Effect of oxidation treatment of multi-walled carbon nanotubes on the adsorption of pentachlorophenol from aqueous solution: Kinetics study

2.1 Oxidation and characterization of MWCNTs

MWCNTs were purchased from Sun Nanotech (China) and were used as received. The MWCNTs (2.0 g) were added to 200 mL solutions of 8 M HNO₃, 18% H₂O₂ and 1 M KMnO₄ (acidified), separately. The HNO₃ suspension was refluxed at 140 °C for 4 h. The H₂O₂ and KMnO₄ suspensions were heated at 80 °C for 4 h. MWCNTs oxidized with KMnO₄ were further treated with H₂SO₄ in order to remove MnO₂ from the product. The three suspensions were then washed with distilled water until the wash water was pH neutral and a centrifuge (3500 rpm for 30 min) was used for solid-liquid separation. The solid products were dried in air at 50 °C. Infrared spectra measurements of the oxidized MWCNTs were performed on an ABB Bomem MB Series FTIR spectrometer using KBr pellets. The specific surface area of the pristine and oxidized MWCNTs were measured by nitrogen adsorption at 77 K using the single point BET method with a Micromeretics model 2200 surface area analyzer. All the chemicals used were purchased from Sigma–Aldrich.

2.2 Analytical procedure

For the determination of PCP concentration in aqueous using gas chromatograph with an electron capture detector (GC–ECD) an acetylation step is required. For acetylation of PCP, 3.0 mL of 5% K₂CO₃ solution was added to the aqueous solution of PCP and stirred for 1 min, followed by the addition of 2.0 mL of acetic anhydride and stirred vigorously for 2 min to ensure complete acetylation of the PCP. The acetylated PCP was extracted into 4.0 mL n-hexane with 2 min of shaking. The acetylated PCP was analyzed using a Varian 3600 GC-ECD and a 30 m × 0.25 μm id × 0.25 μm film thickness DB-5 silica capillary column (Varian). The GC conditions were as follows: helium as the carrier gas at a flow rate of 1 mL/min; nitrogen as the make-up gas at a flow rate of 27 mL/min; injector temperature, 300 °C; column initial temperature, 70 °C; and final temperature, 280 °C; with heating rate, 15 °C/min; detector temperature, 300 °C. The sample size injected was 1 μL.

2.3 Solid phase extraction (SPE) of PCP

Stock solutions were prepared by dissolving pentachlorophenol (Aldrich 99.0 + %) in deionized water followed by serial dilutions. Three milliliters inert glass tubes (Supelco) were used as the packing cartridges for SPE and 5.0 mg of the MWCNTs were packed into the cartridge. Teflon upper and lower frits (20 μm pore size) at each end of the cartridge were used to hold the MWCNTs packing in place. The PCP solution (1.0 mg/L) was passed through the cartridge at a flow rate of 2 mL/min using suction. The aqueous solution was then analyzed for the remaining PCP.

2.4 Kinetics study

The kinetics study was done using 14 mL glass vials with Teflon-lined caps. In each vial, 10 mL of the 1.0 mg/L PCP solution was added to 5.0 mg of MWCNTs. The vials were mounted on a shaker and shaken at 23 ± 2 °C for different intervals of time. Each suspension was centrifuged at 3000 rpm and the supernatant solution was analyzed for PCP.

3.1 Infrared spectroscopy

The IR measurements were performed in order to confirm the formation of acidic functional groups upon oxidation and the spectra are presented in Fig. 1. The figure shows that the oxidized MWCNTs spectra have absorption peaks at 1730 cm⁻¹ corresponding to the stretching vibration of C⚌O from the carboxylic acid groups (–COOH) (Sun et al., 2003; Chen et al., 1998, 2001; Shaffer et al., 1998; Hamon et al., 2001; Koshio et al., 2001). This absorption peak was very clear and strong for the MWCNTs oxidized with HNO₃, whereas it was much weaker with the MWCNTs oxidized with H₂O₂ and KMnO₄. A broad peak at approximately 3500 cm⁻¹ was much stronger in the H₂O₂ and HNO₃-treated MWCNTs. This peak is characteristic of an O–H stretch. It is due to either moisture, alcohol or phenol, OH⁻ or carboxylic groups in these samples. The aromatic C⚌C stretch is observed at 1645 cm⁻¹ and at 1450 cm⁻¹ for all of the spectra with comparable intensities. The peak at 1450 cm⁻¹ is more intense for the MWCNTs oxidized with HNO₃ and H₂O₂ than in the pristine and MWCNTs oxidized with KMnO₄. The weak, broad peaks at 1130 cm⁻¹ are found in all the oxidized MWCNTs which could be due to a C–O stretch from either a phenol or lactone. These results show clearly that the MWCNTs can be oxidized and that carboxylic, phenolic and lactone groups are likely obtained.

Close

Figure 1

IR spectra of MWCNTs before and after oxidation.

Export to PPT

3.2 Adsorption of PCP on pristine and oxidized MWCNTs

The adsorption of PCP to pristine and oxidized MWCNTs was studied after packing the CNTs in a solid-phase extraction (SPE) cartridge and the results are shown in Fig. 2. It is clear from the figure that oxidation of the MWCNTs decreased the adsorption of PCP in general. The amount of PCP adsorbed was found to be 91.9% for the pristine MWCNTs, and 76.4%, 77.8%, and 43.4% for the MWCNTs oxidized with HNO₃, H₂O₂, and KMnO₄, respectively. It was expected that PCP, with its phenolic group would strongly interact with the oxygen-containing acidic functional groups; carboxylic, phenolic and lactonic, that were created upon oxidation of the MWCNTs, and consequently the adsorption of PCP would be increased. However, the adsorption of PCP on the oxidized MWCNTs actually decreased compared with the pristine MWCNTs. This may be explained as follows. Generally, in the case of adsorption from the liquid phase to solid phase, the adsorption capacity of any carbonueous adsorbent for aromatic compounds, such as PCP, depends on the following factors (Haghseresh et al., 2002):

• –

  the physical nature of the adsorbent, such as pore structure, and the presence of surface functional groups,

• –

  the nature of the adsorbate, such as its pK_a, functional groups present, polarity, molecular weight, and size,

• –

  the conditions of the solution, such as pH, ionic strength as well as the adsorbate concentration.

Close

Figure 2

Adsorption of PCP by pristine MWCNTs and different oxidized MWCNTs. The error bars represent three standard deviations, 5 mg of the adsorbent, 2 mL/min flow rate, PCP concentration is 20 ng/mL.

Export to PPT

From the physical nature of the adsorbent point of view, one of the important features for most of carbon adsorbent is their aromaticity due to the presence of the delocalized π-electrons. Pristine MWCNTs, mostly, have a uniform surface with many delocalized π-electrons, which increases their adsorption capability in comparison with other carbonueous materials. It was found that the adsorptive properties of carbonueous material are determined mostly by its surface chemical composition. The presence of oxygen and hydrogen within the surface groups crucially affected the adsorptive properties of the adsorbent. Functional groups and delocalized electrons of the MWCNTs structure determine the chemical character of the MWCNTs surface (León and Radovic, 1994). It was reported that the presence of the carboxyl and hydroxyl groups inhibited the adsorption of phenolic compounds. This effect can be explained by PCP adsorption that is governed by “π–π dispersion interaction” between the MWCNTs and the aromatic ring of the PCP. Oxygen-containing functional groups can localize the π-electrons, and consequently, remove them from the π-electron system of the MWCNTs (Couglin and Ezra, 1968), lowering the dispersive forces with PCP π-electron. Another explanation is the adsorption of water molecules to the oxygen-containing function groups on the surface of the oxidized MWCNTs by the means of hydrogen bonding (Dąbrowski et al., 2005) which is the “solvent effect”.

It can be concluded here that the oxidation of MWCNTs and the introduction of oxygen-containing functional groups inhibit the adsorption of PCP due to the localization of the π-electron of the MWCNTs as well as the solvent effect.

3.3 Kinetics study of the adsorption of PCP on pristine and oxidized MWCNTs

The kinetics of PCP adsorption in aqueous solution on pristine or oxidized MWCNTs were studied and the results are presented in Fig. 3. Within the first 5 min 88% of the PCP was adsorbed by the pristine MWCNTs whereas 88.9%, 87.1%, and 75.0% by the oxidized MWCNTs; H₂O₂, HNO₃, and KMnO₄, respectively. The adsorption capacity reached 1.76 mg/g for the pristine MWCNTs and 1.78, 1.74, and 1.50 mg/g for the MWCNTs oxidized with H₂O₂, HNO₃, and KMnO₄, respectively. The adsorption of PCP by pristine MWCNTs and the oxidized MWCNTs (HNO₃, H₂O₂) achieved equilibrium gradually at about 60 min and the final adsorption capacity reached 1.92, 1.88, and 1.89 mg/g, respectively. For the other oxidized MWCNTs it took 120 min to reach the equilibrium concentration of 1.69 mg/g and for the oxidized MWCNTs KMnO₄. Comparing the kinetics results of the %PCP adsorbed by the oxidized MWCNTs with the SPE application (Fig. 2) showed that the oxidized MWCNTs adsorbed less if they are packed inside the SPE cartridge. The explanation for that may be attributed to the fact that the diffusion of the PCP molecules during the kinetics experiment was fast due to the shaking of the solution during the experiment. This shaking enhanced the transfer of the PCP from the liquid phase to the solid phase as well as exposed most of the MWCNTs to the PCP molecules.

Close

Figure 3

The kinetic model for the adsorption of PCP to pristine and oxidized MWCNTs and the comparison between the experimental data and the predicted data (one component, two component).

Export to PPT

3.4 Kinetic modelling

The transport of the PCP from the aqueous solution to the surface of the adsorbent may occur in several steps (Srivastava et al., 2006). The overall adsorption process may be controlled by one or more steps, e.g. film or external diffusion, pores diffusion, surface diffusion and adsorption on the pore surface, or a combination of more than one step.

In order to distinguish between the different adsorption steps, the following kinetic model was adopted.

The adsorption of PCP molecules from aqueous solution to an active site S, on the adsorbent (MWCNTs or AC) can be represented as: $${\text{Analyte}}_{(\text{aq}.)}+{\text{Absorbent}}_{(\text{aq}.)}\underset{\text{desorption}}{\overset{\text{adsorption}}{\rightleftarrows }}\text{Analyte-}{\text{adsorbent complex}}_{(\text{aq}.)}$$ or

The rate of formation of the PCP − S is determined by the following equation:

In the case where [S] ⋙ [PCP] and before significant adsorption has occurred, the adsorption is considered a pseudo-first-order, and the rate of adsorption can be simplified as:

By assuming that the adsorption of PCP from the aqueous phase to the solid phase occurs in different steps; film or external diffusion, pores diffusion, surface diffusion and adsorption on the pore surface, and if PCP molecules undergo a first-order or pseudo-first-order adsorption to the solid adsorbent, the sum of the concentrations of all PCP adsorbed at the solid adsorbent at time t, $C_{\text{PCP}-\text{S}}(t)$ increases exponentially as:

The experimental data were analyzed for discrete values of the adsorption rate constants by nonlinear regression. One of the important conditions for the application of the kinetic model is the speed convergence of the regressions, and there should be a reasonable fit (as measured by correlation coefficient, R²) with the kinetic data. The kinetic adsorption curves of PCP may suggest that adsorption onto the surface occurs quickly, followed by a slower diffusion step. The short time needed to reach equilibrium also suggests that both the pristine and oxidized MWCNTs, excluding the tubes oxidized with KMnO₄, are superior adsorbents to remove PCP from aqueous solution. The experimental data were analyzed for discrete values of the adsorption rate constants by a nonlinear regression. Fig. 3 shows the relation between the % of PCP adsorbed from the aqueous solution to the adsorbent and time for pristine and oxidized MWCNTs. These data were fitted using the one component model and the values for kinetic parameters of the PCP adsorption are presented in Table 1. From this table, it is clear that applying the one component kinetic model for the adsorption of PCP to MWCNTs converges well and the value of the R² was acceptable (more than 0.95), whereas in the case of the CNTs oxidized with KMnO₄, the R² was not acceptable with a value of 0.75. This may be due to the fact that KMnO₄ was a harsh oxidizing agent, which destroyed the crystalline structure of the MWCNTs and led to the formation of amorphous carbon. Comparing the one component predicted data (dotted line) with the experimental data of the PCP adsorption to the pristine and the oxidized tubes shows significant differences. Again, the fitted data did not match the experimental data at all. This may imply that the adsorption of PCP mainly in more than one step. The two component kinetic model was used to fit the experimental data and the results are presented in Table 1 and Fig. 3. It was obvious from the graph and the table that the two component kinetic model fits much better than the one component kinetic model as the predicted data fitted well with the experimental data with very good correlation coefficients (R² = 0.998 and 0.999). In general, the %PCP adsorbed during the first step (%C₁) was significantly higher than the second step ((%C₂) in the case of pristine MWCNTs; 89.2%, and the oxidized MWCNTs; 90.0% and 88.7%, for the MWCNTs oxidized with H₂O₂ and HNO₃, respectively. The differences between the first rate constant for the pristine and the oxidized MWCNTs were not significant. The first component may be attributed to the fast diffusion of the PCP from the aqueous phase to the surface of the MWCNTs. The fact that the difference between the pristine and the oxidized MWCNTs for the first component was not significant may indicate that the oxidation of the MWCNTs with H₂O₂ and HNO₃ was successful. This oxidation increases the hydrophilic character of the MWCNTs through the introduction of carboxylic and phenolic groups on the surface of the tubes without destroying the tube like what happened in the case of KMnO₄. Analysis of the second component shows that nearly 10% of the PCP was adsorbed on the pristine and the oxidized MWCNTs. This step may be attributed to intra-particle diffusion in aggregates of MWCNTs. The probability of intra-particle diffusion was explored by using the intra-particle diffusion model (Weber and Morris, 1963).

Close

Figure 4

Weber–Morris intra-particle diffusion plots for the adsorption of PCP to pristine and oxidized MWCNTs.

Export to PPT