pubs.acs.org/NanoLett
Graphene-Based Supercapacitor with an Ultrahigh Energy Density Chenguang Liu,†,|,§ Zhenning Yu,‡,§ David Neff,† Aruna Zhamu,‡ and Bor Z. Jang*,† †
Nanotek Instruments, Inc. and ‡ Angstron Materials, Inc., Dayton, Ohio 45404, United States, | School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, People’s Republic of China ABSTRACT A supercapacitor with graphene-based electrodes was found to exhibit a specific energy density of 85.6 Wh/kg at room temperature and 136 Wh/kg at 80 °C (all based on the total electrode weight), measured at a current density of 1 A/g. These energy density values are comparable to that of the Ni metal hydride battery, but the supercapacitor can be charged or discharged in seconds or minutes. The key to success was the ability to make full utilization of the highest intrinsic surface capacitance and specific surface area of single-layer graphene by preparing curved graphene sheets that will not restack faceto-face. The curved morphology enables the formation of mesopores accessible to and wettable by environmentally benign ionic liquids capable of operating at a voltage >4 V. KEYWORDS Graphene, supercapacitor, energy density, ionic liquid
S
upercapacitors are very attractive power sources. Compared with batteries, they are essentially maintenance-free, possess a longer cycle-life, require a very simple charging circuit, experience no memory effect, and are generally much safer.1,2 Physical rather than chemical energy storage is the key reason for their operational safety and exceptionally long cycle-life. Perhaps most importantly, supercapacitors can be charged and discharged at high rates. One of the most promising supercapacitor applications is in electric vehicles (EV). Supercapacitors can be coupled with fuel cells or batteries to deliver the high power needed during acceleration and to recover the energy during braking. However, a major shortcoming of current supercapacitors is their low energy density (typically 5-10 Wh/kg), which is significantly lower than the 20-35 Wh/kg of lead-acid, 40-100 Wh/kg of Ni metal hydride, and 120-170 Wh/kg of lithium-ion cells. Despite their relatively high specific energies, batteries suffer from a severe drawback: low charge-discharge rates. For instance, the current lithiumion battery for plug-in hybrid vehicles requires 2-6 h of recharge time, as opposed to just seconds for supercapacitors. Hence, a supercapacitor that is capable of storing as much energy as a battery and yet can be fully recharged in one or two minutes would be considered a revolutionary advancement in energy technology. Supercapacitors have two energy storage mechanisms: electrical double-layer (EDL) capacitance and pseudocapacitance.1 Currently, EDL capacitors contain activated carbon (AC) with a high surface area as the electrode material, and the capacitance comes from the charge accumulated at the
electrode/electrolyte interface. In contrast, a pseudocapacitor uses a conducting polymer or metal oxide as an electrode material, which undergoes reversible Faradic redox reactions. Conducting polymers, such as polyanilines,3 have been shown to exhibit high pseudocapacitance, but poor stability during the charge/discharge cycling. Further, the typical response times of pseudocapacitors are significantly longer than those of EDL capacitors. Carbon materials, such as ACs and carbon nanotubes (CNTs), usually exhibit good stability, but limited EDL capacitance.4-6 For ACs, typically only about 10-20% of the “theoretical” capacitance was observed due to the presence of micropores that are inaccessible by the electrolyte, wetting deficiencies of electrolytes on electrode surface, and/or the inability of a double layer to form successfully in the pores. CNTs do not exhibit satisfactory capacitance values unless a conducting polymer is used to form a pseudocapacitance pair.5 A strong pseudocapacitance component is not desirable in many applications due to the slow response time and high capacitance decay rate. Furthermore, CNTs remain expensive and “vertically grown CNTs” are difficult to produce and handle in a real supercapacitor manufacturing environment. An outstanding candidate electrode material is graphene. A single-layer graphene sheet is basically a 2D hexagonal lattice of sp2 carbon atoms covalently bonded along two plane directions.7,8 Graphene has recently been found to exhibit exceptionally high thermal conductivity, electrical conductivity, and strength.9-12 Another outstanding characteristic of graphene is its exceptionally high specific surface area up to 2675 m2/g. Most significantly, the intrinsic capacitance of graphene was recently found to be 21 µF/ cm2,13 which sets the upper limit of EDL capacitance for all carbon-based materials. This study asserts that graphene is the ideal carbon electrode material for EDL supercapacitors because it is capable of storing an EDL capacitance value of
* To whom correspondence should be addressed. E-mail: Bor.Jang@ NanotekInstruments.com. § These authors contributed equally to this work Received for review: 07/29/2010 Published on Web: 11/08/2010
© 2010 American Chemical Society
4863
DOI: 10.1021/nl102661q | Nano Lett. 2010, 10, 4863–4868
up to 550 F/g, provided the entire 2675 m2/g is fully utilized. Another advantage of graphene in a supercapacitor electrode is the notion that both major surfaces of a graphene sheet are exterior surfaces readily accessible by electrolyte. The outstanding capacitance performance of graphene-based supercapacitors was first reported in 2006 by our research group led by Dr. Jang,14,15 and its scientific and technological significance is now being more widely recognized.16-37 A supercapacitor based on chemically modified graphene electrodes exhibited specific capacitances of 135 and 99 F/g with aqueous and organic electrolytes, respectively.17 With microwave irradiation20 or directly heating a suspension of graphene oxide platelets in propylene carbonate21 to exfoliation and reduction of graphene oxide, capacitance values as high as 190 F/g in aqueous and 120 F/g in organic electrolytes have been achieved.37 Graphene made by thermally expanding graphene oxide at high temperatures16 or at relatively low temperatures but under vacuum19 has been used in supercapacitor electrodes. Several groups have reported graphene-based supercapacitors using metal oxide/ graphene,25,26 CNTs/graphene,27 and polymer/graphene composites for electrodes.28-33 However, to the best of our knowledge, there has been no report on a graphene-based supercapacitor using an ionic liquid as electrolyte to achieve a high energy density at room temperature; however, Rao et al. used graphene with an ionic liquid but obtained an energy density of only 31.9 Wh/kg at 5 mV/s and 60 °C.16 The compatibility between graphene sheets and ionic liquids, which are significantly larger than the molecular sizes of aqueous and other organic liquid electrolytes, remains poorly understood. This subject has been largely overlooked by the supercapacitor research community. Specifically, graphene sheets have a high tendency to restack themselves during all phases of graphene preparation and subsequent electrode production procedures, leaving behind intergraphene pore sizes that are not sufficient for accessibility to the electrolyte and the formation of EDL charges. The pore accessibility and surface compatibility of graphene by the ionic liquid electrolyte is a critically important issue further due to the fact that water-based electrolytes cannot be operated at a voltage higher than 1.2 V for a symmetric supercapacitor. A larger working voltage implies a much higher specific energy density according to E ) (1/ 2)CcellV2, where Ccell ) specific capacitance of the cell and V ) voltage. In most of these earlier studies, high energy density values were obtained from the pseudocapacitance mechanism of a conducting polymer-graphene redox pair.28,32 The redox mechanism is not very desirable since it typically involves a chemical reaction, which is not conducive to fast charging and discharging. Although the high pseudocapacitance value of the graphene-conductive polymer pair could push the energy density to above 100 Wh/kg (based on the electrode weight),28 this value was obtained at an extremely low current density or scan rate (1 mV/s) unsuitable for practical use. Such an extremely low rate defeats the primary purpose © 2010 American Chemical Society
of using a supercapacitor for high charge/discharge rate applications. Ionic liquids exhibit high ionic conductivity, large electrochemical windows (up to 7 V), excellent thermal stability (-40 to +200 °C typical), and characteristics of being nonvolatile, nonflammable, and nontoxic. Herein reported are the results of a study on a mesoporous graphene structure that is ionic liquid electrolyteaccessible and, hence, achieves an exceptionally high EDL capacitance even though ionic liquids have large molecules and high viscosity. With no major pseudocapacitance contribution, this mesoporous graphene electrode enables fast charging and discharging of supercapacitors. The outstanding specific capacitance of the graphene electrode, coupled with the use of an ionic liquid capable of operating at a high voltage (e.g., 4 V), pushes the specific energy density of an EDL supercapacitor to an unprecedented level of nearly 90 Wh/kg at a current density of 1 A/g at room temperature and a level of 136 Wh/kg at 80 °C. The electrodes were made of graphene, mixed with 5 wt % Super-P and 10 wt % polytetrafluoroethylene (PTFE) binder. The mass of each electrode was 6.6 mg with a diameter of 13 mm. Coin-size capacitor cells were assembled in a glovebox. A supercapacitor unit cell comprised two electrodes that were electrically isolated from each other by a Celguard-3501 porous membrane. The ionic liquid electrolyte was 1-ethyl-3-methylimidazolium tetrafluoroborate (EMIMBF4). A representative scanning electron microscopy (SEM) image of the curved graphene sheets is shown in Figure 1a. This curved graphene sheet morphology appears to be capable of preventing graphene sheets from closely restacking with one another when they are packed or compressed into an electrode structure, thereby maintaining a mesoporous structure having a pore size in the range of 2 to 25 nm. The pore size data was obtained from low-temperature nitrogen sorption experiments by using a volumetric adsorption apparatus (Quantachrome Instruments, U.S.A.). The nitrogen adsorption isotherm of graphene (Figure 2a) shows type IV isotherm characteristics (characterized by a hysteresis loop), based on which, pore size distributions were obtained by means of the Barret-Joyner-Halenda (BJH) equation, as shown in Figure 2b. In contrast, the graphene sheets (e.g., Figure 1b) prepared by conventional chemical routes tend to restack with one another, effectively reducing the useful surface area. The intergraphene gaps are clearly less than 1 nm. Evaluation of more than 200 samples of flat-shaped graphene sheetbased supercapacitors in our lab indicates that the specific capacitance can reach 100-150 F/g in aqueous electrolyte (KOH + water), but typically
