PROGRESS ARTICLE NATURE MATERIALS DOL:10.1038/NMAT4170 Sodium-ion batteries.As explained previously for the negative lid th terial forS is en rmance have rece y been devel undoped gr 62 to the to the lov graphene in SABs Graphene as an inactive compon yehben propossod candidate to 5 being actively i -sulphu th cases of epox nal conduct atmteotsordi ng performanc hnpo into six during h g has not yet ave been es.LiCoO,.LiMn,O nd LiFeP ost common used cath ot higher thar odes,such a ordered me carbor rom these ene interface in the ation of graph or has spurd el syste ers of the m been ropose how very poor electrochemical b they could represent ea graphene hasalso been considered for cou ega ively affect ng t n redox-flow erally reported toh trolyti rap phite felts,are used as electrodes bec %of th Ne carbon blacks,which are much cheaper and easier to hand al properties".Accordingly,the use of graphene-based materials has 6 NATURE MATERIALS|ADVANCE ONLINE/ 2014 Macmillan Publishers Limited All rights reserved 6 NATURE MATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials have been increasingly proposed owing to their low production cost and the availability of the required raw materials73. In contrast with lithium, sodium is capable of reversibly forming during discharge a stable superoxide (NaO2) with low overpotentials73. This enables SABs to cycle with a charge efficiency of 80–90% after the first cycle. The formation of peroxide (Na2O2), however, is kinetically hindered as it requires a suitable catalyst. RGO has demonstrated, under dry air conditions, remarkable catalytic properties towards the forma￾tion of Na2O2 (ref. 74), which are not exhibited by conventional car￾bon. As reported by Liu et al.74, the micro- and nanostructures of the graphene air electrode enable one of the highest specific discharge capacities for SABs. These results suggest that RGO can efficiently function as a catalyst for both oxygen-reduction and oxygen-evo￾lution reactions. Nitrogen-doped RGO nanosheets have also been investigated in this respect. The defective sites introduced by nitro￾gen doping enable a more uniform and smaller size distribution of the discharge products and, therefore, a higher specific discharge capacity with respect to the undoped graphene75. Even though this technology is in the very early stages of development, the few reports available in this field depict a quite promising scenario for graphene in SABs. Graphene as an inactive component Graphene could play an important role in EESDs, even without being actively involved in the electrochemical reaction. Owing to its impressive electrical conductivity (Table 1), graphene was pro￾posed as a conductive agent in metal-ion battery electrodes as well as an encapsulating carbon matrix in, for example, lithium–sulphur batteries. Besides enabling efficient electron transport, its superior thermal conductivity (Table 1) may be advantageous for dissipating the heat generated in the case of high current loads or/and abuse conditions5 . This would result in devices with improved intrinsic safety. The variety of structures reported can be classified into six different models (Fig. 4a)6 . Lithium-ion batteries. LiCoO2, LiMn2O4 and LiFePO4 (hereafter referred to as LMO) are some of the most commonly used cath￾ode materials in LIBs. The cycle life and rate capability of these materials are generally limited by their poor electrical conductiv￾ity. Introducing low-cost conductive additives (for example, carbon black) into the composite electrodes commonly solves this issue. Nevertheless, owing to their amorphous structure, carbon blacks possess a rather low electrical conductivity, with respect to more crystalline carbons such as graphene5,76. Recently, Kucinskis  et  al.76 reviewed state-of-the-art graphene￾based composite cathode materials. Among the vast number of reports, most employ GO as a source for the formation of graphene conductive networks. Additionally, in a large part of these works, GO is reduced to RGO simultaneously with the LMO precur￾sors (one-pot synthesis) to produce graphene-based composites. This approach, which is different from simply mixing the carbon additive with the LMO active material during electrode preparation, promotes the formation of small-size LMO particles (leading to improved Li+ diffusion) directly onto the RGO matrix76. Moreover, the RGO 3D network is reportedly capable of preventing the disso￾lution of some LMOs39, thus extending the cycle life of the batteries. However, it was also suggested that when RGO is mixed in a manner similar to conventional carbon additives during electrode prepara￾tion, it could negatively affect the Li+ mobility, thus worsening the electrochemical performance of the composite cathode76 (Fig. 4b). Regardless of this fact, RGO is generally reported to enhance the rate capability of the cathode with respect to conventional carbon additives. Depending on the active material, improvements of up to 160% of the discharge capacity (at the same current rate) have been observed76. Nevertheless, it is not yet clear whether RGO may replace carbon blacks, which are much cheaper and easier to handle76. Sodium-ion batteries. As explained previously for the negative electrode, the larger size of Na+ with respect to Li+ restricts the choice of available cathode material for SIBs47. Several layered oxides with promising electrochemical performance have recently been devel￾oped50. However, like their LIB analogues, they usually possess poor electrical conductivity and thus limited rate capability. So far, only a few reports are available on graphene-containing composite cath￾odes for SIBs77–80. However, in all cases, the RGO matrix seems to enhance the electrical conductivity of the composite, thus improv￾ing the rate capability compared with bare cathode materials77–80. Lithium–sulphur batteries. Lithium–sulphur batteries (LSBs), through the redox reaction of metallic lithium (anode) and elemen￾tal sulphur (cathode), could provide a remarkably high theoretical specific energy of up to 2,600 Wh kg−1 (ref. 39). Despite the intrinsic advantages of sulphur in terms of low cost, abundance, low toxic￾ity and high theoretical specific capacity (1,672 mAh g–1), LSBs are affected by several drawbacks: (1) slow kinetics owing to the low electrical conductivity of the redox reaction products; (2) low energy efficiency; (3) poor cycle life as a direct result of the dissolution of the intermediate reaction products (polysulphides) in the electro￾lyte; and (4) large volume changes during the electrochemical reac￾tion69. Graphene has been proposed as a good candidate to address these issues because of its high electrical conductivity and capabil￾ity of trapping the charge/discharge products39. Several studies have reported that RGO and GO are suitable substrates for the deposi￾tion of sulphur micro- and nanoparticles39,69. Good encapsulation seems to be achieved in both cases; however, the presence of epoxy and hydroxyl groups in GO promotes the immobilization of sulphur and thus prevents its dissolution39. Promising performance — with specific capacities >1,000 mAh g–1 — has also been obtained with hybrid graphene–polymer–sulphur composites, although an accept￾able capacity retention during cycling has not yet been achieved41. Recently, different graphene–sulphur composites have been syn￾thesized and tested in LSBs81–83. They exhibit a good performance in terms of capacity, Coulombic efficiency and stability during cycling81–84, even if the values reported are not higher than those obtained with other carbon-encapsulated sulphur cathodes, such as ordered mesoporous carbon84. From these results, graphene might be a possible candidate for encapsulating sulphur on LSB cathodes. However, real advances in the field require improvements to the sulphur/graphene interface in order to achieve stable electrochemical performance39. Developing applications of graphene The recent outbreak of graphene in the field of electrochemical energy storage has spurred research into its applications in novel systems such as magnesium-ion batteries (MIBs), which is one of the newest members of the metal-ion battery family. MIBs have been proposed as a high-energy-density and environmentally friendly replacement for LIBs85. Although research in this field is still at an early stage, a few graphene-based composites have recently been proposed as MIB cathode materials86,87. Although the results obtained so far show very poor electrochemical behaviour, they could represent the first attempts to use graphene in MIBs. Interestingly, Wang et al. have already patented a rechargeable magnesium-ion cell based on graphene active materials88. The employment of graphene has also been considered for the improvement of vanadium redox-flow batteries (VRFBs). Patented in 198689, VRFBs enable energy storage using V3+/V2+ and [VO2+]/[VO2 +] redox couples as negative and positive acid electrolytic solutions, respectively90. Carbon-based materials, such as cloths or graphite felts, are used as electrodes because of their electrochemical stability and wide operating potential91. Unfortunately, despite their high SSAs, these electrodes do not exhibit satisfactory electrochemi￾cal properties92. Accordingly, the use of graphene-based materials has PROGRESS ARTICLE NATURE MATERIALS DOI: 10.1038/NMAT4170 © 2014 Macmillan Publishers Limited. All rights reserved