Graphene has recently enabled the dramatic improvement of portable electronics and electric vehicles by providing better means for storing electricity. In this Review, we discuss the current status of graphene in energy storage and highlight ongoing research activities, with specific emphasis placed on the processing of graphene into electrodes, which is an essential step in the production of devices. We calculate the maximum energy density of graphene supercapacitors and outline ways for future improvements. We also discuss the synthesis and assembly of graphene into macrostructures, ranging from 0D quantum dots, 1D wires, 2D sheets and 3D frameworks, to potentially 4D self-folding materials that allow the design of batteries and supercapacitors with many new features that do not exist in current technology.
Since its discovery a decade ago, dozens of potential uses for graphene have been proposed, from faster computer chips and flexible touchscreens to hyper-efficient solar cells and desalination membranes. One exciting property that has sparked significant interest is its ability to store electrical charge. A single sheet of graphene sufficient in size to cover an entire American football field would weigh just a fraction of a gram. This huge surface area associated with this small amount of graphene can be squeezed inside an AA battery, enabling the design of new energy-storage devices with the ability to store massive amounts of charge. In this Review, we discuss the inherent properties of graphene and what it has to offer for energy storage. Much of the Review covers the synthesis and assembly of graphene into macrostructures that exploit the unique features of individual graphene sheets to build new materials for various applications. Particular attention is paid to the processing of graphene into electrodes, which is an essential step in the production of devices. Graphene can be useful by itself, but it is also promising — as we discuss — for composites with superior performance compared with existing materials.
Currently, graphene is the most studied material for charge storage and the results from many laboratories confirm its potential to change today's energy-storage landscape. Specifically, graphene could present several new features for energy-storage devices, such as smaller capacitors, completely flexible and even rollable energy-storage devices, transparent batteries, and high-capacity and fast-charging devices. These and other features are explored in this Review. Despite notable progress, the future of graphene in the energy-storage market is uncertain because of several challenges. We discuss and propose solutions to these challenges and also briefly discuss the potential of other emerging 2D materials for energy-storage applications.
Graphene for energy storage
The fundamental properties of graphene make it promising for a multitude of applications. In particular, graphene has attracted great interest for supercapacitors because of its extraordinarily high surface area of up to 2,630 m2 g−1. Recently, the intrinsic capacitance of single-layer graphene was reported to be ∼21 μF cm−2; this value sets the upper limit for electric double-layer (EDL) capacitance for all carbon-based materials1. Thus, supercapacitors based on graphene could, in principle, achieve an EDL capacitance as high as ∼550 F g−1 if the entire surface area can be fully utilized. However, to understand the limits of graphene in supercapacitors, it is important to know the energy density of a fully packaged cell and not just the capacitance of the active material. In addition to the capacitance of graphene, the maximum energy density of graphene-based supercapacitors depends on several other parameters, such as the thickness and density of the graphene film and other cell components, including the current collector and the separator, the nature and density of the electrolyte, the operating voltage window of the cell and the packaging efficiency.
As illustrated in Fig. 1, when using a standard current collector, separator and acetonitrile-based electrolyte, the two key parameters that control the energy density of graphene supercapacitors are the density of the graphene film and the voltage of the cell. For an electrochemical cell using 200-μm-thick graphene electrodes with a density of 1.5 g cm−3 and an operating voltage of 4 V, the maximum theoretical energy density is ∼169 Wh kg−1 on a gravimetric basis and ∼303 Wh l−1 on a volumetric basis. If achieved, these values could put graphene supercapacitors on par with nickel metal hydride and lithium-ion batteries (LIBs) with the additional advantages of higher power density and longer cycle life. In practice, the energy density of graphene supercapacitors achieved so far is between 15 and 35 Wh kg−1, and less than 60 Wh l−1 — far below the theoretical values.
Graphene is also very useful in a wide range of batteries including redox flow, metal–air, lithium–sulfur and, more importantly, LIBs. For example, first-principles calculations indicate that lithium ions can be stored on both sides of graphene, forming C3Li (Ref. 2). Accordingly, graphene can deliver a theoretical capacity of 744 mAh g−1, which is about twice the capacity of conventional graphite electrodes. While these calculations were carried out on perfect and crystalline graphene, real graphene electrodes use chemically modified and defective forms of graphene. The performance of these graphene electrodes varies between 100 and 1,000 mAh g−1 depending on the properties of graphene, electrode processing, mass loading, structural defects and surface area. A few reports show that specially designed graphene electrodes can exceed 1,000 mAh g−1, mainly when tested in a half cell, but normally, smaller values are obtained in full-cell measurements. Graphene can be chemically processed into various forms suitable for both the positive and negative electrodes, enabling the fabrication of an all-graphene battery with an ultrahigh energy density3. Graphene is a great substrate for anchoring LIB anode and cathode materials to create high-energy-density, flexible, stretchable, fast-charging and longer-lasting batteries. In addition to LIBs, graphene hybrids have also been shown to achieve excellent performance in a range of other batteries: for example, serving as electrodes in Na+ and Al3+ batteries, and as a high-efficiency catalyst in metal–air batteries. Although graphite is an excellent anode in LIBs, it cannot be utilized in Na+ and Al3+ batteries because these ions are too large to effectively intercalate into graphite under ambient conditions, so alternative anode materials are required, such as porous graphene composites, to achieve high energy-storage performance.
Graphene synthesis and assembly
Since it was first isolated in 2004, researchers have developed a plethora of methods to synthesize graphene, which can be classified as either ‘bottom-up’ strategies, including SiC epitaxial growth, chemical vapour deposition and chemical synthesis, or ‘top-down’ methods, including micromechanical cleavage, liquid-phase exfoliation and the reduction of graphene oxide4. However, extending the excellent properties of individual graphene sheets to a macroscopic scale is the most important issue for realizing its practical applications. Sheet assembly is a thermodynamic process in which the interactions between single sheets, such as van der Waals forces, π–π stacking, electrostatic interactions, hydrogen bonding and hydrophobic interactions, have decisive roles. Various unique macroscopic assemblies of graphene can be prepared that cover all ‘four’ dimensions, in the form of quantum dots (0D), wires (1D), films (2D), monoliths (3D), and potentially 4D self-healing and/or self-folding structures, as illustrated in Fig. 2. We now discuss how each of these forms affects the properties and applications of graphene.
Dimensionality is an important parameter and defines, to a large extent, the properties of a material. In fact, graphene can exhibit different properties in different dimensions. For example, interfacial self-assembly of graphene oxide at the water/air interface is an interesting strategy for the design of 0D graphene materials, such as nanospheres, hollow capsules and even crumpled balls. In contrast to 2D graphene sheets, which readily aggregate during electrode fabrication, these structures have proven to be very useful in preventing the restacking of graphene, thereby leading to supercapacitors with larger specific capacitances and higher rate capabilities5. Graphene can also be made into smaller fragments — less than 30 nm — called graphene quantum dots6. Their high surface area and biocompatibility make them promising for biosensors, bioimaging and drug delivery. They also have large open surfaces and free edges that allow the design of supercapacitors with high specific capacitances7 and micro-supercapacitors with extremely fast response times (with a time constant of approximately 100 μs)8. Moreover, doping graphene oxide quantum dots with nitrogen results in a photocatalyst with both p- and n-type conduction that can drive overall water splitting spontaneously9.
One-dimensional graphene fibres are attractive because of their ability to integrate the fascinating properties of individual graphene sheets into macroscopic assemblies with mechanical and electrical properties beyond those of conventional carbon fibres10. Graphene fibres can reach an electrical conductivity of up to 104 S m−1 and a tensile strength of 400 MPa; these values are comparable to those of carbon nanotube (CNT) fibres and far exceed those of the carbon fibres widely used today in the aircraft and automotive industries. Compared with films and bulk structures, graphene fibres are favourable for the construction of flexible and wearable supercapacitors because they can be directly shaped into fabrics and textiles11. In addition, graphene fibres can be tailored to yield batteries and supercapacitors with unconventional architectures12. Another form of 1D graphene — nanowires — shows promise for flexible transparent conducting electrodes with excellent optical and electrical properties and high durability13. By contrast, graphene nanotubes exhibit extremely low sheet resistance down to 0.11 Ω sq−1 and a thermal conductivity of 8.28 W m−1 K−1, which allow the design of efficient reservoirs for thermal energy storage14.
Graphene film is one of the most applicable macroscopic forms of graphene because of the high aspect ratio along its lateral direction. Laminar graphene sheets readily assemble into 2D macroscopic structures, such as thin films, membranes and paper, with orderly packed lamellar sheets. Graphene films are particularly promising in electrochemical energy-storage devices that already use film electrodes. Graphene batteries and supercapacitors can become viable if graphene films can equal or surpass current carbon electrodes in terms of cost, ease of processing and performance. The unique properties of graphene sheets provide graphene films with excellent electrical conductivity, mechanical strength, thermal conductivity and optical transparency, depending on the quality of the individual graphene layers and the thickness of the resulting film. In addition, graphene sheets are highly flexible and can be readily functionalized to create new materials with unique properties. They can also be processed into a library of other 2D structures that may find use in different applications (Fig. 2).
Perhaps one of the most effective ways to apply the unique features of 2D graphene sheets in practical applications is to assemble them into 3D structures, as shown in Fig. 2. Various 3D structures have been reported that include aerogels, hydrogels, sponges, foams, frameworks with periodic structures, honeycomb-like structures and 3D-printed graphene. These structures possess unique features, such as exceptionally high specific surface area, hierarchically interconnected pore structure and excellent electrical conductivity. They can support fast electron transfer in 3D and provide accessible surfaces for the adsorption and desorption of ions, leading to batteries and supercapacitors with very high energy densities. Not only does 3D graphene offer exceptionally high surface areas that are required for effective charge storage, but it also has large internal pores that are suitable for anchoring specific materials for fuel cells, catalysis and sensing applications. These characteristics give graphene sponges the properties needed for cleaning oil spills. Experiments have shown that graphene sponges are capable of absorbing petroleum products up to 86 times their own weight, which is much higher than other common absorbents15. They also exhibit excellent electrical conductivity and biocompatibility, enabling their use as scaffolds for tissue engineering and regeneration16. Tremendous research efforts are ongoing to control the morphology and the properties of 3D graphene for integration into practical devices17 (Fig. 2).
The future of the field lies in the development of responsive graphene structures that can adapt to their surroundings. Normally, real-world objects do not change shape or properties once produced; however, a material that can transform and change shape when subjected to a change in its environment would be incredibly useful. This may lead to batteries that can self-heal in case of an accident or smart windows that can change colour to block out sunlight. These 3D objects that change structure over time are termed ‘4D materials’. It was recently shown that 4D printing of a ‘smart object’ may, one day, lead to the development of such devices18.
Processing of graphene
To realize the commercial potential of graphene, it is necessary to develop reliable, cost-effective and facile processes for the industry-scale fabrication of graphene electrodes with good performance. Solution processing offers a simple yet effective strategy for the fabrication of graphene electrodes and is compatible with the current production protocols adopted by industry for the preparation of battery electrodes. There are two requirements for solution-processing graphene-based dispersions in scalable devices: first, a homogeneous dispersion of high-quality graphene must be formed in a suitable solvent; second, a scalable technique for the continuous fabrication of graphene-based films must be used. Liquid-phase exfoliation of graphene in organic solvents is an efficient and low-cost method to produce graphene dispersions. However, the drawbacks of this process are the high boiling points and toxicity of the organic solvents required. Thus, aqueous dispersions of high-quality graphene sheets are highly desirable. In 2008, a facile method was developed to make chemically converted graphene sheets to form stable aqueous dispersions19, which has proven to be an excellent candidate method for fabricating graphene films.
Graphene dispersions can be processed into electrodes using various coating techniques, including dip coating, rod coating, spray coating, inkjet printing, spin coating, screen printing, gravure printing, blade coating, electrospinning, electrodeposition, vacuum filtration, drop casting, interfacial deposition, Langmuir–Blodgett deposition and layer-by-layer assembly20,21. The most appropriate coating technique depends on the desired properties and size of the final electrodes, and the desired film thickness and amount of graphene in the dispersion. Another important parameter is the design of the device (that is, whether a traditional sandwich structure or a planar energy storage device is needed). The setups used in the techniques listed above are evaluated in Table 1. Because many of these techniques involve similar mechanisms or properties, we focus on just a few of them.
Vacuum filtration is one of the most commonly used methods for preparing graphene and graphene-based composite films because of its general applicability and easy manipulation. One advantage of this process is that it affords convenient control of the density and thickness of the as-filtered film simply by adjusting the filtration time, concentration and/or volume of the graphene dispersion. The filtered graphene film can be transferred onto essentially any substrate. Alternatively, a freestanding graphene film can be produced by peeling off or dissolving the filter membrane. For example, by using such a method, the properties of graphene-based LIB anodes can be controlled, which allows the rapid transport of both electrons and ions throughout the electrode, resulting in a high material tap density (1.3 g cm−3)22. However, there are drawbacks of the filtration method. Most importantly, the practical size of the filtered graphene film is limited by the filtration apparatus, which means it is difficult to continuously fabricate graphene films by filtration processing on a large scale.
Doctor blading, also known as knife coating, is an established method for processing large-area films over rigid or flexile substrates. This technique is widely used in research laboratories to produce graphene films for battery and supercapacitor applications. In a typical experiment, the slurry is transferred to the current collector by moving the blade over a flat substrate. Variable film thicknesses can be obtained by controlling the size of the gap between the blade and the substrate. Recently, this technique was used for the direct preparation and processing of graphene composite electrodes with ultrahigh specific capacitances in excess of 1,100 F g−1 (Ref. 23).
Mayer-rod coating is one of the most popular coating techniques. In this method, a stainless steel wire of varying diameter is tightly wound around a metal rod. The grooves made by the wire allow the right amount of the slurry to pass through, thereby controlling the coating thickness. This technique was used to produce large-area graphene electrodes with excellent optical properties, with potential for use in flexible devices24. A Mayer rod was also shown to be effective for fabricating flexible energy-storage devices by integrating electrodes and separators into a single sheet of paper25.
Spray coating may be the simplest method to realize the scalable fabrication of graphene films from graphene dispersions. In general, spray coating involves the atomization of a dispersion and deposition of the formed droplets onto a heated substrate, which evaporates the solvent to form a film. The spray-coating process usually results in a more uniform deposition than other methods, such as drop casting or dip coating. The quantity and density of the film can be controlled by adjusting the dispersion concentration and spray time. Low-temperature spray coating can be used to fabricate high-performance LIBs26. The cycling stability of lithium–sulfur batteries was also improved by coating a thin layer of graphene oxide-oxidized CNTs on the separator27. In addition, some extended approaches, including supersonic kinetic spraying28 and electrospray deposition29, have been developed to improve the quality of the final films.
Roll coating is the standard technology used today in the battery and supercapacitor industries, whereby a slurry is continuously coated onto the substrate of choice at high speed using a coating roll. This is the technique of choice in these industries because it offers several advantages, including large-volume production, high line speeds and lower manufacturing costs. Unfortunately, this technique is rarely used in academic institutions to make graphene electrodes, although it would be beneficial because the equipment required is already commercially available, which would lead to decreased capital investment and time-to-market for new products.
With the continuous shrinking of electronic devices, research on microbatteries and micro-supercapacitors has recently flourished30. In these cases, the most important requirement for the processing method is the ability to print patterned microelectrodes with large storage capacity. Techniques such as inkjet printing31, screen printing32, gravure printing33 and electrodeposition34 have successfully been used in the fabrication of miniaturized energy-storage devices.
Energy-storage devices with new features
Now that smartphones and other portable electronic devices have become ubiquitous, one major problem remains: battery technology has not kept up with the demands placed on them. In addition, the lack of a reliable battery still holds back electric vehicles from being readily adopted over vehicles using an internal combustion engine. Soon, graphene could establish a new generation of energy-storage devices with 12 new features not possible with current technology, as summarized in Table 2 and discussed in the following 12 subsections.
Supercapacitors with AC line filtering. Capacitors are essential components in almost all electronic devices. They are primarily used as electronic filters to smooth out the ripples from power supplies because they have a very fast response time — on the submillisecond timescale. However, unlike transistors and other electronic components that are constantly shrinking over time, capacitors are still bulky and often limit the miniaturization of the entire system. Commercial-grade carbon-based supercapacitors charge and discharge in seconds, which is much faster than batteries but not quite fast enough to be used as electronic filters.
In 2010, it was reported that an EDL capacitor based on vertically oriented graphene sheets could be charged and discharged in less than a millisecond, which is similar to the performance of capacitors but with a capacitance per volume at least ten times higher35. This ultrafast supercapacitor could replace the large electrolytic capacitors used in today's electronics and may someday help make electronic devices smaller and lighter. The key to achieving ultrafast charging and discharging is to vertically orient graphene electrodes with external rather than internal surface area. This leads to a high-frequency response and eliminates the slow diffusion of ions, which has so far limited the speed of carbon-based supercapacitors. This work has inspired several research efforts to design new electrodes with ultrahigh power density by controlling pore size and distribution. Various materials, including electrochemically reduced graphene oxide36, graphene quantum dots8, a graphene–CNT carpet37, graphene–PEDOT:PSS38 and conducting polymers39, have been developed and tested for alternating current (AC) line filtering.
Flexible, rollable and twistable energy-storage devices. The market for flexible and printed electronics is rapidly growing, with products now ranging from flexible solar-cell arrays to flexible displays and wearable electronics40. The rise of this technology is mainly due to rapid progress in the production of flexible electronic devices over large areas at a fraction of the cost of traditional semiconductor technology41. In the field of energy storage, various devices, such as flexible supercapacitors and thin LIBs with variable sizes, shapes and mechanical properties, are being developed.
Current batteries and supercapacitors come in several sizes and shapes but are all rigid; therefore, bending them may cause cell damage and electrolyte leakage. In brief, this is because the current technology relies on particulate-like energy-storage materials, which lose their mechanical integrity upon bending. With its 2D one-atom-thick structure, graphene can adapt to mechanical stress by deforming in the direction normal to its surface. This inherent mechanical flexibility together with exceptional electrical properties and large surface area makes graphene attractive for flexible energy-storage devices. Interesting research activities in this area are discussed in detail in recent reviews42,43.
Stretchable batteries and supercapacitors. Stretchable electronics is an emerging field featuring materials and devices that can sustain large deformations while maintaining normal function and reliability44. Stretchable devices will help monitor body functions, make portable electronics more comfortable and perhaps even wearable. Inspired by human skin, crumpled graphene was used to develop supercapacitors that can be stretched up to eight times their length. This was accomplished by attaching graphene paper to an elastomer that had been pre-stretched up to five times its original dimensions. When the pre-stretched parts were released, the lateral dimensions of the graphene paper were reduced by the same ratio. An alternative approach was adopted in another study, in which a wrinkled graphene film was first synthesized by chemical vapour deposition. Transfer of the graphene film to a polydimethylsiloxane substrate resulted in supercapacitors that were both stretchable and transparent45.
Energy-storage devices for wearable electronics. Portable electronics have revolutionized our everyday lives and have driven the development of other future electronic devices (for example, wearable electronics that can be incorporated into clothing and accessories, and worn comfortably on the body). These smart garments are expected to need to communicate with the user to give them access to information in real time46. In 2011, the first textile-based batteries and supercapacitors were obtained by coating graphene and CNTs onto cotton fabrics47. Recently, research efforts have been directed towards the development of coaxial and core–sheath electrodes that combine the current collector and electrode material in a single yarn, which can be knitted or woven directly into textiles. Graphene is a great candidate as it can be readily tailored and assembled into multifunctional microfibres and woven into fabrics. For example, graphene core–sheath microfibres were used to demonstrate spring-like supercapacitors that are flexible, stretchable, all solid-state and can be incorporated into textiles using conventional weaving technology48. Similarly, a coaxial wet-spinning strategy was used to produce polyelectrolyte-wrapped graphene/CNT core–sheath yarns, but unlike most yarn electrodes that suffer from low areal capacitance, this one produced up to 269 mF cm−2 (Ref. 11).
Transparent batteries and supercapacitors. Imagine an electronic display nearly as clear as a window or a flexible smartphone that you can see through. This may not be too far off owing to recent advances in transparent electronics that have focused on the use of wide-bandgap semiconductors for the realization of invisible electronic circuits49. Although researchers have shown that transistors, organic light-emitting diodes and other electronic components can be made transparent, the development of a transparent power source remains a challenge. This is because of the traditional multilayer structure of a battery comprising two electrodes, a separator, an electrolyte and packaging to hold these layers together. This means that each of these layers will have to be clear for the battery to be transparent. Unfortunately, only the electrolyte is inherently transparent, whereas the other components are opaque, with the electrodes being the most difficult component to make transparent. One way to make a transparent electrode is to make it very thin — on the nanometre scale45. But thin electrodes do not store much charge. Another approach is to make patterned electrodes with design features smaller than the detection limit of the human eye50. In either case, graphene can have a key role in making transparent batteries more efficient because of its high conductivity and good transparency (up to 97.7% transmittance). These characteristics make graphene an ideal electrode material not only for transparent energy-storage devices, but also for solar cells, smart windows and other optoelectronic devices.
Fast-charging lithium-ion batteries. First commercialized by Sony in 1991, LIBs represent a major technological advancement from lead–acid batteries, which have been dominant since the late 1850s. Owing to their low weight and high energy density, LIBs have become the technology of choice for consumer electronics. However, they suffer from long recharge times (typically hours), whereas battery users are looking for a battery that recharges in minutes or even seconds.
The use of graphene allows faster electron and ion transport in the electrodes, which controls the speed over which the battery can be charged and discharged. For example, a fast-charging LIB was developed by loading nanoscale Li4Ti5O12 anode and LiFePO4 cathode materials on flexible graphene foam without the use of conducting additives or binders51. The excellent electrical conductivity and the open pore structure of the hybrid electrodes allowed the battery to become fully charged in only 18 s. Pure graphene can also be used as the anode for LIBs with an improved capacity, and ultrafast charge and discharge rate.
Few-layered graphene may have an important role in the progress of other battery chemistries. For example, an aluminium-ion battery was recently reported that can be fully charged in under one minute because of the ultrafast ion intercalation into 3D graphene foams52.
Ultrathin current collectors for lightweight devices. The continuing demand for lighter and smaller mobile electronics, especially notebook personal computers and cell phones, has driven the development of higher-capacity LIBs. Current efforts focus mainly on developing new battery materials with higher capacities; however, reducing the weight of the existing battery components is also very useful. Existing battery technology uses metal foil current collectors, such as aluminium, copper or nickel, with thicknesses between 20 and 80 μm to allow for the transfer of electrons between electrodes and external circuits. This accounts for at least 30% of the weight of the battery, and because it does not store any charge, it reduces the overall energy density. In addition, these metal foils often suffer from corrosion during operation, which has a negative effect on the internal resistance of the cell and the lifetime of the battery.
Graphene has been proposed as an alternative current collector because of its low density, high electrical conductivity and chemical stability under harsh operating conditions. Graphene can be processed into films with wrinkles and ripples on the surface, resulting in better electrical contact with the active materials and thus may further reduce the equivalent series resistance of the cell53. Interestingly, it was found that a LIB with a graphene paper current collector exhibits a better cycle life than a battery with a traditional metal current collector54. Owing to its good electrical conductivity and high surface area, graphene can be used as both the active material and the current collector, leading to lightweight, flexible supercapacitors55.
Graphene oxide as solid electrolyte and separator. Batteries are generally treated as hazardous waste because of the potential leakage of the liquid electrolyte. Therefore, there is the need to develop alternative, solid-type electrolytes. Graphene oxide is an excellent electronic insulator; however, when a substantial amount of moisture is entrapped in the layered graphene oxide structure, it becomes an ion conductor with ion-transport features similar to those observed in the Nafion membranes used in fuel cells56. The hydrophilic sites on the surface of graphene oxide, such as −OH, −O− and −COOH functional groups, attract protons that are propagated on the graphene oxide surface by the adsorbed water molecules. This contributes to the ionic connection between the positive and negative electrodes and enables the use of graphene oxide as both a viable solid electrolyte and electrode separator simultaneously. This has led to the development of in-plane and sandwich supercapacitors with many rGO–GO–rGO (GO, graphene oxide; rGO, reduced graphene oxide) configurations without the use of any external electrolyte56. In one experiment, the proton conductivity of three different forms of graphene oxide was measured: graphene oxide nanosheets, bulk graphite oxide and a graphene-oxide/proton hybrid. Surprisingly, it was found that the in-plane conductivity of graphene oxide nanosheets is two to three orders of magnitude higher than that of bulk graphene oxide57. Further investigation of the proton conductivity of individual graphene oxide sheets led to the development of multilayered graphene oxide films with high proton conductivity that may be useful for energy storage and fuel cells58.
In another study, it was found that a rGO–GO–rGO sandwich structure, in which a graphite oxide film acts as a solid electrolyte, shows a capacitance (up to 0.86 mF cm−2) comparable to that of electrochemical capacitors but with undetectable ionic diffusion similar to dielectric capacitors, indicating a new ‘third’ category of electrical capacitor59. These interesting results may lead to the development of lightweight, energy-dense and ultrafast solid-state capacitors that do not suffer from ion diffusion, which limits the kinetics of traditional electrochemical capacitors, or the hazards associated with electrolyte leakage.
Supercapacitors with the energy density of batteries. Most currently available supercapacitors feature activated-carbon electrodes and an organic electrolyte that operates at voltages between 2.3 and 2.8 V. Although they offer higher power density and exceptional cycling stability compared with batteries, their energy density (5–7 Wh kg−1) is one order of magnitude lower. As described earlier, graphene has the potential to produce EDL capacitors with similar energy density to batteries without compromising the high power density and cycling stability. However, in most cases, graphene sheets tend to agglomerate during the processing of electrodes, resulting in devices that provide little performance advantage over traditional activated-carbon supercapacitors. To solve this problem, a colloidal approach was adopted for the assembly of graphene using a mixture of a volatile and a non-volatile solvent60. In this case, the non-volatile solvent prevents the sheets from touching by acting as a nanospacer. This makes it possible to control the density of the graphene electrodes and thus improve the volumetric performance. These supercapacitors demonstrated ultrahigh energy densities of up to 60 Wh l−1, which is comparable to lead–acid batteries. In a different approach, porous and dense graphene foams were developed for efficient capacitive energy storage by digging nanoholes in the basal planes of graphene and using a hydraulic press to control the density of the foam61. Recently, graphene-based hybrid supercapacitors capable of providing up to 42 Wh l−1 have been reported62. The advantage of these hybrid supercapacitors is that they work with aqueous electrolytes and can be produced in air without the need for expensive ‘dry room’ assembly. An overview of the different strategies used to improve the energy density of supercapacitors is provided in Box 1.
Permselective graphene oxide membranes. The separator is an essential component of electrochemical energy-storage devices. According to classic battery textbooks, a separator is a permeable membrane placed between the anode and cathode of a battery63. The main function of a separator is to provide a physical barrier between the two electrodes to prevent electrical short circuits, while also allowing the transport of ions during charge and discharge cycles. Separators are normally porous polymer membranes, which physically keep the two electrodes apart without affecting the transport of ions through the membrane. This nonselective nature of the separator can, in some cases, have a large influence on the cycling stability and rate capability of the battery. For example, in lithium–sulfur batteries, the polysulfide species generated at the sulfur cathode can diffuse freely through the membrane and react with the lithium-metal anode, which induces rapid capacity degradation and poor cycling stability. An ideal battery separator should, thus, be permselective; that is, it should allow specific species to be transported while restricting the passage of other species that can reduce the reversibility of the battery.
Graphene oxide membranes exhibit great mechanical strength and are electrically insulating. Owing to their negative charge, graphene oxide membranes are expected to be permeable for lithium ions but reject anions. This electrostatic repulsion has been shown to largely inhibit the diffusion of polysulfides through graphene oxide membranes in lithium–sulfur batteries, which leads to a stable electrochemical response with greatly improved self-discharge rates64. Graphene oxide membranes consist of micrometre-sized graphene oxide crystallites that form an interlocked layered structure. These characteristics can also be very useful in LIBs, in which the pore size and the non-uniform permeability across the separator may result in the growth of metallic lithium dendrites that can eventually lead to short circuiting between the anode and cathode. We note that graphene oxide membranes exhibit a wide range of unique barrier properties. It was shown that graphene oxide membranes, in the dry state, are impermeable to everything except water vapour65. When immersed in water, graphene oxide membranes act as molecular sieves, allowing transport of small ions while blocking large ones66. These features offer opportunities for the design of a new generation of ion-selective membranes for batteries, supercapacitors and fuel cells67.
Longer-lasting energy-storage devices. One of the main reasons why consumers choose one smartphone over another is not the processor speed or data storage; it is how long the smartphone will last on a single charge. Although the next big breakthrough may still be a few years away, recent advances in graphene research are bringing us closer to arguably the most important feature of all: longer-lasting batteries. Researchers have shown that it is possible to fabricate such batteries by replacing the graphite anodes used in today's LIBs with graphene electrodes in the form of folded graphene paper68, porous graphene films69 and solvated graphene frameworks70. This has led to LIB anodes with specific capacities of more than 1,000 mAh g−1 — more than three times the capacity of standard graphite electrodes — which promises longer run times for portable electronics and a longer range for electric vehicles. It has also been discovered that nitrogen and boron doping are crucial in improving the capacity and power density of graphene anodes71.
In addition, graphene can serve as a support for metal oxide active materials, while providing high electrical conductivity, mechanical strength and greater surface area. For example, with a reduced graphene oxide wrapping, an Fe2O3 composite anode exhibited a specific capacity three times higher than that of a pure Fe2O3 electrode72. Graphene can effectively increase the capacity of LIB cathodes as well. For example, LiFePO4 is widely used in LIB cathodes because of its long cycle life and low cost. In this case, graphene can act as a structural backbone on which LiFePO4 can be attached, creating a hybrid material with improved conductivity and better cycling durability. Because this hybrid allows for the maximum utilization of LiFePO4 in charge storage, it stores twice as much charge as commercially available LiFePO4 (Ref. 73).
Binder and additive-free electrodes. In both batteries and supercapacitors, the electrodes are generally made by mixing particulate active materials with carbon additives and polymeric binders (such as polyvinylidene fluoride and polytetrafluoroethylene) with the help of an appropriate solvent. The resulting slurry is coated on the current collector of choice, dried and then exposed to a mechanical press to control the pore volume and density of the electrodes. The binder is used to glue the materials together with the current collector, whereas the conductive agent improves the electrical properties of the laminate. Although this process is widely used in industry, the use of insulating polymer binders increases the contact resistance between particles, which increases the equivalent series resistance of the cell74. The binder and additive together makes up 20–40% of the mass of the electrode. This is called ‘dead mass’ because it does not contribute to charge storage, but instead decreases the energy density of the electrodes. In addition, the diffusion pathways of ions and electrons are unclear because of the inhomogeneous nature of the blend, which causes low power density and slow rate capability.
Recent studies have demonstrated that graphene can be fabricated into electrodes directly without the need for binders and conductive agents, as is the case for conventional particulate materials55. This is possible because of the ability to assemble graphene into self-standing 2D and 3D structures with excellent electrical conductivity. Compared with other battery and supercapacitor electrodes, graphene-based materials exhibit additional advantages, such as low weight, diverse macroscopic structures, controllable pore size and distribution, and mechanical flexibility75.
Other 2D materials for charge storage
The discovery of graphene has led to a new class of materials known as ‘2D crystals’ (Ref. 76). In addition to graphene, the family of 2D materials includes many transition metal dichalcogenides (such as MoS2, TiS2 and WS2), layered metal oxides (such as Nb2O5, MnO2 and TiO2), hydroxides, nitrides and carbides, and several other compounds (for example, BN and Bi2Te3). Although graphene has unveiled a set of fascinating properties, the other 2D materials cover a wide variety of properties as well; for example, they range from electrically insulating to superconducting and from transparent to optically active77. By combining graphene with other 2D materials, researchers hope to create ‘heterostructures’ with functionalities capable of delivering a new generation of superconductors, metallic materials, semimetals, semiconductors and insulators78. These materials have the potential for use in a wide range of applications, such as high-performance electronics, sensors, solar cells, gas separation, catalysis and energy storage.
Because of the reduced dimensionality of these 2D crystals, they exhibit electrochemical properties that are distinct from their bulk counterparts79. Two-dimensional transition metal dichalcogenides and oxides can support the ultrafast ion transport of lithium, hydrogen and sodium ions — a characteristic that makes them ideal for ultrafast battery and capacitor electrodes80. These materials are promising alternatives to currently used 3D bulk materials with slow ion diffusion that have so far limited the power density of traditional battery electrodes. In addition, MoS2 nanosheets show excellent capacitive behaviour by storing charge via intersheet and intrasheet double layers over individual MoS2 layers and through Faradaic reactions on the Mo centre because of its multiple oxidation states. Combining MoS2 with graphene results in supercapacitors with higher capacitance and better cycling stability81. These hybrid electrodes can also be used as LIB anodes, offering stable cycling performance and a high specific capacity of up to ∼1,100 mAh g−1 at a current density of 100 mA g−1 (Ref. 82).
Recently, a new class of 2D transition metal carbides or carbonitrides, known as MXenes, was synthesized by selective etching of the so-called MAX phases83. These materials offer an unusual combination of metallic conductivity and hydrophilicity, and are attractive for electrochemical energy storage. They can be shaped like clay and processed into films achieving an ultrahigh volumetric capacitance of 900 F cm−3 (Ref. 84). Exfoliated layers can be suspended in water and filtered into flexible MXene papers that can be used in LIBs to achieve a high capacity at high charging rates85.
Limitations of graphene and possible solutions
Graphene batteries and supercapacitors have witnessed rapid developments and substantial achievements during the past few years, but there are still many challenges to be addressed to realize their practical applications. Although researchers have demonstrated graphene-based batteries and supercapacitors with performance characteristics far exceeding those of commercially available ones, the lack of feasible techniques for the mass production of high-quality graphene limits their potential. Estimates for the cost of production of graphene vary depending on the quality of the material from tens to thousands of dollars per kilogram86, but it is still not competitive with state-of-the-art materials. For example, the very low cost of activated carbon currently used in supercapacitors (US$15 per kilogram) presents a difficult barrier to the entry of other materials87.
Most techniques that have shown potential for large-scale production depend on the exfoliation of expanded graphite or the reduction of graphite oxide. The former often results in multilayered graphene platelets with limited surface area, which is a critical parameter for supercapacitors88, and the latter form of graphene has relatively low electronic conductivity caused by residual oxygen and structural defects. In addition, reduction of graphite oxide is still challenging as it involves complicated procedures and purification methods, and often employs toxic and corrosive chemicals leading to environmental concerns. The conductivity of reduced graphene oxide needs to be further enhanced with more efficient reduction methods.
Ultrathin graphene flakes, when processed, result in electrodes with low densities and large pore volumes. The addition of an electrolyte causes huge ‘dead mass’ that adds weight and volume, but does not contribute to charge storage; this kills the performance of graphene electrodes on the device level. It is necessary to find simple methods for making graphene electrodes with large packing densities without affecting their accessible surface area. Some recent studies have addressed this issue, but the solutions are not yet practical.
The processing of electrodes is another important parameter that defines the viability of new technologies. Current battery and supercapacitor technologies use roll coating for manufacturing electrodes. This process is inexpensive, efficient and fast, producing up to several metres of electrodes per minute. Most devices made out of graphene use techniques that are limited to research laboratories and difficult to scale up. Future studies on graphene should focus on competitive processing techniques and not just laboratory-feasible preparation methods.
Finally, the properties of graphene electrodes are determined by their microstructure; however, precise control of the pore size and porosity of graphene is a big technical challenge. Most of the 3D graphene electrodes reported thus far feature pores that vary in size between hundreds of nanometres and several micrometres. To address this issue and to increase the capacity of graphene electrodes, it would be interesting to introduce micropores into the structure of graphene through assembly from porous graphene sheets or by using a chemical activation process of the type that is widely used nowadays for the preparation of activated carbon.
Graphene is widely used in batteries either as the active component or inactive conductive additive. In the latter case, graphene forms a 3D electron conducting network offering electron ‘superhighways’ that promote the charge transfer exchange rate of active materials. This increases electron conductivity and therefore rate capability of the electrode under investigation. Graphene can also buffer the volume changes during charge and discharge, and therefore increase the cycling stability of the active materials. However, in most cases, large amounts of graphene (10–20% w/w) are normally used in these hybrid electrodes. It should be noted that too much graphene does not help because of its low packing density, which can reduce the energy density of the battery. It is thus advisable to reduce the amount of graphene in the hybrid electrodes while maintaining good electrochemical performance. The low electrical conductivity of chemically derived graphene — which is almost always lower than carbon black — is another challenge. Improving the conductivity of graphene for conductive additive applications should be a priority at this stage. Solution processing of graphene oxide offers a great opportunity for large-scale production of chemically derived graphene, although stronger reducing agents are needed. Graphitization of the produced graphene through chemical and thermal methods holds great promise too.
There are several methods in which graphene can be hybridized with battery materials to produce composites with improved electrochemical performance. Specifically, the battery materials can be anchored to the graphene surface, wrapped by graphene sheets, encapsulated in a graphene shell or sandwiched between two graphene monolayers. In addition, graphene can be assembled with 2D battery materials forming layered heterogeneous structures90. However, it is not clear which of these structures produces graphene composites with the best electrochemical performance. More efforts are needed to gain a fundamental understanding of the molecular interaction between graphene and other battery components in both the bulk and at the electrode/electrolyte interface. This will allow the design of novel materials and composites with custom properties and could enable the practical use of graphene-based materials in energy-storage devices.
Another issue to be considered in graphene composites is the accessibility of the active materials to the electrolyte. Although ion diffusion may occur through surface defects and grain boundaries, the diffusion through the basal plane is rather limited. A small amount of graphene may thus provide a good balance between improved electron transport and fast ion diffusion, whereas too much graphene may isolate the active material from reaching the electrolyte because of complete wrapping91. This also depends on the structure of the graphene; for example, the narrow width of graphene nanoribbons provides more grain boundaries and therefore faster ion diffusion than normal graphene sheets. Another strategy to combat this problem is to open nanopores in the basal plane of graphene through chemical activation.
Although graphene is attractive as an anode for LIBs with a capacity of more than 1,000 mAh g−1 already achieved, it suffers from low first-cycle Coulombic efficiency92. This is mainly attributed to the high surface area of graphene and structural defects associated with the different synthetic routes. The storage of lithium ions at defects causes very high initial irreversible capacity, which results in poor energy efficiency. Unless a solution is found, this problem may hinder the practical application of graphene anodes. One possible solution is to carefully design the structure of graphene to lower the surface area contacting the electrolyte and reduce the defects on graphene to avoid side reactions. Knowing that reduced graphene oxide is the most common form of graphene currently explored in LIBs, the progressive reduction of oxygenated functional groups should produce fewer defects and may help improve the Coulombic efficiency of graphene anodes. Another strategy is to make hybrid electrodes from graphene and other battery materials to reduce or dilute the negative effect of graphene on the Coulombic efficiency.
There is no doubt that graphene has changed the landscape of energy storage because of its outstanding electrochemical properties and unique combination of large surface area, high electronic conductivity and excellent mechanical properties. Innovative techniques have been demonstrated for the fabrication of energy-storage devices with new features not possible with current state-of-the-art technology. However, the full potential of energy-storage devices built from graphene has yet to be realized. Many challenges remain, particularly with regard to feasible techniques for the low-cost mass production of graphene with controlled microstructure and low residual oxygen content. More research efforts should be made for resolving the existing challenges using theoretical calculations together with experimental investigations. Further understanding of the interaction of graphene sheets at the nanoscale to form self-assembled structures of different shapes and dimensionalities will drive more potential applications for graphene. Graphene research is expected to continue to expand rapidly during the next decade with the promise that it will make a real impact on our lives.
The authors thank Nanotech Energy for financial support.