cr900070d

Honeycomb Carbon: A Review of Graphene Matthew J. Allen,† Vincent C. Tung,‡ and Richard B. Kaner*,†,‡ Department of Chemistry and Biochemistry and California NanoSystems Institute, and Department of Materials Science and Engineering, University of California, Los Angeles, Los Angeles, California 90095 Received February 20, 2009 Contents 1. Introduction 132 2. Brief History of Graphene 133 2.1. Chemistry of Graphite 134 3. Down to Single Layers 134 3.1. Characterizing Graphene Flakes 136 3.1.1. Scanning Probe Microscopy 136 3.1.2. Raman Spectroscopy 136 4. Extraordinary Devices with Peeled Graphene 136 4.1. High-Speed Electronics 137 4.2. Single Molecule Detection 138 5. Alternatives to Mechanical Exfoliation 138 5.1. Chemically Derived Graphene from Graphite Oxide 139 5.1.1. Depositions 139 5.1.2. Defect Density in Chemically Derived Graphene 139 5.1.3. Field-Effect Devices 139 5.1.4. Practical Sensors 140 5.1.5. Transparent Electrodes 141 5.2. Total Organic Synthesis 141 5.3. Epitaxial Graphene and Chemical Vapor Deposition 142 6. Graphene Nanoribbons 143 7. Future Work 143 8. Conclusions 144 9. Acknowledgments 144 10. References 144 1. Introduction Graphene is the name given to a two-dimensional sheet of sp2-hybridized carbon. Its extended honeycomb network is the basic building block of other important allotropes; it can be stacked to form 3D graphite, rolled to form 1D nanotubes, and wrapped to form 0D fullerenes. Long-range π-conjugation in graphene yields extraordinary thermal, mechanical, and electrical properties, which have long been the interest of many theoretical studies and more recently became an exciting area for experimentalists. While studies of graphite have included those utilizing fewer and fewer layers for some time,1 the field was delivered a jolt in 2004, when Geim and co-workers at Manchester University first isolated single-layer samples from graphite (see Figure 1).2 This led to an explosion of interest, in part because two-dimensional crystals were thought to be ther- modynamically unstable at finite temperatures.3,4 Quasi-two- dimensional films grown by molecular beam epitaxy (MBE) are stabilized by a supporting substrate, which often plays a significant role in growth and has an appreciable influence on electrical properties.5 In contrast, the mechanical exfo- liation technique used by the Manchester group isolated the two-dimensional crystals from three-dimensional graphite. Resulting single- and few-layer flakes were pinned to the substrate by only van der Waals forces and could be made free-standing by etching away the substrate.6-9 This mini- mized any induced effects and allowed scientists to probe graphene’s intrinsic properties. The experimental isolation of single-layer graphene first and foremost yielded access to a large amount of interesting physics.10,11 Initial studies included observations of graphene’s ambipolar field effect,2 the quantum Hall effect at room temperature,12-17 measurements of extremely high carrier mobility,7,18-20 and even the first ever detection of single molecule adsorption events.21,22 These properties generated huge interest in the possible implementation of graphene in a myriad of devices. These include future generations of high-speed and radio frequency logic devices, thermally and electrically conductive reinforced composites, sensors, and transparent electrodes for displays and solar cells. Despite intense interest and continuing experimental success by device physicists, widespread implementation of graphene has yet to occur. This is primarily due to the difficulty of reliably producing high quality samples, espe- cially in any scalable fashion.23 The challenge is really 2-fold because performance depends on both the number of layers present and the overall quality of the crystal lattice.19,24-26 So far, the original top-down approach of mechanical exfoliation has produced the highest quality samples, but the method is neither high throughput nor high-yield. In order to exfoliate a single sheet, van der Waals attraction between exactly the first and second layers must be overcome without disturbing any subsequent sheets. Therefore, a number of alternative approaches to obtaining single layers have been explored, a few of which have led to promising proof-of- concept devices. Alternatives to mechanical exfoliation include primarily three general approaches: chemical efforts to exfoliate and stabilize individual sheets in solution,27-32 bottom-up meth- ods to grow graphene directly from organic precursors,33-36 and attempts to catalyze growth in situ on a substrate.37-43 Each of these approaches has its drawbacks. For chemically derived graphene, complete exfoliation in solution so far requires extensive modification of the 2D crystal lattice, which degrades device performance.31,44 Alternatively, bot- tom-up techniques have yet to produce large and uniform † Department of Chemistry and Biochemistry and California NanoSystems Institute. ‡ Department of Materials Science and Engineering and California Nano- Systems Institute. Chem. Rev. 2010, 110, 132–145132 10.1021/cr900070d  2010 American Chemical Society Published on Web 07/17/2009 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 single layers. Total organic syntheses have been size limited because macromolecules become insoluble and the occur- rence of side reactions increases with molecular weight.36 Substrate-based growth of single layers by chemical vapor deposition (CVD) or the reduction of silicon carbide relies on the ability to walk a narrow thermodynamic tightrope.40 After nucleating a sheet, conditions must be carefully controlled to promote crystal growth without seeding ad- ditional second layers or forming grain boundaries. Despite tremendous progress with alternatives, mechanical exfoliation with cellophane tape still produces the highest quality graphene flakes available. This fact should not, however, dampen any interest from chemists. On the contrary, the recent transition from the consideration of graphene as a “physics toy” to its treatment as a large carbon macromolecule offers new promise. Years of carbon nano- tube, fullerene, and graphite research have produced a myriad of chemical pathways for modifying sp2 carbon structures,45-50 which will undoubtedly be adapted to functionalize both the basal plane of graphene and its reactive edges. This not only promises to deliver handles for exploiting graphene’s intrinsic properties but also should to lead to new properties altogether. This review will discuss the field of graphene from a materials chemistry standpoint. After a brief history of the topic, the exciting progress made since 2004, in both the production of graphene and its implementation in devices, will be discussed. For a thorough discussion focused on the physics of graphene, see refs 10, 11, 51, and 52. 2. Brief History of Graphene To understand the trajectory of graphene research, it is useful to consider graphene as simply the fewest layer limit of graphite. In this light, the extraordinary properties of honeycomb carbon are not really new. Abundant and naturally occurring, graphite has been known as a mineral for nearly 500 years. Even in the middle ages, the layered morphology and weak dispersion forces between adjacent sheets were utilized to make marking instruments, much in the same way that we use graphite in pencils today. More recently, these same properties have made graphite an ideal material for use as a dry lubricant, along with the similarly structured but more expensive compounds hexagonal boron Matthew J. Allen is a graduate student in the Kaner laboratory at the University of California, Los Angeles (UCLA). He received his B.S. in physics at Rice University, where he researched carbon nanostructures in the laboratories of Richard Smalley and Robert Curl. Vincent C. Tung is a graduate student in the Yang laboratory coadvised by Prof. Kaner at the University of California, Los Angeles (UCLA). He received his M.S. in chemistry from the National Tsing-Hua University in Hsinchu, Taiwan. His previous work was on the photochemistry of organic light emitting diodes (OLEDs). Richard B. Kaner received a Ph.D. in inorganic chemistry from the University of Pennsylvania in 1984. After carrying out postdoctoral research at the University of California, Berkeley, he joined the University of California, Los Angeles (UCLA), in 1987 as an Assistant Professor. He was promoted to Associate Professor with tenure in 1991 and became a Full Professor in 1993. Professor Kaner has received awards from the Dreyfus, Fulbright, Guggenheim, and Sloan Foundations, as well as the Exxon Fellowship in Solid State Chemistry and the Buck-Whitney Research Award from the American Chemical Society for his work on refractory materials, including new synthetic routes to ceramics, intercala- tion compounds, superhard materials, graphene, and conducting polymers. Figure 1. Single layer graphene was first observed by Geim and others at Manchester University. Here a few layer flake is shown, with optical contrast enhanced by an interference effect at a carefully chosen thickness of oxide. (Reprinted with permission from Science (http://www.aaas.org), ref 2. Copyright 2006 American Association for the Advancement of Science.) Honeycomb Carbon: A Review of Graphene Chemical Reviews, 2010, Vol. 110, No. 1 133 nitride and molybdenum disulfide. High, in-plane electrical (∼104 Ω-1 cm-1) and thermal conductivity (∼3000 W/mK) enable graphite to be used in electrodes and as heating elements for industrial blast furnaces.53,54 High mechanical stiffness of the hexagonal network (1060 GPa) is also utilized in carbon fiber reinforced composites. These uses and others generate an annual demand of more than 1 million tons of graphite worldwide.55 The anisotropy of graphite’s material properties continues to fascinate both scientists and technologists. The s, px, and py atomic orbitals on each carbon hybridize to form strong covalent sp2 bonds, giving rise to 120 ° C-C-C bond angles and the familiar chicken-wire-like layers. The remaining pz orbital on each carbon overlaps with its three neighboring carbons to form a band of filled π orbitals, known as the valence band, and a band of empty π* orbitals, called the conduction band. While three of the four valence electrons on each carbon form the σ (single) bonds, the fourth electron forms one-third of a π bond with each of its neighbors producing a carbon-carbon bond order in graphite of one and one-third. With no chemical bonding in the c-direction, out-of-plane interactions are extremely weak. This includes the propagation of charge and thermal carriers, which leads to out-of-plane electrical and thermal conductivities that are both more than 103 times lower than those of their in-plane analogues.56 2.1. Chemistry of Graphite Graphite has a rich chemistry in which it can participate in reactions as either a reducing agent (electron donor) or an oxidizer (electron acceptor). This is a direct consequence of its electronic structure, which results in both an electron affinity and an ionization potential of 4.6 eV.53 A large number of experiments for graphite focus on the insertion of additional chemical species between the basal planes, or intercalation. Shaffault is credited with the first intercalation compound using potassium, dating back to 1841.57 Graphite intercalation compounds (GICs) appear to be the only layered compounds sufficiently ordered to exhibit “staging” in which the number of graphitic layers in between adjacent intercalants can be varied in a controlled fashion. The stage of a compound refers to the number of graphitic layers in between adjacent planes of intercalant. The inter- layer spacing can increase from 0.34 nm (3.4 Å) in native graphite to more than 1 nm in some GICs, which further exaggerates the anisotropy of many properties.56,58 The increased interlayer spacing in GICs also means a significant reduction in the van der Waals forces between adjacent sheets, which leads one to consider their exfoliation as a possible route to single layers of graphene. Our group tried just that in 2003 by violently reacting a stage-1 potassium intercalation compound (KC8) with various sol- vents such as alcohols, but exfoliation produced only metastable slabs around 30 layers thick that had a tendency to scroll under high-powered sonication (see Figure 2).53,59,60 The interlayer spacing in GICs can be further increased by thermal shock to produce “expanded” graphite, which has now served as a starting material for recent techniques, including a nanoribbon synthesis developed by Dai (see Figure 3).53,61 A second focus of experiments on graphite has been substitutional doping by the replacement of carbon with other elements. This includes work by Bartlett and co-workers at Berkeley in which substitution of carbon with boron and nitrogen resulted in p- and n-type graphite, respectively.62,63 In light of recent progress with CVD of single layer graphene, such work will almost certainly be revisited as an alternative to external gating for controlling electronic behavior in graphene-based devices, or perhaps to form graphene-only p-n junctions. It is also important to mention a few points about progress in the chemistry of carbon nanotubes. Among the most important observations have been of the differences in reactivity between the different crystallographic directions (zigzag or armchair).64-66 This knowledge should transfer directly to the “unrolled” or “flattened” case of planar graphene. A myriad of techniques have also been developed to selectively modify either the sidewalls of carbon nanotubes or their end-caps. Such reactions are important looking forward because they correspond to modification of the basal plane of graphene and its edges. In fact, in situ TEM was recently used to study reactions on graphene’s zigzag edge by Zettl and others at Berkeley.67 3. Down to Single Layers Researchers have used mechanical exfoliation of layered compounds to produce thin samples for some time. In 1999, Ruoff’s group presented one such approach for graphite by using an atomic force microscope (AFM) tip to manipulate small pillars patterned into highly oriented pyrolytic graphite (HOPG) by plasma etching (see Figure 4).1 The thinnest slabs observed at that time were more than 200 nm thick or the equivalent of ∼600 layers. Kim’s group at Columbia later improved the method by transferring the pillars to a tipless cantilever, which successively stamped down slabs as thin as 10 nm, or ∼30 layers, on SiO2.68 Electrical measurements made on the thin crystallites foreshadowed a wealth of work to come. Other early groups working toward graphene included Enoki’s in Tokyo, who used temperatures around 1600 °C to convert nanodiamonds into nanometer-sized regions of graphene atop HOPG in 2001.69 While these elegant methods produced thin samples, it was ultimately a much simpler approach that led to the first isolation of single layer graphene in 2004 by a Manchester group led by Geim (see Figures 5).2 In its most basic form, the “peeling” method utilizes common celluphene tape to successively remove layers from a graphite flake. The tape is ultimately pressed down against a substrate to deposit a Figure 2. Schematic diagram showing the intercalation and exfoliation process to produce thin slabs of graphite. Potassium is inserted between the layers and reacted violently with alcohols. The exfoliated slabs are ∼30 layers thick. (Reprinted with permis- sion by The Royal Society of Chemistry from ref 60.) 134 Chemical Reviews, 2010, Vol. 110, No. 1 Allen et al. sample (see Figure 1). Although the flakes present on the tape are much thicker than one layer, van der Waals attraction to the substrate can delaminate a single sheet when the tape is then lifted away. The method requires a great deal of patience, as depositions put down by inexperienced scientists are often a mess of thick slabs in which locating a single layer can be extremely difficult. With practice, the technique results in high-quality crystallites, which can be more than 100 µm2 in size. Perhaps the most important part of isolating single layer graphene for the first time was the ability to spot an atomically thin specimen in some readily identifiable fashion. Optical absorbance of graphene has since been measured at just 2.3%, ruling out direct visual observation (see Figure 6).70,71 In order to visualize single flakes, the Manchester group took advantage of an interference effect at a specially chosen thickness (300 nm) of SiO2 on Si to enhance the optical contrast under white-light illumination.72 Although seemingly a simple idea, this was a major step forward and has contributed a great deal toward progress in this field. Figure 3. Scanning electron micrographs of natural graphite before (a) and after (b) expansion by acid intercalation and thermal shock. (Reprinted with permission by The Royal Society of Chemistry from ref 60.) Figure 4. Scanning electron microscope images of early attempts at mechanical exfoliation using graphite pillars. (a and b) Ruoff’s group peeled away layers with an AFM tip. (Reprinted with permission from ref 1. Copyright 1999 Institute of Physics.) (c and d) Kim’s group transferred the pillars to a tipless cantilever and deposited thin slabs onto other substrates in tapping mode. A series of scanning electron microscope images show thin samples cleaved onto the Si/SiO2 substrate and a typical mesoscopic device. (Reprinted with permission from ref 68. Copyright 2005 American Institute of Physics.) Figure 5. Mechanical exfoliation produced the very first single layer graphene flakes. (a) An atomic force microscopy image shows the substrate-graphene step height of <1 nm and a folded step height of 0.4 nm. (Reprinted with permission from ref 9. Copyright 2005 PNAS.) (b) TEM image of a free-standing graphene film after etching of the underlying substrate. [Reprinted with permission from Nature (http://www.nature.com), ref 6. Copyright 2007 Nature Publishing Group.] Honeycomb Carbon: A Review of Graphene Chemical Reviews, 2010, Vol. 110, No. 1 135 Groups have since adapted the same effect to image graphene on a variety of substrates and under nonwhite-light condi- tions.72-75 3.1. Characterizing Graphene Flakes With new access to 2D crystallites, experimentalists scrambled to confirm results long predicted by theory. Before they could do so, techniques needed to be developed for the characterization of deposited flakes. While optical micros- copy using the interference effect was a good method for identifying thin candidates, it could not provide conclusive evidence that a given flake was single, double, or multilay- ered. This is an important issue because some of the more interesting properties of graphene are dependent on crystallite thickness. The most obvious example is electronic band structure. Single-layer graphene is a zero band gap semi- conductor or semimetal in which the highest occupied molecular orbital (HOMO) touches the lowest unoccupied molecular orbital (LUMO) at a single Dirac point. For thicker flakes, stacking of multiple layers leads to some overlap of their carrier wave functions and the overall behavior becomes metallic. To match observations with theory, reliable iden- tification of the number of layers present in a given sample became imperative. 3.1.1. Scanning Probe Microscopy Scanning probe microscopy was perhaps the most obvious choice for verification of crystallite thickness. The method is relatively slow, but the 0.34 nm (3.4 Å) step height for each successive layer is well within the detection limits for modern atomic force microscopes (AFMs). Resolving the substrate-graphene height profile proved difficult, however, due to th