- Email: [email protected]
Adsorption mechanism of different organic chemicals on fluorinated carbon nanotubes
Chemosphere 154 (2016) 258e265
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Adsorption mechanism of different organic chemicals on fluorinated carbon nanotubes Hao Li, Nan Zheng, Ni Liang, Di Zhang, Min Wu, Bo Pan* Faculty of Environmental Science & Engineering, Kunming University of Science & Technology, Kunming, 650500, PR China
h i g h l i g h t s Carbon nanotubes were fluorinated using PTFE as a fluorine source. Organic contaminant adsorption on CNTs was enhanced after fluorination. BPA butterfly structure resulted in its strong sorption and desorption hysteresis. OFL exothermic sorption resulted in its strong desorption hysteresis.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 November 2015 Received in revised form 12 March 2016 Accepted 21 March 2016
Multi-walled carbon nanotubes (MC) were fluorinated by a solid-phase reaction method using polytetrafluoroethylene (PTFE). The surface alteration of carbon nanotubes after fluorination (MC-F) was confirmed based on surface elemental analysis, TEM and SEM. The incorporation of F on MC surface was discussed as F incorporation on carbon defects, replacement of carboxyl groups, as well as surface coating of PTFE. The adsorption performance and mechanisms of MC-F for five kinds of representative organic compounds: sulfamethoxazole (SMX), ofloxacin (OFL), norfloxacin (NOR), bisphenol a (BPA) and phenanthrene (PHE) were investigated. Although BET-N2 surface area of the investigated CNTs decreased after fluorination, the adsorption of all five chemicals increased. Because of the glassification of MC-F surface coating during BET-N2 surface area measurement, the accessible surface area of MC-F was underestimated. Desorption hysteresis was generally observed in all the sorption systems in this study, and the desorption hysteresis of MC-F were stronger than the pristine CNTs. The enhanced adsorption of MC-F may be attributed the pores generated on the coated PTFE and the dispersed CNT aggregates due to the increased electrostatic repulsion after fluorination. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. The butterfly structure of BPA resulted in its high sorption and strong desorption hysteresis. The exothermic sorption character of OFL on CNTs resulted in its strong desorption hysteresis. © 2016 Elsevier Ltd. All rights reserved.
Handling Editor: X. Cao Keywords: Aggregation Desorption hysteresis Fluorination Nanoparticles Sorption thermodynamics
1. Introduction Since carbon nanotubes (CNTs) was discovered by Iijima in 1991 (Iijima, 1991), which has aroused extensive attentions due to their unique and superior physicochemical properties. CNTs have potential applications in many areas including manufacturing, electronics, construction, and medicine (Eatemadi et al., 2014). CNTs exhibited a very strong affinity for organic contaminates and heavy metal ions, thus it is proposed as a superior and promising sorbent applied in wastewater treatment (Ren et al., 2011; Yu et al., 2014). * Corresponding author. E-mail address: [email protected] (B. Pan). http://dx.doi.org/10.1016/j.chemosphere.2016.03.099 0045-6535/© 2016 Elsevier Ltd. All rights reserved.
In order to maximize the selectivity of adsorption and to improve the adsorption performance of CNTs, surface modification of CNTs has drawn a widely attention (Lu et al., 2008; Ghobadi et al., 2014). In most cases, acidic or oxidizing reagents were applied to treat CNTs, and thus oxygen containing functional groups (such as hydroxyl and carboxyl groups) were incorporated on the surface of CNTs (Liao et al., 2008; Chand et al., 2009). The oxygen-containing functional groups will facilitate the wettability properties of the surface, and then enhance the adsorption for polar compounds (Patino et al., 2015). And also, the oxygen-containing functional groups will make the CNTs more negatively charged in neutral condition, which could enhance the dispersion of CNTs in solutions due to the electrostatic repulsion, leading to a higher adsorption than pristine
H. Li et al. / Chemosphere 154 (2016) 258e265
CNTs. However, the interactions between oxide-carbon nanotubes and hydrophobic compounds will dramatic decrease in adsorption as the increasing extent of CNTs oxidation (Zhang et al., 2009). Li et al. reported that PHE exhibited a much lower adsorption capacity on oxidized CNTs than on pristine CNTs (Li et al., 2013). The oxygencontaining groups on surface of CNTs can form strong hydrogen bond with water molecules, resulting in decreasing adsorption of organic compounds, especially hydrophobic organic compounds (Wu et al., 2012). Therefore, CNT surface modification may result in the increased sorption of certain chemicals, but will also decrease its sorption with other chemicals. In most of the circumstances, various organic contaminants with different properties simultaneously present in wastewater or the environment. Effective removal of organic contaminants with a wide range of chemical physical properties is a major challenge for researchers. Fluorination is one of the most effective routes to modify and control physico-chemical properties of carbon materials (Lee, 2007; Lu et al., 2016). Fluorinated CNTs was reported to enhance surface polarity, electroconductivity and capacitive character, and is widely used in lithium batteries, solid lubricants and nano-composites (Zhang et al., 2010c; Maiti et al., 2012). Recent study has reported that fluorinated CNTs exhibited great performance in removing chromium ions (Osikoya et al., 2014). However, the study using fluorinated CNTs as adsorbent to retain organic compounds from aqueous is relatively rare. Due to the utmost electronegativity of F, the CeF bond on fluorinated CNTs enables CNTs to act as a strong pacceptor, which could enhance the adsorption controlled by pacceptor and p-donor interaction. The strong electronegativity of F will also make the CNTs strongly negatively charged, and thus facilitate the dispersion of CNTs. In addition, different from oxygen containing groups, the CeF bond can hardly form hydrogen bond with water molecules (Dunitz and Taylor, 1997). Therefore, incorporating F into CNTs will not suppress the adsorption of hydrophobic compounds. We thus hypothesized that fluorination will enhance the adsorption of organic compounds on CNTs. In this study, fluorinated CNTs was synthesized by the solidphase reaction method. Five representative organic contaminants including sulfamethoxazole (SMX), ofloxacin (OFL), norfloxacin (NOR), bisphenol A (BPA) and phenanthrene (PHE) were selected as adsorbates. Batch adsorption, desorption and thermodynamics experiments were conducted to evaluate the adsorption characteristics and to investigate the adsorption mechanisms of MC-F. In addition, with the wide applications of fluorinated CNTs, it will be inevitably released into the aqueous environment. Therefore, investigating the interactions between fluorinated CNTs and organic contaminants is of great environmental significance. The aim of this work is to illustration the sorption mechanisms of organic contaminants on fluorinated CNTs, which will be an important input for CNT applications in wastewater treatment. 2. Materials and methods 2.1. Materials Carboxylized multiwall carbon nanotubes (MC, purity > 95%, diameter: 8e15 nm) were purchased from Chengdu Organic
259
Chemistry Co., Chinese Academy of Sciences. It was synthesized in the CH4/H2 mixture at 700 C by the chemical vapor deposition method and was purified by mixed HNO3 and H2SO4 solutions to reduce the contents of metal catalyst and amorphous carbon. Fluorinated multiwall carbon nanotubes were synthesized by the solid-phase reaction method. Briefly, the MC and PTFE (5 mm, Aladdin-reagent Co. Ltd., China) fine particles were mixed in an enclosed vessel according weight ratio: 1:2 (MC:PTFE), then the mixture was put in a temperature-controlled furnace which was purged continuously with N2. The temperature was controlled at 460e500 C for 4e6 h to generate the source of fluorine for fluorination. The resulted particles were chilled and grinded. This particle is noted as MC-F. SMX (AR, 99.0%), OFL (AR, 99.0%) and NOR (AR, 99.0%) were obtained from Bio Basic Inc. PHE (AR, 97%) was obtained from Acros Co., Belgium. BPA (AR, 99.0%) was obtained from Sinopharm Chemical Reagent Co., Ltd, China. Both MC and MC-F were characterized for their surface area (Autosorb-1C, Quantachrome, US), elemental composition (MicroCube, Elementar, Germany), surface elemental composition (X-ray photoelectron spectroscopy, PHI5000 VersaProbe II, ULVACPHI, INC.) scanning electron microscopy (SEM, Tescan, VEGA3 SBH) equipped with energy-dispersiveX-ray spectrometer (EDAX, Thermo Fisher Scientific, NORAN System 7), transmission electron microscopy (TEM, JEM 2100) and zero point of charge (pHzpc, ZetaPlus, Brookhaven, US). For SEM-EDAX tests, MC or MC-F particles were mounted on aluminum stubs, and then gold-coated using a sputter coater. The accelerating voltage was set at 20 kV. For TEM tests, the MC and MC-F solutions were ultra-sonicated for 5 min, and then a drop of solution containing MC or MC-F was put onto a copper TEM grid. After air-dried, the measurements were operated with the accelerating voltage of 200 kV. The properties of compounds studied are listed in Table 1. 2.2. Batch adsorption experiments Batch adsorption experiments were separately conducted for SMX, BPA, OFL, NOR and PHE in 4 mL (for MC) and 15 mL (for MC-F) amber glass vials with Teflon-lined screw caps. According to our preliminary work, the aqueous:solid ratios were 2000:1 and 15000:1 (w/w) for MC and MC-F sorption experiments, respectively. The different aqueous:solid ratios were to ensure comparable equilibrated aqeuous phase concentrations in two sorption systems. The initial concentration ranges were 0.2e1.24 mg L1 for PHE and 1e50 mg L1 for the other four chemicals, and all the adsorbates were dissolved in background solution containing 0.02 M NaCl and 200 mg L1 NaN3. We did not use CaCl2 in the background solution to avoid possible impact of Ca2þ on SMX sorption. The adsorbate solutions without adsorbents were referred to as the initial concentration references. All vials were kept in the dark and were continuously shaken in an air-bath shaker (80 rpm) at 25 C or in a refrigerator at 4 C for 7 d. According to our preliminary study, 7 d was enough to reach apparent sorption equilibrium. All vials were then centrifuged at 1000 g for 10 min, and the supernatants were sampled for adsorbates quantification. For desorption experiment, the equilibrium supernatant was removed from the vials and then re-filled with the same volume of
Table 1 Property comparison of MC and MC-F. CNTs
MC MC-F
SSA (m2 g1)
118.8 17.5
Pore volume (cc g1)
0.454 0.083
Pore radius (Å)
9.6 12.2
XPS
Elemental analysis
pHzpc
C
O
F
C
O
H
95.8 57.5
4.2 3.3
0 39.3
97 56.7
0.5 0.2
2.7 2.1
4.0 2.1
260
H. Li et al. / Chemosphere 154 (2016) 258e265
background solution. After equilibration, the vials were centrifuged and the supernatants were sampled for adsorbates quantification as described before. All the samples have three replications.
The thermodynamic parameters were calculated using the following equations:
DG ¼ RT lnKd
(7)
2.3. Quantification of adsorbates
DG ¼ DS T DH
(8)
The aqueous-phase equilibrium concentrations of all the adsorbates were quantified by HPLC (Agilent Technologies 1200) equipped with a reversed-phase C18 column (5 mm, 4.6 150 mm). For SMX analysis, the wavelength for the UV detector was 265 nm, and the mobile phase was 40:60 (v:v) ratio of acetonitrile and deionized water with 0.1% acetic acid. The UV detector wavelength for BPA analysis was 280 nm, and the mobile phase was 40:60 (v:v) of acetonitrile and deionized water. For OFL analysis, the wavelength for the UV detector was 286 nm, and the mobile phase was 10:90 (v:v) of acetonitrile and deionized water with 0.8% acetic acid. The UV detector wavelength for NOR analysis was 280 nm, and the mobile phase was 5:95 (v:v) of acetonitrile and deionized water with 0.1% acetic acid. PHE was detected using a fluorescence detector with the excitation wavelength of 265 nm and emission wavelength of 332 nm. The mobile phase was 80:20 (v:v) of methanol and deionized water. The flow rate for all the mobile phase was 1 mL min1.
where DG, DH, and DS are the standard Gibbs free energy, standard enthalpy and standard entropy change, respectively. DH and DS were obtained from the slope and intercept of the linear plot of T against DG, respectively. R is the universal gas constant [8.314 103 kJ (mol K)1]. T is the absolute temperature (K), and Kd is the single-point adsorption coefficient.
2.4. Data analysis The logarithmic forms of the Freundlich model (FM) and Polanyi-Manes model (PMM) were used for data processing:
FM : logQe ¼ logKF þ n logCe
(1)
PMM : logQe ¼ logQ 0 þ ZðRTlnðCs =Ce ÞÞa 1
(2) 1
In these equations, Qe (mg kg ) and Ce (mg L ) represent the equilibrium solid-phase and aqueous-phase concentrations, respectively. KF [(mg kg 1)/(mg L1)n] is the Freundlich sorption coefficient, and n is the Freundlich nonlinearity factor. Q0 (mg kg 1) is the PMM adsorption capacity. Z and a are fitting parameters of PMM. RTln(Cs/Ce) (kJ mol 1) is the effective adsorption potential, in which R is the universal gas constant [8.314 103 kJ (mol K)1], and T is the absolute temperature. Cs (mg L1) is the solubility of adsorbates. The site energy distribution, F (E*), is obtained by differentiating Qe (E*) to E*
FðE Þ ¼ dQ eðE Þ=dE
(3)
1
E* (kJ mol ) is the difference between the sorption energy at Ce and Cs. The relationship between E* and Ce can be described by the following equation:
Ce ¼ Cs expðE =RTÞ
(4)
Thus, for PMM model, Equation (5) is obtained by Equations (2)e(4):
FðE*Þ ¼ lnð10Þ Q 0 Z a E*ða1Þ 10ðZE
a
Þ
(5)
Adsorption/desorption hysteresis was investigated by index of irreversibility (TII). In this study, a Freundlich form of TII was adopted:
. TII ¼ 1 ndesorption nadsorption
(6)
where ndesorption and nadsorption are the nonlinear factors for desorption and adsorption isotherms, respectively.
3. Results and discussion 3.1. Characterization of CNTs The C and O contents in the original CNTs were 97% and 0.5%, respectively (Table 1). The surface elemental composition was also analyzed using XPS. The surface C and O contents were 95.8% and 4.2%, respectively. The much higher O content on CNTs surface in comparison to that of the bulk CNTs indicated that most O-containing functional groups were located on CNTs surface. The fluorination process greatly altered CNT properties. The first distinct change is the F content, which increased to 39.3% after fluorination. Consequently, C content decreased from 95.8% to 57.5% after fluorination. Considering the structure of MC, F could be incorporated on CNT surface structure in three ways: 1) directly react with C in benzene rings or with the C in defect sites of CNTs to form CeF bond. 2) Replace the carboxyl group on the surface of CNTs to form CeF bond. 3) Coat on the surface of CNTs as original PTFE. It is unlikely that CNT surface F was solely from the replacement of carboxyl groups, because the incorporated F was 10 times higher than O content. TEM as well as SEM was conducted to compare CNTs before and after fluorination. Two obvious differences could be noted in Fig. 1: 1) the outer diameter and the thickness of the CNTs were increased notably after fluorination. 2) Some of the CNT aggregates were “fused” together. According to measurement using EDAX connected with SEM, CNTs of the increased outer diameter (OD) and “fused” aggregates showed comparable F contents with PTFE (Fig. S1). This result clearly indicated that PTFE was coated on CNT surface. FTIR analysis provided further evidence of the abundance of surface CeF bond (Fig. 1). For example, the strong peaks at 1180 cm1 and 1220 cm1 are assigned to CeF stretching vibrations (Zhang et al., 2010b; Maiti et al., 2012). The surface coating of PTFE on CNT surface also resulted in decreased specific surface area, decreased pore volume, and increased pore radius (Table 1). The decreased surface areas after fluorination may be understood from the following two reasons: 1) PTFE on the surface of MC contributed to the weight of CNTs, and thus resulted in smaller specific surface area. 2) Fluorination may have “fused” some of the CNTs and decreased the surface area. 3.2. Fitting of adsorption isotherms The adsorption isotherms of SMX, OFL, BPA, NOR and PHE on MC and MC-F are shown in Fig. 2. The adsorption isotherms of these five chemicals were fitted with FM and PMM. The r2adjs of FM values were in the range of 0.951e0.999, and were in the range of 0.971e0.999 for PMM. Both models have a satisfactory fitting performance. Considering the obvious curvature of some of the isotherms (such as PHE at 25 C in Fig. 2 and OFL at 4 C in Fig. S3), the main discussions were based on the fitting results of PMM. The fitted parameters of PMM are listed in Table 2, and the fitted parameters of FM are listed in Table S1. Single point adsorption coefficients, Kds, were calculated at
H. Li et al. / Chemosphere 154 (2016) 258e265
261
Fig. 1. SEM images of MC (a) and MC-F (b); TEM images of MC (c) and MC-F (d). The FTIR spectra of MC-F and MC (e). The strong peaks at 1180 cm1 and 1220 cm1 are assigned to CeF stretching vibrations.
logQe (mg kg-1)
4
6.5
A
SMX BPA OFL NOR PHE
5
B 5.5 4.5
3
3.5
2
2.5 -2
-1
0
1
2
-2
-1
0
1
2
-1
logCe (mg L ) Fig. 2. Adsorption isotherms of SMX, OFL, BPA, NOR and PHE on MC (A) and MC-F (B) at 25 C. The sorption isotherms of PHE were in a smaller range than the other chemicals, because of its low solubility in aqueous phase. The adsorption isotherms at 4 C are presented in Fig. S3.
Table 2 Isotherm fitting results for SMX, OFL, BPA, NOR and PHE using PMM. Solids
T ( C)
r2adja
logQ0
Z
Kd (L kg1)b
a
Ce ¼ 0.5 mg L1 SMX
MC MC-F
OFL
MC MC-F
BPA
MC MC-F
NOR
MC MC-F
PHE
MC MC-F
a b
25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4
0.997 0.990 0.997 0.990 0.996 0.996 0.999 0.989 0.994 0.997 0.996 0.998 0.998 0.997 0.996 0.993 0.995 0.994 0.971 0.993
r2adj ¼ adjusted coefficient of determination. Kd ¼ Qe/Ce (calculated from the fitted model equations).
6.081 4.067 6.238 4.728 8.275 5.803 7.700 6.015 7.393 6.237 7.086 6.386 6.015 4.832 7.091 5.446 3.208 3.150 3.461 3.574
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.279 0.054 0.176 0.065 1.031 0.175 0.274 0.112 0.715 0.215 0.271 0.097 0.170 0.069 0.293 0.100 0.038 0.027 0.021 0.009
0.521 0.028 0.222 0.073 0.585 0.010 0.149 0.001 0.426 0.377 0.150 0.180 0.239 0.134 0.322 0.121 0.130 0.128 0.041 0.094
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.146 0.009 0.062 0.021 0.408 0.006 0.053 0.001 0.276 0.100 0.066 0.036 0.059 0.024 0.117 0.034 0.025 0.020 0.012 0.008
0.674 1.622 0.925 1.264 0.699 1.937 1.039 3.098 0.766 0.795 1.033 1.006 0.891 1.127 0.814 1.152 1.003 1.065 1.480 1.121
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.071 0.119 0.079 0.099 0.160 0.191 0.091 0.325 0.160 0.074 0.120 0.061 0.067 0.059 0.096 0.094 0.080 0.074 0.163 0.052
933 621 4092 2172 3318 37,840 21536 200,909 14,908 10,544 65,241 50,526 3058 661 20,687 3322 1592 1727 4132 5170
262
H. Li et al. / Chemosphere 154 (2016) 258e265
Ce ¼ 0.5 mg L1 according to the fitting result of PMM (Equation (2)) and are also listed in Table 2). 3.3. Possible adsorption mechanisms of the five chemicals on MC For organic chemical adsorption on the hydrophobic surfaces of CNTs, hydrophobic effects may play an important role in the overall adsorption (Apul and Karanfil, 2015). PHE possesses the highest logKow (Table 3), indicating that PHE was more hydrophobic than the other four chemicals. However, PHE exhibited much lower adsorption than BPA, OFL and NOR. SMX exhibited the lowest adsorption even though the logKow of SMX was not the lowest one. Therefore, the dominant role of hydrophobic effects could be excluded. Several studies have reported that oxygen containing functional groups on carbon nanotubes may form water cluster (Yang and Xing, 2010; Wu et al., 2012), which will decrease the hydrophobicity of carbon nanotubes, resulting in inhibiting adsorption of hydrophobic organic contaminants (HOCs). Literatures (Wang et al., 2010; Li et al., 2011) have also reported that hydrogen bonds formed between functional groups of CNTs and chemicals could enhance the adsorption. The hydrogen bond can form between a hydrogen atom and an electronegative atom such as N and O. The functional groups, such as eOH on BPA, NH2- and eNHe on SMX, eCOOH on OFL, as well as eCOOH and eNHe on NOR, can form hydrogen bond with the carboxyl group on MC. However, PHE has no functional groups, which cannot form hydrogen bond with MC. BPA exhibited the highest adsorption as presented in Fig. 2. It has been previously reported that the “butterfly” molecular structure of BPA could wedge into the groove region of CNT bundles, leading to a high adsorption of BPA on CNTs (Pan et al., 2008). The morphology benefit of BPA may account for its high adsorption on MC. In comparison to the other chemicals, SMX displayed a much lower pKa (Table 2). The pH values of the solutions after adsorption equilibrium were 6.8 ± 0.3. SMX molecules will be dissociated and negatively charged. Both OFL and NOR have two pKas, and they are zwitterionic at pH around 7. BPA molecule has pKa of 10.1, and thus most BPA molecules will not dissociate at pH around 7. PHE has no pKa because it has no dissociable functional groups. The MC will also be negatively charged in the sorption systems because of the low pHzpc of MC (which is 4.0). Therefore, the electrostatic
interactions between the SMX and MC likely contributed to the lowest adsorption of SMX. The sorption of OFL and NOR on CNTs was comparable, probably because of their similar molecular structure and chemical properties. Our previous study have suggested that structure-controlled sorption is important for OFL and NOR sorption, which is the reason of why these two chemicals have solubilities of one order of magnitude difference, but have the similar sorption (Peng et al., 2014).
3.4. Enhanced adsorption of the chemicals on MC-F As shown in Fig. 2, the adsorption of all five chemicals was increased when the CNTs were fluorinated. Previous studies have emphasized the specific surface area of CNTs played an important role in the overall adsorption (Zhang et al., 2010a). Comparing the specific surface area of MC and MC-F, the specific surface area of CNTs was actually decreased after fluorination. However, the adsorption coefficients increased up to 7 times. It is also very interesting to notice that the adsorption sequence did not change for five chemicals on MC and MC-F. We thus speculate that after modification, some adsorption interactions were enhanced or new adsorption sites appeared after fluorination. The most distinct difference in their elemental compositions between MC and MC-F was F. FTIR results have indicated the abundance of CeF bond on the surface of MC-F. The high electronegativity of F in comparison to O, and the abundance of F on MC-F surface may enable MC-F to act as a strong p-acceptor. Thus strong pp interaction is expected for chemicals with p-donor properties (Pan and Xing, 2008). Because of two hydroxyl groups connected with benzene rings, BPA molecular will act as p-donor. The pp interaction between BPA and MC-F will be stronger than BPAeMC pairs, leading to a higher adsorption of BPA on MC-F. However, OFL and NOR are both p-acceptors because of F and N-heteroaromatic rings connected with benzene ring. SMX is also p-acceptor because of the N-heteroaromatic ring and sulfonyl group. Therefore, the adsorption of OFL, NOR and SMX on MC-F will not be higher than that on MC if pp interaction is dominant. The high adsorption of MC-F may be attributed to some other mechanisms. As discussed in Section 3.1, the original PTFE may coat on the outer surface of MC during the fluorination process, resulting in the increased outer diameter. Studies have systematically investigated
Table 3 Selected chemical properties of the adsorbates. Adsorbates
Chemical formula
MW (g mol1)
pKa
Molar volume (cm3 mol1)
Solubility (mg L1)
BPA
C15H16O2
228
10.1
199.5
285a, 56b
2.2
SMX
C10H11N3O3S
253
1.7, 5.7
152.6
356a, 93b
0.9
OFL
C18H20FN3O4
361
6.1, 8.3
288.4
3400a, 197b
0.35
NOR
C16H18FN3O3
319
6.2, 8.6
256.3
320a, 114b
0.6
PHE
C14H11
178
e
160.6
1.3a, 1.01b
a b
Measured at 25 C. Measured at 4 C.
logKow
4.74
Molecular structure
H. Li et al. / Chemosphere 154 (2016) 258e265
generated during PTFE coating may provide a large number of adsorption sites for different types of chemicals. In addition, CNT aggregation will be partially dispersed because of the surface coating of PTFE and the increased electrostatic repulsion after fluorination (decreased pHzpc). All the above reasons contribute to the increased adsorption on CNTs after fluorination.
the thermal degradation of PTFE under nitrogen atmosphere (Odochian et al., 2011, 2014). The dominant thermal degradation product of PTFE in the range of 460e500 C is gaseous tetrafluoroethylene (C2F4), which was acted as the F source for fluorination of MC. The melting point of PTFE is 327 C, and the heating process resulted in PTFE polymer transit from a crystalline state to an amorphous state. In addition, the generation of gaseous product may make the compact texture of PTFE porous. However, this hypothesis seems controversy to the surface area measured using N2. It should be noted that the specific surface area and pore analysis were determined by BET with N2 in 77 K (196 C), and the glass-transition temperature of PTFE is about 100 C (Rae and Dattelbaum, 2004). The rubbery structure of PTFE will shrink in this low temperature and thus transited from rubbery state to glassy state. This process may result in underestimate of the surface area as well as the porosity of the coated PTFE. Similar phenomenon was reported for soil organic matter, which is considered a dual-mode sorbent with both glassy and rubbery state (Pignatello, 1998). Therefore, the porous structure
1.2
1
1
0.8
TII
0.8
3.5. Desorption hysteresis of chemicals adsorbed on the MC-F and MC Desorption hysteresis of the adsorbed chemicals was described by TII. The Qe vs. TII plots are presented in Fig. 3. Desorption hysteresis (with TII values higher than 0) was generally observed in all the sorption systems in this work. This desorption hysteresis may be understood from the rearrangement of CNT bundles or aggregates (Pan et al., 2008) or the strong interactions between chemicals and functional groups on CNT surfaces (Yang and Xing, 2010). The TII values of MC-F were higher than that of MC, indicating the stronger
0.8
0.6
0.6 OFL-MC OFL-MC-F SMX-MC SMX-MC-F
0.4 0.2 0
263
1.5
2.5
3.5
4.5
5.5
0.4
0.6
0
NOR-MC NOR-MC-F
BPA-MC BPA-MC-F
0.2 2
3 4 5 logQe (mg kg-1)
6
0.4
PHE-MC PHE-MC-F
2
2.5
3
3.5
4
Fig. 3. TII at different solid-phase concentrations.
-1
E* (kj mol )
30
20
20
10
10
0
F(E*) (×103 mol mg kJ-1kg-1)
30
A
0 2
3
4
2.5
5
logQe (mg kg-1)
3.5
4.5
5.5
25
5
C
SMX BPA OFL NOR PHE
D
4
20
3
15
2
10
1
5
0
B
0 0
10
20
30
0
10
20
30
-1
E* (kj mol ) Fig. 4. Adsorption free energy of chemicals at different solid-phase concentrations on MC (A) and MC-F (B) at 25 C. Site energy distribution of chemicals adsorption on MC (C) and MC-F (D) at 25 C.
264
H. Li et al. / Chemosphere 154 (2016) 258e265
desorption hysteresis of the former. As discussed earlier, MC-F aggregates are less compact than MC. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. Among the five chemicals investigated, OFL and BPA exhibited very high TII values. Although these two chemicals showed relatively high sorption than the other chemicals, it was previous reported that the strength of sorption could not be determined from their apparent sorption (Wang et al., 2014). Thus, energy analysis was conducted to provide further information to understand desorption hysteresis.
3.6. Adsorption energy analysis and thermodynamic studies The sorption site energy distribution of chemicals on MC and MC-F were calculated and are presented in Fig. 4. E* was calculated as the sorption energy difference between Ce and Cs. According to Equation (4), E* calculation was independent of adsorbent properties, but adsorbate solubility. This is why the calculated E* was comparable for MC and MC-F. The E* values decreased with the increased of Qe, indicating that the adsorbate may occupy high
energy sorption sites firstly and then spreads to low energy adsorption sites. The density of sorption energy distribution, F(E*), was about 5 times higher for MC-F than MC at the same E* (comparing Fig. 4C and D). This difference suggested that although the sorption energy was comparable, the abundance of sorption sites was much higher for MC-F. This difference is consistent with the higher sorption of MC-F, which was resulted from the additional sorption sites by PTFE surface coating and CNT disaggregate as discussed earlier. A very important observation is that both F(E*) and E* were higher for BPA and OFL than the other three chemicals. Therefore, the higher energy at the same Qe and the high energy distribution density at the same E* probably one of the important reasons for the high sorption and strong desorption hysteresis of BPA and OFL. Thermodynamic parameters including standard Gibbs free energy change (DG), standard enthalpy change (DH), and standard entropy change (DS) were calculated based on Equations (7)e(8), and were presented in Fig. 5. As shown in Fig. 5, the negative DG values for all the chemicals indicated that all the adsorption processes are thermodynamically favorable and spontaneous. The less
Fig. 5. Gibbs free energy change (DG) of chemicals sorption on MC (A) and MC-F (B) at 25 C. The standard enthalpy change (DH) of chemicals adsorption on MC (C) and MC-F (D) and standard entropy change (DS) of chemicals adsorption on MC (E) and MC-F (F).
H. Li et al. / Chemosphere 154 (2016) 258e265
negative the DG value, the weaker the driving force of the adsorption. The DH values of BPA, SMX and NOR were positive, indicating that adsorption of BPA, SMX and NOR on these sorbents were endothermic; whereas, the DH values of OFL were negative, suggesting exothermic. DS values also suggested that OFL is the only chemical that showed negative DS values, inferring decreased orderliness of the sorption system. The system with OFL as the adsorbate showed higher orderliness (more negative DS values) than the other systems. Considering the very high solubility of OFL in water, the strong interactions between OFL and water molecules resulted in a highly disordered system. The adsorption of OFL on CNTs may have shrunk the water shells and the packing of OFL molecules on the sorption sites increased the orderliness of the system, which consequently led to negative DS values. Combining the information of energy analysis, the butterfly structure of BPA resulted in its high sorption and strong desorption hysteresis. The exothermic sorption character of OFL on CNTs resulted in its strong desorption hysteresis. 4. Conclusions The main conclusions of this study were reached combining the results of batch adsorption, desorption and thermodynamics experiments of MC-F adsorption for five kinds of organic compounds. The adsorption of all five chemicals was increased when the CNTs were fluorinated. The enhanced adsorption of MC-F may be attributed the pores generated on the coated PTFE and the dispersion of CNT aggregates due to their increased electrostatic repulsion after fluorination. Desorption hysteresis was generally observed in all the sorption systems in this study, and the desorption hysteresis of MC-F were stronger than the pristine CNTs. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. The results of this study are expected to offer fundamental information for application of fluorinated CNTs in water treatment and understanding the environmental behavior between fluorinated CNTs and organic contaminants. Acknowledgments This research was supported by National Natural Scientific Foundation of China (41222025, 41273138 and 41361086), and Recruitment Program of Highly-Qualified Scholars in Yunnan (2010CI109). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2016.03.099. References Apul, O.G., Karanfil, T., 2015. Adsorption of synthetic organic contaminants by carbon nanotubes: a critical review. Water Res. 68, 34e55. Chand, R., Watari, T., Inoue, K., Luitel, H.N., Torikai, T., Yada, M., 2009. Chemical modification of carbonized wheat and barley straw using HNO3 and the adsorption of Cr(III). J. Hazard. Mater. 167, 319e324. Dunitz, J.D., Taylor, R., 1997. Organic fluorine hardly ever accepts hydrogen bonds. Chemistry-A Eur. J. 3, 89e98. Eatemadi, A., Daraee, H., Karimkhanloo, H., Kouhi, M., Zarghami, N., Akbarzadeh, A., Abasi, M., Hanifehpour, Y., Joo, S.W., 2014. Carbon nanotubes: properties, synthesis, purification, and medical applications. Nanoscale Res. Lett. 9.
265
Ghobadi, J., Arami, M., Bahrami, H., Mahmoodi, N.M., 2014. Modification of carbon nanotubes with cationic surfactant and its application for removal of direct dyes. Desalination Water Treat. 52, 4356e4368. Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 354, 56e58. Lee, Y.-S., 2007. Syntheses and properties of fluorinated carbon materials. J. Fluor. Chem. 128, 392e403. Li, M., Hsieh, T.-C., Doong, R.-A., Huang, C.P., 2013. Tuning the adsorption capability of multi-walled carbon nanotubes to polar and non-polar organic compounds by surface oxidation. Sep. Purif. Technol. 117, 98e103. Li, X., Zhao, H., Quan, X., Chen, S., Zhang, Y., Yu, H., 2011. Adsorption of ionizable organic contaminants on multi-walled carbon nanotubes with different oxygen contents. J. Hazard. Mater. 186, 407e415. Liao, Q., Sun, J., Gao, L., 2008. Adsorption of chlorophenols by multi-walled carbon nanotubes treated with HNO3 and NH3. Carbon 46, 553e555. Lu, C., Su, F., Hu, S., 2008. Surface modification of carbon nanotubes for enhancing BTEX adsorption from aqueous solutions. Appl. Surf. Sci. 254, 7035e7041. Lu, Y., Zhou, Y., Yan, Q., Fong, E., 2016. Bio-inspired synthesis of N, F co-doped 3D graphitized carbon foams containing manganese fluoride nanocrystals for lithium ion batteries. J. Mater. Chem. A 4, 2691e2698. Maiti, J., Kakati, N., Lee, S.H., Yoon, Y.S., 2012. Fluorination of multiwall carbon nanotubes by a mild fluorinating reagent HPF6. J. Fluor. Chem. 135, 362e366. Odochian, L., Moldoveanu, C., Maftei, D., 2014. TG-FTIR study on thermal degradation mechanism of PTFE under nitrogen atmosphere and in air. Influence of the grain size. Thermochim. Acta 598, 28e35. Odochian, L., Moldoveanu, C., Mocanu, A.M., Carja, G., 2011. Contributions to the thermal degradation mechanism under nitrogen atmosphere of PTFE by TGFTIR analysis. Influence of the additive nature. Thermochim. Acta 526, 205e212. Osikoya, A.O., Wankasi, D., Vala, R.M.K., Afolabi, A.S., Dikio, E.D., 2014. Synthesis, characterization and adsrotpion studies of fluorine-doped carbon nanotubes. Dig. J. Nanomater. Biostructures 9, 1187e1197. Pan, B., Lin, D., Mashayekhi, H., Xing, B., 2008. Adsorption and hysteresis of bisphenol A and 17a-ethinyl estradiol on carbon nanomaterials. Environ. Sci. Technol. 42, 5480e5485. Pan, B., Xing, B., 2008. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42, 9005e9013. Patino, Y., Diaz, E., Ordonez, S., Gallegos-Suarez, E., Guerrero-Ruiz, A., RodriguezRamos, I., 2015. Adsorption of emerging pollutants on functionalized multiwall carbon nanotubes. Chemosphere 136, 174e180. Peng, H., Li, H., Wang, C., Zhang, D., Pan, B., Xing, B., 2014. Sorption and solubility of ofloxacin and norfloxacin in water-methanol cosolvent. Chemosphere 103, 322e328. Pignatello, J.J., 1998. Soil organic matter as a nanoporous sorbent of organic pollutants. Adv. Colloid Interface Sci. 76, 445e467. Rae, P.J., Dattelbaum, D.M., 2004. The properties of poly (tetrafluoroethylene) (PTFE) in compression. Polymer 45, 7615e7625. Ren, X., Chen, C., Nagatsu, M., Wang, X., 2011. Carbon nanotubes as adsorbents in environmental pollution management: a review. Chem. Eng. J. 170, 395e410. Wang, C., Li, H., Liao, S., Zhang, D., Wu, M., Pan, B., Xing, B., 2014. Sorption affinities of sulfamethoxazole and carbamazepine to two sorbents under co-sorption systems. Environ. Pollut. 194, 203e209. Wang, F., Yao, J., Sun, K., Xing, B., 2010. Adsorption of dialkyl phthalate esters on carbon nanotubes. Environ. Sci. Technol. 44, 6985e6991. Wu, W., Chen, W., Lin, D., Yang, K., 2012. Influence of surface oxidation of multiwalled carbon nanotubes on the adsorption affinity and capacity of polar and nonpolar organic compounds in aqueous phase. Environ. Sci. Technol. 46, 5446e5454. Yang, K., Xing, B., 2010. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: polanyi theory and its application. Chem. Rev. 110, 5989e6008. Yu, J.-G., Zhao, X.-H., Yang, H., Chen, X.-H., Yang, Q., Yu, L.-Y., Jiang, J.-H., Chen, X.-Q., 2014. Aqueous adsorption and removal of organic contaminants by carbon nanotubes. Sci. Total Environ. 482, 241e251. Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2009. The Impacts of aggregation and surface chemistry of carbon nanotubes on the adsorption of synthetic organic compounds. Environ. Sci. Technol. 43, 5719e5725. Zhang, S., Shao, T., Kose, H.S., Karanfil, T., 2010a. Adsorption of aromatic compounds by carbonaceous adsorbents: a comparative study on granular activated carbon, activated carbon fiber, and carbon nanotubes. Environ. Sci. Technol. 44, 6377e6383. Zhang, W., Dubois, M., Guerin, K., Bonnet, P., Kharbache, H., Masin, F., Kharitonov, A.P., Hamwi, A., 2010b. Effect of curvature on C-F bonding in fluorinated carbons: from fullerene and derivatives to graphite. Phys. Chem. Chem. Phys. 12, 1388e1398. Zhang, W., Spinelle, L., Dubois, M., Guerin, K., Kharbache, H., Masin, F., Kharitonov, A.P., Hamwi, A., Brunet, J., Varenne, C., Pauly, A., Thomas, P., Himmel, D., Mansot, J.L., 2010c. New synthesis methods for fluorinated carbon nanofibres and applications. J. Fluor. Chem. 131, 676e683.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Adsorption mechanism of different organic chemicals on fluorinated carbon nanotubes Hao Li, Nan Zheng, Ni Liang, Di Zhang, Min Wu, Bo Pan* Faculty of Environmental Science & Engineering, Kunming University of Science & Technology, Kunming, 650500, PR China
h i g h l i g h t s Carbon nanotubes were fluorinated using PTFE as a fluorine source. Organic contaminant adsorption on CNTs was enhanced after fluorination. BPA butterfly structure resulted in its strong sorption and desorption hysteresis. OFL exothermic sorption resulted in its strong desorption hysteresis.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 November 2015 Received in revised form 12 March 2016 Accepted 21 March 2016
Multi-walled carbon nanotubes (MC) were fluorinated by a solid-phase reaction method using polytetrafluoroethylene (PTFE). The surface alteration of carbon nanotubes after fluorination (MC-F) was confirmed based on surface elemental analysis, TEM and SEM. The incorporation of F on MC surface was discussed as F incorporation on carbon defects, replacement of carboxyl groups, as well as surface coating of PTFE. The adsorption performance and mechanisms of MC-F for five kinds of representative organic compounds: sulfamethoxazole (SMX), ofloxacin (OFL), norfloxacin (NOR), bisphenol a (BPA) and phenanthrene (PHE) were investigated. Although BET-N2 surface area of the investigated CNTs decreased after fluorination, the adsorption of all five chemicals increased. Because of the glassification of MC-F surface coating during BET-N2 surface area measurement, the accessible surface area of MC-F was underestimated. Desorption hysteresis was generally observed in all the sorption systems in this study, and the desorption hysteresis of MC-F were stronger than the pristine CNTs. The enhanced adsorption of MC-F may be attributed the pores generated on the coated PTFE and the dispersed CNT aggregates due to the increased electrostatic repulsion after fluorination. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. The butterfly structure of BPA resulted in its high sorption and strong desorption hysteresis. The exothermic sorption character of OFL on CNTs resulted in its strong desorption hysteresis. © 2016 Elsevier Ltd. All rights reserved.
Handling Editor: X. Cao Keywords: Aggregation Desorption hysteresis Fluorination Nanoparticles Sorption thermodynamics
1. Introduction Since carbon nanotubes (CNTs) was discovered by Iijima in 1991 (Iijima, 1991), which has aroused extensive attentions due to their unique and superior physicochemical properties. CNTs have potential applications in many areas including manufacturing, electronics, construction, and medicine (Eatemadi et al., 2014). CNTs exhibited a very strong affinity for organic contaminates and heavy metal ions, thus it is proposed as a superior and promising sorbent applied in wastewater treatment (Ren et al., 2011; Yu et al., 2014). * Corresponding author. E-mail address: [email protected] (B. Pan). http://dx.doi.org/10.1016/j.chemosphere.2016.03.099 0045-6535/© 2016 Elsevier Ltd. All rights reserved.
In order to maximize the selectivity of adsorption and to improve the adsorption performance of CNTs, surface modification of CNTs has drawn a widely attention (Lu et al., 2008; Ghobadi et al., 2014). In most cases, acidic or oxidizing reagents were applied to treat CNTs, and thus oxygen containing functional groups (such as hydroxyl and carboxyl groups) were incorporated on the surface of CNTs (Liao et al., 2008; Chand et al., 2009). The oxygen-containing functional groups will facilitate the wettability properties of the surface, and then enhance the adsorption for polar compounds (Patino et al., 2015). And also, the oxygen-containing functional groups will make the CNTs more negatively charged in neutral condition, which could enhance the dispersion of CNTs in solutions due to the electrostatic repulsion, leading to a higher adsorption than pristine
H. Li et al. / Chemosphere 154 (2016) 258e265
CNTs. However, the interactions between oxide-carbon nanotubes and hydrophobic compounds will dramatic decrease in adsorption as the increasing extent of CNTs oxidation (Zhang et al., 2009). Li et al. reported that PHE exhibited a much lower adsorption capacity on oxidized CNTs than on pristine CNTs (Li et al., 2013). The oxygencontaining groups on surface of CNTs can form strong hydrogen bond with water molecules, resulting in decreasing adsorption of organic compounds, especially hydrophobic organic compounds (Wu et al., 2012). Therefore, CNT surface modification may result in the increased sorption of certain chemicals, but will also decrease its sorption with other chemicals. In most of the circumstances, various organic contaminants with different properties simultaneously present in wastewater or the environment. Effective removal of organic contaminants with a wide range of chemical physical properties is a major challenge for researchers. Fluorination is one of the most effective routes to modify and control physico-chemical properties of carbon materials (Lee, 2007; Lu et al., 2016). Fluorinated CNTs was reported to enhance surface polarity, electroconductivity and capacitive character, and is widely used in lithium batteries, solid lubricants and nano-composites (Zhang et al., 2010c; Maiti et al., 2012). Recent study has reported that fluorinated CNTs exhibited great performance in removing chromium ions (Osikoya et al., 2014). However, the study using fluorinated CNTs as adsorbent to retain organic compounds from aqueous is relatively rare. Due to the utmost electronegativity of F, the CeF bond on fluorinated CNTs enables CNTs to act as a strong pacceptor, which could enhance the adsorption controlled by pacceptor and p-donor interaction. The strong electronegativity of F will also make the CNTs strongly negatively charged, and thus facilitate the dispersion of CNTs. In addition, different from oxygen containing groups, the CeF bond can hardly form hydrogen bond with water molecules (Dunitz and Taylor, 1997). Therefore, incorporating F into CNTs will not suppress the adsorption of hydrophobic compounds. We thus hypothesized that fluorination will enhance the adsorption of organic compounds on CNTs. In this study, fluorinated CNTs was synthesized by the solidphase reaction method. Five representative organic contaminants including sulfamethoxazole (SMX), ofloxacin (OFL), norfloxacin (NOR), bisphenol A (BPA) and phenanthrene (PHE) were selected as adsorbates. Batch adsorption, desorption and thermodynamics experiments were conducted to evaluate the adsorption characteristics and to investigate the adsorption mechanisms of MC-F. In addition, with the wide applications of fluorinated CNTs, it will be inevitably released into the aqueous environment. Therefore, investigating the interactions between fluorinated CNTs and organic contaminants is of great environmental significance. The aim of this work is to illustration the sorption mechanisms of organic contaminants on fluorinated CNTs, which will be an important input for CNT applications in wastewater treatment. 2. Materials and methods 2.1. Materials Carboxylized multiwall carbon nanotubes (MC, purity > 95%, diameter: 8e15 nm) were purchased from Chengdu Organic
259
Chemistry Co., Chinese Academy of Sciences. It was synthesized in the CH4/H2 mixture at 700 C by the chemical vapor deposition method and was purified by mixed HNO3 and H2SO4 solutions to reduce the contents of metal catalyst and amorphous carbon. Fluorinated multiwall carbon nanotubes were synthesized by the solid-phase reaction method. Briefly, the MC and PTFE (5 mm, Aladdin-reagent Co. Ltd., China) fine particles were mixed in an enclosed vessel according weight ratio: 1:2 (MC:PTFE), then the mixture was put in a temperature-controlled furnace which was purged continuously with N2. The temperature was controlled at 460e500 C for 4e6 h to generate the source of fluorine for fluorination. The resulted particles were chilled and grinded. This particle is noted as MC-F. SMX (AR, 99.0%), OFL (AR, 99.0%) and NOR (AR, 99.0%) were obtained from Bio Basic Inc. PHE (AR, 97%) was obtained from Acros Co., Belgium. BPA (AR, 99.0%) was obtained from Sinopharm Chemical Reagent Co., Ltd, China. Both MC and MC-F were characterized for their surface area (Autosorb-1C, Quantachrome, US), elemental composition (MicroCube, Elementar, Germany), surface elemental composition (X-ray photoelectron spectroscopy, PHI5000 VersaProbe II, ULVACPHI, INC.) scanning electron microscopy (SEM, Tescan, VEGA3 SBH) equipped with energy-dispersiveX-ray spectrometer (EDAX, Thermo Fisher Scientific, NORAN System 7), transmission electron microscopy (TEM, JEM 2100) and zero point of charge (pHzpc, ZetaPlus, Brookhaven, US). For SEM-EDAX tests, MC or MC-F particles were mounted on aluminum stubs, and then gold-coated using a sputter coater. The accelerating voltage was set at 20 kV. For TEM tests, the MC and MC-F solutions were ultra-sonicated for 5 min, and then a drop of solution containing MC or MC-F was put onto a copper TEM grid. After air-dried, the measurements were operated with the accelerating voltage of 200 kV. The properties of compounds studied are listed in Table 1. 2.2. Batch adsorption experiments Batch adsorption experiments were separately conducted for SMX, BPA, OFL, NOR and PHE in 4 mL (for MC) and 15 mL (for MC-F) amber glass vials with Teflon-lined screw caps. According to our preliminary work, the aqueous:solid ratios were 2000:1 and 15000:1 (w/w) for MC and MC-F sorption experiments, respectively. The different aqueous:solid ratios were to ensure comparable equilibrated aqeuous phase concentrations in two sorption systems. The initial concentration ranges were 0.2e1.24 mg L1 for PHE and 1e50 mg L1 for the other four chemicals, and all the adsorbates were dissolved in background solution containing 0.02 M NaCl and 200 mg L1 NaN3. We did not use CaCl2 in the background solution to avoid possible impact of Ca2þ on SMX sorption. The adsorbate solutions without adsorbents were referred to as the initial concentration references. All vials were kept in the dark and were continuously shaken in an air-bath shaker (80 rpm) at 25 C or in a refrigerator at 4 C for 7 d. According to our preliminary study, 7 d was enough to reach apparent sorption equilibrium. All vials were then centrifuged at 1000 g for 10 min, and the supernatants were sampled for adsorbates quantification. For desorption experiment, the equilibrium supernatant was removed from the vials and then re-filled with the same volume of
Table 1 Property comparison of MC and MC-F. CNTs
MC MC-F
SSA (m2 g1)
118.8 17.5
Pore volume (cc g1)
0.454 0.083
Pore radius (Å)
9.6 12.2
XPS
Elemental analysis
pHzpc
C
O
F
C
O
H
95.8 57.5
4.2 3.3
0 39.3
97 56.7
0.5 0.2
2.7 2.1
4.0 2.1
260
H. Li et al. / Chemosphere 154 (2016) 258e265
background solution. After equilibration, the vials were centrifuged and the supernatants were sampled for adsorbates quantification as described before. All the samples have three replications.
The thermodynamic parameters were calculated using the following equations:
DG ¼ RT lnKd
(7)
2.3. Quantification of adsorbates
DG ¼ DS T DH
(8)
The aqueous-phase equilibrium concentrations of all the adsorbates were quantified by HPLC (Agilent Technologies 1200) equipped with a reversed-phase C18 column (5 mm, 4.6 150 mm). For SMX analysis, the wavelength for the UV detector was 265 nm, and the mobile phase was 40:60 (v:v) ratio of acetonitrile and deionized water with 0.1% acetic acid. The UV detector wavelength for BPA analysis was 280 nm, and the mobile phase was 40:60 (v:v) of acetonitrile and deionized water. For OFL analysis, the wavelength for the UV detector was 286 nm, and the mobile phase was 10:90 (v:v) of acetonitrile and deionized water with 0.8% acetic acid. The UV detector wavelength for NOR analysis was 280 nm, and the mobile phase was 5:95 (v:v) of acetonitrile and deionized water with 0.1% acetic acid. PHE was detected using a fluorescence detector with the excitation wavelength of 265 nm and emission wavelength of 332 nm. The mobile phase was 80:20 (v:v) of methanol and deionized water. The flow rate for all the mobile phase was 1 mL min1.
where DG, DH, and DS are the standard Gibbs free energy, standard enthalpy and standard entropy change, respectively. DH and DS were obtained from the slope and intercept of the linear plot of T against DG, respectively. R is the universal gas constant [8.314 103 kJ (mol K)1]. T is the absolute temperature (K), and Kd is the single-point adsorption coefficient.
2.4. Data analysis The logarithmic forms of the Freundlich model (FM) and Polanyi-Manes model (PMM) were used for data processing:
FM : logQe ¼ logKF þ n logCe
(1)
PMM : logQe ¼ logQ 0 þ ZðRTlnðCs =Ce ÞÞa 1
(2) 1
In these equations, Qe (mg kg ) and Ce (mg L ) represent the equilibrium solid-phase and aqueous-phase concentrations, respectively. KF [(mg kg 1)/(mg L1)n] is the Freundlich sorption coefficient, and n is the Freundlich nonlinearity factor. Q0 (mg kg 1) is the PMM adsorption capacity. Z and a are fitting parameters of PMM. RTln(Cs/Ce) (kJ mol 1) is the effective adsorption potential, in which R is the universal gas constant [8.314 103 kJ (mol K)1], and T is the absolute temperature. Cs (mg L1) is the solubility of adsorbates. The site energy distribution, F (E*), is obtained by differentiating Qe (E*) to E*
FðE Þ ¼ dQ eðE Þ=dE
(3)
1
E* (kJ mol ) is the difference between the sorption energy at Ce and Cs. The relationship between E* and Ce can be described by the following equation:
Ce ¼ Cs expðE =RTÞ
(4)
Thus, for PMM model, Equation (5) is obtained by Equations (2)e(4):
FðE*Þ ¼ lnð10Þ Q 0 Z a E*ða1Þ 10ðZE
a
Þ
(5)
Adsorption/desorption hysteresis was investigated by index of irreversibility (TII). In this study, a Freundlich form of TII was adopted:
. TII ¼ 1 ndesorption nadsorption
(6)
where ndesorption and nadsorption are the nonlinear factors for desorption and adsorption isotherms, respectively.
3. Results and discussion 3.1. Characterization of CNTs The C and O contents in the original CNTs were 97% and 0.5%, respectively (Table 1). The surface elemental composition was also analyzed using XPS. The surface C and O contents were 95.8% and 4.2%, respectively. The much higher O content on CNTs surface in comparison to that of the bulk CNTs indicated that most O-containing functional groups were located on CNTs surface. The fluorination process greatly altered CNT properties. The first distinct change is the F content, which increased to 39.3% after fluorination. Consequently, C content decreased from 95.8% to 57.5% after fluorination. Considering the structure of MC, F could be incorporated on CNT surface structure in three ways: 1) directly react with C in benzene rings or with the C in defect sites of CNTs to form CeF bond. 2) Replace the carboxyl group on the surface of CNTs to form CeF bond. 3) Coat on the surface of CNTs as original PTFE. It is unlikely that CNT surface F was solely from the replacement of carboxyl groups, because the incorporated F was 10 times higher than O content. TEM as well as SEM was conducted to compare CNTs before and after fluorination. Two obvious differences could be noted in Fig. 1: 1) the outer diameter and the thickness of the CNTs were increased notably after fluorination. 2) Some of the CNT aggregates were “fused” together. According to measurement using EDAX connected with SEM, CNTs of the increased outer diameter (OD) and “fused” aggregates showed comparable F contents with PTFE (Fig. S1). This result clearly indicated that PTFE was coated on CNT surface. FTIR analysis provided further evidence of the abundance of surface CeF bond (Fig. 1). For example, the strong peaks at 1180 cm1 and 1220 cm1 are assigned to CeF stretching vibrations (Zhang et al., 2010b; Maiti et al., 2012). The surface coating of PTFE on CNT surface also resulted in decreased specific surface area, decreased pore volume, and increased pore radius (Table 1). The decreased surface areas after fluorination may be understood from the following two reasons: 1) PTFE on the surface of MC contributed to the weight of CNTs, and thus resulted in smaller specific surface area. 2) Fluorination may have “fused” some of the CNTs and decreased the surface area. 3.2. Fitting of adsorption isotherms The adsorption isotherms of SMX, OFL, BPA, NOR and PHE on MC and MC-F are shown in Fig. 2. The adsorption isotherms of these five chemicals were fitted with FM and PMM. The r2adjs of FM values were in the range of 0.951e0.999, and were in the range of 0.971e0.999 for PMM. Both models have a satisfactory fitting performance. Considering the obvious curvature of some of the isotherms (such as PHE at 25 C in Fig. 2 and OFL at 4 C in Fig. S3), the main discussions were based on the fitting results of PMM. The fitted parameters of PMM are listed in Table 2, and the fitted parameters of FM are listed in Table S1. Single point adsorption coefficients, Kds, were calculated at
H. Li et al. / Chemosphere 154 (2016) 258e265
261
Fig. 1. SEM images of MC (a) and MC-F (b); TEM images of MC (c) and MC-F (d). The FTIR spectra of MC-F and MC (e). The strong peaks at 1180 cm1 and 1220 cm1 are assigned to CeF stretching vibrations.
logQe (mg kg-1)
4
6.5
A
SMX BPA OFL NOR PHE
5
B 5.5 4.5
3
3.5
2
2.5 -2
-1
0
1
2
-2
-1
0
1
2
-1
logCe (mg L ) Fig. 2. Adsorption isotherms of SMX, OFL, BPA, NOR and PHE on MC (A) and MC-F (B) at 25 C. The sorption isotherms of PHE were in a smaller range than the other chemicals, because of its low solubility in aqueous phase. The adsorption isotherms at 4 C are presented in Fig. S3.
Table 2 Isotherm fitting results for SMX, OFL, BPA, NOR and PHE using PMM. Solids
T ( C)
r2adja
logQ0
Z
Kd (L kg1)b
a
Ce ¼ 0.5 mg L1 SMX
MC MC-F
OFL
MC MC-F
BPA
MC MC-F
NOR
MC MC-F
PHE
MC MC-F
a b
25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4 25 4
0.997 0.990 0.997 0.990 0.996 0.996 0.999 0.989 0.994 0.997 0.996 0.998 0.998 0.997 0.996 0.993 0.995 0.994 0.971 0.993
r2adj ¼ adjusted coefficient of determination. Kd ¼ Qe/Ce (calculated from the fitted model equations).
6.081 4.067 6.238 4.728 8.275 5.803 7.700 6.015 7.393 6.237 7.086 6.386 6.015 4.832 7.091 5.446 3.208 3.150 3.461 3.574
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.279 0.054 0.176 0.065 1.031 0.175 0.274 0.112 0.715 0.215 0.271 0.097 0.170 0.069 0.293 0.100 0.038 0.027 0.021 0.009
0.521 0.028 0.222 0.073 0.585 0.010 0.149 0.001 0.426 0.377 0.150 0.180 0.239 0.134 0.322 0.121 0.130 0.128 0.041 0.094
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.146 0.009 0.062 0.021 0.408 0.006 0.053 0.001 0.276 0.100 0.066 0.036 0.059 0.024 0.117 0.034 0.025 0.020 0.012 0.008
0.674 1.622 0.925 1.264 0.699 1.937 1.039 3.098 0.766 0.795 1.033 1.006 0.891 1.127 0.814 1.152 1.003 1.065 1.480 1.121
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.071 0.119 0.079 0.099 0.160 0.191 0.091 0.325 0.160 0.074 0.120 0.061 0.067 0.059 0.096 0.094 0.080 0.074 0.163 0.052
933 621 4092 2172 3318 37,840 21536 200,909 14,908 10,544 65,241 50,526 3058 661 20,687 3322 1592 1727 4132 5170
262
H. Li et al. / Chemosphere 154 (2016) 258e265
Ce ¼ 0.5 mg L1 according to the fitting result of PMM (Equation (2)) and are also listed in Table 2). 3.3. Possible adsorption mechanisms of the five chemicals on MC For organic chemical adsorption on the hydrophobic surfaces of CNTs, hydrophobic effects may play an important role in the overall adsorption (Apul and Karanfil, 2015). PHE possesses the highest logKow (Table 3), indicating that PHE was more hydrophobic than the other four chemicals. However, PHE exhibited much lower adsorption than BPA, OFL and NOR. SMX exhibited the lowest adsorption even though the logKow of SMX was not the lowest one. Therefore, the dominant role of hydrophobic effects could be excluded. Several studies have reported that oxygen containing functional groups on carbon nanotubes may form water cluster (Yang and Xing, 2010; Wu et al., 2012), which will decrease the hydrophobicity of carbon nanotubes, resulting in inhibiting adsorption of hydrophobic organic contaminants (HOCs). Literatures (Wang et al., 2010; Li et al., 2011) have also reported that hydrogen bonds formed between functional groups of CNTs and chemicals could enhance the adsorption. The hydrogen bond can form between a hydrogen atom and an electronegative atom such as N and O. The functional groups, such as eOH on BPA, NH2- and eNHe on SMX, eCOOH on OFL, as well as eCOOH and eNHe on NOR, can form hydrogen bond with the carboxyl group on MC. However, PHE has no functional groups, which cannot form hydrogen bond with MC. BPA exhibited the highest adsorption as presented in Fig. 2. It has been previously reported that the “butterfly” molecular structure of BPA could wedge into the groove region of CNT bundles, leading to a high adsorption of BPA on CNTs (Pan et al., 2008). The morphology benefit of BPA may account for its high adsorption on MC. In comparison to the other chemicals, SMX displayed a much lower pKa (Table 2). The pH values of the solutions after adsorption equilibrium were 6.8 ± 0.3. SMX molecules will be dissociated and negatively charged. Both OFL and NOR have two pKas, and they are zwitterionic at pH around 7. BPA molecule has pKa of 10.1, and thus most BPA molecules will not dissociate at pH around 7. PHE has no pKa because it has no dissociable functional groups. The MC will also be negatively charged in the sorption systems because of the low pHzpc of MC (which is 4.0). Therefore, the electrostatic
interactions between the SMX and MC likely contributed to the lowest adsorption of SMX. The sorption of OFL and NOR on CNTs was comparable, probably because of their similar molecular structure and chemical properties. Our previous study have suggested that structure-controlled sorption is important for OFL and NOR sorption, which is the reason of why these two chemicals have solubilities of one order of magnitude difference, but have the similar sorption (Peng et al., 2014).
3.4. Enhanced adsorption of the chemicals on MC-F As shown in Fig. 2, the adsorption of all five chemicals was increased when the CNTs were fluorinated. Previous studies have emphasized the specific surface area of CNTs played an important role in the overall adsorption (Zhang et al., 2010a). Comparing the specific surface area of MC and MC-F, the specific surface area of CNTs was actually decreased after fluorination. However, the adsorption coefficients increased up to 7 times. It is also very interesting to notice that the adsorption sequence did not change for five chemicals on MC and MC-F. We thus speculate that after modification, some adsorption interactions were enhanced or new adsorption sites appeared after fluorination. The most distinct difference in their elemental compositions between MC and MC-F was F. FTIR results have indicated the abundance of CeF bond on the surface of MC-F. The high electronegativity of F in comparison to O, and the abundance of F on MC-F surface may enable MC-F to act as a strong p-acceptor. Thus strong pp interaction is expected for chemicals with p-donor properties (Pan and Xing, 2008). Because of two hydroxyl groups connected with benzene rings, BPA molecular will act as p-donor. The pp interaction between BPA and MC-F will be stronger than BPAeMC pairs, leading to a higher adsorption of BPA on MC-F. However, OFL and NOR are both p-acceptors because of F and N-heteroaromatic rings connected with benzene ring. SMX is also p-acceptor because of the N-heteroaromatic ring and sulfonyl group. Therefore, the adsorption of OFL, NOR and SMX on MC-F will not be higher than that on MC if pp interaction is dominant. The high adsorption of MC-F may be attributed to some other mechanisms. As discussed in Section 3.1, the original PTFE may coat on the outer surface of MC during the fluorination process, resulting in the increased outer diameter. Studies have systematically investigated
Table 3 Selected chemical properties of the adsorbates. Adsorbates
Chemical formula
MW (g mol1)
pKa
Molar volume (cm3 mol1)
Solubility (mg L1)
BPA
C15H16O2
228
10.1
199.5
285a, 56b
2.2
SMX
C10H11N3O3S
253
1.7, 5.7
152.6
356a, 93b
0.9
OFL
C18H20FN3O4
361
6.1, 8.3
288.4
3400a, 197b
0.35
NOR
C16H18FN3O3
319
6.2, 8.6
256.3
320a, 114b
0.6
PHE
C14H11
178
e
160.6
1.3a, 1.01b
a b
Measured at 25 C. Measured at 4 C.
logKow
4.74
Molecular structure
H. Li et al. / Chemosphere 154 (2016) 258e265
generated during PTFE coating may provide a large number of adsorption sites for different types of chemicals. In addition, CNT aggregation will be partially dispersed because of the surface coating of PTFE and the increased electrostatic repulsion after fluorination (decreased pHzpc). All the above reasons contribute to the increased adsorption on CNTs after fluorination.
the thermal degradation of PTFE under nitrogen atmosphere (Odochian et al., 2011, 2014). The dominant thermal degradation product of PTFE in the range of 460e500 C is gaseous tetrafluoroethylene (C2F4), which was acted as the F source for fluorination of MC. The melting point of PTFE is 327 C, and the heating process resulted in PTFE polymer transit from a crystalline state to an amorphous state. In addition, the generation of gaseous product may make the compact texture of PTFE porous. However, this hypothesis seems controversy to the surface area measured using N2. It should be noted that the specific surface area and pore analysis were determined by BET with N2 in 77 K (196 C), and the glass-transition temperature of PTFE is about 100 C (Rae and Dattelbaum, 2004). The rubbery structure of PTFE will shrink in this low temperature and thus transited from rubbery state to glassy state. This process may result in underestimate of the surface area as well as the porosity of the coated PTFE. Similar phenomenon was reported for soil organic matter, which is considered a dual-mode sorbent with both glassy and rubbery state (Pignatello, 1998). Therefore, the porous structure
1.2
1
1
0.8
TII
0.8
3.5. Desorption hysteresis of chemicals adsorbed on the MC-F and MC Desorption hysteresis of the adsorbed chemicals was described by TII. The Qe vs. TII plots are presented in Fig. 3. Desorption hysteresis (with TII values higher than 0) was generally observed in all the sorption systems in this work. This desorption hysteresis may be understood from the rearrangement of CNT bundles or aggregates (Pan et al., 2008) or the strong interactions between chemicals and functional groups on CNT surfaces (Yang and Xing, 2010). The TII values of MC-F were higher than that of MC, indicating the stronger
0.8
0.6
0.6 OFL-MC OFL-MC-F SMX-MC SMX-MC-F
0.4 0.2 0
263
1.5
2.5
3.5
4.5
5.5
0.4
0.6
0
NOR-MC NOR-MC-F
BPA-MC BPA-MC-F
0.2 2
3 4 5 logQe (mg kg-1)
6
0.4
PHE-MC PHE-MC-F
2
2.5
3
3.5
4
Fig. 3. TII at different solid-phase concentrations.
-1
E* (kj mol )
30
20
20
10
10
0
F(E*) (×103 mol mg kJ-1kg-1)
30
A
0 2
3
4
2.5
5
logQe (mg kg-1)
3.5
4.5
5.5
25
5
C
SMX BPA OFL NOR PHE
D
4
20
3
15
2
10
1
5
0
B
0 0
10
20
30
0
10
20
30
-1
E* (kj mol ) Fig. 4. Adsorption free energy of chemicals at different solid-phase concentrations on MC (A) and MC-F (B) at 25 C. Site energy distribution of chemicals adsorption on MC (C) and MC-F (D) at 25 C.
264
H. Li et al. / Chemosphere 154 (2016) 258e265
desorption hysteresis of the former. As discussed earlier, MC-F aggregates are less compact than MC. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. Among the five chemicals investigated, OFL and BPA exhibited very high TII values. Although these two chemicals showed relatively high sorption than the other chemicals, it was previous reported that the strength of sorption could not be determined from their apparent sorption (Wang et al., 2014). Thus, energy analysis was conducted to provide further information to understand desorption hysteresis.
3.6. Adsorption energy analysis and thermodynamic studies The sorption site energy distribution of chemicals on MC and MC-F were calculated and are presented in Fig. 4. E* was calculated as the sorption energy difference between Ce and Cs. According to Equation (4), E* calculation was independent of adsorbent properties, but adsorbate solubility. This is why the calculated E* was comparable for MC and MC-F. The E* values decreased with the increased of Qe, indicating that the adsorbate may occupy high
energy sorption sites firstly and then spreads to low energy adsorption sites. The density of sorption energy distribution, F(E*), was about 5 times higher for MC-F than MC at the same E* (comparing Fig. 4C and D). This difference suggested that although the sorption energy was comparable, the abundance of sorption sites was much higher for MC-F. This difference is consistent with the higher sorption of MC-F, which was resulted from the additional sorption sites by PTFE surface coating and CNT disaggregate as discussed earlier. A very important observation is that both F(E*) and E* were higher for BPA and OFL than the other three chemicals. Therefore, the higher energy at the same Qe and the high energy distribution density at the same E* probably one of the important reasons for the high sorption and strong desorption hysteresis of BPA and OFL. Thermodynamic parameters including standard Gibbs free energy change (DG), standard enthalpy change (DH), and standard entropy change (DS) were calculated based on Equations (7)e(8), and were presented in Fig. 5. As shown in Fig. 5, the negative DG values for all the chemicals indicated that all the adsorption processes are thermodynamically favorable and spontaneous. The less
Fig. 5. Gibbs free energy change (DG) of chemicals sorption on MC (A) and MC-F (B) at 25 C. The standard enthalpy change (DH) of chemicals adsorption on MC (C) and MC-F (D) and standard entropy change (DS) of chemicals adsorption on MC (E) and MC-F (F).
H. Li et al. / Chemosphere 154 (2016) 258e265
negative the DG value, the weaker the driving force of the adsorption. The DH values of BPA, SMX and NOR were positive, indicating that adsorption of BPA, SMX and NOR on these sorbents were endothermic; whereas, the DH values of OFL were negative, suggesting exothermic. DS values also suggested that OFL is the only chemical that showed negative DS values, inferring decreased orderliness of the sorption system. The system with OFL as the adsorbate showed higher orderliness (more negative DS values) than the other systems. Considering the very high solubility of OFL in water, the strong interactions between OFL and water molecules resulted in a highly disordered system. The adsorption of OFL on CNTs may have shrunk the water shells and the packing of OFL molecules on the sorption sites increased the orderliness of the system, which consequently led to negative DS values. Combining the information of energy analysis, the butterfly structure of BPA resulted in its high sorption and strong desorption hysteresis. The exothermic sorption character of OFL on CNTs resulted in its strong desorption hysteresis. 4. Conclusions The main conclusions of this study were reached combining the results of batch adsorption, desorption and thermodynamics experiments of MC-F adsorption for five kinds of organic compounds. The adsorption of all five chemicals was increased when the CNTs were fluorinated. The enhanced adsorption of MC-F may be attributed the pores generated on the coated PTFE and the dispersion of CNT aggregates due to their increased electrostatic repulsion after fluorination. Desorption hysteresis was generally observed in all the sorption systems in this study, and the desorption hysteresis of MC-F were stronger than the pristine CNTs. The rearrangement of the bundles or diffusion of the adsorbates in MC-F inner pores were the likely reason for the strong desorption hysteresis of MC-F. The results of this study are expected to offer fundamental information for application of fluorinated CNTs in water treatment and understanding the environmental behavior between fluorinated CNTs and organic contaminants. Acknowledgments This research was supported by National Natural Scientific Foundation of China (41222025, 41273138 and 41361086), and Recruitment Program of Highly-Qualified Scholars in Yunnan (2010CI109). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2016.03.099. References Apul, O.G., Karanfil, T., 2015. Adsorption of synthetic organic contaminants by carbon nanotubes: a critical review. Water Res. 68, 34e55. Chand, R., Watari, T., Inoue, K., Luitel, H.N., Torikai, T., Yada, M., 2009. Chemical modification of carbonized wheat and barley straw using HNO3 and the adsorption of Cr(III). J. Hazard. Mater. 167, 319e324. Dunitz, J.D., Taylor, R., 1997. Organic fluorine hardly ever accepts hydrogen bonds. Chemistry-A Eur. J. 3, 89e98. Eatemadi, A., Daraee, H., Karimkhanloo, H., Kouhi, M., Zarghami, N., Akbarzadeh, A., Abasi, M., Hanifehpour, Y., Joo, S.W., 2014. Carbon nanotubes: properties, synthesis, purification, and medical applications. Nanoscale Res. Lett. 9.
265
Ghobadi, J., Arami, M., Bahrami, H., Mahmoodi, N.M., 2014. Modification of carbon nanotubes with cationic surfactant and its application for removal of direct dyes. Desalination Water Treat. 52, 4356e4368. Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 354, 56e58. Lee, Y.-S., 2007. Syntheses and properties of fluorinated carbon materials. J. Fluor. Chem. 128, 392e403. Li, M., Hsieh, T.-C., Doong, R.-A., Huang, C.P., 2013. Tuning the adsorption capability of multi-walled carbon nanotubes to polar and non-polar organic compounds by surface oxidation. Sep. Purif. Technol. 117, 98e103. Li, X., Zhao, H., Quan, X., Chen, S., Zhang, Y., Yu, H., 2011. Adsorption of ionizable organic contaminants on multi-walled carbon nanotubes with different oxygen contents. J. Hazard. Mater. 186, 407e415. Liao, Q., Sun, J., Gao, L., 2008. Adsorption of chlorophenols by multi-walled carbon nanotubes treated with HNO3 and NH3. Carbon 46, 553e555. Lu, C., Su, F., Hu, S., 2008. Surface modification of carbon nanotubes for enhancing BTEX adsorption from aqueous solutions. Appl. Surf. Sci. 254, 7035e7041. Lu, Y., Zhou, Y., Yan, Q., Fong, E., 2016. Bio-inspired synthesis of N, F co-doped 3D graphitized carbon foams containing manganese fluoride nanocrystals for lithium ion batteries. J. Mater. Chem. A 4, 2691e2698. Maiti, J., Kakati, N., Lee, S.H., Yoon, Y.S., 2012. Fluorination of multiwall carbon nanotubes by a mild fluorinating reagent HPF6. J. Fluor. Chem. 135, 362e366. Odochian, L., Moldoveanu, C., Maftei, D., 2014. TG-FTIR study on thermal degradation mechanism of PTFE under nitrogen atmosphere and in air. Influence of the grain size. Thermochim. Acta 598, 28e35. Odochian, L., Moldoveanu, C., Mocanu, A.M., Carja, G., 2011. Contributions to the thermal degradation mechanism under nitrogen atmosphere of PTFE by TGFTIR analysis. Influence of the additive nature. Thermochim. Acta 526, 205e212. Osikoya, A.O., Wankasi, D., Vala, R.M.K., Afolabi, A.S., Dikio, E.D., 2014. Synthesis, characterization and adsrotpion studies of fluorine-doped carbon nanotubes. Dig. J. Nanomater. Biostructures 9, 1187e1197. Pan, B., Lin, D., Mashayekhi, H., Xing, B., 2008. Adsorption and hysteresis of bisphenol A and 17a-ethinyl estradiol on carbon nanomaterials. Environ. Sci. Technol. 42, 5480e5485. Pan, B., Xing, B., 2008. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42, 9005e9013. Patino, Y., Diaz, E., Ordonez, S., Gallegos-Suarez, E., Guerrero-Ruiz, A., RodriguezRamos, I., 2015. Adsorption of emerging pollutants on functionalized multiwall carbon nanotubes. Chemosphere 136, 174e180. Peng, H., Li, H., Wang, C., Zhang, D., Pan, B., Xing, B., 2014. Sorption and solubility of ofloxacin and norfloxacin in water-methanol cosolvent. Chemosphere 103, 322e328. Pignatello, J.J., 1998. Soil organic matter as a nanoporous sorbent of organic pollutants. Adv. Colloid Interface Sci. 76, 445e467. Rae, P.J., Dattelbaum, D.M., 2004. The properties of poly (tetrafluoroethylene) (PTFE) in compression. Polymer 45, 7615e7625. Ren, X., Chen, C., Nagatsu, M., Wang, X., 2011. Carbon nanotubes as adsorbents in environmental pollution management: a review. Chem. Eng. J. 170, 395e410. Wang, C., Li, H., Liao, S., Zhang, D., Wu, M., Pan, B., Xing, B., 2014. Sorption affinities of sulfamethoxazole and carbamazepine to two sorbents under co-sorption systems. Environ. Pollut. 194, 203e209. Wang, F., Yao, J., Sun, K., Xing, B., 2010. Adsorption of dialkyl phthalate esters on carbon nanotubes. Environ. Sci. Technol. 44, 6985e6991. Wu, W., Chen, W., Lin, D., Yang, K., 2012. Influence of surface oxidation of multiwalled carbon nanotubes on the adsorption affinity and capacity of polar and nonpolar organic compounds in aqueous phase. Environ. Sci. Technol. 46, 5446e5454. Yang, K., Xing, B., 2010. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: polanyi theory and its application. Chem. Rev. 110, 5989e6008. Yu, J.-G., Zhao, X.-H., Yang, H., Chen, X.-H., Yang, Q., Yu, L.-Y., Jiang, J.-H., Chen, X.-Q., 2014. Aqueous adsorption and removal of organic contaminants by carbon nanotubes. Sci. Total Environ. 482, 241e251. Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2009. The Impacts of aggregation and surface chemistry of carbon nanotubes on the adsorption of synthetic organic compounds. Environ. Sci. Technol. 43, 5719e5725. Zhang, S., Shao, T., Kose, H.S., Karanfil, T., 2010a. Adsorption of aromatic compounds by carbonaceous adsorbents: a comparative study on granular activated carbon, activated carbon fiber, and carbon nanotubes. Environ. Sci. Technol. 44, 6377e6383. Zhang, W., Dubois, M., Guerin, K., Bonnet, P., Kharbache, H., Masin, F., Kharitonov, A.P., Hamwi, A., 2010b. Effect of curvature on C-F bonding in fluorinated carbons: from fullerene and derivatives to graphite. Phys. Chem. Chem. Phys. 12, 1388e1398. Zhang, W., Spinelle, L., Dubois, M., Guerin, K., Kharbache, H., Masin, F., Kharitonov, A.P., Hamwi, A., Brunet, J., Varenne, C., Pauly, A., Thomas, P., Himmel, D., Mansot, J.L., 2010c. New synthesis methods for fluorinated carbon nanofibres and applications. J. Fluor. Chem. 131, 676e683.