Novel Graphene-Like Electrodes for Capacitive Deionization

Oct 21, 2010 - of South Australia, Adelaide, SA 5095, Australia, and. Engineering Research ... However, most water desalination technologies are energ...
0 downloads 0 Views 4MB Size
Environ. Sci. Technol. 2010, 44, 8692–8697

Novel Graphene-Like Electrodes for Capacitive Deionization H A I B O L I , †,‡ L I N D A Z O U , * ,† L I K U N P A N , ‡ AND ZHUO SUN‡ SA Water Centre for Water Management and Reuse, University of South Australia, Adelaide, SA 5095, Australia, and Engineering Research Center for Nanophotonics & Advanced Instrument, Ministry of Education, Department of Physics, East China Normal University, Shanghai, 200062 China

Received June 7, 2010. Revised manuscript received July 28, 2010. Accepted October 7, 2010.

Capacitive deionization (CDI) is a novel technology that has been developed for removal of charged ionic species from salty water, such as salt ions. The basic concept of CDI, as well as electrosorption, is to force charged ions toward oppositely polarized electrodes through imposing a direct electric field to form a strong electrical double layer and hold the ions. Once the electric field disappears, the ions are instantly released back to the bulk solution. CDI is an alternative low-energy consumption desalination technology. Graphene-like nanoflakes (GNFs) with relatively high specific surface area have been prepared and used as electrodes for capacitive deionization. The GNFs were synthesized by a modified Hummers’ method using hydrazine for reduction. They were characterized by atomic force microscopy, N2 adsorption at 77 K and electrochemical workstation. It was found that the ratio of nitric acid and sulfuric acid plays a vital role in determining the specific surface area of GNFs. Its electrosorption performance was much better than commercial activated carbon (AC), suggesting a great potential in capacitive deionisation application. Further, the electrosorptive performance of GNFs electrodes with different bias potentials, flow rates and ionic strengths were measured and the electrosorption isotherm and kinetics were investigated. The results showed that GNFs prepared by this process had the specific surface area of 222.01 m2/g. The specific electrosorptive capacity of the GNFs was 23.18 µmol/g for sodium ions (Na+) when the initial concentration was at 25 mg/ L, which was higher than that of previously reported data using graphene and AC under the same experimental condition. In addition, the equilibrium electrosorption capacity was determined as 73.47 µmol/g at 2.0 V by fitting data through the Langmuir isotherm, and the rate constant was found to be 1.01 min-1 by fitting data through pseudofirst-order adsorption. The results suggested that the chemically synthesized GNFs can be used as effective electrode materials in CDI process for brackish water desalination.

1. Introduction With the rapid development of modern industry and population growth, fresh water shortages and the energy crisis have become the major issues in many countries worldwide. * Corresponding author phone: +61 8 830 25489; fax: +61 8 830 23386; e-mail: [email protected]. † University of South Australia. ‡ East China Normal University. 8692

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

Among all of alternative approaches for water purification, desalination is one important option for producing potable water from brackish water and seawater in many parts of the world. However, most water desalination technologies are energy and capital intensive. The widely applied desalination technologies include reverse osmosis, thermal processes, electrodialysis and ion exchange. During the past few decades, great efforts have been devoted to improve these technologies and make them more efficient and cost-effective, but shortcomings still exist. Regarding the ion exchange and reverse osmosis process, for example, the main disadvantage of the ion exchange process is the secondary chemical waste resulting from the regeneration of the saturated resin by acid. In the conventional reverse osmosis system, significant energy is needed to drive the water through the membrane, and the membrane will also undergo degradation with time (1, 2). In recent years, new technology has been developed for removal of charged ionic species from brackish water: capacitive deionization (CDI), which is a novel water purification technology that has no secondary pollution, is cost-effective and energy efficient (3-6). The basic concept of CDI, as well as electrosorption, is to force charged ions toward oppositely polarized electrodes through imposing a direct electric field. If such an electrostatic field is introduced, electrodes of high conductivity and high specific surface area will form a strong electrical double layer to hold the ions, and once the electric field disappears, the ions are instantly released back to the bulk solution. Thus, CDI offers several advantages over the above-mentioned water treatment technologies, such as lower energy consumption (1.2 V in charge process), easy regeneration of electrodes and environmental friendliness because no harsh chemicals are used in cleaning process. From the energy efficiency point of view, CDI aims to remove the salt ions, which are only a small percentage of the feed solution, as compared to most other technologies that aim to separate water, which accounts of 90% of the feed solution. The former requires less energy. Suitable carbon materials are the most important component in CDI devices, as they are used as the electrodes that play a significant role in electrosorptive process. Graphene is an ideal two-dimensional carbon material and has attracted much research attention due to several breakthroughs in fundamental research and some promising practical applications (7-12). The recent rapid improvement in chemical methodology for synthesizing graphene has made large-scale preparation feasible. Further, its potential applications in various fields have been explored (13, 14). As a low-dimensional material, graphene has a large theoretical specific area of 2600 m2/g, which is twice that of the finely divided activated carbon (15). Graphene presents an excellent tensile modulus up to 35 GPa and a superior room temperature electrical conductivity of 7200 s/m (16). These intriguing mechanical and electrical properties enable graphene to be used as a freestanding electrode for energy storage as well as in electrosorption electrode. The possibility of using graphene in electrosorption of NaCl from aqueous solutions has been reported by this research group previously (17). It was found the specific electrosorptive capacity of graphene was 20.06 µmol/g, which was a good result for a material that has relatively low specific surface area (14.2 m2/g). It is believed that this electrosorption capacity can be further improved by increasing the surface area of graphene. In this paper, GNFs have been synthesized by a modified Hummers’ method using hydrazine for reduction. The working conditions of the capacitive desalting tests, such as 10.1021/es101888j

 2010 American Chemical Society

Published on Web 10/21/2010

on the electrosorptive isotherm and kinetics of GNFs electrodes were carried out and the obtained data are validated.

2. Materials and Methods

FIGURE 1. Schematic of electrosorptive unit (a) and cell batch-mode experiment (b), respectively. The CDI unit was consisted of graphite plate, GNFs film, and separator. flow rate and bias potential, are first discussed. The electrosorptive behavior of GNFs, as well as commercial activated carbon (AC), was investigated under the same experimental conditions for comparison. In addition, further investigations

2.1. Fabrication of GNFs. Graphite oxide (GO) powders were synthesized by a modified Hummers method (17, 18). More detailed information can be found in the Supporting Information (SI). In order to investigate the impact of acid pretreatment on the specific surface area of graphene, we defined the samples as GNFs-1, GNFs-2, and GNFs-3, corresponding to the volume ratio of nitric acid and sulfuric acid as 0:1, 1:1, and 2:1, respectively. Exfoliation was carried out by mixing with hydrazine for 24 h at a temperature of 353 K. The as-prepared graphene suspension was washed several times and filtered carefully, followed by air-drying at 313 K. The GNFs were then ready to use. 2.2. Fabrication of Graphene Based Electrode. The carbon materials (GNFs and AC), graphite powder as conductive material and PTFE as binder, were used to fabricate the electrode and their percentage in the final electrode were 72%, 20%, and 8%, respectively. Each electrode was 70 mm wide × 140 mm long × 0.3 mm thick, and had a flow-through hole with a diameter of 4 mm. To achieve adhesion between the carbon mixture and graphite layer, the raw mixture of powders (carbon materials, graphite powder and PTFE powder) were grinded several hours. Ethanol (10-20 mL) was added dropwise to the mixture to increase the moisture and then the mixture was pressed on a graphite sheet. Finally, the electrodes were assembled into a CDI device for testing, as shown in Figure 1(a). 2.3. Characterization. The structure of the electrode material was characterized by scanning electron microscopy (SEM, XL30) and transmission electron microscopy (TEM, CM200) imaging. The thickness of as-prepared GNFs sample was determined by atomic force microscopy (AFM, SPA-

FIGURE 2. (a) TEM image of GNFs-2. The folded edge exhibits a relative height of 1.22 nm obtained from (b) and (c), which depict the AFM image and cross sectional analysis of GNFs-2, respectively (the scale bar is 10 µm). VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8693

400) analysis. The N2 adsorption-desorption isotherms of the carbon materials were performed at -196 °C on Belsorp system (BEL JAPAN, Inc., calculation method in SI). Electrochemical measurements of GNFs were examined by the potentiostat analyzer (Autolab PGSTAT128N) in a typical three-compartment cell at ambient temperature (see SI). Film resistivity of GNFs and AC was measured by a dip coating method (see SI). 2.4. Batch Mode Electrosorptive Experiment. To investigate the electrosorptive capacity of GNFs and AC electrodes, batch-mode experiments were conducted in a continuously recycling system including an electrosorptive unit cell and conductivity monitor as shown in Figure 1(b). In each experiment, the solution was continuously pumped by a peristaltic pump into the unit cell and the effluent returned to the feed tank. To determine the optimum flow rate and working voltage for GNFs-based CDI, the solution volume was maintained at 50 mL and the solution temperature was kept at 298 K, respectively. Meanwhile, the applied voltage was adjusted to 2.0 V at which no electrolysis of water occurred and the flow rate was varied in the range from 15 to 45 mL/min. In other experiments, the flow rate was kept at a constant value but the applied voltage was increased in steps of 0.2 V from 1.0 to 2.0 V to find the optimum working voltage. The above-mentioned experiments were performed in a NaCl solution, which had an initial conductivity of 55 µS/cm. The relationship between conductivity and concentration was obtained according to a calibration table made prior to the experiment. The concentration variation of NaCl solution was continuously monitored and measured at the outlet of the unit cell using conductivity meter. In our experiment, the electrosorptive capacity was defined as follows: electrosorptive capacity(µmol/L) )

(C-C0)V 38.44 × M

(1)

Where C and C0 (mg/L) represent the final and initial concentration, respectively, and V is the volume of the container (mL), and M is the mass of GNFs or AC.

3. Results and Discussion 3.1. Morphology and Surface Area of GNFs. Figure 2(a) displays the TEM image of GNFs. It can be seen that the GNFs sheets shown a transparent character due to its thin layered structure with some wrinkles and folds on the surface and edge. An AFM image of a typical GNFs film with severalnanometers thickness deposited on the silicon nitride-onsilicon substrate is shown in Figure 2(b) and (c). This film is made up of overlapped graphite oxide nanoflakes so that a given region might consist of one layer (about 0.86 nm thick) or multiple layers (1.62 nm thick, not shown in Figure 2), the overlapped regions typically have three or more nanoflake layers. By observing the layers’ structure from the TEM, SEM, and AFM image, it is found that the GNFs used in this experiment were mainly made up of several grapheme layers. The film with a small area shows conjugated characteristics that may be induced from the acid pretreatment in the first step (19). In addition, the dependence of electrosorption capacities on the characteristics of GNF electrodes such as specific surface area is the focus in the experiments. As discussed elsewhere (19), the acid treatment in the first step plays an important role in the preparation of graphene, especially the amount of nitric acid. Thus, the ratio of nitric acid and sulfuric acid was increased from 0:1 to 2:1 to achieve the result of higher BET surface area. It is found that the measured pore size associated with GNFs by BJH method was mainly contributed by the interlayer gaps between the GNFs. If the original graphite was not adequately exfoliated during the acid pretreatment, it increased the 8694

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

FIGURE 3. (a) N2 adsorption-desorption isotherm of AC and GNFs sample at 77 K. (b) Pore size distribution of AC and different GNFs samples, inset presents the SBET and pore volume of corresponding sample. possibility of serious aggregation and therefore resulted in the decreased specific surface. In addition, no peak was found in pore size distribution suggested such gaps between GNFs varied widely in size. Conversely, once the raw graphite was treated by an excessive acid in oxidation process, it could also induce the aggregation due to the recombine between graphite oxides and thus lead to the decrease of the specific surface area. Figure 3(a) shows the typical N2 adsorptiondesorption isotherm of AC and GNFs samples at 77 K, respectively. The pore size distribution of the AC and GNFs samples are shown in Figure 3(b). The inset presents the BET surface area and pore volume of the AC and GNFs samples. The BET specific surface area of AC is 989.54 m2/g, whereas the maximum BET specific surface area of GNFs is 222.01 m2/g, corresponding to the ratio of nitric acid and sulfuric acid of 1:1. As shown in Figure 3(b), the BET surface area of GNFs-1 is 71.76 m2/g and the total pore volume is 0.16 cm3/g, indicating that the severe aggregation in GNFs-1 sample. Once the nitric acid was added into the mixture of graphite and sulphuric acid at the ratio of nitric acid and sulphuric acid 1:1, the BET surface area for GNFs-2 reached the higher value of 222.01 m2/g and the total pore volume increased to 51.01 cm3/g as shown in Figure 3(b). Further, if we increase the ratio further from 1:1 to 2:1, the BET surface area of GNFs-3 was decreased to 44.7 m2/g with a lower total pore volume of 10.27 cm3/g, indicating that the excessive nitric acid could dilute the concentration of sulfuric acid and thereby reduce its strength. In addition, the obtained surface areas still less than 10% of theoretical value indicate the presence of graphitized carbon that remained from the incomplete reaction of graphite precursor, therefore im-

FIGURE 5. The electrosorption isotherm of Na+ onto GNFs electrode at 2.0 V.

TABLE 1. Determined Parameters and Regression Coefficients r2, KL and KF of Langmuir and Freundlich Isotherms of GNFs-based CDI at 297 K isotherm

Langmuir

model equation

q)

qmKLC 1 + KLC

parameter

value

qm

73.4721

KL

0.0215

2

0.9753

r

Freundlich

q ) KFC1/n

FIGURE 4. The electrosorption of Na+ onto GNFs electrode at (a) different bias potentials, (b) different flow rates, respectively. proving the bulk conductivity of GNFs. Taking into account the importance of specific surface area to electrosorptive capacity, the GNFs-2 with the highest specific surface area among the GNFs samples was employed to fabricate the CDI electrodes. Due to the similar working principle of supercapacitor and CDI, cyclic voltammetry (CV) is often considered as an important tool to evaluate the potential of electrode materials used in supercapacitor as well as CDI and thus the electrochemical properties such as specific capacitance can be obtained. The CV analysis was depicted in SI Figure S1, showing that the GNFs electrodes have greater potential as effective electrode materials than AC in the CDI application. 3.2. Effect of Working Conditions on Electrsorptive Performance of GNFs. The key factors that affect the electrosorptive performance of Na+ on the surface of GNFs include the flow rate and bias potential. As expected, higher voltage leads to higher EC because of stronger Coulombic interaction. Figure 4(a) shows the electrosorptive performance of GNFs at different bias potentials, and the inset represents the electrosorptive capacity as a function of voltage. With the voltage increased from 0.8 to 2.0 V, the electrosorptive capacity is gradually increased from 4.44 µmol/g to 23.18 µmol/g, which is nearly a 5-fold increase. When the applied voltage is at 2.0 V, no visible gas bubbles were observed, indicating that no water electrolysis was taking place. When taking both electrosorptive capacity and energy consumption into account, the optimum working voltage for GNFs based CDI was 2.0 V. Figure 4(b) depicts the variation of conductivity in solution with flow rate, and the inset demonstrates to the correlation between the electrosorptive

KF

11.9049

n

3.4742

r2

0.9717

capacity and flow rate. It is clear from the figures that at a lower flow rate, for example, 15 mL/min, GNFs has a lower electrosorptive capacity of 9.03 µmol/g. When the flow rate was increased from 15 mL/min to 25 mL/min, the electrosorptive capacity reached the maximum value of 13.9 µmol/g. However, the electrosorptive capacity of GNFs electrode was reduced from 13.9 µmol/g to 8.49 µmol/g when the flow rate was further increased. This was due to the equilibrium between electrostatic force and the driving force in the flow rate of 25 mL/min. 3.3. Effect of Initial Concentration on Electrsorptive Performance of GNFs. To investigate the electrosorption behavior, the experiment was carried out at different initial concentrations, and thereby the electrosorption isotherm was obtained. The electrosorption of Na+ onto the GNFs electrode was evaluated at a constant temperature of 298 K for the isotherm as well as the kinetic models. The initial concentrations of the NaCl solutions were 25, 50, 100, 250, 400, and 500 mg/L, respectively, as shown in Figure 5. Langmuir eq 2 and Freundlich eq 3 isotherms were used to validate the experimental data for electrosorption of Na+ onto GNFs, respectively. q)

qmKLC 1 + KLC

(2)

q ) KFC1/n

(3)

Where C is the equilibrium concentration (mg/L), q is the amount of adsorbed Na+ (in micromoles per gram of GNFs), qm is the maximum adsorption capacity corresponding to VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8695

2 V and the flow rate of 25 mL/min, was 73.47 µmol/g, which is higher than previously reported data (16). 3.4. Electrosorption Kinetics of GNFs. Adsorption kinetics, indicating the adsorption rate, is an important characteristic of adsorbents. The adsorption rate constants of Na+ were determined by the Lagergren equation, which is often called the pseudofirst-order adsorption kinetics eq 4: log(qe - q) ) log qe -

FIGURE 6. The electrosorption kinetics of Na+ onto GNFs electrode at 2.0 V. complete monolayer coverage. Table 1 shows the determined parameters and regression coefficients r2, KL and KF of Langmuir and Freundlich isotherms for CDI, respectively. It was found that the Langmuir isotherm correlated better with the experimental data according to the r2. This phenomenon suggested that the monolayer adsorption was dominant during the electrosorption process. The parameter qm in the Langmuir isotherm model is considered as the maximum adsorption capacity, as mentioned above. Thus the equilibrium electrosorption capacity, at polarization of

kt 2.303

(4)

Where k is the adsorption rate constant (min-1), and qe and q the amount of Na+ adsorbed at equilibrium (µmol/g) and time t (min), respectively. The adsorption kinetics was fitted with the Lagergren law by nonlinear regression using the method of least-squares. The result is shown in Figure 6 and the inset displays the fitting parameters, including rate constant k and regression coefficient r2. It can be seen that the r2 is close to 1, indicating that the experimental data fit to the pseudofirst-order adsorption kinetics equation very well. Additionally, the rate constant from the figure is of 1.01 min-1, which is higher than that of reference data (16). That means the GNFs with high specific surface area better facilitate the ion diffusion. 3.5. Comparison. At the applied voltage of 2.0 V, comparative results of the electrosorptive capacity as well as BET surface area between the GNFs and AC based CDI in the same experimental conditions are shown in Figure 7(a). Although having the larger surface area (989.54 m2/g) than GNFs (222.01 m2/g), AC has an electrosorptive capacity of only 13.73 µmol/g, which is much lower than that of GNFs

FIGURE 7. (a) Comparison of electrosorptive performance by employing GNFs and AC at the same experimental condition, the pictures at top-left and top-right depict the pore size distribution of AC and GNFs below 10 nm, respectively. (b) Mechanism of CDI employing AC and GNFs electrode. TEM observation images of AC (c) and GNFs in low (d) and high (e) magnification, respectively. 8696

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

(whose electrosorptive capacity is 23.18 µmol/g). This can be attributed to the fact that GNFs have an interlayer structure that is more accessible for ions, while AC has a large fraction of inaccessible small micropores. As a result, the effective surface area of GNFs is higher than that of AC. Figure 7(b) illustrates the principle of CDI and the mechanism of how ions are electrosorbed onto AC and GNFs electrodes, respectively. TEM and SEM images of AC and CNFs are used to confirm the hypothesis. The TEM image of GNFs indicates that GNFs are aggregated together, showing a semitransparent flower shape interlayer pattern. It also shows that GNFs are homogeneous flakes with microsize that are beneficial to ions accessing and are adsorbed on the surface of the flakes. In contrast, the structure of AC on the TEM image shows that it presents a beehive-type pore structure so that the ions cannot gain access to the inner pores and therefore a high electrosorptive capacity is difficult to achieve. In addition, it is believed that the conductivity of carbon materials also plays a vital role in the electrosorptive process. Having higher bulk conductivity is equivalent to having a higher applied voltage between the two electrodes. It is reported that the conductivity of graphene prepared via chemical approaches is normally above 200 s/m (20). It is much higher than that of commercial AC whose electrical conductivity is normally between the orders of 10-6 and 10-3 sm-1. Further, several experiments regarding conductivity measurement were also performed by measuring GNFs and AC film, respectively and found that the conductivity of GNFs was much better than that of AC (see SI). Another possible reason for the high conductivity for GNFs is the presence of conductive graphitized chunks in the GNFs, which was caused by the incomplete grapheme preparation from the graphite precursor. Thus, considering both effective specific surface area and electrical conductivity, it is believed that the GNFs with high specific surface area have the potential as an excellent candidate electrode material for the CDI. In this paper, GNFs with a relatively high specific surface area of 222.01 m2/g were employed to fabricate an electrode for Na+ and Cl- removal by CDI. The electrosorption experimental results indicated that the GNFs had a better CDI performance than that of AC. It was found that the optimum working conditions of flow rate and bias potential are 25 mL/min and 2.0 V, respectively. Based on the abovementioned conditions, the specific electrosorptive amount of the GNFs is 23.18 µmol/g for sodium ions. The electrosorptive performance of GNFs is in agreement with the Langmuir adsorption isotherm and pseudofirst-order adsorption kinetics. The equilibrium electrosorption capacity and rate constant at 2.0 V were 73.47 µmol/g and 1.01 min-1 respectively. In addition, in comparison to AC and previously reported graphene under the same experimental condition, the electrosorptive performance of GNFs showed a promising potential for applications in brackish water desalination and drinking water purification.

Acknowledgments We acknowledge the financial support from ARC Linkage grant (LP0883282). H.B.L. appreciates the China Scholarship Council (CSC, File No. 2009614107) for finical support. L.K.P. acknowledges the support from Shanghai Pujiang Program (No. 08PJ14043).

Supporting Information Available Additional information on sample preparation, characterization, electrochemical performance analysis and resistivity measurement. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Zarra, M. A. Mass transfer characteristics during the removal of dissolved heavy metals from waste solutions in gas spared fluidized beds of ion exchange resins. Chem. Eng. Technol. 1999, 22 (6), 533–537. (2) Matsuura, T. Progress in membrane science and technology for seawater desalinationsA review. Desalination 2001, 134 (1-3), 47–54. (3) Gabelich, C.; Tran, T.; Melsuffet, A. H. Electrosorption of inorganic salts from aqueous solution using carbon aerogels. Environ. Sci. Technol. 2002, 36 (13), 3010–3019. (4) Farmer, J. C.; Fix, D. V.; Mack, G. V.; Pekala, R. W.; Poco, J. F. Capacitive deionization of NaCl and NaNO3 solutions with carbon aerogel electrodes. J. Electrochem. Soc. 1996, 143 (1), 159–169. (5) Zou, L. D.; Song, H. H.; Li, L. X.; Morris, G. Using mesoporous carbon electrodes for water desalination. Water Res. 2008, 42 (8-9), 2340–2348. (6) Li, H. B.; Gao, Y.; Pan, L. K.; Zhang, Y. P.; Chen, Y. W.; Sun, Z. Electrosorptive desalination by carbon nanotubes and nanofibres electrodes and ion-exchange membranes. Water Res. 2008, 42 (20), 4923–4928. (7) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y. S.; Dubonos, V.; Grigorieva, I. V.; Firsov, A. A. Electric field effect in atomically thin carbon films. Science 2004, 306 (5696), 666–669. (8) Li, D.; Mu ¨ ller, M. B.; Gilje, S.; Kaner, R. B.; Wallace, G. G. Processable aqueous dispersions of graphene nanosheets. Nat. Nanotechnol. 2008, 3 (1), 101–105. (9) Geim, A. K.; Novoselov, K. S. The rise of graphene. Nat. Mater. 2007, 6 (30), 183–191. (10) Gilje, S.; Han, S.; Wang, M. S.; Wang, K. L.; Kaner, R. B. A chemical route to graphene for device applications. Nano Lett. 2007, 7 (11), 3394–3398. (11) Lui, C. H.; Liu, L.; Mak, K. F.; Flynn, G. W.; Heinz, T. F. Ultraflat graphene. Nature 2009, 462 (11), 339–341. (12) Fan, X. B.; Peng, W. C.; Li, Y.; Li, X. Y.; Wang, S. L.; Zhang, G. L.; Zhang, F. B. Deoxygenation of exfoliated graphite oxide under alkaline conditions: a green route to graphene preparation. Adv. Mater. 2008, 20 (23), 4490–4493. (13) Choucair, M.; Thordarson, P.; Stride, J. A. Gram-scale production of graphene based on solvothermal synthesis and sonication. Nat. Nanotechnol. 2009, 4 (12), 30–33. (14) Hao, R.; Qian, W.; Zhang, L. H.; Hou, Y. L. Aqueous dispersions of TCNQ-stabilized graphene sheets. Chem. Commun. 2008, 48, 6576–6578. (15) Stoller, M. D.; Park, S. J.; Zhu, Y. W.; An, J. B.; Ruoff, R. S. Graphene-Based Ultracapacitors. Nano Lett. 2008, 8 (10), 3498– 3502. (16) Li, X. L.; Wang, X. R.; Zhang, L.; Lee, S. W.; Dai, H. J. Chemically derived ultrasmooth graphene nanoribbon semiconductors. Science 2008, 319 (5867), 1229–1232. (17) Li, H. B.; Lu, T.; Pan, L. K.; Zhang, Y. P.; Sun, Z. Electrosorption of behaviour of graphene in NaCl solutions. J. Mater. Chem. 2009, 19 (37), 6773–6779. (18) William, S.; Hummers, J.; Offeman, R. E. Preparation of graphitic oxide. J. Am. Chem. Soc. 1958, 80 (6), 1339–1341. (19) Dreyer, D. R.; Park, S. G.; Bielawski, C. W.; Ruoff, R. S. The chemistry of graphene oxide. Chem. Soc. Rev. 2010, 39 (1), 228– 240. (20) Park, S. J.; Ruoff, R. S. Chemical methods for the production of graphene. Nat. Nanotechnol. 2009, 4 (3), 217–224.

ES101888J

VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8697