. W.; Gubbins, K. E. Molecular simulation of fluid adsorption in buckytubes. Langmuir 1995, 11, 3988–3996. Cho, H. H.; Smith, B. A.; Wnuk, J. D.; Fairbrother, D. H.; Ball, W. P. Influence of surface oxides on the adsorption of naphthalene onto multiwalled carbon nanotubes. Environ. Sci. Technol. 2008, 42, 2899–2905. Zhu, D.; Pignatello, J. J. Characterization of aromatic compound sorptive interactions with black carbon (charcoal) assisted by graphite as a model. Environ. Sci. Technol. 2005, 39, 2033–2041. Nguyen, T. H.; Cho, H.-H.; Poster, D. L.; Ball, W. P. Evidence for a pore-filling mechanism in the adsorption of aromatic hydrocarbons to a natural wood char. Environ. Sci. Technol. 2007, 41, 1212–1217. Sander, M.; Lu, Y. F.; Pignatello, J. J. A thermodynamically based method to quantify true sorption hysteresis. J. Environ. Qual. 2005, 34, 1063–1072.

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010

(34) Sander, M.; Lu, Y. F.; Pignatello, J. J. On the reversibility of sorption to black carbon: Distinguishing true hysteresis from artificial hysteresis caused by dilution of a competing adsorbate. Environ. Sci. Technol. 2007, 41, 843–849. (35) Yang, X.; Shi, B.; Lu, J.; Feng, C.; Wang, H.; Tang, H. Adsorption and desorption of atrazine on carbon nanotubes. J. Colloid Interface Sci. 2008, 321, 30–38. (36) Abraham, M. H.; Chadha, H. S.; Whiting, G. S.; Mitchell, R. C. Hydrogen Bonding. 32. An analysis of water-octanol and wateralkane partitioning and the delta log P parameter of seiler. J. Pharm. Sci. 1994, 83, 1085–1100. (37) Borisover, M.; Graber, E. R. Relationship between strength of organic sorbate interactions in NOM and hydration effect on sorption. Environ. Sci. Technol. 2002, 36, 4570–4577. (38) 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.

ES1014828