Graphene/polymer composites for energy applications



Graphene has wide potential applications in energy-related systems, mainly because of its unique atom-thick two-dimensional structure, high electrical or thermal conductivity, optical transparency, great mechanical strength, inherent flexibility, and huge specific surface area. For this purpose, graphene materials are frequently blended with polymers to form composites, especially when fabricating flexible devices. Graphene/polymer composites have been explored as electrodes of supercapacitors or lithium ion batteries, counter electrodes of dye-sensitized solar cells, transparent conducting electrodes and active layers of organic solar cells, catalytic electrodes, and polymer electrolyte membranes of fuel cells. In this review, we summarize the recent advances on the synthesis and applications of graphene/polymer composites for energy applications. The challenges and prospects in this field have also been discussed. © 2012 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys, 2013


Energy is one of the most important recent topics of concern to human society. To replace conventional fossil fuels, clean and renewable energies based on sunlight, wind, new chemicals, and biofuels are urgently demanded. Therefore, materials that can directly convert or store renewable energies are being extensively studied. Among them, graphene has recently attracted a great deal of attention because of its unique atom-thick two-dimensional (2D) structure1,2 and excellent electrical,2 thermal,3 optical, and mechanical properties.4,5 Graphene is a promising material for applications in various energy-related systems such as supercapacitors,6–9 secondary batteries,10–14 solar cells,15–17 and fuel cells.18–20 For this purpose, graphene materials are frequently blended with polymers to form functional composites.21–23 The polymer component can improve the processability and/or flexibility of graphene materials, and also possibly provide them with new functions. To date, various graphene composites with insulating or conducting polymers (CPs) have been prepared through noncovalent22 or covalent approaches.24 They have also been applied as electrodes for supercapacitors and lithium ion batteries (LIBs),25–29 counter electrodes of dye-sensitized solar cells (DSSCs),15,30,31 and transparent conducting electrodes (TCEs)32,33 or active layers of organic solar cells,34 catalytic electrodes, and polymer electrolyte membranes of fuel cells.35–37 This review focuses on summarizing the recent advances in the synthesis of graphene/polymer composites and their energy applications.


Graphene was first obtained by mechanical cleavage of graphite with Scotch tape.2 Although exfoliated graphene has an almost perfect structure with excellent properties, this method is not suitable for large-scale production. In addition to micromechanical exfoliation, epitaxial growth38–40 and chemical vapor deposition (VD)41–43 can also produce graphene with relatively high quality, and even single-crystal graphene. These graphene materials have potential applications in microelectronics and solar cells as transparent conductive electrodes because of their excellent electrical properties and superior optical transparency.17,42,43 However, the methods listed above still suffer from low yield, high synthesis temperature, and complex procedures to transfer the graphene sheets to desired substrates. Chemical synthesis of graphene using graphite,44 graphite oxide, or other graphite derivatives45,46 as the starting materials has also been developed. This technique can not only be scalable but can also provide graphene with processability and new functions. The resulting graphene materials are described as chemically converted (or modified or derived) graphene (CCGs).47 CCGs are the most widely used graphene materials for synthesizing graphene/polymer composites. However, they are different from perfect graphene, having many structural defects including holes and defective clusters on graphene basal planes and edges. Thus, their conductivities were reported to be in the scale of 10–1000 S cm−1, and these values are at least three orders of magnitude lower than that of ideal graphene. Nevertheless, in many cases, CCGs satisfy the requirements for energy applications.

The most common method for preparing CCGs is the reduction of graphene oxide (GO) by chemical or thermal reduction. GO is usually synthesized by exfoliation of graphite oxide, which is produced by the oxidation of graphite using a variety of oxidants.48–51 Hummers' method is the most classical technique, in which graphite oxide was produced by the oxidation of graphite in concentrated H2SO4 using KMnO4 and NaNO3 as oxidants.48 GO was then obtained by stirring or sonication of graphite oxide in aqueous media and it can be purified by dialysis. Changing the ratio of KMnO4 and graphite can considerably influence the degree of oxidation. Our group prepared mildly oxidized GO by decreasing the feeding mass ratio of KMnO4 and graphite.51 The CCGs prepared from this mildly oxidized GO showed much higher electrical conductivities than those of the counterparts synthesized by conventional Hummers' method. To enhance the yield of GO, a modified Hummers' method was also developed.49 In this case, preoxidation of graphite by K2S2O8, and P2O5 was conducted.

CCGs can be prepared from GO via chemical, thermal, or electrochemical reduction processes.52–54 Chemical reduction is one of the most important methods. Many chemical reducing agents, such as hydrazine monohydrate,52,55,56 sodium borohydride,57,58 and hydrogen iodide59 have been investigated to reduce GO. Among them, hydrazine monohydrate is the most widely used reducing agent; it can remove most oxygen-containing functional groups of GO and partially restore the conjugation areas of graphene sheets. However, the enhanced hydrophobic and π–π interactions make reduced GO (rGO) sheets heavily aggregate into nanoparticles.52 They cannot be dispersed and processed into desired structures or shapes by conventional material processing techniques such as molding, spin coating, and printing. To overcome this problem, Li et al. synthesized water-soluble rGO using hydrazine monohydrate as the reducing agent in ammonia solution with a pH value around 10.56 In this case, the carboxylic acid groups of rGO sheets were ionized and electrostatic repulsion increased their solubility. We also functionalized rGO sheets with pyrenebutyrate (PB) to improve their dispersibility.60 The large aromatic pyrene rings led PB molecules to be assembled on the surfaces of rGO sheets via π-stacking interaction and their butylate groups introduced a large amount of hydrophilic carboxyl groups. Consequently, the PB-functionalized rGO can be stably dispersed in water. A lightly sulfonated graphene (SG) was reported to be dispersible in an acidic aqueous medium.61 SG was synthesized through following procedures: first, GO was prereduced to remove the majority of its oxygenated functional groups; then, prereduced GO was sulfonated with aryl diazonium salt of sulfanilic acid; finally, sulfonated GO (SGO) was further reduced by hydrazine.

In addition to chemical reduction, thermal reduction is another effective approach to reduce GO.53 For example, Aksay's group has exfoliated and reduced graphite oxide sheets by heating them at 1050°C.53 Generally, high-temperature treatment of chemically rGO can further enhance the reduction reaction and increase its electrical conductivity. However, the CCGs prepared by treating GO at high temperatures could not be redispersed in solutions. Solvothermal reduction can avoid this drawback. Dubin et al. mixed GO aqueous solution into N-methyl-2-pyrrolidinone and used a solvothermal method to evaporate its water and reduce GO.62 The resulting CCG was tested to be redispersible in various polar organic solutions such as dimethylsulfoxide, ethyl acetate, acetonitrile, ethanol, tetrahydrofuran, dimethylformamide (DMF), chloroform, and acetone with minimal precipitation at 1 mg mL−1. Ruoff and coworkers synthesized stable CCG by heating GO in propylene carbonate (PC) solution at 150°C. Unlike the Dubin's method, GO can be directly exfoliated and dispersed in PC without adding water.63 Thus, the resulting CCG can be directly used in nonaqueous electronic devices.

Graphene/Polymer Composites

Graphene/polymer composites have attracted a great deal of attention because of their wide applications in high-strength and conductive materials, catalysts, and energy-related systems, especially flexible energy conversion and storage devices.15,22,35,64 Therefore, a variety of methods of preparing graphene/polymer composites have been developed. According to the interactions between graphene materials and polymers, these methods can be classified into two categories: noncovalent and covalent strategies.

Noncovalent Strategies

Noncovalent strategies include mixing7,65 and in situ polymerization.66,67 Mixing is the simplest approach to prepare graphene/polymer composites and it can be further subclassified into solution mixing68–71 and melting mixing.65,72 Solution mixing requires both graphene material and polymer to be stably dispersed in a common solvent. One of the most widely used precursors for preparing the graphene/polymer composites is GO because of its good dispersibility in water or polar organic solvents such as PC, DMF, and 1-methyl-2-pyrrolidinone. A GO/polymer composite can be easily formed by mixing the aqueous solutions of GO and a water-soluble polymer such as poly(vinyl alcohol) (PVA), poly(ethylene oxide) (PEO) or poly(sodium 5-styrensulfonate) (PSS).73 The GO component in the composite can be converted to conductive rGO upon chemical or thermal reduction. After removing the solvent, a solid rGO/polymer composite was obtainable. For example, we prepared a GO/PVA composite film by vacuum filtrating the uniform dispersion of GO and PVA [Fig. 1(A)].68 The structure of the resulting composite film has been examined by X-ray diffraction [Fig. 1(B)]. The disappearance of the diffraction associated with the interlamellar spacing of GO sheets suggests that the GO component was uniformly dispersed in the PVA matrix. A flexible film of graphene/polyaniline (PANI) nanofibers (NFs) with a layered structure was also prepared through the mixing and filtering approach; this film exhibited an excellent performance as the electrodes of supercapacitors.7 The well-ordered structure of this composite was formed by flow-directed assembly of the unique 2D structured graphene sheets.

Figure 1.

(A) Photograph of a 120-μm-thick GO/PVA composite film. (B) XRD patterns of GO, PVA, and GO/PVA composite. Reproduced from ref. 68, with permission from Elsevier B.V.

Solution mixing can also be carried out in organic media.22,74 Modification of CCGs with small organic molecules or polymers can improve their dispersibility in organic solvents. For example, isocyanate-modified GO can be stably dispersed in DMF,22 and polystyrene was added to its dispersion under stirring to form a mixture. After the reduction of the GO component with dimethylhydrazine, a graphene/polystyrene composite was coagulated by adding methanol dropwise. A free-standing composite film can also be obtained by vacuum filtration of the coagulated composite powder.

Solution mixing techniques cannot be applied to prepare uniform composites as the graphene component has a strong interaction with the polymer. In this case, aggregation of both graphene and polymer would occur. Furthermore, the mixtures of graphene derivatives and polymers frequently tend to form gels if their concentrations are sufficiently high.75 On the other hand, in melt mixing process, graphene derivatives are blended into molten polymer matrixes under dramatic shearing. Thus, there is no requirement to predisperse the graphene and polymer components in solution. For instance, Kim et al. prepared a graphene/thermoplastic polyurethane (TPU) composite by directly feeding thermally rGO and TPU into a recirculating, conical twin-screw extruder at 180 °C and blended under dry nitrogen for 6 min at a screw speed of 360 rpm.76

In situ polymerization method involves the polymerization of monomer in a system containing the graphene material. Many graphene/polymer composites, such as graphene/nylon667 and graphene/CPs,64,77,78 can be prepared using this approach. The morphology of the graphene/polymer composite can be controlled by tuning the polymerization conditions. The typical example is the synthesis of a hierarchical nanocomposite of PANI nanowire arrays on GO sheets by dilute polymerization of aniline monomer in GO aqueous solution at 10°C [Fig. 2(A,B)].64 The morphologies of GO/PANI nanocomposites were heavily affected by the ratio of aniline monomer to GO. At a given concentration of GO (0.36 mg mL−1), the optimized concentration of aniline was tested to be 0.05 M. A lower concentration of aniline (e.g., 0.04 M) resulted in the growth of sparser and shorter PANI nanowires than those obtained at 0.05 M. When the concentration of aniline was increased over 0.06 M, both randomly connected PANI nanowires and aligned PANI nanowires were produced. The formation mechanism of GO/PANI nanocomposites is illustrated in Figure 2(C). In the chemical oxidative polymerization process, two possible nucleation sites (heterogeneous nucleation and homogeneous nucleation) compete with each other. As a dilute aniline solution was used (less than 0.05 M), most active nucleation sites were generated on GO nanosheet surfaces at the beginning of polymerization through heterogeneous nucleation. However, when the concentration of aniline was higher, homogeneous nucleation will take place after the initial nucleation on the solid surface. Consequently, homogeneous nucleation will produce randomly connected PANI nanowires.

Figure 2.

(A, B) SEM images of GO/PANI nanocomposites. (C) Schematic illustration of nucleation and growth mechanism of PANI nanowires: (a) heterogeneous nucleation on GO nanosheets; (b) homogeneous nucleation in bulk solution. Reproduced from Ref. 64, with permission from American Chemical Society.

Among the composites of graphene/polymer, graphene/CPs can be produced not only by in situ chemical polymerization but also by in situ electrochemical polymerization.79,80 Different from the powdery products prepared through chemical approaches, electrochemical polymerization yields mechanically stable composite films and they can be directly used as the electrodes of energy storage devices. Furthermore, electrochemical polymerization can be precisely controlled by the applied potential, current density, and polymerization time. Wang et al. synthesized graphene/PANI paper by in situ anodic electrochemical polymerization of the adsorbed aniline monomers on the surfaces of graphene sheets to form PANI between neighboring graphene sheets.80 After polymerization, the graphene paper still maintained its layered structure. The mechanical and electrochemical properties of the composite can be modulated by simply adjusting the polymerization time because that changes its PANI content. In many cases, nanostructures of CPs can be grown along the direction of the electric field to form oriented structures. For example, highly aligned PANI nanorod arrays were grown on patterned rGO electrodes (Fig. 3), making them potentially useful in nanoelectronics.81 In this process, the polymerization was performed in an aqueous solution of aniline (0.05 M) and H2SO4 (0.5 M) at a potential 0.75 V versus Ag/AgCl. The PANI nanorods grown on patterned rGO electrodes have smaller diameters and smoother surfaces than those grown on gold electrodes. The morphology of PANI nanorod arrays, including the length and diameter of single nanorod and the surface roughness of rGO/PANI composites, can also be modulated by controlling the polymerization time. The ordered morphologies of the GO/PANI and rGO/PANI composites described above are mainly due to the preferentially adsorption of aniline monomers to graphene sheets through π–π interaction to form PANI nuclei during polymerization.

Figure 3.

(A) Fabrication of rGO/PANI microelectrodes. (B, C) SEM images of PANI nanorods on rGO pattern. Reproduced from ref. 81, with permission from Wiley-VCH.

Covalent Modification

Recently, several excellent reviews have summarized the functionalization of graphene sheets with polymers via covalent approaches.73,82–84 Here, we just briefly introduce this technique. In cases of covalent modifications, polymer chains are stably grafted to graphene sheets and surround them to prevent their aggregation caused by π-stacking. Two main approaches have been developed to covalently modify graphene materials with polymers: “graft to”85 and “graft from.”82 “Graft to” is chemically connecting polymer chains to the surfaces of CCG or GO sheets. In this case, the functional groups of polymer chains react with those of graphene sheets to form chemical connects. For example, the carboxylic groups of GO can react with hydroxyl or amine groups of a polymer via esterification85 or amidation reactions86,87 to form graphene/polymer composites. Click cycloaddition was also used to graft PS chains onto CCG sheets.88 “Graft from” refers to immobilizing initiators onto graphene sheets to initiate polymerization. Atom transfer radical polymerization (ATRP) is the most widely studied method for this purpose, and it has been used for synthesizing a graphene/PS.24 In this process, CCG sheets were first prepared by the reduction of well-exfoliated graphite oxide with the aid of a surfactant (sodium dodecylbenzene sulfonate, SDBS). Then, aryl diazonium salt as an ATRP initiator was covalently linked to CCG sheets. PS chains were then grown from the surfaces of graphene sheets by AIRP polymerization. Methyl 2-bromopropionate was used as a grafting initiator to control the chain propagation on the surfaces of graphene sheets (Fig. 4). The length and grafting density of polymer chains were controlled by tuning the amount of the grafting initiator and its molar ratio with styrene monomer.

Figure 4.

Synthetic routes for achieving controllable functionalization of graphene/PS composites. Reproduced from ref. 24, with permission from Royal Society of Chemistry.


As discussed above, graphene/polymer composites can be synthesized by mixing, in situ polymerization, and covalent modification. These composites have or combine the high electrical, mechanical, optical properties, and huge specific area of graphene materials with the multiple functions of the polymer components. As a result, they have been explored for the applications in various energy-related systems. Table 1 lists the most graphene/polymer composites, and their electrical conductivity as well as energy applications discussed in the following sections.

Table 1. Synthesis, Electrical properties, and Applications of Graphene/Polymer Composites
inline image

Lithium Ion Batteries

LIBs are the most attractive energy storage devices because of their high energy and power densities, long cycling life, and good environmental compatibility.115,116 The traditional cathode materials of LIBs used in lithium-ion batteries are inorganic materials, such as LiCoO2, LiMn2O4, and LiFePO4.116–119 However, they suffer from limited capacities and nonrenewable mineral resources. As an alternative approach, polymeric cathode materials have several inherent advantages including lightness, environmentally friendly characteristics, mechanical flexibility, and processing compatibility. Furthermore, they are promising for applications in flexible LIBs. However, polymeric cathode materials suffer from poor electrical conductivities and slow redox reactions. Thus, highly conductive graphene materials have been explored as effective nanofillers to improve the performances of polymer cathodes.

Many polymers have been investigated as the cathode electrodes of LIBs, including conductive polymers,120,121 poly (organosulfur),122 and radical polymers.123,124 Among them, CPs have attracted the most intense attention, mainly because of their inherent conductive and electroactive properties. However, the mechanical strengths and conductivities of CPs still need to be improved. More seriously, these polymers are insulating in their neutral states. The poor charge/discharge rate capability and limited cycle stability of CP-based electrodes also limit their practical applications in LIBs. The introduction of carbon nanomaterials such as carbon nanotubes (CNTs) can improve the properties of CP electrodes. Recently, graphene has also been investigated as the conducting additive. The 2D carbon nanostructures and high electrical conductivities of graphene materials mean they can easily form 3D porous structures, which can act as conductivity networks in CP electrodes. For example, polypyrrole (PPy)/rGO composite films were synthesized by electrodeposition and explored as the cathodes of LIBs.29 In this work, PPy/GO film was galvanostatically electrodeposited on a stainless steel mesh from an aqueous solution containing GO (0.5 wt %) and 0.1 M pyrrole monomer. This film was subjected to electrochemical reduction to form PPy/rGO in PBS (pH = 7.4) at a constant potential of −1.1 V (vs. Ag/AgCl) for 40 min. PPy/sodium p-toluenesulfonate (PPy/pTS) was also synthesized for comparison. PPy/pTS exhibited a cauliflower morphology composed of large nodules as reported previously [Fig. 5(A)], whereas a wrinkled or crumpled morphology was observed for the PPy/rGO composite film [Fig. 5(B)]. Cross-sectional scanning electron micrographs revealed that the PPy/rGO composite film has a porous structure [Fig. 5(D)] in contrast to the solid, compact structure of PPy/pTS composite film [Fig. 5(C)]. The improved conductivity of the PPy/rGO composite was indicated by electrochemical impedance spectroscopic results when compared with that of the PPy/pTS counterpart. The rate capability of the electrode based on the PPy/rGO composite was also superior to that of the PPy/pTS electrode [Fig. 5(E,F)]. The PPy/rGO electrode exhibited a slightly higher discharge capacity (74 mA h g−1 at 1 C and 70 mA h g−1 at 2 C) than that of the PPy/pTS electrode (62 mA h g−1 at 1 C and 57 mA h g−1 at 2 C) at low current rates. The PPy/rGO electrode showed 74% capacity retention as the discharging rate increased from 1 C to 20 C. The high performance of the PPy/rGO composite when used as the cathode material in LIB is mainly attributed to its porous structure and the highly conductive rGO component.

Figure 5.

(A–D) Field-emission scanning electron microscope (FESEM) images of surface morphology and cross-sectional views of PPy/pTS (A, C) and PPy/rGO films (B, D). Electrochemical data for PPy/rGO and PPy/pTS films in LIBs in the potential window of 2.0–4.0 V (vs. Li/Li+). (E) Voltage profiles for PPy/pTS film at different current rates. (F) Voltage profiles for PPy/rGO film at different current rates. Reproduced from ref. 29, with permission from Wiley-VCH.

In addition to CPs, poly(organosulfur) and radical polymers have also been studied as cathode materials for LIBs.25,28 These polymers are insulators with electroactive properties. Thus, their conductivity and electrochemical properties can be improved by blending with conductive graphene sheets. Among the radical polymers, poly(2,2,6,6-tetramethylpiperidinyloxy-4-yl methacrylate) (PTMA) was first studied as the cathode of LIBs, exhibiting relatively high capacity, high rate performance, and long cycling life.122 Graphene/PTMA composite materials were prepared through a dispersing–depositing process, in which graphene served as a conductive support.28 This composite displays a reversible two-electron process redox reaction and improved reversible capacity of 222 mA h g−1 at 1 C, and this value is about twice that of pristine PTMA (100 mA h g−1). The capacity of the composite electrode was preserved around 100 mA h g−1 after 20,000 cycles of charging/discharging. These results show that the graphene component greatly improves the performance of radical polymers in LIBs.

Polymers can also be used as stabilizers for growing nanostructures on graphene sheets. For example, Guo et al. synthesized nanocomposites of layered birnessite-type manganese oxide/poly(3,4-ethylenedioxythiophene)/graphene (LMO/PEDOT/graphene) [Fig. 6(A)].89 In this composite, graphene/PEDOT was first synthesized by in situ polymerization and then used as the supporting matrix to guide the growth of uniformly distributed rods of LMO.

Figure 6.

(A) Schematic representation of the hierarchically nanostructured composite fabrication. (B) Capacities of LMO/PEDOT/graphene batteries measured in different charging/discharging cycles. (C) Cycling variation in charge/discharge capacity versus cycle number for different lithium ion batteries. (D) SEM image of the LMO/PEDOT/graphene. Reproduced from ref. 89, with permission from Wiley-VCH.

Manganese oxide (MnO2) is also a promising cathode material for LIBs, which possesses a high theoretical capacity, and is environmentally benign and naturally abundant. LMO consisting of 2D monolayers of edge-shared MnO6 octahedra between which Li+ can be intercalated has been shown to be particularly promising.125,126 However, pure MnO2 material showed poor performance in batteries because of its low electrical conductivity and large volume expansion during the charging/discharging process. The introduction of a graphene component can solve these problems. Nevertheless, Guo et al. also found that clean graphene was not an effective matrix for LMO growth. Theoretical analysis revealed that the distance between two neighboring manganese (or oxygen) atoms in face (001) of MnO6 octahedra is 2.94 Å, whereas the largest distance between carbon atoms in a graphene hexagon carbon ring is only 2.84 Å. Thus, it is difficult for LMO to form nuclei on clean graphene sheets. In comparison, the distance of two oxygen atoms in the six-atom ring of PEDOT is around 2.94 Å, equal to the distance between two manganese atoms in the (001) face of MnO6 octahedra in LMO.127,128 Furthermore, the affinity of manganese atoms for oxygen atoms is stronger than that for carbon atoms because of their native electron characteristics. Thus, the PEDOT component induced the formation of MnO2 nanostructures on graphene sheets. The PEDOT coating not only directed the growth of relatively ordered textured LMO but also enhanced the stability of the resulting LIBs.

A battery based on the LMO/PEDOT/graphene composite showed a discharge capacity of 1835 mA h g−1 in the first cycle at a discharge rate of 50 mA g–1, and a reversible capacity of 1105 mA h g−1 in the second cycle [Fig. 6(B)]. After 15 cycles, it retained a capacity as high as 948 mA h g−1. For comparison, the battery based on LMO exhibited a first discharge capacity of 1249 mA h g−1 at 50 mA g−1, but it dropped sharply to only 71 mA h g−1 after 15 cycles [Fig. 6(C)]. The improved capacity and stability of the composite electrode is mainly due to the graphene/PEDOT composite preventing the aggregation of LMO. Thus, the composite can accommodate a large volume expansion during battery operation for reversible capacity. Similarly, PANI was used to stabilize SnO2 anchored to rGO sheets, and LIBs based on PANI/SnO2/rGO composites showed excellent cycling stability and high capacity with a reversible storage capacity of 573.6 mA h g−1 and a coulombic efficiency of 99.26% after 50 cycles.26

A three-component composite of PEO, molybdenum disulfide, and graphene (PEO/MoO2/graphene) has been prepared by the hydrolysis of lithiated MoO2 in an aqueous solution of PEO and rGO. The rGO sheets and PEO bestow this composite with electrical and ionic conductivities, respectively.27 Compared with the LIBs based on pure MoO2, the LIBs based on the composite exhibited an increased reversible capacity and improved rate performance.


Supercapacitors are important energy storage devices that can be charged/discharged at rapid rates.129,130 A high-performance supercapacitor should have high energy density (1–10 W h kg−1, determined by its capacitance and voltage), power density (103–105 W kg−1, determined by its voltage and internal resistance), and ultra-long cycling life (>100,000 cycles).131 Thus, supercapacitors act as a perfect complement for batteries or fuel cells, and used in cooperation they are considered to be promising power supplies for versatile applications such as environmentally friendly automobiles, artificial organs, and high-performance portable electronics. Supercapacitors can be classified by their storage mechanisms into two types: electrical double-layer capacitors (EDLCs) and pseudo-capacitors.132 EDLCS store energy by the accumulation of electrostatic charges at the interfaces of electrodes and electrolytes. Therefore, the performances of EDLCs strongly depend on their architecture and the specific surface areas of their electrodes.130 A pseudo-capacitor works predominately through fast reversible redox reactions.132 Compared with EDLCs, pseudo-capacitors have much larger specific capacitances, but usually possess lower charge/discharge rates and worse cycling stability.

Graphene/CP Composite-Based Supercapacitors

CPs are typical pseudocapacitive electrode materials. Among them, PANI and PPy have been paid the most intensive attention because of their fast electrochemical reactions, low cost, and high specific capacitances.7,80,90,133–135 However, CP electrodes exhibit large volume changes during the doping/dedoping processes, so the electrodes degrade and the electrochemical performance diminishes during cycling tests.130 Adding graphene to the polymers can improve the mechanical and electrochemical cycling stability of CP electrodes.

Graphene/PANI Composites

In Situ Polymerization

In situ polymerization of aniline in a graphene dispersion can produce graphene/PANI composites with a sandwich structure. PANI prefers to grow on the surface of graphene sheets rather than in solution, probably because of the preferential adsorption of aniline molecules on graphene sheets via electrostatic, π–π stacking, and hydrogen bonding interactions.65 Wang et al. used GO sheets as the substrate for growing PANI.79 The PANI grown on GO sheets has a fibril morphology with diameters around 300 nm, whereas pure PANI synthesized in the system without GO is a granular aggregate. Although GO has poor electrical conductivity, the GO-doped PANI had a conductivity of 10 S cm−1, higher than that of pristine PANI (2 S cm−1). A supercapacitor based on the GO/PANI composite exhibited a high specific capacitance of 531 F g−1 when measuring in the potential range of 0–0.45 V at the current density of 200 mA g−1 with a three-electrode method. This value is much higher than that of PANI-based supercapacitor (216 F g−1) measured under the same conditions.

To further improve the electrical conductivity, rGO was used to replace GO to form a rGO/PANI composite.92 In this composite, PANI particles with diameters around 2 nm were coated on the surfaces of graphene sheets. The specific capacitance of the rGO/PANI composite-based supercapacitor was measured to be as high as 1046 F g−1 using a three-electrode method at a scan rate of 1 mV s−1; this value is about one order of magnitude higher than that of the device based on pure PANI (115 F g−1). Li et al. also synthesized graphene nanosheet/PANI (GNS/PANI) NF composites via in situ polymerization of aniline monomer in HClO4 solution.93 The PANI NFs were homogeneously coated on the surfaces of GNS. The GNS/PANI composite exhibited better electrochemical performance than that of either of its individual components. Its specific capacitance was measured to be 1130 F g−1 using cyclic voltammetric scanning at a scan rate of 5 mV s−1 in 1 M H2SO4 solution. This value is much higher than 402 F g−1 for pure PANI and 270 F g−1 for GNS.

Postreduction of GO/CP composites is an effective method for preparing uniform rGO/CP composites because it avoids the restacking of rGO sheets. Zhang et al. made a homogenous aqueous solution of aniline and GO.90 After the polymerization, GO sheets in the composite were reduced by hydrazine. The specific capacitance of the supercapacitor was measured to be 480 F g−1 at a current density of 0.1 A g−1. A similar rGO/PANI composite was prepared using ethylene glycol to replace water as the reaction medium and then the GO component was reduced by hot sodium hydroxide solution.94 The resulting rGO/PANI composite showed a high specific capacitance of 1126 F g−1 at a CV scan rate of 1 mV s−1.

Sulfonated graphene can be stably dispersed in acidic water and form a uniform composite with PANI by in situ polymerization.95 The composite exhibited a specific capacitance of 763 F g−1 with capacity retention of 96% after a 100-cycle charging/discharging test and good rate performance. Surfactant (tetrabutylammonium hydroxide and sodium dodecyl benzenesulfonate)-stabilized graphene (SSG) is also used in the preparation of graphene/PANI NF composites.96 This composite material was used as supercapacitor electrode and a high specific capacitance of 526 F g−1 was also obtained at a current density of 0.2 A g−1.

The performance of supercapacitor electrodes depends not only on the electrical conductivity of electrode materials but also on their specific surface areas. Therefore, a variety of graphene/PANI nanostructures have been produced to increase the specific surface areas of the electrodes and optimize ionic transport pathways.64,92 Yan et al. synthesized a composite of graphene sheets coated with PANI nanoparticles (2 nm) on both sides of the sheet.92 The graphene sheets served as a charge-transfer layer in the middle of the composite. A high specific capacitance of 1046 F g−1 (based on GNS/PANI composite) was obtained at a scan rate of 1 mV s−1 compared to 115 F g−1 for pure PANI. Xu et al. constructed a hierarchical nanocomposite with PANI nanowire arrays aligned vertically on a GO substrate.64 The specific capacitance of this GO/PANI composite electrode was measured to be 555 F g−1 at a discharge current density of 0.2 A g−1 and it stayed as high as 227 F g−1 even at a discharge current density of 2 A g−1.

Electrochemical Polymerization

Graphene/PANI composites prepared through the chemical routes described above are usually powdery. Therefore, they have to be blended with a polymer binder such as polytetrafluoroethylene to enhance their mechanical strength for application as electrodes of supercapacitors. In contrast, electrochemical polymerization can produce mechanically stable composite films.80,81 Furthermore, the electrochemical polymerization reaction can be precisely modulated using electrochemical parameters such as applied potentials and current densities.81 In many cases, this method can easily synthesize nanostructures grown along the direction of the electric field to form oriented nanostructures. Wang et al. first synthesized free-standing and flexible graphene/PANI composite paper, which exhibited a specific capacitance of 233 F g−1.80 However, this graphene paper was prepared by filtration, which would greatly suppress the specific surface of graphene. A modified synthesis method was proposed by Feng et al.97 In this case, the graphene/PANI composite was prepared using a GO suspension and aniline as the starting materials for electrodeposition; the resulting composite's specific capacitance was measured to be 640 F g−1 with a retention of 90% after 1000 charge/discharge cycles.

A micro-supercapacitor was fabricated by electrodepositing PANI nanorods on the surfaces of rGO patterns.81 This micro-supercapacitor possesses a high specific capacitance of 970 F g−1 at a relatively high discharge current density of 2.5 A g−1 with a retention of 90% after 1700 cycles. This is mainly because the vertical PANI nanorod arrays were susceptible to strain relaxation, which reduced their structural degradation during the doping/dedoping process. Moreover, rGO nanosheets can also resist mechanical deformation caused by the redox reaction of PANI nanorods, which avoids the destruction of the electrode material.

Solution Mixing

To date, only a few graphene/PANI composites have been prepared by solution mixing, because CPs are usually insoluble. We prepared an rGO/PANI NF (G-PNF) composite with a sandwich-like structure using this technique [Fig. 7(A,B)].7 In this case, the rGO sheets were negatively charged and polyaniline nanofibers (PANI-NFs) were positively charged, so the two components can be self-assembled using electrostatic interactions into a uniform composite film by filtration. The conductivity of the composite film containing 44% rGO (5.5 × 102 S m−1) is about one order of magnitude higher than that of a neat PANI film. Supercapacitor devices based on this conductive flexible composite film showed a specific capacitance of 210 F g−1 at a discharge rate of 0.3 A g−1. Furthermore, the electrochemical stability of this supercapacitor is much better than that of the devices based on pure PANI-NF, because the graphene framework supported the PANI-NFs, preventing the fibers from severely swelling and shrinking during cycling.

Figure 7.

(A,B) Cross-sectional SEM images of a rGO/PANI-NF composite film containing 44 wt % rGO. Insert: digital photograph of a flexible rGO/PANI-NF composite film. (C) Cyclic voltammograms (scan rate = 5 mV s−1) and (D) galvanostatic charge/discharge curves (charging/discharging current density = 0.3 A g−1) of the supercapacitors based on G-PNF, as-formed PANI-NF, and CCG films. Reproduced from ref. 7, with permission from American Chemical Society.

A composite of GNS/CNT/PANI has also been fabricated via filtration of a suspension of GO and PANI/CNT.98 The PANI/CNT component was preprepared with a one-dimensional structure of a CNT core and PANI skin. As a result, the GNS/CNT/PANI composite exhibited a specific capacitance of 1035 F g−1 (1 mV s−1) in 6 M of KOH, which is a little lower than that of GNS/PANI composite (1046 F g−1), while much higher than those of pure PANI (115 F g−1) and CNT/PANI composite (780 F g−1).

Graphene/PPy Composites

PPy is also an appropriate electrode material for applications in supercapacitors because of the water solubility of the pyrrole monomer as well as the much lower carcinogenic risks associated with its degradation products compared with PANI.136–138 However, the low cycling stability and poor rate performance of supercapacitors based on pristine PPy drastically limit their practical applications. To improve the performance of PPy-based supercapacitors, we synthesized composite films of SG and PPy through an electrocodeposition process.99 In this case, SG and pyrrole were dispersed in an aqueous electrolyte containing dodecylbenzene sulfonic acid to stabilize SG sheets and improve the ionic conductivity of the electrolyte. Upon electrochemical deposition at a constant potential of 0.7 V (vs. SCE), PPy and SG sheets were codeposited on the surface of the working electrode to form composite films with different morphologies by controlling the total charge densities passing through the electrochemical cell. After polymerization at 0.5 C cm−2, a relatively compact film of the PPy/SG composite was generated. However, semimicrospheres were formed on the composite film when the charge density was increased to 1 C cm−2 [Fig. 8(A)]. In the case of polymerization charge density at >2 C cm−2, the microspheres disappeared and a porous film was produced [Fig. 8(B)]. The formation of the microspheres follows a gas-bubble template mechanism.139,140 Upon increasing charge density, the amount of the semimicrospheres increased, and they finally connected to each other to form a porous film. This porous film exhibited a high specific capacitance of 285 F g−1 at a discharge rate of 0.5 A g−1 [Fig. 8(C)] and greatly improved electrochemical stability and rate performance [Fig. 8(D)]. Wang et al. used a similar method to prepare a rGO/PPy composite for fabricating supercapacitors.100 The specific capacitance of this composite electrode was measured to be 224 F g−1 at a high charge/discharge current density of 240 A g−1. This is mainly due to rGO being embedded in the composite and the resulting formation of a porous structure and a high protonation level on PPy rings. This enhanced the ionic and electronic transport in the composite.

Figure 8.

(A) SEM image of PPy/SG composite film grown at 1 C cm−2. (B) Cross-sectional SEM image of the PPy/SG composite film grown at 2 C cm−2. (C) Galvanostatic charge/discharge curves of (a) a PPy film grown at 2 C cm−2, the PPy/SG composite films grown at (b) 0.5, (c) 2, and (d) 4 C cm−2, respectively, and (e) SG film in the aqueous solution of 1 mol L−1 KCl. Charging/discharging current density = 0.5 A g−1. (D) Cycling stability test of the PPy or PPy/SG composite film grown at 2 C cm−2 upon charging/discharging in an aqueous solution of 1 mol L−1 KCl and at a current density of 2 A g−1. Reproduced from ref. 99, with permission from American Chemical Society.

A flexible free-standing GO/PPy composite film without any supporting electrolyte has also been prepared by electrocodeposition.101 In this system, GO served as a weak electrolyte as well as an effective charge-balancing dopant for PPy. The specific capacitance of this film was tested to be 356 F g−1 at a discharge rate of 0.5 A g−1 and this value is 50% higher than that of pure PPy.

Although electrochemical polymerization has many advantages as described above, in situ polymerization can be more easily scaled up for practical applications. A graphene/PPy composite was prepared via in situ polymerization of pyrrole in the aqueous dispersion of rGO.102 The specific capacitance of this composite was measured to be 482 F g−1 at a current density of 0.5 A g−1. The polymerization of pyrrole should be carried out in an acidic medium, whereas the dispersion of rGO is stable only when pH ≥ 10. Thus, surfactants were frequently used to stabilize the graphene dispersions. Zhang et al. synthesized a sandwich-structured rGO composite with PPy fibrils or spherical particles by controlling the shape of surfactant micelles (cetyltrimethyl ammonium bromide).103 The composite showed a high specific capacitance of over 500 F g−1 at a current density of 0.3 A g−1 and a 70% capacitance retention after 1000 cycles, whereas the pure PPy has a specific capacitance of only 360 F g−1 at the same current density and only 30% capacitance retention after 1000 cycles.

Bose et al. prepared poly(sodium 4-styrenesulfonate)-modified graphene/PPy composites using in situ polymerization.104 However, the specific capacitance of this composite was tested to be only about 267 F g−1 at a scan rate of 100 mV s−1.

Self-assembly has also been applied for the fabrication of GO/PPy composites.105,106 A graphene/PPy nanowire composite film with a porous nanostructure was prepared by multilayer depositon.105 The supercapacitor based on this composite film displayed symmetric charge–discharge characteristics and a nearly ideal rectangular cyclic voltammogram in the scan rate range of 10–100 mV s−1. In another case, water-dispersible CNT/PPy core-shell NFs were prepared by chemical polymerization of pyrrole in a dispersion of CNTs.106 These were then used to fabricate a flexible film of graphene/CNT/PPy using flow-directed assembly of graphene sheets in a dispersion of the CNT/PPy composite. The specific capacitance of this three-component composite was measured to be 211 F g−1 at a CV scanning rate of 0.2 A g−1 with an excellent cycling stability (95% retention after 5000 cycles) because the strain of the PPy chains was released by the flexible graphene sheets.

Graphene/PEDOT Composites

Compared to PANI and PPy, PEDOT has been less investigated for composites with graphene materials as supercapacitor electrodes. Multilayered graphene/PEDOT thin films were prepared by alternating electrochemical deposition of graphene and PEDOT, and this composite showed a specific capacitance of 154 F g−1 and a capacitance retention of 86% after 1000 cycles.107

One work also compared graphene composites of PANI, PPy, or PEDOT.91 rGO/PANI exhibited a specific capacitance of 361 F g−1 at a current density of 0.3 A g−1. The composites consisting of rGO and PPy or PEDOT displayed specific capacitances of 248 or 108 F g−1 at the same current density. Moreover, the composites showed improved electrochemical stability. Taking the rGO/PANI composite as an example, it kept about 82% of its initial capacitance after a 1000 cycle test, which is much better than that of pristine PANI fibers (68% after 600 charge/discharge cycles).

Solar Cells

Solar cells can directly convert solar energy to electrical power. Inorganic solar cells, especially silicon-based solar cell, have been industrialized for a long time because of their high power conversion efficiencies (PCEs). However, they still suffer from high costs and serious pollution problems. Organic solar cells with heterojunction structures and DSSCs are regarded as promising alternatives. As a novel and unique member of the carbon nanomaterials family, graphene is an attractive material for fabricating transparent electrodes27,43,141 and improving the performances of dye-sensitized15,142,143 and heterojunction solar cells.108–110

Transparent Conducting Electrodes

TCEs are a key component of optoelectronic devices. The ideal TECs should possess high transparency (>80%), low resistance (<100 Ω/□), and an appropriate work function (4.5–5.2 eV).144 The cost of TCEs is also an important issue for considering their practical applications. Currently, indium tin oxide (ITO) is the most widely used material for fabricating TCEs, and it performs well in the aforementioned aspects. However, ITO also suffers from some disadvantages such as high production costs, limited indium resources, and ion diffusion into polymer layers.145 In particular, ITO cannot be used for flexible devices because it is mechanically rigid and brittle. Several well-developed high-performance ITO alternatives, such as single-walled CNTs or metal nanowires, are also expensive.146 Therefore, new alternative materials for TCEs are urgently required for the development of high-performance solar cells.

Graphene, with its high theoretical electrical conductivity, transparency (97.7%), and flexibility, is an emerging candidate for next-generation TCEs, especially for ultra-thin or flexible photovoltaic devices.4,147 Transparent graphene thin films can be prepared using a variety of techniques including micromechanical exfoliation,2 epitaxial growth,148 chemical VD,42 and GO reduction at high temperatures.47,149,150 In particular, TCEs based on rGO can be fabricated via cheap solution-based methods, and their work functions can also be engineered via chemical doping or structural engineering. However, these TCEs usually have low conductivities because of the structural defects of each rGO sheet and the contact resistance between neighboring sheets.

Two strategies can be applied to decrease the contact resistances between graphene sheets. One is the synthesis of rGO sheets with large sizes. A TCE based on large-size rGO showed a resistance of 840 Ω/□ (transmittance = 78%), which is much lower than that based on small sheets (19.1 kΩ/□, at T =79%).151 The other strategy is bridging graphene sheets with other conducting additives, such as CNTs and CPs.32,33,69,152–154 A composite film of graphene and CNTs showed an enhanced performance of 636 Ω/□ at T = 92%, which is superior to both single-walled CNT and graphene-based TCEs.152

Composites of graphene and CPs are also considered to have synergic effects.32,33,69 CP themselves are also promising materials for developing new TCEs, especially for potential applications in the flexible electronic devices, because of their unique optical, electrical, and mechanical properties. The composite of PEDOT and PSS is the most widely explored CP material for this purpose because of its good processibility, high conductivity, and transparency. The introduction of PEDOT:PSS is also helpful to reduce the contact resistance between graphene sheets as well as the solution processability of CCGs.

TCEs based on the SG/PEDOT composite prepared by in situ polymerization showed a conductivity of 0.2 S cm−1 and transmittances higher than 80% in the wavelength range of 400–1800 nm.32 This conductivity is much higher than that of a commercial PEDOT:PSS product (Clevios™ P AI 4083, 10−6–10−5 S cm−1). Moreover, when a PMMA sheet coated with this composite was bent inward, it still retained a high electrical conductivity (0.18 S cm−1). As the film was bent outward, its conductivity became 0.13 S cm−1. Finally, the film became flat again after releasing the bending force, and its conductivity returned to around the original value (0.10 S cm−1). Similarly, Chang et al. prepared TECs by just spin coating a mixed solution of SDBS-functionalized graphene and PEDOT:PSS [Fig. 9(A)].69 The surfactant-modified graphene sheets are uniformly distributed in the PEODT:PSS matrix [Fig. 9(B)]. As the weight ratio of SDBS–graphene to PEDOT:PSS increased from 0 to 1.6%, the sheet resistance of this TCE decreased from 227 ± 15 to 80 ± 10 Ω/□ (3 × 104−1 × 105 S m−1) [Fig. 9(C)]. The sheet resistance of 80 ± 10 Ω/□ is lower than that of the commercially available ITO/PET (120 Ω/□). The transparency of the graphene/PEDOT–PSS (weight ratio = 1.6%) composite film was measured to be about 79% at 550 nm, which is only about 5% lower than commercially available ITO/PET (about 84%). More importantly, this electrode can be bent over 1000 cycles with only a 5% increase in its resistance. However, the presence of extra stabilizer is undesirable for most applications. Jo et al. directly synthesized an aqueous graphene suspension stabilized by PEDOT:PSS through the chemical reduction of GO in the presence of PEDOT:PSS without additional stabilizer. The film deposited on PET exhibited a conductivity of 2.3 kΩ/□ with a transmittance of 80% at 550 nm. Moreover, its conductivity was sustained at nearly 100% after 100 bending cycles.33

Figure 9.

(A) Digital pictures of SDBS-graphene/PEDOT:PSS (1.6 wt %) composites on PET substrates as TCEs. (B) SEM of SDBS-graphene/PEDOT:PSS (1.6 wt %) composites on PET substrates. Inset to (B): cross-sectional SEM image of SDBS-graphene/PEDOT:PSS composites on PET. (C) Transparency of TCEs with different doping concentrations of SDBS-graphene (0, 0.2, 0.4, 0.8, and 1.6 wt %) and commercial ITO/PET electrodes. Reproduced from ref. 69, with permission from Wiley-VCH.

Dye-Sensitized Solar Cells

DSSCs have received great attention because of their low cost, convenient manufacturing processes, and comparable efficiencies to those of solid-state silicon solar cells. Each DSSC consists of a working electrode of mesoporous dye-sensitized titania nanocrystals coated on TEC, an electrolyte containing a redox pair (usually I2/I3), and a platinum-based counter electrode.155–157 To date, ruthenium sensitizer-based DSSCs have a maximum PCE reaching 11.9%, which gives an open circuit voltage (Voc) of 965 mV, a short circuit current density (Jsc) of 17.3 mA cm−2, and a fill factor (FF) of 0.71 under standard AM 1.5 sunlight at 995 W m−2 intensity.158 Recently, a PCE record of 12.3% was reported by Graetzel and coworkers.158 However, the high cost of ruthenium-based sensitizers and platinum-based electrodes is still an obstacle to the commercialization of DSSCs. Therefore, it is important to synthesize new dye molecules with high efficiency, long stability, and lower prices as well as non-noble metal-based electrodes.157,159 Enlarging the interfacial area of the dye with the electrolyte and accelerating electron transfer in the semiconductor layer to reduce the possibility of charge recombination are also important factors for improving the performance of DSSCs.

Graphene and its composites have been tested to be effective facilitators of electron transport in TiO2 electrodes160,161 and new dyes for DSSCs with satisfactory photoelectrochemical stability.162 Furthermore, graphene composites with polymers have also been explored as the counter electrodes for DSSCs.15,142 In these composites, both graphene and CP components show effective catalytic activity. For example, our group prepared a catalytic counter electrode of DSSC by spin coating 1-PB-functionalized graphene/PEDOT:PSS hybrid on ITO.15 The electrode has a 69-nm composite layer and its transmittance in the visible wavelength range is higher than 80%. The PCE of the DSSC with this counter electrode was tested to be 4.5%, which is a little lower than that of the cell with platinum counter electrode (6.3%) under the same conditions (100 mW, AM 1.5).

Flexible solar cells are expected to have applications in portable devices and foldable chargers.163,164 Furthermore, they could possibly be produced through low-cost roll-to-roll processes.165 Consequently, the composites of CCGs and CPs have been studied as flexible counter electrodes of DSSCs. A pristine PEDOT film without any transparent conductive oxide blended in was used for this purpose,166 and the corresponding DSSC showed a PCE of 5.08%, close to that of a Pt/FTO (fluorine-doped tin oxide)-based DSSC (PCE = 5.88%). However, the former DSSC has a relatively low FF of 0.6 caused by the high resistance of its counter electrode. Therefore, a counter electrode was made by spin coating a PEDOT layer on a CVD-grown graphene-covered PET sheet [Fig. 10(A)].30 The DSSC with this PEDOT/graphene/PET counter electrode had a PCE of 6.26%. In comparison, the PCE of the DSSC with a Pt/ITO or PEDOT counter electrode was measured to be 6.68 or 5.62%, respectively [Fig. 10(B)]. Moreover, the DSSC with the transparent PEDOT/graphene/PET counter electrode has a high FF value and excellent performance even when bent [Fig. 10(C)].

Figure 10.

(A) Photograph of a graphene-coated PET substrate, and schematic diagram of the fabrication steps involved in preparing a DSSC with a graphene/PEDOT counter electrode on a PET substrate. (B) J–V characteristics of DSSCs using as counter electrode: graphene/PEDOT/PET (black ▪), PEDOT/PET (green ▴), and Pt/ITO/PET (red □). (C) J–V characteristics of bended (•) and pristine (▪) DSSCs using PEDOT/graphene/PET as counter electrode. Reproduced from ref. 30, with permission from Wiley-VCH.

Heterojunction Solar Cells

Heterojunction solar cells convert solar energy to electrical power using the photovoltaic effect of p-n junctions, which can be simply divided into two types: inorganic and organic photovoltaic (OPV) cells.167 Although inorganic (especially silicon-based) cells usually have high PCEs, they still have several issues including high production costs, rigorous processing conditions, and environmental problems that need to be addressed. In comparison, organic photovoltaic (OPV) cells have several advantages such as solution processability, cost-effectiveness, and light weight.168 Moreover, OPV can be designed and fabricated as flexible devices using flexible TCEs. However, relatively low PCEs and poor environmental stabilities hinder their practical application.169 Graphene has therefore been widely explored as a new photoelectronic material to be applied in OPVs because of its excellent optoelectronic properties, environmental stability, and flexibility.

Active Layers

In OPV devices, electron acceptors can form heterojunctions with electron donors for the charge separation of excitons.144 The electron affinity of an electron acceptor should be larger than that of the electron donor and smaller than its ionization potential. To date, the most effective OPV devices are based on the bulk heterojunction structure.170 In this structure, fullerene derivatives such as [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) are the most successful electron acceptors. However, these materials usually have low LUMO (lowest unoccupied molecular orbital) energy levels and only absorb visible light weakly. There have therefore been extensive effects devoted to using other additives to either replace or work with C60 derivatives in a cooperative manner as the electron acceptors of polymer-based OPVs.

CNTs have been used as the acceptor because of their high charge mobility, long π-conjugation lengths, and large aspect ratio. However, CNT-based OPVs have low PCEs (about 0.5%). This is mainly because CNTs have poor solubility in organic polymer matrices and wide distributions in lengths and diameters.171–175 Graphene has the advantage of higher electron mobility than that of C60 derivatives as well as energy levels easily tunable by adjusting the material's size and functionality. Furthermore, graphene's large specific surface area and 2D structure favor the formation of a bicontinuous interpenetrating network of donor and acceptor at nanometer scale with maximum interfacial area. Therefore, graphene is an attractive acceptor for OPVs.

GO is the most widely used precursor of CCG. However, it cannot be directly used as a filler of hydrophobic organic polymers such as polythiophene derivatives because it is hydrophilic.109 Consequently, functionalization of GO is important for improving its dispersibility in organic matrices. For example, phenyl isocyanate (PIC) was used to synthesize organic solution-processable functionalized graphene. The PIC-functionalized graphene was used as the acceptor and poly(3-octylthiophene-2,5-diyl) (P3OT) as the donor of an OPV cell. The as-prepared solar cell showed a PCE of 0.32%, and its efficiency was further increased to 1.4% by annealing treatment because of the removal of functional groups from the graphene component and the improvements in the morphology and crystallinity of the P3OT component. An OPV based on the hybrid of graphene and poly(3-hexylthiophene) (P3HT) was also fabricated and annealed.108 It exhibited a PCE of 1.1%, open circuit voltage of 0.72 V, a short circuit current density of 4.0 mA cm−2, and a FF of 0.38 upon excitation with AM 1.5 simulated sunlight at 100 mW cm−2. The efficiencies are still far from those of the state-of-the-art OPVs, possibly because of the irregular molecular structure of rGO; however, graphene-based cells exhibited higher efficiencies than most of the best OPVs based on other alternative acceptor materials to C60 derivatives. Furthermore, a theoretical study predicts that the graphene-based OPV may have an efficiency of over 12% for a single cell; therefore, there is much room for the improvement of graphene-based OPVs.171

In comparison with large graphene sheets, graphene quantum dots (GQDs) posses the advantages of tunable bandgap, good solubility, and larger specific surface area. Thus, GQDs are considered to be potentially good acceptor materials. GQDs [Fig. 11(A)] synthesized by electrolyzing a graphene film can be directly used as the electron acceptor in OPV devices [Fig. 11(B)] according to its energy level structure [Fig. 11(C)].110 Figure 11(D) compares the current–voltage (J–V) characteristics of typical ITO/PEDOT:PSS/P3HT/Al and ITO/PEDOT:PSS/P3HT:GQDs/Al devices. The annealed ITO/PEDOT:PSS/P3HT:GQD/Al device showed the highest PCE of 1.1%. Aniline-functionalized GQDs (ANI-GQDs) have also been used as the acceptor to fabricate OPV cells.175 Hybrid solar cells with the device structure of ITO/PEDOT:PSS/P3HT:ANI-GQDs/LiF/Al exhibited a PCE of 1.14%.

Figure 11.

(A) TEM images of as-prepared GQDs. (B) Schematic of the ITO/PEDOT: PSS/P3HT: GQDs/Al device. (C) Energy band diagrams of the ITO/PEDOT: PSS/P3HT: GQDs/Al device. (D) J–V characteristic curves for the ITO/PEDOT: PSS/P3HT/Al, ITO/PEDOT: PSS/P3HT: GQDs/Al, and ITO/PEDOT: PSS/P3HT: GQDs/Al devices after annealing at 140°C for 10 min, single log scale. Reproduced from ref.137, with permission from Wiley-VCH.

Functionalized graphene and PCBM were simultaneously blended into a P3HT matrix to form a composite active layer of OPV to combine the properties of both carbon materials. The optimized device displayed a PCE of 1.4% under standard conditions.176 CH2OH-terminated regioregular P3HT-grafted GO (G-P3HT) has also been synthesized via esterification reaction for the same purpose.81 This material can be dispersed uniformly in common organic solvents. The red shifts of the π–π* absorption bands of G-P3HT composites in both the solution and the solid state indicated that the chemical grafting of P3HT onto graphene sheets induced a strong electronic interaction, leading to an enhanced electron delocalization and a slightly reduced bandgap energy for the graphene-bound P3HT. A bilayer photovoltaic device based on the solution-cast G-P3HT/C60 heterostructures showed a 200% increase in PCE (0.61%) compared with that of a P3HT/C60 counterpart under AM 1.5 illumination (100 mW cm−2).

Interfacial Layers

Bulk heterojunctions are one of the most promising structures for fabricating highly efficient OPV cells, because they not only have a greatly increased donor/acceptor interface for efficient exciton dissociation but also have a nanoscale interpenetrating network for charge transport to the electrodes. However, blending the electron donor and electron acceptor will lead to the direct electrical contact of the cathode and anode. To prevent this direct electrical contact, an electron blocking and hole transporting layer (HTL) needs to be deposited between the active layer and anode. The most commonly used material for fabricating HTLs is PEDOT:PSS because of its excellent hole transport properties and solution processibility. The PEDOT:PSS layer not only blocks electrons but also adjusts the work function of the active layers and ITO. However, the acidic PEDOT:PSS solution destroys ITO electrodes and introduces trace amount of water into the devices. Some inorganic materials, such as V2O5 and MoO3, have been reported as the most effective HTLs, providing OPVs with PCEs higher than 5%.177,178 An inorganic material layer can overcome the drawbacks of the PEDOT:PSS because they use a dry VD process. However, the VD technique is expensive and cannot be scaled up to meet commercial requirements. Thus, it is necessary to find new materials to replace current HTL materials.

Li et al. were the first to report that GO thin films deposited from neutral solutions by spin coating could be used as efficient HTLs [Fig. 12(A)].34 The introduction of GO obviously reduced the recombination of electrons and holes in comparison with pristine ITO. A GO film with an optimized thickness of 2 nm exhibited the best results [Fig. 12(C,D)], and an OPV device showed a PCE of 3.5% ± 0.3%, a Voc of 0.57 V, Jsc of 11.4 mA cm−2, and a FF of 0.543. With the thickness increases, the performance of OPVs gradually decreases. This is mainly because the resistance of the HTL increases with the thickness of the GO layer. Furthermore, spin coating such a thin GO film can be extremely challenging when trying to fully cover the underlying surface. To overcome these problems, rGO or its composite with single-walled CNTs was used as an HTL. For example, Kim et al. prepared a mixture of GO and SWCNTs by direct sonication of both components in water.179 The blending of SWCNTs greatly enhances the conductivity of the composite; thus, it also greatly reduced the sensitivity of device performance on the thickness of the GO layer. Without SWCNTs, the PCEs of the cells decreased from 3.28% ± 0.14% to 2.36% ± 0.28% when the thickness of GO increased from around 1 to around 3–4 nm. However, in the case of blending SWCNTs, the PCE difference between the two types of cells (3.66% ± 0.18% and 3.13% ± 0.11%, respectively) became much smaller. Finally, the optimized device based on the GO/SWCNT composite with a mass ratio (1:0.2) has a PCE of 4.1% ± 0.18 %, a Voc of 0.6 V, Jsc of 10.82 mA cm−2, and FF of 0.628, which are superior to those of the device using pristine GO as the HTL (PCE = 3.28% ± 0.14%, FF = 60.1% ± 3.5%, Jsc = 9.3 ± 0.73 mA cm−2, and Voc of 0.60 V).

Figure 12.

(A) Schematic of the photovoltaic device structure: ITO/GO/P3HT:PCBM/Al. (B) Energy level diagrams of the bottom electrode ITO, interlayer materials (PEDOT:PSS, GO), P3HT (donor), and PCBM (acceptor), and top Al electrode. (C) Current–voltage characteristics of photovoltaic devices with no hole transport layer (curve labeled as ITO), with 30-nm PEDOT:PSS layer, and 2-nm-thick GO film. (D) Current–voltage characteristics of ITO/GO/P3HT:PCBM/Al devices with different GO thicknesses. All of the measurements were under simulated A.M. 1.5 illumination at 100 mW cm−2. Reproduced from ref. 34, with permission from American Chemical Society.

rGO prepared by reducing GO with p-toluenesulfonyl hydrazide (p-Tos NHNH2) can be well dispersed in aqueous solution with a concentration of 0.6 mg mL−1.31 Thus, the thickness of HTL can be easily controlled using a simple solution-based thin-film fabrication process, which is highly desirable for manufacturing low-cost, high-performance optoelectronic devices. OPVs with p-Tos NHNH2-reduced rGO HTLs exhibited much better performances than the devices with GO or hydrazine-rGO HTLs. This work also reflects that the chemically modified rGO is a bright prospect for application as the HTLs of OPVs.

Composites of graphene derivatives and PEDOT:PSS have also been exploited as HTLs. Few-layered GNSs (FLGs) were blended to PEDOT:PSS to form a uniform composite. By replacing a conventional PEDOT:PSS HTL with the FLG/PEDOT composite, the PCE of OPV increased from 3.10 to 3.70%.111 Similarly, when a butylamine-modified graphene/PEDOT:PSS composite was used as HTL, the PCE of OPVs increased from 0.42 to 0.74%.112 These results indicate that the introduction of graphene derivatives into PEDOT:PSS can improve the performances of OPVs, while the mechanism behind this phenomenon has not yet been clearly revealed.

Fuel Cells

Fuel cells generate energy by oxidizing fuels at low temperatures catalyzed by catalysts immobilized on electrodes.180 To date, polymer electrolyte membrane fuel cells (PEMFCs) have been extensively studied because of their great potentials as environmentally benign energy-conversion devices for transportation, residential, and portable electronic devices. Graphene/polymer composites have been used as the catalysts and solid electrolytes of fuel cells.


Platinum is the most efficient catalyst used in fuel cells.115,181–183 However, the practical application of this catalyst has been greatly hindered by the high cost of noble platinum, CO poisoning, and agglomeration. Furthermore, the use of platinum-based electrocatalysts in direct methanol fuel cells is blocked by the methanol crossover effect. To date, extensive efforts have been devoted to overcome these problems by using Pt-based alloys,184–189 non-noble metals,190–193 and enzymes194,195 to replace Pt-based electrocatalysts. Nitrogen-doped graphene has also been reported to have high electrocatalytic activities and durability, because of its atom-thick 2D structure and high conductivity.196–198 In addition to nitrogen-doped graphene, the composites of graphene materials and nitrogen-containing polymers have also been tested to be effective catalysts for oxygen reduction reaction (ORR).113,114

We synthesized graphene/polymeric graphitic carbon nitrides (GCNs) by polymerizing melamine molecules adsorbed on CCG at a high temperature of 823 K.113 The composite had a layered structure with an improved electrical conductivity over pristine GCN. Therefore, the immobilization of GCN on the surfaces of CCG sheets strongly enhanced the electrocatalytic activity of this polymeric catalyst for ORR with an onset potential at about −0.14 V, slightly higher than that of a CCG/platinum nanoparticle composite. The CCG/GCN catalyst is insensitive to CO, and thus has good durability in electrocatalysis.

A composite of graphene/poly(diallyldimethylammonium chloride) (PDDA) was prepared for the same purpose.114 PDDA worked as an electron acceptor to functionalize graphene to impart it with electrocatalytic activity for ORR in fuel cells. The composite was prepared by reducing the mixture of PDDA and GO with NaBH4, and it showed an electrocatalytic activity with an onset potential of ORR at −0.15V (vs. SCE), compared to an onset potential of pristine graphene electrode at −0.25V. This composite also exhibited better fuel selectivity, more tolerance to CO poisoning, and improved stability compared to those of commercially available Pt/C catalyst.

Solid Polymer Electrolytes

Among various solid polymer electrolytes, Nafion has been most widely applied in PEMFCs because of its remarkable ionic conductivity and chemical and mechanical stabilities. However, Nafion is expensive and cheaper substitutes are required. GO is a chemically modified graphene material with abundant oxygen-containing functional groups and high surface area, which provide it with proton transport channels and ability of holding water. Thus, it is a promising nanofiller for improving the proton conductivities and mechanical properties of polymer electrolyte membranes.199 Cao et al. prepared a PEO/GO composite membrane via solution mixing for application in low-temperature PEMFCs.37 The [BOND]COOH groups on GO sheets are partially ionized to [BOND]COO and H+ ions, providing protons in the composite membrane. The PEO/GO composite membrane has a proton conductivity of 0.09 S cm−1 at 60 °C, and a hydrogen–PEMFC with this solid electrolyte gave a power density of 53 mW cm−2.

If PEMFCs were operated at high temperatures (>100 °C), it would significantly improve their performance because it would enhance the electrode reaction kinetics, reducing the need for excess precious metal (Pt) catalysts as well as improving their CO tolerance. However, the widely used Nafion membrane cannot be used at temperatures higher than 80 °C, mainly because of the dramatic decrease in its water content. Therefore, polymer acid complexes and protic ionic liquids (PILs) have been widely investigated as high-temperature polymer electrolyte membranes to replace Nafion.200 The polybenzimidazole (PBI)/H3PO4 membrane is one of the effective polymer acid complexes.200,201 In many cases, this membrane exhibited high conductivity only as it has a high acid content (3.0 H3PO4 molecules per repeat unit of PBI); thus, it is mechanically weak. Especially, free acid may be lost with evaporated water at high temperatures. One promising choice is reducing its acid content by adding a solid proton conductor into the PBI membrane. For this purpose, Xu et al. blended GO and SGO into PBI membrane,35 and the resulting composite membranes with low H3PO4 contents (1.8–2.0 molecules per repeat unit of PBI) had high proton conductivities of 0.027–0.052 S cm−1 at 175 °C. GO/PBI and SGO/PBI composite membranes-based hydrogen–oxygen fuel cells gave peak power densities of 380 and 600 mW cm−2, respectively.

Composite membranes consisting of PILs and ionic liquid polymer-modified graphene (PIL(NTFSI)-G) have also been investigated.36 The ionic conductivity of the composite membrane, with 0.5 wt % graphene loading, was 7.5 × 10−3 S cm−1 at 160 °C, which is approximately four times higher than that of a sulfonated polyimide/PIL (SPI/PIL) membrane. The mechanical properties of the composites membrane showed an increase of 127% in Young's modulus, and a 345% increase in tensile strength with the addition of 0.9 wt % graphene, comparing with those of a SPI/PIL membrane.


Carbon nanomaterials including CNTs, fullerene, carbon black, graphene, and their derivatives have been widely applied in energy-related systems. Among them, graphene is a new star energy material for electrical and optical devices, mainly because of its unique atom-thick 2D structure, excellent electrical and thermal conductivity, high optical transparency, great mechanical strength, and huge specific surface area. The inherent flexibility of graphene also makes it is one of the most attractive materials for fabricating flexible energy conversion and storage devices.

Graphene materials frequently need to be blended with polymers to form composites to improve their processibility, mechanical, chemical, and electrochemical properties, and/or extend their functions. In this review, we systematically summarized the recent achievements in the synthesis of graphene/polymer composites and their applications in LIBs, supercapacitors, solar cells, and fuel cells. Although great progress has already been achieved, the research in this field is still in its early stages; at the very least, the following problems remain unsolved.

First, perfect graphene has been considered to a highly conductive material; however, the CCGs widely used to fabricate composites with polymers have much lower conductivities than that of perfect graphene because of their structural defects. Therefore, most graphene/polymer composites still suffer from poor electrical properties and do not satisfy the requirements for transparent conductive electrodes and counter electrodes of solar cells. The electrodes of LIBs and supercapacitors would also benefit from improvements in the conductivities of graphene materials. A technique that enables the synthesis of high-quality graphene on a large scale at relatively low coat is urgently required. Several efforts have already been devoted to overcome this problem. For example, the synthesis of graphene sheets with large lateral dimensions151 or the use of mildly oxidized GO as the precursors.51 However, these techniques are still limited by their low efficiencies.

Second, the large specific surface areas of graphene materials should be kept during the processes of synthesizing graphene/polymer composites. However, the real specific areas of graphene materials are usually much lower than their theoretical values because of the π–π stacking between graphene sheets. In many cases, blending with polymer increases the severity of this problem. New techniques for homogenously blending single-layered graphene sheets into polymer matrices need to be developed. The construction of graphene-based 3D architectures with controlled microstructures can partly overcome this problem.202–204

Third, the application of graphene/polymer in solar cells is still at its initial stages: many mechanisms used to explain the effects of graphene are based on assumptions, and the performances of graphene/polymer composites are usually far below state-of-the-art counterparts. The performances of graphene composites in solar cells are still worse than those of traditional materials. Thus, the compositions and morphologies of the composites and the structures of solar cells need further optimization. Nevertheless, after fully exploring the potential of graphene/polymer composites, it is clear that this class of material will have wide and important applications in energy production and storage systems of the future.


This work was supported by National Basic Research Program of China (973 Program, 2012CB933402) and Natural Science Foundation of China (51161120361, 91027028).




atom transfer radical polymerization


chemically converted grapheme


carbon nanotube


conducting polymers




direct methanol fuel cells


dye-sensitized solar cells


electrical double layer capacitor


electrochemical impedance spectroscopy


fill factor


graphene oxide


graphene quantum dots


indium tin oxide


short circuit current


lithium ion batteries


layered birnessite-type manganese oxide


lowest unoccupied molecular orbital


methyl 2-bromopropionate


organic photovoltaic


oxygen reduction reaction










propylene carbonate


[6,6]-phenyl-C61-butyric acid methyl ester


power conversion efficiencies




polymer electrolyte membrane fuel cells


poly(ethylene oxide)




poly(sodium 5-styrensulfonate)


poly(2,2,6,6-tetramethylpiperidinyloxy- 4-yl methacrylate)


poly(vinyl alcohol)


reduced graphene oxide


sodium dodecylbenzene sulfonate


sulfonated grapheme


transparent conducting electrodes


open circuit voltage

Biographical Information

original image

Yiqing Sun received her B.S. from the Department of Polymer Science and Engineering at Beijing University of Chemical Technology in 2009, and became a PhD candidate under the supervision of Prof. Gaoquan Shi in the Department of chemistry at Tsinghua University. Her research interest focuses on the graphene-based materials.

Biographical Information

original image

Gaoquan Shi received his BS degree (1985) and PhD degree (1992) in the Department of Chemistry at Nanjing University. He then joined Nanjing University and was promoted to full professor in 1995. In 2000, he moved to Tsinghua University as a Professor of Chemistry. His research interests are focused on synthesis and application of conducting polymers and graphene. He received the second-grade award of Natural Science of China and the youth knowledge innovation prize of Chinese Chemical Society and BASF Company in 2004.