Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

25 downloads 0 Views 1MB Size Report
ABSTRACT: Carbon nanotubes (CNTs), a new crystalline form in the carbon family ... Many methods, including chemical precipitation, ion exchange, membrane ...
475

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes Yan-Hui Li1,*, Zhaokun Luan2, Xu Xiao1, Xiangwen Zhou1, Cailu Xu1, Dehai Wu1 and Bingqing Wei3 (1) Department of Mechanical Engineering, Tsinghua University, Beijing 100084, P.R. China. (2) State Key Laboratory of Environmental Aquatic Chemistry, Research Centre for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, P.R. China. (3) Department of Materials Science & Engineering, Rensselaer Polytechnic Institute, Troy, NY 12180, USA. (Received 14 January 2003, revised form accepted 9 April 2003)

ABSTRACT: Carbon nanotubes (CNTs), a new crystalline form in the carbon family, have been shown to be an effective adsorbent for Cu2+ ion removal from aqueous solution. After oxidation with nitric acid, the adsorption capacity of the oxidized CNTs towards Cu2+ ions at a pH of 5.2 reached 27.6 mg/g, compared to a value of only 14.4 mg/g for the as-grown CNTs under the same circumstances. Increasing the pH to 5.4 led to 95% removal of Cu2+ ions with the oxidized CNTs, whereas with the as-grown CNTs it was necessary to increase the pH to 8.6 to achieve the same extent of removal. The Cu2+ ion adsorption capacity increased with increasing CNT dosage for the different initial Cu2+ ion concentrations. The kinetic curve for adsorption of the Cu2+ ions suggested that not only the outer surfaces but also inner cavities and interlayers in the structures of the CNTs were responsible for the removal of the ion from aqueous solutions.

INTRODUCTION Copper is an essential element for human health since small concentrations play an important role in enzyme-mediated systems. However, long-term consumption of water containing high levels of copper can cause many diseases such as stomach ailments, intestinal distress, liver and kidney damage, and anaemia. For this reason, excessive concentrations of Cu2+ ions in drinking water should be reduced to a permitted level of less than 1 mg/l to make the water fit for drinking purposes (Panday et al. 1986). Copper pollution can arise from copper mining and smelting, brass manufacture, petroleum refineries, electroplating and the excessive use of Cu-based agrochemicals (Doula et al. 2000). Many methods, including chemical precipitation, ion exchange, membrane filtration and adsorption, are currently used to remove Cu2+ ions from aqueous solutions. Chemical precipitation is the most widely used process allowing the transformation of Cu2+ ions into sulphides or the insoluble hydroxide at alkaline pH values (Ajmal et al. 2001). However, where alkaline conditions are employed, disposal of the precipitate and difficulty in its treatment for trace elements limit the application of this method to some degree. Adsorption is another promising and widely used method for removing copper pollution. Researchers have devoted much effort to exploring for new adsorbents with high adsorption *Author to whom all correspondence should be addressed. E-mail: [email protected].

476

Y.-H. Li et al./Adsorption Science & Technology Vol. 21 No. 5 2003

capacities and efficiencies. Crystalline minerals (Panday et al. 1986; Doula et al. 2000; Ajmal et al. 2001; Scheinost et al. 2001), fly ash (Lin and Chang 2001), biomaterials (An et al. 2001; Williams et al. 1998; Reddad et al. 2002), silica gel (Park et al. 1995), sawdust (Yu et al. 2001), coconut husk (Low et al. 1995), wheat straw cell wall residue (Merdy et al. 2002), activated carbon (Hasar and Cuci 2000; Kim et al. 2001) and activated carbon fibres (Lee et al. 2002) have all been studied as adsorbents and shown to exhibit different capabilities towards Cu2+ ion removal. Of these, activated carbon exhibited a good performance towards Cu2+ ion adsorption due to its high surface area and the ready modification of its surface by oxidation. Carbon nanotubes (CNTs), a new kind of crystalline material in the carbon family, have been proposed as promising candidates for novel electronic devices and composites due to their unique electrical behaviour (Ebbesen et al. 1996) and exceptional mechanical properties (Treacy et al. 1996). Their hollow and layered structure, large surface area and high chemical stability suggest that CNTs could have considerable potential applications in adsorption. Long and Yang (2001) found that CNTs were more efficient adsorbents than activated carbon in dioxin adsorption, which they attributed to the stronger interactions between dioxins and CNTs. We have shown that the fluoride adsorption capacity of CNTs supported on amorphous alumina is ca. 13.5-times higher than that of AC-300 carbon and four-times greater than that of g-Al2O3 (Li et al. 2001). In addition, CNTs oxidized by nitric acid show exceptional adsorption capabilities and high adsorption efficiencies towards lead removal from water (Li et al. 2002). Our previous reports suggest that CNTs might be a promising candidate as a material for adsorbing both anions and cations from wastewater. In the present work, we have compared the variation in specific surface area, size distribution, functional groups and zeta potential of as-grown and oxidized CNTs, and studied their adsorption characteristics towards the removal of Cu2+ ions from aqueous solution when parameters such as the Cu2+ ion concentration, pH, CNT dosage and agitation time have been varied. EXPERIMENTAL Materials The CNTs used were prepared by chemical vapour deposition (CVD) methods in which methane carried in a hydrogen flow was pyrolyzed at 680°C in the presence of Ni nanoparticles as a catalyst. The transmission electron microscopy image depicted in Figure 1(a) shows the development of CNTs with an average diameter of 30 nm and length ranging from hundreds of nanometers to micrometers. The as-grown CNTs were dispersed in hydrofluoric acid for 24 h to remove the diatomite used as the catalyst support and subsequently placed in concentrated nitric acid and oxidized at 140°C for 1.5 h to remove most of the catalyst particles. Methods Characterization of CNTs The specific surface areas of the as-grown and the oxidized CNTs were measured via nitrogen adsorption/desorption experiments conducted at -196°C using the BET method. The size distributions of the as-grown and oxidized CNTs were analyzed by laser light scattering using a Malvern Mastersizer 2000 that operated on the Mie diffraction theory principle. Thus, suspensions of both kinds of CNTs were prepared by dispersing 0.2 g CNTs into 100 ml deionized water and

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

477

Figure 1. Transmission electron micrographs of carbon nanotubes: (a) the as-grown CNTs with arrow a showing the covered tips; (b) the oxidized CNTs with arrow b showing the opened tips.

the suspensions sonicated for 4 h to ensure a sufficient dispersion of the CNTs in water. They were then introduced into the measurement cell with stirring until an opacity of between 10% and 20% was attained. The relative volumes of the particles and their size distribution could then be estimated from the measurements. Functional groups attached to the CNTs were determined qualitatively via FT-IR spectroscopy or quantitatively using Boehm’s titration method (Shaffer et al. 1998). The latter method involved dispersing 0.2 g respectively of the as-grown and oxidized CNTs in 50 ml deionized water. The suspensions were mixed with 10 ml of 0.1 M base solutions of sodium hydroxide, sodium hydrogen carbonate and sodium carbonate and stirred in a sealed vessel for 48 h. The suspensions were then filtrated and 20 ml of the filtrates added to 15 ml of 0.1 M HCl to neutralize any unreacted bases. The resulting solutions were back-titrated with 0.1 M NaOH using an Automatic Potential Titrator 716 (Metrohm Co., Switzerland). The amounts of the different types of oxygencontaining groups were obtained by computing the volumes of the bases consumed.

478

Y.-H. Li et al./Adsorption Science & Technology Vol. 21 No. 5 2003

The zeta potentials of the as-grown and oxidized CNTs were measured using a Malvern Zetasizer 2000 instrument. Thus, the pH value of the CNT solution was adjusted from 2.0 to 12.0 by the addition of 0.1 M HCl or NaOH solutions to the glass beaker at 25°C. By measuring the zeta potential as a function of pH, the acidity or basicity of the CNT surface and the isoelectric point (IEP) could be determined. Adsorption studies A stock solution containing 1000 mg/l Cu2+ ions was prepared by dissolving copper chloride in deionized water. The solution was further diluted to the required concentration before use. All experiments were carried out at 25°C. Batch adsorption isotherms were measured by adding 0.05 g of the as-grown and oxidized CNTs to 100 ml of solutions with an initial concentration of Cu2+ ions within the range 6–30 mg/l. The initial pH value of the solution was 5.2. After the suspensions had been shaken for 2 h, they were filtered through 0.45 mm membrane filters and the filtrates immediately measured via an atomic absorption spectrometer. The amount of Cu2+ ions adsorbed onto the CNTs was determined from the difference between the initial Cu2+ ion concentration and that attained under equilibrium conditions. To determine the effect of pH on the removal of Cu2+ ions from aqueous solution, 50 mg of the as-grown and oxidized CNTs were dispersed in 100 ml of solutions containing 20 mg/l of Cu2+ ions. The initial pH values of the solutions were adjusted from 2.2 to 12 using 0.5 M HCl and 0.5 M NaOH as appropriate. The effect of the CNT dosage on Cu2+ ion adsorption was also studied. Kinetic studies were conducted by adding 1 g of the oxidized CNTs to 2000 ml of solutions containing Cu2+ ion concentrations of 10, 20 and 30 mg/l without any pH adjustment (pH = 4.2). Samples were then removed from these solutions at predetermined time intervals, separated and analyzed. RESULTS AND DISCUSSION Characterization of as-grown and oxidized CNTs The specific surface area and pore-size distribution can greatly influence the adsorption capacity of carbon materials. The specific surface area of CNTs was found to increase from 122 m2/g to 154 m2/g after oxidation with nitric acid. This small increase suggested a partial opening of the nanotubes. Figure 1 clearly shows that the nanotube tips were partially removed under these conditions. The pore-size distributions of the as-grown and oxidized CNTs are depicted in Figure 2. The major peaks for the as-grown CNTs (curve a) were mainly distributed at 3.0–4.2 nm and 34.2–62.6 nm, the former peaks corresponding to the inner diameter of the as-grown CNTs while the latter peaks may be due to the porosity of the diatomite used as a catalyst support. The peaks for the oxidized CNTs (curve b) occurred at ca. 3.0–4.5 nm and exhibited a greater intensity than those of the as-grown CNTs due to partial opening of the tips of the nanotubes (Hou et al. 2002) and the removal of amorphous carbon and catalyst support. Particle-size distribution is another factor which influences the suspension capabilities and adsorption properties of adsorbents. Figure 3 shows the comparable results obtained for the particle-size distributions of the as-grown and the oxidized CNTs. Curves a and c correspond

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

479

Figure 2. Pore-size distribution of (a) the as-grown and (b) oxidized CNTs.

to the frequency curve, which is useful for displaying the distinct sizes of objects. Curves b and d show the same results displayed as the percentage of the sample below a certain particle size. It will be seen that two peaks occur at 15 mm and 310 mm for the as-grown CNTs (curve c), while the peaks move to 0.12 mm and 8 mm, respectively, for the oxidized CNTs (curve a). Curve b indicates that 100% of the oxidized CNTs were less than 22 mm in length, while ca. 40% of the as-grown CNTs exceeded this size (curve d). These results suggest that oxidation led to a decrease in the sizes of the CNTs, a phenomenon that can be explained by the nature of CNTs. Thus, CNTs prepared by the CVD method possess many defects such as pentagons and heptagons that lead to

Figure 3. Particle-size distribution of the as-grown and oxidized CNTs. Curves a and c represent the frequency curve for the oxidized and as-grown CNTs; curves b and d show the percentage of the sample below a certain particle size for the oxidized and as-grown CNTs.

480

Y.-H. Li et al./Adsorption Science & Technology Vol. 21 No. 5 2003

Figure 4. FT-IR spectra of (a) the CNTs obtained after oxidation with acids and (b) the as-grown CNTs.

bending and deformation of the CNTs. Oxidation causes rupture of the CNT structure at the defects leading to smaller pieces and hence a decrease in CNT sizes. The cation- and anion-exchange capacities of carbon materials depend mainly on the presence of surface functional groups or complexes (Strelko et al. 2002). Oxidation is one effective method for introducing oxygen-containing functional groups such as hydroxy (–OH), carboxy (–COOH) and carbonyl (>C=O) onto the surfaces of CNTs (Jia et al. 1999). The presence or absence of such groups may be established by IR spectroscopy. Figure 4 depicts the FT-IR spectra for the oxidized and as-grown CNTs in which the peaks at ca. 1550 cm-1 and 1200 cm-1 may be assigned to the carbon skeleton while the peaks at 1750 cm-1 and 3500 cm-1 are associated with >C=O and –OH functional groups, respectively. The signal at ca. 1400 cm-1 may be attributed to O–H bending deformation in carboxylic acid and phenolic groups. As shown in the figure, there were obviously more functional groups connected to the carbon atoms on the oxidized CNTs than on the as-grown CNTs. Further information in this regard was provided by the results of the Boehm titration that allowed a quantitative evaluation of the amounts of acidic groups present. In these titrations, it is assumed that sodium hydrogen carbonate only neutralizes carboxy groups, sodium carbonate neutralizes carboxy groups and lactones, while sodium hydroxide can neutralize carboxy groups, lactones and phenols. Since the as-grown CNTs contained only trace amounts of functional groups, these were hardly capable of detection by the titration method. However, the oxidized CNTs exhibited amounts of carboxy groups, lactones and phenols equal to 0.76, 1.83 and 1.45 mmol/g, respectively. The results of zeta potential measurements (Figure 5) indicated that the surfaces of the as-grown CNTs exhibited an acidic character with isoelectric point (IEP) values of ca. 5.0. This may be attributed to the functional groups introduced via oxidation in air when the as-grown CNTs were removed from the CVD chamber at ca. 380°C. The IEP for the oxidized CNTs, which could not be determined within the experimental range employed, was shifted to a lower pH value due to the greater surface functionality ascribed to dissociatable acidic groups introduced by nitric acid oxidation.

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

481

Figure 5. Plots of zeta potential versus pH for the as-grown and oxidised CNTs.

Influence of Cu2+ ion concentration on adsorption The dependence of adsorption on the initial Cu2+ ion concentration was studied to compare the ability of the as-grown and oxidized CNTs for removing such ions from aqueous solution (Figure 6). It should be noted that the as-grown CNTs demonstrated a lower percentage removal of Cu2+ ions relative to the behaviour of the oxidized material. Although Cu2+ ion removal by both the oxidized and as-grown CNTs decreased as the initial Cu2+ ion concentration increased, the

Figure 6. Effect of initial concentration of Cu2+ ions on their adsorption by the as-grown and oxidized CNTs. Experimental conditions: pH = 5.2; CNT concentration = 0.05 g/I; agitation time = 2 h.

482

Y.-H. Li et al./Adsorption Science & Technology Vol. 21 No. 5 2003

amounts of ion adsorbed per unit amount of adsorbent increased under the same circumstances to attain values of 27.6 mg/g and 14.4 mg/g, respectively, for the oxidized and as-grown material at an initial Cu2+ ion concentration of 30 mg/l. Adsorption isotherms The adsorption isotherm data for the removal of Cu2+ ions by the as-grown and oxidized CNTs were fitted to the Langmuir and Freundlich models whose linearized equations may be expressed by equations (1) and (2), respectively: 1 q

=

1 qm

ln q

=

ln K F

1

+

(1)

bq m C e +

1 ln C e n

(2)

where Ce is the equilibrium Cu2+ ion concentration (mg/l), q is the amount adsorbed (mg/g), qm is the amount adsorbed corresponding to the formation of a monolayer on the adsorbent, b is the Langmuir constant, and K F and n are the Freundlich constants. The correlation coefficients R2 (Table 1) for the Freundlich model were 0.93 for the as-grown CNTs and 0.89 for the oxidized CNTs. The corresponding correlation constants for the Langmuir equation were 0.97 for the as-grown CNTs and 0.98 for the oxidized CNTs. These values demonstrate that the equilibrium adsorption data for different initial Cu2+ ion concentrations fitted the Langmuir model best. The amounts of Cu2+ ions adsorbed per unit amount of adsorbent, qm, computed by the linearized Langmuir equation for the as-grown and oxidized CNTs were 16.55 and 28.49 mg/g, respectively, which were very similar to the values deduced from the experimental data depicted in Figure 6. These results also support the view that the adsorption capacity of the CNTs was improved by oxidation. The effect of pH on Cu2+ ion adsorption Figure 7 depicts the dependence of Cu2+ ion removal by the as-grown and oxidized CNTs on the pH value of the aqueous solution. Over the pH region < 4.8 (the IEP of the as-grown CNTs),

TABLE 1. Freundlich and Langmuir Isotherm Constants Freundlich isotherm

As-grown CNT Oxidized CNT

Langmuir isotherm

R2

KF

n

R2

qm

b

0.93 0.89

1.67 0.84

0.59 1.23

0.97 0.98

16.55 28.49

0.11 0.26

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

483

Figure 7. Effect of pH value on Cu2+ ion adsorption onto CNTs. Experimental conditions: CNT concentration = 0.05 g/I.

almost no Cu2+ ions were adsorbed on the as-grown CNTs due to electrostatic repsulsion between the surface of the CNTs and the Cu2+ ions. As the pH value of the Cu2+ ion solution increased, the removal efficiency increased gradually and attained a value of 30% at pH 6.0. This suggests that the Cu2+ ion adsorption capability of the as-grown CNTs was weak. A sharp increase in the removal efficiency occurred at pH > 6.0, to attain a value of 95% removal at pH 8.6. The obviously increased removal efficiency towards Cu2+ ions demonstrated by the as-grown CNTs was mainly due to precipitation, with a precipitate of copper hydroxide being produced within the higher pH region. With the oxidized CNTs, the removal efficiency increased rapidly above a pH value of 2.0 and attained 95% at pH 5.4. This was due to the negative surface of the oxidized CNTs. A comparison of the removal efficiencies exhibited by the oxidized and as-grown CNTs showed that the oxidized CNTs had a higher Cu2+ ion adsorption capacity than the asgrown material which may be due to the functional groups introduced by nitric acid oxidation. Effect of CNT dosage on Cu2+ ion adsorption In order to study the effect of CNT dosage on the adsorption of Cu2+ ions, different masses of oxidized CNTs were weighed and added to 100 ml of an aqueous solution containing a Cu2+ ion concentration within the range 10–40 mg/l. The initial pH of the solution thus obtained was adjusted to 5.2 in all cases. Figure 8 shows that, at the same CNT dosage, the percentage removal of Cu2+ ions increased with decreasing Cu2+ ion concentration from 40 mg/l to 10 mg/l. The figure also shows that the Cu2+ ion adsorption capacity increased with increasing CNT dosage for different initial Cu2+ ion concentrations. This may be due to the greater amount of adsorbent surface and pore volume available at higher CNT doses providing more functional groups and active adsorption sites that result in a higher Cu2+ ion adsorption capacity.

484

Y.-H. Li et al./Adsorption Science & Technology Vol. 21 No. 5 2003

Figure 8. Effect of CNT dosage on the removal of Cu concentrations. Experimental conditions: pH = 5.2.

2+

ions from aqueous solution under different initial Cu2+ ion

Kinetics of adsorption Figure 9 demonstrates that an increase in Cu2+ ion concentration over the range 10–30 mg/l led to an increase in the amount of Cu2+ ions adsorbed on the oxidized CNTs from 14.4 mg/g to 20.4 mg/g with a corresponding decrease in the percentage removal from 74.5% to 36.3%. The adsorption rate was rapid over the first 15 min to gradually attain equilibrium at ca. 70 min. The initial steep adsorption curve suggests that the adsorption occurred rapidly on the surface of

Figure 9. Effect of time and initial concentration of Cu2+ ions on the adsorption of the latter. Experimental conditions: pH = 4.2; CNT concentration = 0.05 g/I.

Removal of Cu2+ Ions from Aqueous Solutions by Carbon Nanotubes

485

the oxidized CNTs. The adsorption rate became slow subsequently due to the longer diffusion range of the Cu2+ ions (ionic radius = 0.072 nm) into the inner cavities (diameter = ca. 3.6 nm) and interlayers (spacing = 0.344 nm) of the oxidized CNTs. Such slow diffusion would lead to a smaller increase in the adsorption curve at later stages. CONCLUSIONS It may be concluded that oxidation treatment with nitric acid led to a considerable increase in both the specific surface area and number of functional groups on the surfaces of CNTs. At the same time, their particle sizes and isoelectric points diminished. The experimental results indicated that oxidation was very significant in improving the Cu2+ ion adsorption capabilities of the CNTs, due to changes in the physical and chemical properties of the CNTs induced by oxidation. The adsorption capacity of oxidized CNTs was 27.6 mg/g, but only 14.4 mg/g for as-grown CNTs at a Cu2+ ion concentration of 20 mg/l and a pH value of 5.2. Setting aside the current costeffectiveness of using CNTs as adsorbents, such CNTs may be effective and potential candidates for the replacement of activated carbon as an adsorbent in water treatment. REFERENCES Ajmal, M., Rao, R.A.K., Ahmad, R., Ahmad, J. and Rao, L.A.K. (2001) J. Hazard. Mater. B 87, 127. An, H.K., Park, B.Y. and Kim, D.S. (2001) Water Res. 35, 3551. Doula, M., loannou, A. and Dimirkou, A. (2000) Adsorption 6, 325. Ebbesen, T.W., Lezee, H.J., Hiura, H., Bennett, J.W., Ghaemi, H.F. and Thio, T. (1996) Nature (London) 382, 54. Hasar, H. and Cuci, Y. (2000) Environ. Technol. 21, 1337. Hou, P.X., Bai, S., Yang, Q.H., Liu, C. and Cheng, H.M. (2002) Carbon 40, 81. Jia, Z., Wang, Z., Liang, J., Wei, B. and Wu, D. (1999) Carbon 37, 903. Kim, J.W., Sohn, M.H., Kim, D.S., Sohn, S.M. and Kwon, Y.S. (2001) J. Hazard. Mater. B 85, 301. Lee, S.M., Ryu, S.K., Jung, C.H., Won, H.J. and Oh, W.Z. (2002) Carbon 40, 329. Li, Y.-H., Wang, S., Wei, J., Zhang, X., Xu, C., Luan, Z., Wu, D. and Wei, B. (2002) Chem. Phys. Lett. 357, 263. Li, Y.-H., Wang, S., Cao, A., Zhao, D., Zhang, X., Xu, C., Luan, Z., Ruan, D., Liang, J., Wu, D. and Wei, B. (2001) Chem. Phys. Lett. 350, 412. Lin, C.J. and Chang, J.E. (2001) Chemosphere 44, 1185. Long, R.Q. and Yang, R.T. (2001) J. Am. Chem. Soc. 123, 2058. Low, K.S., Lee, C.K. and Wong, S.L. (1995) Environ. Technol. 16, 877. Merdy, P., Guillon, E., Dumonceau, J. and Aplilncourt, M. (2002) Environ. Sci. Technol. 36, 1728. Panday, K.K., Prasad, G. and Singh, V.N. (1986) Water, Air, Soil Pollut. 27, 287. Park, Y.J., Jung, K.H. and Park, K.K. (1995) J. Colloid Interface Sci. 171, 205. Reddad, Z., Gerente, C., Andres, Y. and Cloirec, P.L. (2002) Environ. Sci. Technol. 36, 2067. Shaffer, M.S.P., Fan, X. and Windle, A.H. (1998) Carbon 36, 1603. Scheinost, A.C., Abend, S., Pandya, K.I. and Sparks, D.L. (2001) Environ. Sci. Technol. 35, 1090. Strelko, V., Malik, D.J. and Streat, M. (2002) Carbon 40, 95. Treacy, M.M.J., Ebbesen, T.W. and Gibson, J.M. (1996) Nature (London) 381, 678. Williams, C.J., Aderhold, D. and Edyvean, R.G.J. (1998) Water Res. 32, 216. Yu, B., Zhang, Y., Shukla, A., Shukla, S.S. and Dorris, K. (2001) J. Hazard. Mater. B 84, 83.