About the Author
Read an Excerpt
By Vikas Mittal
The Royal Society of ChemistryCopyright © 2012 The Royal Society of Chemistry
All rights reserved.
Graphene Functionalization: A Review
MO SONG AND DONGYU CAI
Department of Materials, Loughborough University, Loughborough LE11 3TU, United Kingdom
Graphene is a carbon sheet one atom thick, consisting of a two-dimensional honeycomb lattice, which is considered to be the thinnest material in the world. It is generally viewed as the basic unit for other carbon materials. Figure 1.1 shows that graphene can become 3D graphite, ID carbon nanotube (CNT) or OD fullerenes by the operations of stacking, rolling and wrapping, respectively. Before the discovery of graphene, there was a long-standing argument that strictly 2D crystals are thermodynamically unstable and should not exist. In 2004, this theory was undermined by two physicists (Geim and Novoselov) from Manchester University, who successfully isolated freestanding graphene from 3D graphite using Scotch tape. However, the existence of graphene can compromise with the theory to some extent. One point of view is that strong sp2 C-C bonds can resist the dislocation of a carbon atom or other crystal defects under thermal fluctuation even at elevated temperature although graphene is a stable material. Another interesting observation was that suspended graphene sheets were not perfectly flat, and out-of-plane deformation in the surface reached 1 nm. The roughening surfaces, with microscopic corrugations in the third dimension, provided the possibility of thermodynamic stability for 2D graphene. This new 2D model has prompted physicists to discover many exceptional physical phenomena which are not found in other materials; we can only present a brief introduction here.
Graphene has some exceptional physical properties. It is the strongest material ever measured, with a Young's modulus of 1 TPa and tensile stress of 130 GPa, 100 times that of steel. Its thermal conductivity of graphene is high, up to 5000 W m-1 K-1, and its theoretical specific surface area is close to 2630 m2/g. Its excellent electrical properties are the major reason why graphene has diverted attention from other 2D materials; ambipolar electric properties and quantum Hall effects have been demonstrated in graphene. The carrier mobility historically hits a high value of 200 000cm2/VS as the electrons transporting across graphene behave like massless relativistic particles. Graphene has been described as a wonder material and may become an exciting multidisciplinary platform attracting physicists, chemists and engineers together to revolutionize the technologies currently used in our society.
Although people have started to dream of seeing graphene-based products, concern remains about future application of graphene, e.g. the possibility of low cost and mass production and the assembly of graphene into complex systems with desirable properties. The most difficult handling problem is that suspended graphene, although it can exist, is unstable and tends to stick together. The functionalization of graphene is considered to be the main route to make suspended graphene stable in a complex environment by introducing a third party into the graphene surface via chemical or physical approaches. We believe that it will play a key role when moving graphene from the laboratory to real-world applications. Progress on functionalizing graphene is making rapid progress, helped by quick learning from experience with CNTs, as both these carbon materials have a similar chemical structure. This chapter attempts to review up-to-date progress in this field and hopefully will benefit those readers who are either already involved with research into graphene or are eager to become involved.
1.2 Fabrication of Graphene
1.2.1 Mechanical Cleavage
The 'Scotch tape' method, as it is sometimes humorously described in the media, is key to the world of graphene. Geim and his coworkersl used oxygen plasma to etch square mesas (5 µm deep and 20 µm-2 mm width) on top of highly oriented pyrolytic graphite (HOPG), and then pressed the treated HOPG against a layer of wet photoresist spun on the [n.sup.+]-doped Si substrate topped with a SiO2 layer. With good adhesion achieved by baking, Scotch tape was used to repeat the action of peeling the mesas to remove thick graphite flakes. Afterwards, the substrate with thin flakes was dipped in acetone and then washed in plenty of water and propanol with the assistance of ultrasound. As a result, the thin flakes (<10 nm thick) attached strongly to the SiO2 layer owing to Van der Waals and capillary forces. Finally, graphene multilayers (<1.5nm) were selectively detected under atom force microscopy. Later, Geim's group reported another mechanical approach, described as 'rubbing' to successfully extract single-layer graphene from graphite in one step. The of rubbing process was simply described as being similar to drawing with chalk on a blackboard.
Awareness of the problem of low yields, Coleman's group from Dublin attempted to improve the efficiency of mechanical exfoliation in the liquid phase. It was demonstrated by Hernandez et al. that graphite could be split into graphene in a well-selected organic solvent by low-power ultrasound, and the concentration of graphene dispersion could reach 0.01 mg/ml with a yield of 1wt%. Repeating the process for recycled graphite sediments potentially allowed improvement of the yield up to 12wt%. The stabilization of graphene was dependent on the surface energy of organic solvents according to the following equation:
[MATHEMATICAL EXPRESSION OMITTED] (1)
where [MATHEMATICAL EXPRESSION OMITTED], is the square root of the surface energy of phase i. The surface energy of graphite is defined as the energy per unit area required to peel two graphene sheets apart. Tflake is the thickness of a graphene flake and φ is the graphene volume fraction. The exfoliation takes place when enthalpy of mixing (ΔHmix) is very small, which ensures the negative free energy of mixing (ΔGmix) according to classical thermodynamic theory:
ΔGmix = ΔHmix - TΔSmix (2)
where ΔSmix is the entropy of mixing per unit volume. These equations tell us that solvents with surface energy matching that of graphene, such as N- methylpyrrolidone, are the best candidates for stabilizing graphene sheets sliced down by ultrasound, ensuring that the graphene-graphene bonding force is balanced by solvent-graphene interaction. Hernandez et al. also measured the dispersibility of graphene in 40 solvents and identified the criteria for selecting good solvents to disperse graphene according to basic solubility theory rather than just making rough judgements using surface energy. They found that suitable solvents should have a Hildebrand solubility parameter ~23 MPa1/2, and Hansen solubility parameters ~18 MPa1/2, ~9.3 MPa1/2 and ~9.3 MPa1/2 for dispersive, polar and hydrogen-bonding components, respectively.
Following this discovery, Coleman's group carried out a series of experiments with the aim of improving the yield. Khan et al. reported that the concentration of graphene solution could be increased to 1.2 mg/ml with up to 4wt% monolayers after treating a graphite/NMP mixture continuously with low-power sonication for ~460 h. The reduction in mean length (L) and width (W) of graphene flake by sonication (L [varies] t-1/2 and W [varies]t-1/2) was the reason for the solvent accommodating more graphene. The length could only be reduced to 1 µm at maximum. A quantitative relation as shown in equation (3) was proposed to relate the concentration (CG) to sonication time (t) as the concentration was defined by dividing mass of average flake by average solvent volume per flake viewed as a sphere with diameter equal to the flake length:
[MATHEMATICAL EXPRESSION OMITTED] (3)
where ρG is graphitic density and τ is flake thickness. This process was significantly refined to achieve concentrations as high as 63 mg/ml, mainly by redispersing a powder of exfoliated graphene, a few layers thick, formed by filtering the graphene dispersion produced by ultrasonication. Surfactant was used to enhance the yield of graphene in aqueous solutions, a technique adapted from work with CNTs. The dispersion of graphene in sodium cholate aqueous solution was reported to have a concentration of ~0.3 mg/ml. The structure of surfactant basically consists of a long tail and an ionic head. Its function here is to charge the surface of graphene flake with ions and generate Coulomb repulsion to prevent the bonding of graphene. The long tail physically attached to graphene creates steric hindrance to prevent aggregation.
It was also reported that a superacid such as chlorosulphonic acid could exfoliate graphite spontaneously without any chemical or physical treatments to increase graphene concentration up to ~2 mg/ml. The mechanism for stabilizing graphene was due to the repulsion between graphene layers induced by acid protonation. Acid strength was key to dispersion quality. The concerns in liquid phase exfoliation are not just the issue of concentration, but include the toxicity and high cost of these organic liquids. Water is an environmentally friendly liquid, but the factors of its wetting ability, surfactants and incompatibility with organic systems will be a bar for future applications. Mechanical exfoliation is simple, but has evident limitations in mass production of graphene for commercial use. This has become a huge challenge for the future of graphene, and it has stimulated massive efforts from chemists to explore a variety of chemical routes to meet the demand for high yields of graphene.
1.2.2 Reduction of Graphene Oxide
22.214.171.124 Chemical Reduction
Apparently, scalability is the catastrophic disadvantage of mechanical exfoliation. Chemists believed that an alternative tool to make graphene in large volume would come from the discovery by Boehm et al. in 1962, who found that extremely thin lamellae with a few carbon layers could be produced by reducing a dispersion of graphite oxide (GO) by chemical or thermal means. Chemical reduction took place in the dispersion of GO in dilute alkali with hydrazine, hydrogen sulfide or iron (II) salts, but chemical reduction was not able to repair the damaged carbon structure completely. Here, it is essential to discuss oxidization of graphite using strong chemical oxidants as it led researchers so close to the perfect graphene monolayer nearly half a century ago. The oxidation of graphite should date back to 1859 when B. C. Brodie used the oxidants of potassium chlorate and nitric acid. Nearly 40 years later, L. Staudenmaier improved Bordie's method by changing single feeding of potassium chlorite to a multiple feeding method, and also adding concentrated sulfuric acid to increase the acidity. Nowadays, the most popular method is the one invented by Hummers and Offeman (Hummers method), in which potassium permanganate and concentrated sulfuric acid are used. The defects in the carbon layers of graphite are commonly viewed as the sites for the formation of functional groups after breakdown of π-conjugated structure by oxidants. It is difficult to discern the mechanism of oxidization precisely, because of the intrinsic complexity of graphite defects. For the same reason, it is very challenging to identify the chemical structure of GO as a key to understanding the mechanism of reduction. Figure 1.2 presents a summary of several models so far proposed for the structure of GO. The evolution of these models has taken more than half a century since 1939 when Hofmann and Hoist proposed the first model that consisted of epoxy groups across a planar carbon structure. In 1946, Ruess modified this model by introducing a hydroxyl group to underpin the observation of hydrogen atoms in GO. The hydroxyl groups repeatedly stood in the fourth position of the aromatic ring and epoxy groups occupied the first and third positions. After another two decades, Scholz and Boehm created a new model that introduced quinoidal species instead of epoxy and ether groups into the cleaved sites of the aromatic backbone. Szabó et al. attempted to combine the Ruess and Scholz-Boehm models. Furthermore, the formation of phenolic groups was used to explain the acidity of GO. The fluorination of GO inspired two models analogous to structure of poly(carbon monofluoride) (CF)n and poly(dicarbon monofluoride) (C2F)n. Nakajima and Matsuo proposed a stage 2 type (C2F)n model containing quasi-pentacovalent carbon in which carbonyl groups hold partial negative charges. This model was proposed as negatively charged carbonyl groups are not stable and tend to react with an adjacent charged carbonyl group to form 1,3-ethers. Currently, the most accepted model is that proposed by Lerf, Klinowski and their coworkers. The key features of this model include: (1) the basal plane structure of GO contains unoxidized benzene rings and aliphatic six-membered rings with hydroxyl and epoxy (1,2-ether) groups (no 1,3-ethers are detected), and the size of these two regions is dependent on the degree of oxidation; (2) The location of hydroxyl and epoxy groups is very close, and the distribution of functional groups is random; (3) carboxyl groups are formed at the edges of carbon sheets. As presented in Figure 1.3, Ajayan and his coworkers recently confirmed the existence of five- and six-membered ring lactols at the edge of GO in addition to hydroxyl and epoxy groups.
After 2004, Ruoff s group pioneered the preparation of monolayer graphene via chemical reduction of GO. The first paper, published in 2006, described the basic route to monolayer graphene in aqueous solution including (1) graphite powder was strongly oxidized to form GO; (2) ultrasonication treatment was further taken to split GO into single layers in aqueous solution; (3) exfoliated GO then went into the reaction with hydrazine to form graphene. Another popular paper from Ruoff s group discussed the possible mechanism of chemical reduction based on the Lerf-Klinowski model of GO. The essence of reduction is the transformation of sp3-carbon to sp2-carbon. Figure 1.4 shows that opening epoxide with hydrazine results in further reaction to form an aminoaziridine moiety, and the removal of diimide follows by thermal elimination. In additional, it was believed that the Wharton reaction was one pathway to remove oxygen from quinones next to an epoxide. Several issues should be clarified here. First, the purpose of oxidation is to introduce oxygen-containing functional groups into the basal structure of graphene sheets. The oxygen groups are introduced to reduce the Van der Waals force between the carbon layers within graphite, as well as to improve the stabilization of graphene in water and organic solvents. Secondly, the key difference, compared to Boehm et al.'s work, is that ultrasonication is used to enforce the exfoliation of GO into monolayers for subsequent chemical reduction. Thirdly, this type of graphene appears in publications with a more specific name, reduced graphene oxide (rGO). rGO is not perfect graphene with lower electricity, because small amount of functional groups remain in the basal structure. The stabilization of rGO is the fourth consideration to be taken into account after the functional components are reduced. A polymeric surfactant such as poly(sodium 4-styrensulfonate) needs to be used to stabilize GO in aqueous solution. Finally, Ruoff's work demonstrated a box of tools for testifying the success of reduction. It can readily be observed that the reduction causes a change in colour from yellow (GO) to black (rGO) in water. X-ray photoelectron spectroscopy (XPS) is used to conduct elemental analysis to quantitatively assess the degree of reduction. Other useful tools include measurement of electrical conductivity, Raman spectroscopy, thermogravi-metric analysis and solid-state NMR.
The chemical reduction technique still requires improvement aiming at a facile method of producing of high-quality rGO. Owing to the high toxicity of hydrazine and dimethylhydrazine, a great deal of work is going on to explore new reductants, as summarized in Table 1.1. Ascorbic acid (vitamin C) was claimed to be an environmentally friendly reductant, which restored the electrical conductivity of rGO up to ~7.7x103 S/m. It seems that when hydroiodic acid or the mixture of hydriodic and acetic acid is used as the reductant the maximum electrical conductivity of rGO is ~3.0 x 104 S/m. Gao et al reported an effective reduction protocol as which involves the use of two reductants including sodium borohydride for the deoxygenation of GO followed by concentrated sulfuric acid for dehydration. Thermal annealing is finally used to further reduce the remaining functional groups to obtain highly conductive rGO (2.02 x 104 S/m) with less than 0.5wt% of sulfur and nitrogen impurities. The possible reduction mechanisms proposed for individual cases can be found in the publications listed in Table 1.1. However, precise understanding remains an open question for chemists.
Excerpted from Polymer-Graphene Nanocomposites by Vikas Mittal. Copyright © 2012 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.