Relation between the Charge Efficiency of Activated Carbon Fiber and

Article pubs.acs.org/Langmuir

Relation between the Charge Efficiency of Activated Carbon Fiber and Its Desalination Performance Zheng-Hong Huang,* Ming Wang, Lei Wang, and Feiyu Kang Laboratory of Advanced Materials, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, China S Supporting Information *

ABSTRACT: Four types of activated carbon fibers (ACFs) with different specific surface areas (SSA) were used as electrode materials for water desalination using capacitive deionization (CDI). The carbon fibers were characterized by scanning electron microscopy and N2 adsorption at 77 K, and the CDI process was investigated by studying the salt adsorption, charge transfer, and also the charge efficiency of the electric double layers that are formed within the micropores inside the carbon electrodes. It is found that the physical adsorption capacity of NaCl by the ACFs increases with increasing Brunauer−Emmett−Teller (BET) surface area of the fibers. However, the two ACF materials with the highest BET surface area have the lowest electrosorptive capability. Experiments indicate that the charge efficiency of the double layers is a key property of the ACF-based electrodes because the ACF material which has the maximum charge efficiency also shows the highest salt adsorption capacity for CDI.

■

INTRODUCTION The water crisis is expected to become more serious in the next few decades because of severe water shortage occurring globally, even in currently considered water-rich regions, caused by environmental contamination.1 To solve such a problem, tremendous amounts of efforts in research are demanded, including identifying new methods of purifying water at a low cost and with less energy and concurrently minimizing the use of chemicals and their impact on the environment.1,2 Capacitive deionization (CDI), which uses ion electrosorption onto the surface of porous materials to remove ions from aqueous solutions, can be used as an energy-efficient alternative for the desalination of brackish water when comparing with more conventional desalination methods like membrane separation and thermal distillation.2,3 The CDI technology works at a low voltage, not exceeding the decomposing potential of water. The ions adsorbed are held in the electric double layers (EDL) near the charged surface of a flow-through capacitor and can be released back into the bulk solution by canceling the potential difference between the electrodes. 2 No additional chemicals are required for regeneration. In fact, the part of discharging energy can be restored for the sake of energy recovery. A prerequisite for the high deionization efficiency of CDI is the demand for the high electrosorption capacity endowed by the electrode materials. To attain the objective, it is important that a large surface area is created, and thus typical materials such as high specific surface area (SSA) materials are used.4−6 Carbon aerogels,7−11 activated carbons,12−15 multiwall carbon nanotubes and nanofibers,16−18 ordered mesoporous carbons,19,20 graphenes,21,22 and oxide-incorporated carbons23−25 © 2012 American Chemical Society

have been investigated as the CDI electrodes. The high SSA, a key factor of the electrosorption capacity, is mainly determined by the contribution of the surface areas of mesopores (pores of width between 2 and 50 nm) and micropores (pores of width smaller than 2 nm) present in the carbon materials. Because the SSA is explicitly related to the pore size, it is especially important to understand its effect on the electrosorption capacity. A latest research related to the electric double layer capacitors has shown an anomalous increase in carbon capacitance at pore sizes less than 1 nm.26 In the CDI, the pore openings of activated carbons were controllably tuned to exhibit an excellent electrochemical selectivity for ions.27,28 The electrosorptive performance with modified activated carbon cloth as the CDI electrodes was also investigated.29,30 Up to now, the research emphasis is more focused on the high SSA of porous carbon materials with the order of 103 m2/g. Activated carbon fibers (ACFs) are typical microporous materials with large surface areas and with a good conductivity along the fiber axis. Thus, they can serve as a candidate for the active component of CDI electrodes. The aim in this work is to identify the effect of SSA on the capacitive electrosorption using ACF electrodes. The experiments demonstrate that the desalination capacity of CDI is highly affected by the SSA of ACFs, and the characteristics of electrosorption under electric field modulation are very different from those of physical adsorption. A high electrosorptive performance is capable of Received: November 28, 2011 Revised: February 27, 2012 Published: February 28, 2012 5079

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084

Langmuir

Article

mL according to the CDI configuration. The volume is 2.7 mL. The cell voltage applied between the two electrodes in these experiments was 1.2 V. The solution conductivity was online monitored with 5 s interval by a conductivity meter (type 308A, Leici Co.). The pH measurements of the solution before and after experiments was conducted, and the pH value keeps the same as 5.6. Galvanostatic cycling was carried out by a Land Cell Measurement system. Cyclic voltammetry (CV) was performed with Ag/AgCl as the reference electrode in a 50 mL beaker. Determination of the Charge Efficiency. The current and voltage were measured online as well as the solution conductivity in the beaker. The experiment was run sufficiently long (1−2 h) for the current drop to 0 (except for a leakage current). The charge efficiency Λ is defined as the ratio of equilibrium salt adsorption Γsalt over electrode charge Σ.

being achieved by the ACF electrode with the maximum charge efficiency of the double layers.31

■

EXPERIMENTAL SECTION

Materials. The pitch-based ACFs (A5, A7, A10, and A15) were kindly provided by Osaka Gas Co., Ltd., Japan. Their Brunauer− Emmett−Teller (BET) surface areas were 500, 700, 1100, and 1600 m2/g, respectively, according to the corporation’s data. The character-

Table 1. Pore Structure Parameters of ACFs sample

BET SSA (m2/g)

pore vol (cm3/g)

av pore width (nm)

A5 A7 A10 A15

480 670 1050 1890

0.21 0.32 0.47 0.92

1.8 1.9 1.8 1.9

istics of ACFs measured in experiments are presented in Table 1. Sodium chloride and N-methyl-2-pyrrolidone were analytical grade without further purification before use. Poly(vinylidene fluoride) (PVDF) 761 was employed as a binder. Highly purified graphite powder was added to enhance the conductivity of CDI electrodes. Graphite paper was used as a current collector. Materials Characterization. The ACFs were observed by a LEO1530 scanning electron microscopy (SEM). The pore structures of ACFs were measured by nitrogen sorption method at 77 k with a Micromeritics ASAP 2020 analyzer. Fourier transform infrared spectra (FTIR) were measured by a spectrophotometer (Nicolet AVATAR 360) using KBr pressed-disk method. Raman spectra were obtained in Raman spectrometer 93 (RM2000, England). Fabrication of CDI Electrodes. The electrodes were fabricated as the following procedures. 0.20 g of ACF, 0.08 g of PVDF binder, and 0.12 g of graphite powder were mixed with 2 mL of N-methyl-2pyrrolidone. A 5.0 × 5.4 cm graphite paper was coated with the mixture and dried in vacuum at 50 °C for 3 h and then pressed by a roller squeezer to make each plate as thick as 0.5 mm. To make sure removing the solvent thoroughly, the electrodes were dried again in vacuum at 130 °C for 3 h. CDI Configuration and Electrochemical Measurements. The CDI configuration is illustrated in Figure 1. A CDI cell consists of two parallel electrode plates separated by a 1.0 mm gap for solution flow. A sodium chloride aqueous solution was pumped into the bottom and exited from the top of the cell by a peristaltic pump with a flow rate of 6 mL/min. The effluent went to a beaker from which the cell was fed again. The total water volume in the system (beaker + tubes + cell) was 11 mL. The volume between the electrodes was calculated as 2.7

Λ = Γsalt/Σ

(1)

Γsalt = (C initial − Cequilibrium)V

(2)

Σ=

∫ I dt

(3)

The Λ was calculated from the independent measurement of the equilibrium salt adsorption Γsalt and equilibrium charge Σ. The salt adsorption could be calculated from (Cinitial − Cequilibrium) and multiplying with the total water volume V. The current signal was integrated with time to obtain the electrode charge Σ for the current to drop to zero. Here, a leakage current, observed as a constant small current in the salt removal step, was subtracted from the data. Γsalt and Σ are given per mass of all electrodes (the mass of electrode is sum of ACF materials, binder, and current collector).

■

RESULTS AND DISCUSSION Four ACFs with different SSA were selected for comparison in the desalination of CDI cell. Figure S1 (Supporting Information) shows the SEM images of the four ACFs with a smooth surface and a similar fiber diameter of ca. 10−18 μm. Figure S2 shows the isotherms of nitrogen adsorption/ desorption for the ACFs. All samples exhibit type I isotherms, which are the typical adsorption characteristics of microporous materials. It could be suggested from the opening of the knee of the isotherm and the higher slope of the plateau that there is a wider pore diameter distribution with the increasing SSA.32 No hysteresis phenomenon in the adsorption/desorption process was observed. The detailed parameters of ACFs in Table 1 show that the pore volume increases with increasing the BET SSA, but the four ACFs have almost the same average pore width of 1.8 nm. The pore size distribution of ACFs is presented in Figure S3. A5-A15 ACFs are mainly microporous. The pore size distribution becomes broader with the increased surface area. A5 ACF has the pore size distribution of less than 1 nm. The pore size distribution of A7 and A10 ACFs are less than 2 nm, while A15 ACF has micro- and mesopores less than 3 nm. Carbon structures and surface functional groups of ACFs were characterized with Raman spectroscopy and infrared spectroscopy (Figures S4 and S5). The Raman spectrum of carbon has two peaks: one at 1580 cm−1 (G peak) near the inherent characteristics of graphite scattering peak and another at 1360 cm−1 (D peak) near the graphite lattice by defects, edge of the disordered arrangement of carbon, and low-symmetry structure factors. The R value (R = ID/IG, where ID and IG are the intensity of D peak and G peak) was employed to evaluate the graphitization degree. It can be calculated that R value increases from A5 to A15, indicating the decreased graphitization degree. FTIR spectra of ACFs show the peaks at 3431,

Figure 1. Schematic of the CDI configuration with ACFs as the electrodes. 5080

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084

Langmuir

Article

1631, and 1066 cm−1, corresponding to the O−H, CO, and C−O vibration, respectively. The peak around at 2888 cm−1 is attributed to C−H vibration. It indicates that there are similar oxygen-containing functional groups on the carbon fiber surface. The desalination performances of A5-A15 ACFs were evaluated by using the CDI apparatus to treat a sodium chloride aqueous solution at a salt concentration of 1 and 16 mM, respectively. The Stokes diameter is 0.184 nm for Na+ and 0.121 nm for Cl−. The electrode assembly was first washed by pure water until the conductivity of solution approached zero and then dried, and the desalination capacity of CDI electrodes in salt solution was recoded. Repeated wash/record process indicates the desalination performance of CDI electrodes was almost kept the same. Figure 2a shows the CDI result of A5-

NaCl solution decreased sharply due to the electrosorption. Then the change in the concentration became gradually gentle before the electrosorption reached equilibrium. In general, it would take more time to reach equilibrium in the case of a higher electrosorption. A remarkable phenomenon could be observed in the experiment that the ACF with a high BET surface area has low electrosorption capability. For example, A15 had the lowest electrosorption capability but possessed the highest physical adsorption capacity. It is noted that A7 is the most capable of adsorbing NaCl in the CDI. The order of electrosorption capacity is A7 > A5 > A10 > A15. Figure 2b indicates the CDI result of A5-A15 ACFs at a higher concentration of 16 mM. During the physical adsorption there is no obvious change in solution concentration for A5A15 ACFs after a long run (1 h); thus, only the electrosorption data are shown. The CDI capacity at a salt concentration of 16 mM shares the same rule with that at a salt concentration of 1 mM. A15 and A10 possess a low CDI capacity. A7 has the highest CDI capacity, and A5 falls in between them. It can be identified that the BET SSA is not a sole significant factor for the deionization in the CDI. The electrochemical characterization of A5-A15 ACFs was performed with the galvanostatic cycling measurement and CV measurement (Figures S6 and S7). It is noted that the electrochemical characterization of A5-A15 ACFs via the galvanostatic cycling measurement and CV measurement focuses on capacitance (or electrode charge). However, the experiment data are very difficult to illustrate that the two ACF materials with the highest BET surface area have the lowest electrosorptive capability. Recently, Zhao and co-workers propose an effective tool, charge efficiency, to probe the double-layer structure inside of porous electrode in the modeling of CDI.31 They introduce an alternative procedure rather than only considering data for electrode charge. The procedure combines the data for electrode charge with the data for the equilibrium salt adsorption from solution into the double layers inside of the electrode. Theoretically, each electron charge will be fully charge-balanced by counterion adsorption. In fact, co-ions are excluded from the double layer simultaneously with counterions adsorption. The effect of coion exclusion reduces the ratio of equilibrium salt adsorption Γsalt over electrode charge Σ, a ratio defined as the charge efficiency Λ. The detailed calculation procedures are given in the Experimental Section. Γsalt and Σ are given per mass of all electrodes. The data in Figure 3 are not for constant background salt concentration, but the salt concentration is different for each data point because of the use of a small recycle volume-batch experiment. Figure 3a presents the electrode charge as a function of cell potential at an initial salt concentration of 16 mM. The electrode charge increases with increasing the cell potential for each ACF. A15 ACF has the largest electrode charge among the four ACFs in the range from 0.2 to 1.2 V. A5 and A7 ACFs have the smallest electrode charge and A10 falls in between them. This is in agreement with the trend obtained by the galvanostatic cycling measurement and CV measurement. Figure 3b presents the salt adsorption as a function of the cell potential at a salt concentration of 16 mM. The salt adsorption increases with increasing the cell potential for each ACF. In contrast to the electrode charge, the capacity is relatively low for the salt adsorption of A15 and A10 ACFs, while relatively high for that of A7. A5 falls in between them. Figure

Figure 2. Desalination behavior of ACFs with different BET SSA at a salt concentration of 1 (a) and 16 mM (b).

A15 ACFs at a salt concentration of 1 mM. ACFs are capable of adsorbing an amount of NaCl by themselves due to the interaction between polar groups of the carbon surface and ions.33 Thus, the CDI can be divided into two stages: the physical adsorption and the electrosorption. There was no an applied electric field during the first hour in the beginning of the CDI process, and the change in the concentration of the solution due to the physical adsorption was recorded. After that, the CDI electrodes were put at a constant voltage of 1.2 V. At the stage of the physical adsorption, it is commonly observed that the amount of NaCl adsorbed by the ACFs increased in the order from sample A5 to A15, which was in agreement with the BET SSA of the fibers. Once an electric field was applied to the CDI electrodes, the concentration of 5081

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084

Langmuir

Article

Figure 4. Desalination cycling of A5 and A7 with initial concentration of 1 mM at 1.2 V cell voltage.

regeneration (discharge process) occurred as the imposed voltage was removed. When the concentration returned to the setting value in the discharge process, the next charge process started. From the cycling data one can calculate that the discharge rate for A7 is slower than that for A5. The charge is obviously influenced by the electric field, while the discharge depends on the concentration gradient from the electrode surface to bulk solution and the interaction between the ACF surface and the adsorbate. A larger electric field force results in a faster charge. During the charge process both A5 and A7 have almost the same rates under the same voltage before reaching the saturated region. By contrast, during the discharge process the rate of A5 is higher than that of A7, which is mainly controlled by the physical adsorption/desorption. The physical adsorption holds ions due to the interaction between polar groups of the carbon surface and ions, which may affect the regeneration by the desorption. The reduced physical adsorption indicates a weaker binding force that makes the desorption more efficient.13 A5 has a weaker physical adsorption than A7, as shown in Figure 2a, which makes A5 capable of desorbing more quickly. Freshwater generally contains an amount of total dissolved solids up to 1000 mg/L, brackish water from 1000 to 10 000 mg/L, and seawater above 35 000 mg/L. To investigate the maximum capacitive adsorption capacity of one unit CDI cell, the concentration variation as a function of an initial feeding solution concentration ranging from 1.7 to 69 mM (100 to 4000 mg/L) for the ACFs was conducted. The electrosorption capacity increases gradually with increasing the initial feeding solution concentration at 1.2 V cell voltage (Figure 5). These are data using the same setup as before with 11 mL of total water. For A5-A15 ACFs, the isotherms show a trend of saturated adsorption at the extremely high ion concentration. A7 exhibits the largest adsorption capacity in the whole ion concentration range. A5 shows better performance than A10 and A15, which is in agreement with the previous observation at the low ion concentration. The adsorption data of A5, A7, A10, and A15 is fitted according to the Langmuir equation:

Figure 3. (a) Equilibrium electrode charge as a function of cell potential with the starting salt concentration of 16 mM. (b) Total equilibrium salt adsorption as a function of cell potential. (c) Charge efficiency as a function of cell potential.

3c indicates the charge efficiency of the double layers for A5A15 ACFs according the definition by Λ = Γsalt/Σ. It can be determined that the charge efficiency for each ACF increases with increasing the cell potential and demonstrates a saturated trend at high potentials. A7 has the largest charge efficiency of the double layers with a charge efficiency of 0.8 at 1.0−1.2 V, while A15 and A10 have the lowest charge efficiency and A5 falls in between them. Such a fact indicates that the charge efficiency of the double layers is a key property of the ACF electrode. The ACF, which has the maximum charge efficiency of the double layers, possesses the optimum CDI capacity. To investigate the regeneration of CDI, the desalination cycling of A5 and A7 was performed (Figure 4). The

qe = qmKLCe/(1 + KLCe)

(4)

where Ce is the equilibrium concentration of solution, qm is the largest mass of NaCl adsorbed per gram single electrode, and KL is the Langmuir constant. The fitting results are shown in Figure 5. For A7, qm is 0.076 mmol/g and KL is 0.32 mM−1. It is 5082

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084

Langmuir

■

Article

AUTHOR INFORMATION

Corresponding Author

*Fax: +86-10-62771160; e-mail: [email protected]. cn. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors gratefully acknowledge the National Natural Science Foundation of China (Grants 51072091 and 40806042), the cooperative project JST-MOST (No. 2011DFA50430), the Postdoctoral Science Foundation of China (Grant 20090460269), and the Program for New Century Excellent Talents in University (NCET-10-0496) for the financial support.

■

Figure 5. Langmuir isotherm fits of salt adsorption per gram single electrode at 1.2 V cell voltage for A5, A7, A10, and A15.

(1) Shannon, M. A.; Bohn, P. W.; Elimelech, M.; Georgiadis, J. G.; Marinas, B. J.; Mayes, A. M. Nature 2008, 452, 301−310. (2) Oren, Y. Desalination 2008, 228, 10−29. (3) Qi, D.; Zou, L.; Hu, E. Res. J. Chem. Environ. 2007, 11, 92−95. (4) Bouhadana, Y.; Avraham, E.; Soffer, A.; Aurbach, D. AIChE J. 2010, 56, 779−789. (5) Biesheuvel, P. M.; Bazant, M. Z. Phys. Rev. E 2010, 81, 031502. (6) Biesheuvel, P. M.; van der Wal, A. J. Membr. Sci. 2010, 346, 256− 262. (7) Biesheuvel, P. M.; van Limpt, B.; van der Wal, A. J. Phys. Chem. C 2009, 113, 5636−5640. (8) Xu, P.; Drewes, J. E.; Heil, D.; Wang, G. Water Res. 2008, 42, 2605−2617. (9) Jung, H. H.; Hwang, S. W.; Hyun, S. H.; Kang-Ho, L.; Kim, G. T. Desalination 2007, 216, 377−385. (10) Gabelich, C. J.; Tran, T. D.; Suffet, I. H. Environ. Sci. Technol. 2002, 36, 3010−3019. (11) Farmer, J. C.; Fix, D. V.; Mack, G. V.; Pekala, R. W.; Poco, J. F. J. Electrochem. Soc. 1996, 143, 159−169. (12) Lee, J. B.; Park, K. K.; Yoon, S. W.; Park, P. Y.; Park, K. I.; Lee, C. W. Desalination 2009, 237, 155−161. (13) Zou, L.; Morris, G.; Qi, D. Desalination 2008, 225, 329−340. (14) Wang, F. Y.; Wang, H.; Ma, J. W. J. Hazard. Mater. 2010, 177, 300−306. (15) Choi, J. H. Sep. Purif. Technol. 2010, 70, 362−366. (16) Pan, L. K.; Wang, X. Z.; Gao, Y.; Zhang, Y. P.; Chen, Y. W.; Sun, Z. Desalination 2009, 244, 139−143. (17) Gao, Y.; Pan, L. K.; Li, H. B.; Zhang, Y. P.; Zhang, Z. J.; Chen, Y. W.; Sun, Z. Thin Solid Films 2009, 517, 1616−1619. (18) Wang, X. Z.; Li, M. G.; Chen, Y. W.; Cheng, R. M.; Huang, S. M.; Pan, L. K.; Sun, Z. Appl. Phys. Lett. 2006, 89, 3. (19) Li, L. X.; Zou, L. D.; Song, H. H.; Morris, G. Carbon 2009, 47, 775−781. (20) Wang, X. Q.; Lee, J. S.; Tsouris, C.; DePaoli, D. W.; Dai, S. J. Mater. Chem. 2010, 20, 4602−4608. (21) Li, H. B.; Lu, T.; Pan, L. K.; Zhang, Y. P.; Sun, Z. J. Mater. Chem. 2009, 19, 6773−6779. (22) Humplik, T.; Lee, J.; O’Hern, S. C.; Fellman, B. A.; Baig, M. A.; Hassan, S. F.; Atieh, M. A.; Rahman, F.; Laoui, T.; Karnik, R.; Wang, E. N. Nanotechnology 2011, 22, 292001. (23) Yang, C. M.; Choi, W. H.; Na, B. K.; Cho, B. W.; Cho, W. I. Desalination 2005, 174, 125−133. (24) Ryoo, M. W.; Seo, G. Water Res. 2003, 37, 1527−1534. (25) Leonard, K. C.; Genthe, J. R.; Sanfilippo, J. L.; Zeltner, W. A.; Anderson, M. A. Electrochim. Acta 2009, 54, 5286−5291. (26) Chmiola, J.; Yushin, G.; Gogotsi, Y.; Portet, C.; Simon, P.; Taberna, P. L. Science 2006, 313, 1760−1763. (27) Noked, M.; Avraham, E.; Bohadana, Y.; Soffer, A.; Aurbach, D. J. Phys. Chem. C 2009, 113, 7316−7321.

found that the Langmuir isotherm correlates well with the experimental data according to the correlation coefficient (R2 = 0.98). This phenomenon suggests that the monolayer adsorption is primary during the electrosorption process. Other researchers also conducted similar experiments to plot the isotherms of different materials. Those data were in agreement with Langmuir isotherms as well, indicating monolayer coverage of the electrode surface area. Thus, we can compare the maximum adsorption capacity (Qm) of different electrode materials. The categories specific surface area (SSA), applied voltage (V), and Qm of different materials are shown in Table 2. Table 2. Maximum Adsorption Capacity (Qm) According to Langmuir Isotherm of Different Electrode Materials samples 9

carbon aerogel disk A7 ACF (this paper) CNT powders34 CNT-CNFs35 graphene21 TiO2-ACC33

applied voltage (V)

SSA (m2/g)

Qm (mg/g)

1.5 1.2 1.2 1.2 2.0 1.2

610 670 153 210.6 14.2 1180

7.02 8.9 8.91 12.3 21 27.5

■

CONCLUSIONS

■

ASSOCIATED CONTENT

REFERENCES

The CDI characteristics of ACFs are significantly different from their physical adsorption. Experiments indicate the charge efficiency of the double layers is a key property of the ACF electrode. The ACF, which has the maximum charge efficiency of the double layers, possesses the optimum CDI capacity. The ACF with not so high activation degree (A7 in present case) possesses a high CDI capacity, which suggests that ACF could be used as CDI electrode materials in large scale for practical application of the CDI method.

S Supporting Information *

Additional figures. This material is available free of charge via the Internet at http://pubs.acs.org. 5083

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084

Langmuir

Article

(28) Noked, M.; Avraham, E.; Soffer, A.; Aurbach, D. J. Phys. Chem. C 2010, 114, 13354−13361. (29) Avraham, E.; Noked, M.; Bouhadana, Y.; Soffer, A.; Aurbach, D. J. Electrochem. Soc. 2009, 156, P157−P162. (30) Avraham, E.; Bouhadana, Y.; Soffer, A.; Aurbach, D. J. Electrochem. Soc. 2009, 156, P95−P99. (31) Zhao, R.; Biesheuvel, P. M.; Miedema, H.; Bruning, H.; van der Wal, A. J. Phys. Chem. Lett. 2010, 1, 205−210. (32) Alcanizmonge, J.; Cazorlaamoros, D.; Linaressolano, A.; Yoshida, S.; Oya, A. Carbon 1994, 32, 1277−1283. (33) Ryoo, M. W.; Kim, J. H.; Seo, G. J. Colloid Interface Sci. 2003, 264, 414−419. (34) Wang, S.; Wang, D. Z.; Ji, L. J.; Gong, Q. M.; Zhu, Y. F.; Liang, J. Sep. Purif. Technol. 2007, 58, 12−16. (35) Wang, X. Z.; Li, M. G.; Chen, Y. W.; Cheng, R. M.; Huang, S. M.; Pan, L. K.; Sun, Z. Electrochem. Solid State 2006, 9, E23−E26.

5084

dx.doi.org/10.1021/la204690s | Langmuir 2012, 28, 5079−5084