Scalable production of graphene via wet chemistry: progress and challenges

Materials Today  Volume 00, Number 00  September 2014

RESEARCH: Review

RESEARCH

Scalable production of graphene via wet chemistry: progress and challenges Yu Lin Zhong*, Zhiming Tian, George P. Simon and Dan Li* Department of Materials Engineering, Monash University, Clayton, VIC 3800, Australia

Although enormous scientific progress has been made in the application of graphene and its related materials, the cost-effective and scalable production of graphene still holds the key to its commercialization. If this aspect cannot be successfully addressed, it may eventually struggle for widespread use, such as has occurred for its allotrope, the carbon nanotubes. Ease of graphene production is especially important if it is to be used in bulk applications such as energy storage in automobiles where the large scale and low cost production of the active materials is required. Fortunately, graphene can be produced not only from a cheap and abundant source (graphite), but also can be produced using a variety of low cost methods. This focus review article will examine three promising, scalable methods of graphene production, namely the graphite oxide, liquid-phase exfoliation (LPE) and electrochemical routes, with focus on their recent progress and remaining challenges. The perspective on these routes will be mainly taken from the industrial viewpoint, thus highlighting the pressing issues for graphene commercialization. Some of the main concerns regarding the quality or crystallinity of the graphene sheet produced from such methods and the importance of a comprehensive evaluation of the final bulk graphene materials will also be discussed.

Introduction The most crucial factor for any material to be commercially viable for industrial scale application is its cost-effectiveness. Currently, platinum may exhibit the highest performance for electrocatalysis and tin-doped indium oxide may display the highest conductivity and transparency, but their low abundance and costs will always fuel the search for a cheaper alternative [1]. In comparison, the large available amount of carbon has ensured its continued usage because the dawn of mankind and has exponentially increased with the discovery of the nanocarbon allotropes. This is especially true for two-dimensional graphene which not only exhibits incredible electronics properties [2] but also many outstanding mechanical properties which are highly sought after by a broad spectrum of applications ranging from ultrafast electronics to mechanical actuators [3,4]. More importantly, there are currently numerous ways to produce graphene (Fig. 1) [5] such that one can *Corresponding authors:. Zhong, Y.L. ([email protected]), Li, D. ([email protected])

always identify a well-suited graphene production method for each and every application [6]. For example, the chemical vapour deposition (CVD) [7–11] and epitaxial growth [12–15] methods are ideal for the fabrication of flexible transparent electrodes and graphene electronics that capitalize on the transparency and high electron mobility of pristine single layer graphene, respectively. However, these production methods might face difficulties in growing or forming graphene materials with porous network structures for increased permeability and bulk surface area. On the other end of the spectrum, there are potentially scalable and low cost methods available to mass produce graphene for applications such as conductive ink or energy storage. However, it is especially crucial that in these applications, graphene must not only outperform, but also cost less than existing materials such as carbon black and activated carbon, to motivate the industries to adopt the new material. The key for graphene commercialization lies in its production method which essentially determines its cost-effectiveness.

1369-7021/ß 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.mattod.2014.08.019

1 Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019

MATTOD-387; No of Pages 6 RESEARCH

Materials Today  Volume 00, Number 00  September 2014

RESEARCH: Review FIGURE 1

Schematic illustration of the main graphene production techniques. (a) Micromechanical cleavage. (b) Anodic bonding. (c) Photoexfoliation. (d) Liquid-phase exfoliation. (e) Growth on SiC. Gold and grey spheres represent Si and C atoms, respectively. At elevated T, Si atoms evaporate (arrows), leaving a carbonrich surface that forms graphene sheets. (f ) Segregation/precipitation from carbon containing metal substrate. (g) Chemical vapour deposition. (h) Molecular beam epitaxy. (i) Chemical synthesis using benzene as building block. Reproduced from Ref. [5] with permission from the Elsevier.

In terms of mass production of graphene, the main factors are the production cost, scalability, reproducibility, processability and the quality of the graphene products. Considering the low cost and abundance of graphite flakes, the wet chemical approaches in exfoliation of graphite to graphene seem to fit all the requirements, except that there is question on the quality of graphene. However, we should take note that the definition on the quality of graphene or rather the efficacy of graphene is highly dependent on its application. For electrocatalysis and storage of capacitive charges, it has been shown that the graphene edges are more superior compared to the graphene basal planes [16] and this is probably due to the localization of HOMO and LUMO levels at the graphene edges [17]. For application as carbocatalysts in the oxidative coupling of organic molecules, it is the edge defects in porous graphene that are responsible for high catalytic activities [18]. Hence, keeping those key factors in mind, this focus review will examine three promising wet chemical graphite exfoliation routes (graphite oxide, liquid-phase exfoliation (LPE) and electrochemical) and highlighting their recent progress and challenges.

Graphite oxide route The graphite oxide route is currently the most popular wet chemical method to produce graphene materials, namely graphene oxide (GO) [19] and reduced graphene oxide (rGO) [20], based on the enormous amount of research activities employing it. This is due to its potential scalability, high yield and most importantly, its excellent dispersibility in various solvents [21–25] which facilitates processability towards many applications. The main exfoliation mechanism lies in the oxidative intercalation and production of oxygen-containing functional groups on the graphene layers which helps the dispersion and stabilization of GO sheet in water. It is also due to the oxygen-containing functional groups that extended coupling and functionalization chemistry can be applied and helped to propel the use of graphite oxide route. The formation of graphite oxide can be traced back to 1859 when Brodie oxidized graphite in the presence of potassium chlorate and fuming nitric acid [26] and it slowly evolved to the widely used Hummers method that uses a combination of sodium nitrate, potassium permanganate and sulfuric acid [27]. Over the years, efforts have been made to improve on the Hummers method by eliminating the use of sodium

2 Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019

MATTOD-387; No of Pages 6 RESEARCH

conductivity by the production of large area GO flakes [44] and the use of milder oxidation conditions by simply reducing the concentration of oxidants [45] Although Eigler and co-workers recently showed the production of GO with larger region of pristine network via low temperature oxidation [46], these methods are still based on the Hummers method which will inevitably give rise to a variety of oxidized functional groups. In particular, the carboxylic acid groups generated on the graphene edges are notoriously difficult to reduce [47]. There are also distinct drawbacks in producing mildly oxidized large GO flakes which are the marked decrease in production yield and their poorer dispersibility in water. Hence, there is a need to seek a fundamental change in the oxidative exfoliation mechanism and the Holy Grail in the graphite oxide route is to develop a novel controlled oxidation mechanism in which the specificity and density of the oxygencontaining functional groups could be well controlled.

LPE route

SCHEME 1

Schematics of conversion of bulk graphite into GO with corresponding micrographic images or sample appearances at each phase. The three steps signify formation of the two intermediate products (stage-1 GIC and PGO) and the final GO product. The solid black lines represent graphene layers; dotted black lines represent single layers of GO; wide blue lines represent H2SO4/HSO4 intercalant; wide purple lines represent a layer of the mixture of H2SO4/HSO4 intercalant with the reduced form of oxidizing agent. Reproduced from Ref. [29] with permission from the American Chemical Society. (Boxed) Formation of dimanganese heptoxide (Mn2O7) from KMnO4 in the presence of strong acid. Reproduced from Ref. [32] with permission from the Royal Society of Chemistry.

nitrate, thus excluding the production of toxic nitrous gas [28]. More significantly, the addition of phosphoric acid led to a higher yield in GO with higher degree of oxidation. The mechanism of graphite oxide formation was recently defined by Dimiev and Tour in three distinct, independent steps as shown in Scheme 1 [29]. The focus in this area has been on trying to achieve a greener method [30] and a recent paper reported a safer (and thus more scalable) improved Hummers method by using expanded graphite as the starting source [31]. Nevertheless, the active oxidizing species is dimanganese heptoxide, which is formed when potassium permanganate reacts with sulfuric acid (Scheme 1, boxed [32]), and it is known to detonate when heated to temperatures greater than 558C or when in contact with organic compounds [33]. Hence, the control of temperature in such a high risk reaction is crucially important, and even more so at a large industrial scale. There are other concerns associated with GO such as its chemical inhomogeneity, batch-to-batch reproducibility and the inevitable generation of irreparable hole defects on the graphene sheets during oxidation which will set a limit to the conductivity of rGO [34]. The reduction of GO is a huge field of study [35] by itself and it can range from simple thermal reduction [36–38] to multi-step reduction process [39]. In addition, the different methods of graphite oxide synthesis [40–42] and even the washing steps after synthesis [43] can have significant effect on graphene properties. There have been attempts to achieve rGO thin films with higher

In contrast to the oxidative mechanism of GO route, the LPE route seeks to directly disperse graphite flake in a suitable solvent via mechanical, solvothermal or sonochemical assistance thereby retaining the crystallinity of the graphene sheet. The simplicity of the LPE route guarantees its scalability as demonstrated from the recent 100 l production scale (max. concentration <0.1 mg/ml) [48]. The choice of solvent is crucial, as the right solubilization parameters determine the effectiveness of graphite exfoliation and determine how well the graphene sheet can be thermodynamically stabilized [49]. Although there were attempts to facilitate the exfoliation process by using expanded graphite [50–53] or graphite intercalation compounds [54–56], the bottleneck in the LPE route lies in the very limited dispersibility of the pristine graphene which behaves like an intractable macromolecule. Moreover, the most effective dispersing solvent for pristine graphene has either high boiling point (NMP) [57] or is too corrosive and atmospherically unstable (chlorosulphonic acid) [25] which set a roadblock for further processing. Subsequently, there have been strategies targeted to increase the concentration of graphene solution by prolonged sonication in solvent [58] and to disperse pristine graphene in low boiling point solvents [59] and even in water with the aid of surfactants [60–63] or polymeric agents [64– 67] (Fig. 2). With prolonged sonication time, reduction in the size of graphene flakes was observed and this limitation contradicts the original intention of producing pristine graphene sheets. The use of surfactants raises issues of added cost if the process is to be scaled up, as only the specialized surfactants (disc-like, planar) are effective in delivering a colloidal solution with decent graphene concentration [68] and most surfactants are typically insulating in nature, which necessitates the need for an additional washing step after film forming or device processing. Because of the inherently poor dispersibility of pristine graphene, the main role of the LPE route should be distinctly different from the GO route in which the latter can generate the highly processable GO for many different applications. The main advantages of LPE route are its simplicity and amenability to suit specialized applications, and it is these advantages should be capitalized on. The ideal situation would be to exfoliate graphite in suitable surfactants/ polymeric system with high yield and directly employ the exfoliated graphene nanocomposites in applications by means of 3D printing, 3

Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019

RESEARCH: Review

Materials Today  Volume 00, Number 00  September 2014

MATTOD-387; No of Pages 6 RESEARCH

Materials Today  Volume 00, Number 00  September 2014

RESEARCH: Review FIGURE 2

Schematic representation of the liquid-phase exfoliation process of graphite in the absence (top-right) and presence (bottom-right) of surfactant molecules. Reproduced from Ref. [57] with permission from the Royal Chemical Society. FIGURE 3

doctor-blading or other delivery methods. In such an ideal case, the exfoliating system would constitute the same composition required in the final graphene nanocomposites, hence making the LPE route extremely simple and cost-effective.

Electrochemical route Glassy carbon and graphite electrodes are conventionally used and well-studied electrodes for electrochemistry and their surfaces can be subjected to various electrochemical modifications [69,70]. Moreover, the availability of large-scale industrial electrolyzers presents a platform for scaling up electrochemical reactions. However, it was not until the recent years that electrochemical methods were seriously employed to produce graphene materials. In comparison to GO and LPE route, the electrochemical route utilizes intercalation of ions and electrochemical initiated reaction with the electrolyte for the exfoliation mechanism. From a broader point of view, it offers a whole branch of electrochemistry to be applied on the functionalization and exfoliation of the graphite electrodes. Depending on the applied voltage (Fig. 3) [71] and the electrolyte (e.g. aqueous inorganic salt solution [72], sulfuric acid [73], lithium perchlorate in propylene carbonate [74], ionic liquid [75,76], among others) employed, various types of graphene materials could be produced ranging from oxidized nanographite [77], carbon nanoribbons and nanoparticles [78] to low-defect graphene [73]. Out of the rapidly increasing reports on electrochemical exfoliation of graphite, the work by Li and co-worker seems promising as they were able to produce relatively low-defect graphene from fast electrochemical exfoliation of graphite in low cost sulfuric acid electrolyte [79]. By contrast, by applying non-oxidative cathodic bias to the graphite electrode followed by in situ electrochemical functionalization, highly electrochemically functionalized graphene were obtained which showed the capability to be transformed back to near pristine graphene via laser ablation [80].

Electrochemical approaches (a) oxidation, intercalation and exfoliation and (b) reduction, intercalation and exfoliation to produce single and multi-layer GN flakes. Reproduced from Ref. [71] with permission from the Elsevier.

The major limitation on the electrochemical route is the need to provide an unbroken voltage bias to the graphite flakes and most of the reported methods employed a single, continuous piece of graphite electrode in the form of graphite rod, graphite foil or highly oriented pyrolytic graphite slab. In the ideal case, the electrochemical exfoliation should only occur on the most outer graphite flakes thus electrochemically peeling off the graphene layers, individually. In reality, the ions intercalation and exfoliation will occur simultaneously at all graphitic edges in contact with the electrolyte hence this will typically result in the production of few-layered graphene or even multi-layered graphite chunks. Furthermore, this irregular exfoliation effect will be even more pronounced during the scaling up of the graphite electrode due to higher area of exposed graphitic edges. Hence, there is a fundamental need to re-engineer the electrochemical setup so as to allow effectively application of the electrochemical driving force to the graphite materials. It is only after solving this limitation can the rich field of electrochemistry be applied gainfully in this way.

Evaluation of bulk graphene materials From the standpoint of industries that require bulk quantity of graphene materials, it is most desirable that the graphene materials be supplied and processed in the form of powder, foam, film or high concentration solution [81–84]. Because of the large interfacial van der Waals force of pristine graphene, this poses a large challenge as the two-dimensional graphene has high tendency to restack during processing and assembly [85]. As a consequence, the excellent properties of individual graphene sheets cannot be translated to bulk graphene assemblies. This can be seen from the fact that the obtainable surface area of functionalized graphene sheets from

4 Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019

MATTOD-387; No of Pages 6

experiments is around 700 m2/g, which is far from its high theoretical specific surface area of 2630 m2/g [86]. As such, effective characterization techniques on both the individual graphene sheets and bulk graphene assemblies are required to realize the full potential of graphene-based material. The basic characterization techniques for individual graphene sheets are well-established and they provide crucial information on the graphene lateral size (optical microscopy [21] and secondary electron microscopy [87]), number of layers (transmission electron microscopy [88], atomic force microscopy [89] and Raman spectroscopy [90]), level of defects (Raman spectroscopy [91]) and identification as well as quantification of functional groups (combustion elemental analysis [92], infra-red spectroscopy [93] and X-ray photoelectron spectroscopy [94]). To obtain information of bulk graphene assemblies such as their accessible surface areas and the interlayer interactions, applications of some of the existing bulk analysis techniques may not be so straightforward. For example, surface area measurement via BET nitrogen gas adsorption is useful for traditional porous activated carbons as their porous structures are rigid. The same cannot be applied for chemically converted graphene (CCG) hydrogels where the CCG layers are sandwiched and separated by electrolyte and water molecules. The need for dried samples for BET measurement precludes its use on graphene hydrogels as the drying process causes the collapse of the CCG layers, yielding meaningless results. Even with a wet analysis technique based on methylene blue adsorption, the limitation will set when the interlayer spacing of the CCG hydrogel is smaller than the size of methylene blue [82,95]. Likewise, X-ray diffraction (XRD) is useful for looking at the interlayer spacing of graphite oxide or graphite intercalation compounds where their interlayer spacings are very well-defined, but for multi-layered CCG films, the large variation in the corrugation of the CCG layers often results in inconclusive XRD patterns [82,96] (Fig. 4d). Hence, taking CCG as just one example, it is a challenging task to characterize and to give a comprehensive evaluation of the bulk graphene assemblies. Intuitively, the yardstick to evaluate of a bulk material depends on its application and retrospectively, its performance parameters in the targeted application can be used to relate to the properties of individual graphene sheets or the methods in bulk assembly/ processing. In the application as supercapacitors, their capacitive performances depend largely on the electrochemical properties of the materials in which there are strong structure–property relationships [97]. On the basis of this concept, Li and co-workers proposed the use of dynamic electrosorption analysis (DEA) by measuring the material’s capacitive performance at various charging and discharging rate and/or frequency response behaviour to evaluate the corrugation levels of CCG films (Fig. 4) [96]. It was demonstrated that the capacitive results correlated to the structure–property relationship of the CCG films in which the CCG films that had undergone reduction at higher temperature, showed marked decrease in charge storage capacity at high charging/discharging rate, even though the XRD pattern looked very similar (Fig. 4). It implied that the CCG films had suffered a higher degree of collapsed pores which caused an increased ionic resistance. Furthermore, it was suggested that DEA could play a complementary role in providing qualitative analysis to the existing characterization techniques of porous carbon [98]. It is an attractive

RESEARCH

RESEARCH: Review

Materials Today  Volume 00, Number 00  September 2014

FIGURE 4

Multi-layered CCG films as an experimental platform for electrosorption study. (a) Structural tuning of CCG films through control of the CCG corrugation level. (b) Image of a synthesized CCG films. (c) SEM image of the cross-section of a CCG film. The scale bar is 1 mm. (d) XRD patterns of different CCG films. (e) Gravimetric capacitances of different CCG films measured at various current densities. Reproduced from Ref. [96] with permission from the Royal Chemical Society.

feature when a technique such as this both can be used to probe microstructure, as well as give a property that is useful in the application of these materials.

Concluding remarks Considering the main factors for industrialization of graphene production (namely the production cost, scalability, reproducibility, processability and the performance of the graphene products), the current promising methods of production (GO, LPE and electrochemical route) each have its own advantages and limitations, as summarized in Table 1. Moreover, each of these production methods has its niche area of applications and hence it is difficult to achieve a one-method-suits-all graphene production TABLE 1

Summary of advantages and challenges associated with each graphene production route. Production routes

Advantages

Challenges

Graphite oxide

High yield Down to single layer High dispersibility

Potentially explosive process Structural inhomogeneity Batch-to-batch reproducibility

Liquid-phase exfoliation

Simple Non-oxidative Potentially scalable

Low graphene yield Low dispersibility Low processability

Electrochemical

Cost-effectiveness High yield Rich field of electrochemistry

Electrical connectivity to graphite materials

5 Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019

MATTOD-387; No of Pages 6 RESEARCH

method unless there is a fundamental breakthrough in its limitations. As a rule of thumb, the ideal production method should be simple and effective for low production cost, scalability and reproducibility. For processability and performance, the production method should ideally create versatile functional groups that aid the graphene dispersibility and the conjugation of functional materials, or be amenable to post-production modifications. There is also the need to find a reliable benchmark for the evaluation of graphene bulk materials supplied from all over the world. RESEARCH: Review

Acknowledgements The authors acknowledge the funding support from BaosteelAustralia Joint Research and Development Centre on the project (BA11006). Y.L.Z. and D.L. acknowledge the financial support from the Australian Research Council. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]

A. Kumar, C. Zhou, ACS Nano 4 (2010) 11. A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183. M. Liu, et al. Nature 474 (2011) 64. J. Liang, et al. J. Phys. Chem. C 113 (2009) 9921. F. Bonaccorso, et al. Mater. Today 15 (2012) 564. K.S. Sivudu, Y.R. Mahajan, Nanotech Insights 6 (2012) 3. Q. Yu, et al. Appl. Phys. Lett. 93 (2008) 11. S. Bae, et al. Nat. Nanotechnol. 5 (2010) 574. X. Li, et al. Science 324 (2009) 1312. K.S. Kim, et al. Nature 457 (2009) 706. L.G. De Arco, et al. IEEE Trans. Nanotechnol. 8 (2009) 135. K.V. Emtsev, et al. Nat. Mater. 8 (2009) 203. P.W. Sutter, et al. Nat. Mater. 7 (2008) 406. T. Ohta, et al. Science 313 (2006) 951. C. Berger, et al. Science 312 (2006) 1191. W. Yuan, et al. Sci. Rep. 3 (2013) 2248. D.A.C. Brownson, et al. RSC Adv. 1 (2011) 978. C. Su, et al. Nat. Commun. 3 (2012) 1298. S. Stankovich, et al. Nature 442 (2006) 282. S. Stankovich, et al. Carbon 45 (2007) 1558. V.C. Tung, et al. Nat. Nanotechnol. 4 (2009) 25. E.M. Milner, et al. J. Am. Chem. Soc. 134 (2012) 8302. D. Li, et al. Nat. Nanotechnol. 3 (2008) 101. Y. Liang, et al. Adv. Mater. 21 (2009) 1679. N. Behabtu, et al. Nat. Nanotechnol. 5 (2010) 406. B.C. Brodie, Trans. R. Soc. Lond. 149 (1859) 249. W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339. D.C. Marcano, et al. ACS Nano 4 (2010) 4806. A.M. Dimiev, J.M. Tour, ACS Nano 8 (2014) 3060. J. Chen, et al. Carbon 64 (2013) 225. L. Sun, B. Fugetsu, Mater. Lett. 109 (2013) 207. D.R. Dreyer, et al. Chem. Soc. Rev. 39 (2010) 228. K.R. Koch, P.F. Krause, J. Chem. Ed. 59 (1982) 973. S. Eigler, et al. Carbon 50 (2012) 3666. S. Pei, H.-M. Cheng, Carbon 50 (2012) 3210. W. Lv, et al. ACS Nano 3 (2009) 3730. I. Jung, et al. Nano Lett. 8 (2008) 4283. C. Zhang, et al. Carbon 62 (2013) 11.

Materials Today  Volume 00, Number 00  September 2014

[39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98]

W. Gao, et al. Nat. Chem. 1 (2009) 403. H.L. Poh, et al. Nanoscale 4 (2012) 3515. L. Buglione, et al. Electrochem. Commun. 14 (2012) 5. C.K. Chua, et al. Chem. Eur. J. 18 (2012) 13453. A. Dimiev, et al. J. Am. Chem. Soc. 134 (2012) 2815. J. Zhao, et al. ACS Nano 4 (2010) 5245. Y. Xu, et al. J. Mater. Chem. 21 (2011) 7376. S. Eigler, et al. Adv. Mater. 25 (2013) 3583. Y. Zhou, et al. Chem. Mater. 21 (2009) 2950. K.R. Paton, et al. Nat. Mater. 13 (2014) 624. Y. Hernandez, et al. Nat. Nanotechnol. 3 (2008) 563. W. Qian, et al. Nano Res. 2 (2009) 706. R. Hao, et al. Chem. Commun. 48 (2008) 6576. L. Zhu, et al. Mater. Chem. Phys. 137 (2013) 984. G.S. Bang, et al. J. Mater. Chem. 22 (2012) 4806. J.H. Lee, et al. Adv. Mater. 21 (2009) 4383. W. Gu, et al. J. Mater. Chem. 19 (2009) 3367. L.M. Viculis, et al. J. Mater. Chem. 15 (2005) 974. A. Ciesielski, P. Samori, Chem. Soc. Rev. 43 (2014) 381. U. Khan, et al. Small 6 (2010) 864. A. O’Neill, et al. J. Phys. Chem. C 115 (2011) 5422. M. Lotya, et al. ACS Nano 4 (2010) 3155. L. Guardia, et al. Carbon 49 (2011) 1653. S.M. Notley, Langmuir 28 (2012) 14110. M. Lotya, et al. J. Am. Chem. Soc. 131 (2009) 3611. A.B. Bourlinos, et al. Solid State Commun. 149 (2009) 2172. Y.T. Liang, M.C. Hersam, J. Am. Chem. Soc. 132 (2010) 17661. L. Xu, et al. J. Phys. Chem. C 117 (2013) 10730. P. May, et al. J. Phys. Chem. C 116 (2012) 11393. D. Parviz, et al. ACS Nano 6 (2012) 8857. R.S. Deinhammer, et al. Langmuir 10 (1994) 1306. C.E. Banks, et al. Chem. Commun. 7 (2005) 829. C.T.J. Low, et al. Carbon 54 (2013) 1. K. Parvez, et al. J. Am. Chem. Soc. 136 (2014) 6083. K. Parvez, et al. ACS Nano 7 (2013) 3598. J. Wang, et al. J. Am. Chem. Soc. 133 (2011) 8888. N. Liu, et al. Adv. Funct. Mater. 18 (2008) 1518. M. Mao, et al. Chem. Commun. 49 (2013) 5301. Y.-F. Li, et al. J. Chem. Phys. 139 (2013) 064703. J. Lu, et al. ACS Nano 3 (2009) 2367. C.-Y. Su, et al. ACS Nano 5 (2011) 2332. Y.L. Zhong, T.M. Swager, J. Am. Chem. Soc. 134 (2012) 17896. L. Qiu, et al. Nat. Commun. 3 (2012) 1241. X. Yang, et al. Science 341 (2013) 534. M. Segal, Nat. Nanotechnol. 4 (2009) 612. X. Yang, et al. Adv. Mater. 23 (2011) 2833. C. Cheng, D. Li, Adv. Mater. 25 (2013) 13. M.J. McAllister, et al. Chem. Mater. 19 (2007) 4396. S. Watcharotone, et al. Nano Lett. 7 (2007) 1888. A. Hashimoto, et al. Nature 430 (2004) 870. C.H. Lui, et al. Nature 462 (2009) 339. A.C. Ferrari, et al. Phys. Rev. Lett. 97 (2006) 187401. M.A. Pimenta, et al. PCCP 9 (2007) 1276. D.R. Dreyer, et al. J. Mater. Chem. 21 (2011) 3443. Y. Si, E.T. Samulski, Nano Lett. 8 (2008) 1679. D. Yang, et al. Carbon 47 (2009) 145. J.J. Kipling, R.B. Wilson, J. Appl. Chem. 10 (1960) 109. J. Zhu, et al. Chem. Eur. J. 19 (2013) 3082. A. Ambrosi, et al. Chem. Eur. J. 17 (2011) 10763. C. Cheng, et al. J. Mater. Chem. A 1 (2013) 9332.

6 Please cite this article in press as: Y.L. Zhong, et al., Mater. Today (2014), http://dx.doi.org/10.1016/j.mattod.2014.08.019