High energy density asymmetric pseudocapacitors fabricated
by graphene/carbon nanotube/MnO
2
plus carbon nanotubes
nanocomposites electrode
Chung Jung Hung
a, Pang Lin
a, Tseung Yuen Tseng
b,*aDepartment of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
bDepartment of Electronics Engineering and Institute of Electronics, National Chiao Tung University, Hsinchu 300, Taiwan
h i g h l i g h t s
The nanocomposite electrode achieves the specific capacitance of 964 F g1at 1 A g1.
The asymmetric pseudocapacitor possesses energy density of record high 304 Wh kg1.
The asymmetric pseudocapacitor shows excellent cycling stability.
a r t i c l e i n f o
Article history:
Received 30 November 2013 Received in revised form 1 February 2014
Accepted 23 February 2014 Available online 6 March 2014 Keywords: Graphene Pseudocapacitor Nanocomposite Cycling stability
a b s t r a c t
Novel graphene/carbon nanotubes (CNTs)/manganese oxide (MnO2) nanocomposites plus CNTs
(GMCþ C) and graphene/CNTs hybrid (GC) thin-film electrodes are prepared by electrophoretic depo-sition (EPD). These nanocomposite electrodes exhibit high surface area and interconnected pore net-works. The GMCþ C nanocomposite electrode shows excellent specific capacitance of 964 F g1at 1 A g 1, rate capability with the residual capacitance of 529 F g1at 500 mV s1, and fast Naþdiffusion with
intercalation value of 6.34 107 cm2s1, and deintercalation value of 8.86 107 cm2s1. Such
excellent pseudocapacitive performances are attributed to low ion/electron transport resistances and short ion/electron diffusion lengths. Furthermore, novel aqueous electrolyte-based asymmetric pseu-docapacitor having 1.8 V cell voltage is successfully fabricated using GMCþ C nanocomposite as a cathode and GC nanocomposite as an anode. The optimized asymmetric pseudocapacitor possesses superior performance with a maximum energy density of record high 304 Wh kg1and retaining 56.2% of its initial specific energy density at the power density up to 242 kW kg1. In addition, the asymmetric
cell configuration also shows excellent cycling stability with 89% specific capacitance maintained after 10,000 cycles. These results suggest that our designed asymmetric pseudocapacitors have a high po-tential for practical applications.
Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction
To achieve high power and long life cycle devices than batteries, electrochemical capacitors or supercapacitors (SCs) attracts much more attention for future energy storage applications. In terms of charge storage mechanisms, SCs are mainly classified into electrical double-layer capacitors (EDLCs) and pseudocapacitors [1,2]. In general, EDLCs have high power density and long life cycle but suffer from relatively low energy density; however,
pseudocapacitors exhibit higher specific capacitance and energy density than EDLCs, but poor cycle ability. In other words, their limitation is that the energy density value of SCs is still far below that battery, which has restricted their potential applications[3e5]. Hence, it is crucial to increase the energy density of SCs to meet the future energy demands without sacrificing excellent power density. Technically, the effective way to improve the energy density for SCs depends on enhancing the intrinsic capacitance of active ma-terials or increasing the operating voltage of the system [6]. Conventionally, many of the envisioned applications have emerged in the replacement of aqueous electrolyte with organic electrolytes or ionic liquids to offer better electrochemical stability and extend the cell potential window, which result in a higher energy density
* Corresponding author.
E-mail address:[email protected](T.Y. Tseng).
Contents lists available atScienceDirect
Journal of Power Sources
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m/ l o ca t e / j p o w s o u r
http://dx.doi.org/10.1016/j.jpowsour.2014.02.094 0378-7753/Ó 2014 Elsevier B.V. All rights reserved.
than aqueous electrolytes [7e9]. However, these non-aqueous electrolytes are generally more expensive and not eco-friendly in comparison to aqueous electrolytes. In addition, non-aqueous SCs usually exhibit low capacitive and power capabilities, attributing to low ionic conductivity and higher specific resistance, which limit the cell performance[10].
In other words, using SCs based on aqueous electrolytes have excellent ion conductivity to minimize the internal resistance leading to faster charging/discharging properties. In particular, they also offer advantages of low-cost green manufacturing and enhanced safety in operation, which make these aqueous electro-lytes becoming a very attractive for electrochemical capacitors[11]. However, it is well-known that aqueous electrolytes limit the in-dividual SCs cell voltage to roughly 1.2 V (theoretical decomposi-tion voltage of water, 1.23 V).
In order to extend the operating voltage, much research effort has been focused on asymmetric pseudocapacitors for a two-electrode setup working in well-separated potential windows in the same electrolyte[12e17]. It is anticipated that, this designed asymmetric two-electrode configuration operating in neutral electrolytes can bring both the EDLC (anode, as the power source) and pseudocapacitive (cathode, as the energy source) behavior together, leading to increase operating voltage and thus signi fi-cantly improve the energy density while retaining their intrinsic high specific power[18e21].
We previously reported that hybrid graphene/MnO2/CNTs ternary nanocompositesfilms were directly deposited on a nickel plate collector as high performance electrodes of pseudocapacitor via EPD method without adding any binder or surfactants during processing[22,23]. Moreover, EPD, in contrast to other processing methods, offers several significant advantages such as low cost, rapid fabrication, non-vacuum approach, room-temperature pro-cessing, eco-friendly, additive free, scalable and suitability for mass production. In addition, it can also control the mass of the depos-ited particles, namely, thickness of thefilms can be easily controlled by the concentration of the suspension, applied potential, and deposition time. Regarding the engineering feasibility study, EPD method is the most feasible option for the practical application of pseudocapacitor.
As far as our knowledge, there are only a very few reports explored the fabrication of hybrid nanocomposites thin-film elec-trode for asymmetric pseudocapacitors cell assemblies by EPD method. We used EPD method to fabricate novel hybrid functional nanocomposite electrodes for asymmetric pseudocapacitors with improved electrochemical properties, such as achieving high en-ergy density, high power density and long cycle lifetime. Herein, we report a straightforward approach to fabricate an aqueous electrolyte-based asymmetric pseudocapacitors based on a designed GMCþ C ternary nanocomposite cathode, and a GC binary nanocomposite anode. Combining the unique properties of each functional material, the present strategy employs hybrid nano-composite thin-film electrodes with porous network structure, which makes the SCs exhibiting excellent power capability and lower equivalent series resistance.
2. Experimental
Designed hybrid GMCþ C and GC nanocomposites as asym-metric pseudocapacitor electrode materials were directly deposited onto 10 mm 30 mm 1 mm flat Ni substrates by EPD method. Before electrophoresis, Ni substrate was etched with 10% nitric acid solution for 30 min to obtain rough surface and thoroughly cleaned with deionized (DI) water in an ultrasonic bath, and dried in an oven at 100C for 12 h. The commercial CNTs (specific surface area: 40e300 m2g1, length: 5e20
m
m) was purified by boiling at 70%nitric acid solution for 24 h, followed by rinsed with DI water and dried in an oven at 100 C for 12 h before synthesizing nanocomposites.
2.1. Synthesis of graphene
Graphene was prepared by reducing the graphene oxide syn-thesized by chemical oxidation and exfoliation of natural graphite fine powder using a modified experiment by Rakhi et al.[24]. In a typical synthesis, we first employed the modified Hummers methods to prepare graphene oxide. The detailed synthesis pro-cedure is reported elsewhere[23]. 100 mg of graphene oxide was dispersed in 50 mL DI water having 30 min ultrasonic agitation, to achieve brown color suspension. Subsequently, graphene oxide suspension was reacted with a given amount of sodium hydroxide (NaOH) (pH¼ 14.0) in a water bath at 95C for 12 h, the brown color suspension now turned into completely black. After the sus-pension cooled down to room temperature, the as-synthesized product of the reacting suspension was collected and then centri-fuged followed by thoroughly rinsed with DI water. Finally, the obtained product, graphene was dried in air at 110C for overnight followed by a 750C thermal annealing under a H2/Ar mixed gas for 30 min.
2.2. Preparation of GMCþ C cathode and GC anode nanocomposites
Pseudocapacitor cathode electrode material, GMCþ C nano-composites were synthesized by using the combination of redox titration reaction and EPD technique. GMCþ C nanocomposites were formed from the mixtures of graphene/MnO2/CNTs nano-composites and purified CNTs. Details of the apparatus and experimental procedures were described elsewhere [23]. The electrolyte used in EPD was a mixture of 0.095 g graphene/MnO2/ CNTs nanocomposites, 0.005 g CNTs powders (5 wt%), and hy-drochloric acid (37%, 0.2 mL) in the isopropyl alcohol (80 mL). Powders were dispersed well in the solution using ultrasonic agitation for 30 min. The GMCþ C nanocomposites were deposited using an EPD procedure. Nickel and platinum foils were dipped in this solution as cathode and anode, respectively at a distance of 1 cm. The nanocomposite powders were electrophoretically deposited on Ni substrate at a 50 V direct current (DC) voltage for 2 min. After EPD, the GMCþ C nanocomposite film formed on Ni substrate (as-grown nanocomposites electrode) was dried in an oven at 110C for 12 h.
Pseudocapacitor anode electrode material, GC nanocomposite was synthesized by using the same technique. Briefly, purified CNTs (0.05 g) were dispersed in the mixture solution of hydrochloric acid (37%, 0.2 mL) and isopropyl alcohol (80 mL) with as-synthesized graphene (0.05 g), and then ultrasonicated for 30 min to produce a uniform suspension. Nickel and platinum foil were kept parallel at 1 cm apart in the suspension as cathode and anode, respectively. The EPD was carried out by applying a constant 50 V DC voltage for 1.5 min. After EPD, the GC nanocompositefilm formed on Ni sub-strate was dried in an oven at 110C for 12 h.
2.3. Fabrication of asymmetric pseudocapacitor system
Asymmetric pseudocapacitor system was successfully con-structed with a simple two-electrode configuration which was composed of GC anode and GMCþ C cathode. Here, the anode and cathode were assembled with a separator (DuPontÔ NafionÒ NRE211 membrane) sandwiched in between. After pressing, the cathode and anode collectors were wrapped with Teflon tape and then infiltrated with 0.1 M Na2SO4as the electrolyte solution, until
all air was replaced. To fabricate two-electrode asymmetric pseu-docapacitor, the loading mass ratio of active materials (GMCþ C/ GC) was estimated to be 0.2 from the specific capacitances, calcu-lated from the CV curves in a three-electrode cell. Its assembly consists of a cathode coated with approximately 0.55 mg GMCþ C nanocomposites, and an anode coated with 2.7 mg GC nano-composites. The mass of every as-deposited sample was measured by a high precision microbalance (PRECISA XR125SM-FR) with an accuracy of 1
m
g.2.4. Structural and properties characterization
Field emission scanning electron microscope (FE-SEM, JEOL JSM-6700) and X-ray diffractometer (XRD, Bede D1) were used to analyze surface morphology and crystallization of the film, respectively. The oxidation states of manganese oxidefilm were examined by X-ray photoelectron spectroscopy (XPS, ULVAC-PHI Quantera SXM). Field emission transmission electron micro-scope and EDS analysis (FE-TEM, JEOL JEM-2100F) were used to observe the coaxial structure of nanocomposite. The specific surface area of the nanocomposite was determined by BET (Bru-nauer, Emmett and Teller theory) surface area analyzer (ASAP 2020). Raman spectra of nanocomposite material were recorded using a high resolution confocal Raman microscope (HOROBA, Lab RAM HR). The laser excitation was provided by a regular model laser operating at 632.8 nm. Thermal analysis was performed from 50 to 950C at 3C min1ramping rate under air atmo-sphere using thermogravimetric analyzer (TGA, TA Instruments Q500).
2.5. Electrochemical testing
The electrochemical properties and specific capacitance mea-surements were carried out in a three-electrode cell with a satu-rated calomel reference electrode (SCE), platinum sheet as a counter electrode, and the prepared electrodes (GMCþ C or GC nanocomposites) as an working electrode. The electrochemical characterization of asymmetric pseudocapacitors, composed of GMCþ C cathode and GC anode, was carried out in a two-electrode configuration. All electrochemical measurements were carried out in a 0.1 M Na2SO4solution as electrolyte. Cyclic voltammetry (CV), galvanostatic charge/discharge cycling (CC), chronoamperometry (CA) and electrochemical impedance spectroscopy (EIS) were conducted using CH Instruments 618B electrochemical analyzer at room temperature.
The specific capacitance (F g1) was calculated from CV curve using equation C ¼ Z Ef Ei IðEÞdE m v Ef Ei (1)
where m is the mass of the active material, v the scan rate of CV curves (V s1), Eithe initial potential, Efthefinal applied potential, and (Ef Ei) the potential window width.
The specific capacitance (F g1) from CC curve was calculated using the formula
C ¼
D
ID
tE m (2)
where I is the discharge current,
D
t the time for a full discharge, m the active material mass, andD
E the width of the potential window for a full discharge.3. Results and discussion
The microstructural properties of as-synthesized graphene were shown inFig. 1. Here, we provided the modified route to synthesize graphene using a very simple refluxing process under alkaline conditions, as a replacement for highly toxic and dangerously unstable hydrazine. Under deoxygenated and heated alkaline conditions, the graphene oxide suspension was changed the color, from brown to black. Finally, graphene products were obtained by H2/Ar gas annealing, as shown inFig. 1a. In literature, the optical changes were observed due to the reduction of graphene oxides
[25,26]. Our results demonstrate the alkaline deoxygenated treat-ment has the unique ability enough to remove oxygen-containing functional groups on the surface of graphene oxide, which is the evidence for reduction process.
Fig. 1b shows a surface view of as-synthesized graphene, indi-cating highly porous wrinkle structure.Fig. 1c illustrates the TEM image of our synthesized graphene. It can be clearly observed that as-synthesized graphene exhibits a crumpled paper-like external morphology and graphene domain area exceeds a few hundreds of square nanometers.
Fig. 1d depicts the typical nitrogen adsorption/desorption isotherm of as-synthesized graphene. The isotherm exhibits a sig-nificant hysteresis loop, indicating major nitrogen adsorption to the large porous structure. The specific surface area of as-synthesized graphene exhibits as high as 808.3 m2 g1, and the sizes of the pores are between 34.3 and 1315.7 A with an average pore size of 191.6 A (as shown in the inset of Fig. 1d). Previous studies have shown that the most probable number of graphene layers could be determined according to specific surface area[27,28]. According to our experimental results, as-synthesized graphene sheets have an average thickness of approximately 1 nm, corresponding to approximately 3 stacked graphene sheets.
Raman spectroscopy has been utilized to elucidate the vibration properties of as-synthesized graphene, as shown inFig. 1e. It can be clearly observed that there are three peaks containing 2D, G, and D bands at 2657.2, 1580.3, and 1330.4 cm1, respectively. As-synthesized graphene exhibits a weak D band which is generally considered to be the defect and disorder carbon in the graphene corresponding to the unavoidable oxidative damage during chemical exfoliation. The ID/IG area ratio of as-synthesized gra-phene is calculated to be 0.27, suggesting that the sp2carbon atom network becomes more and more stable within as-synthesized graphene. Nevertheless, this number is much lower than those calculated from the reduction by hydrazine [23]. In addition, the I2D/IGarea ratio can be used to determine the degree of graphiti-zation for sp2(CaC bonds) of graphitic materials and the number of graphene layers[29]. The I2D/IG area ratio is approximately 2.11, demonstrating that the thickness of graphene about 3e4 layers. This result is well consistent with that obtained from the BET sur-face area data. This result shows that the modified approach can effectively produce the graphene with less defect nature.
Designed GC and GMCþ C nanocomposites are employed as electrode materials for asymmetric pseudocapacitor applications.
Fig. 2a displays the morphology of GC nanocomposites electro-phoretically deposited on Ni substrate, indicating that 3D hybrid carbon architecture is formed by self-assembly of the dispersions of graphene and CNT. This result implies that, it forms an open porous structure for electrolyte ions easily accessing the surface of GC.
The GMCþ C nanocomposites retain porous wrinkling structure and CNTs are incorporated into the GMCþ C matrix, as shown in
Fig. 2b. In addition, GMCþ C nanocomposites have a rough surface, which indicate the growth of MnO2on porous network surface. It is worth to mention that CNTs act as a bridge connecting the large graphene sheets together and filling the voids, in which such a
network structure formation can serve as fast electronic and ionic conducting channels. Detailed observations of the surface mor-phologies are confirmed by TEM. The flake-like MnO2 is well deposited on the GMC matrix surface, as shown in Fig. 2c. EDS spectrum corresponding to the marked region indicates that the coating is only composed of two elements, Mn and O, as shown in inset of Fig. 2c. Interestingly, the appearance of K peaks in the sample indicates Kþco-existing in the MnO2matrix, arising from the precursor, KMnO4, which always happens during the synthesis of MnO2[30].
The selected area electron diffraction pattern (SAED) of GMCþ C nanocomposites is shown in the inset ofFig. 2c. SAED pattern ex-hibits six-fold hexagonal symmetry related to the basal plane of graphene, corresponding to diffraction patterns of (100) and (110).
It indicates successful restoration of the graphite structure in the reduced graphene oxide[31]. In addition, the diffraction rings can be indexed to the birnessite-type (002), (200), (112) crystal planes of MnO2 nanosheets (JCPDS 43-1456). Continuous ring at SAED pattern also confirms their nanocrystalline nature.
XRD patterns of GC and GMCþ C nanocomposites are depicted inFig. 2d, indicating that the GC and GMCþ C nanocomposites both exhibit the major peak at 2
q
around 26.5corresponding to the C (002) plane. This peak is common to both CNTs and graphene[32]. Moreover, the two weak characteristic peaks of GMCþ C nano-composites at 12.3(001) and 24.9(002) can be indexed to the characteristic reflections of birnessite-type MnO2phase. This result corroborates with the SAED results. However, broad diffraction peaks demonstrate that flake-like MnO2 exhibits a poorlyFig. 1. (Color online) (a) Images of the suspension before and after the deoxygenation of exfoliated graphene oxide under alkaline (NaOH) conditions, andfinal black product (graphene), (b) SEM image, (c) TEM image, (d) N2adsorptionedesorption isotherm, and (e) Raman analysis of the as-synthesized graphene.
crystallized structure. The superior pseudocapacitive performance was obtained for MnO2-based electrode containing higher amount of poorly crystallized birnessite phase[33,34]. Poorly crystallized birnessite phase has large interlayer distance and surface area-to-volume ratio, as a result pseudocapacitance is improved due to proton and cation of the electrolytes are easily inserted into the matrix.
Detailed surface chemical composition of GC and GMC þ C nanocomposites is further analyzed by XPS, as shown inFig. 2e. A scanning spectrum of the GC nanocomposites shows the presence of C 1s and O 1s signal. Compared with GC, Mn species is observed in the XPS wide spectrum of GMCþ C, indicating the evidence of MnO2deposition on the porous network surface. In addition, the small amount of K 2s peak corresponds to KMnO4, which is in
agreement with EDS result. To further investigate the oxidation state of Mn and confirm the composition, the Mn 2p spectra are individually shown in the inset of Fig. 2e. The XPS spectra are almost symmetrical, peaks of Mn 2p3/2at 642.4 eV and Mn 2p1/2at 654.0 eV, with a spin-energy separation of 11.6 eV, which are in good agreement with the previous reports of MnO2[35,36].
In order to understand the loading amount of MnO2 on the matrix, the TGA curve was measured and used to determine the weight percentage of MnO2in the as-synthesized GMCþ C nano-composites. Data were recorded in air at a ramp rate of 3C min1 from 50 to 950 C. Based on the TGA curve of GMC þ C nano-composites, as shown inFig. 2f, the weight loss of adsorbed and lattice water is about 6.9% between 50 and 390C. The additional 48.3% weight loss between 390 and 550 C corresponds to the
Fig. 2. (Color online) (a) SEM image of the deigned GC nanocomposite, (b) SEM image and (c) TEM image (the inset is corresponding selected area electron diffraction pattern and EDS spectrum of selected part) of the deigned GMCþ C nanocomposite; (d) XRD, (e) wide XPS survey spectra (the inset is Mn 2p spectra of GMC þ C nanocomposite), and (f) TGA curve of the designed GC (lower line) and GMCþ C nanocomposite (upper line).
removal of graphene/CNTs hybrid material. Furthermore, thefinal weight loss approximately 2.4% between 860 and 910C is asso-ciated with the conversion of MnO2to Mn3O4[37].
In addition, TGA result of GC nanocomposites shows 96% loss (almost total weight) between 450 and 750C, which is attributed to the oxidation of the graphene/CNTs in air[24]. Thus, the actual percentages of MnO2and graphene/CNTs hybrid material matrix in the initial GMC þ C nanocomposites are 40.8% and 48.3%, respectively.
The CV, CC, and CA measurements were employed to evaluate the pseudocapacitive behavior of GMCþ C nanocomposites.Fig. 3a il-lustrates the typical CV curves of the GMC þ C nanocomposite electrode in neutral aqueous electrolyte (0.1 M Na2SO4) at 25C in the potential range of 0< ESCE< 0.8 V vs. SCE with different scan rate. All the CV curves appear almost a near-ideal symmetric rectangular profile and a rapid current response to a voltage reversal at each end potential even at a high scan rate of 500 mV s1. According to Eq.(1), the mean calculated specific capacitances are 860, 697, 634, 572, and 529 F g1as a function of different scan rates of 5, 50, 100, 300, and 500 mV s1respectively. It is shown that the specific capacitance decreases with increasing scan rate, as shown in inset ofFig. 3a, up to 61.5% specific capacitance can be retained when the scan rate in-creases 100 times from 5 to 500 mV s1. These results show excellent power capability of the designed GMCþ C hybrid electrode.
In comparison with the CV results, the linear charge/discharge behavior is observed of GMC þ C electrode in neutral aqueous electrolyte (0.1 M Na2SO4) at 25C in the range of 0< ESCE< 0.8 V at various current densities, as shown inFig. 3b. All these CC curves are very symmetric, and the charge/discharge duration decreases with an increase of current density. In addition, the slope of every
curve is potentially independent with a constant value; IR drops of all curves are dependent on various current densities from 1 to 10 A g1. According to Eq.(2), the specific capacitance of GMC þ C electrode at current densities of 1, 2, 3.5, 5 and 10 A g1can be calculated as 964, 867, 746, 697, and 644 F g1, respectively. About 67% of specific capacitance is retained when the charge/discharge rate changes from 1 to 10 A g1. In addition, GMCþ C electrode exhibits low resistance even at high current densities. The rela-tionship is plotted in the inset ofFig. 3b.
To further understand the correlation between the mobilities of ions and electrons, CA experiment was used to measure the diffu-sion coefficient D. The D is calculated using following the Cottrell equation:
it ¼ nFAC0D 1=2
p
1=2t1=2 (3)where n is stoichiometric number of electrons involved in the re-action; F, Faraday’s constant (96,485 C mol1); A, area of the elec-trode surface (cm2); C0, is the concentration of electroactive species (mol cm3); and D, the diffusion coefficient (cm2s1). A Cottrell plot of the current, i vs. 1/square root of time, t½transformed the data into a linear relationship whose slope is nFAC0D½
p
½[38]. The Cottrell plot of i vs. 1/sqrt(T) is obtained and a typical representation is made inFig. 3c. Using the values of the forward and reverse slopes in Eq.(3), the calculated values of D for intercalation and deintercalation of Naþinto/from the GMCþ C nanocomposite in 0.1 M Na2SO4are 6.34 107cm2 s1and 8.86 107cm2s1, respectively, which are significantly better than those reported in previous works[39,40].Fig. 3. (Color online) (a) The CV curves in 0.1 M Na2SO4at scan rates of 5, 50, 100, 300 and 500 mV s1(the inset is the summary plot of specific capacitance from CV results), (b) the
CC curves in 0.1 M Na2SO4at current densities of 1, 2, 3.5, 5, and 10 A g1(the inset is the summary plot of specific capacitance from CC results), and (c) Cottrell plot for the CA
From the above mentioned data, to enhance pseudocapacitive performance it is expected to use GMCþ C hybrid porous structure, leading to that the electrolyte canflood into pores of its structure and its inner surface is suitable for ion diffusion. Moreover, CNTs plays a critical role in promoting the performance by acting as a spacer to prevent the re-stacking of individual MnO2/graphene/ CNTs nanocomposite sheets.
In general, asymmetric pseudocapacitor that incorporates different materials for both anode and cathode electrodes promises a wider operating voltage and thus achieves higher energy density. One electrode is based on the high specific capacitance of the redox reaction character, and the other one is mostly based on the fast ion transport property of the electric double-layer storage. At this point, we assemble asymmetric pseudocapacitor using GMCþ C nanocomposite as cathode and GC nanocomposite as anode in aqueous electrolyte (0.1 M Na2SO4), as shown inFig. 4a. To further
evaluate the electrochemical potential window and chemical sta-bility of GMC þ C and GC nanocomposites, we performed CV measurements on the two-electrode materials using a three-elec-trode cell configuration before evaluating the asymmetric pseudocapacitor.
Fig. 4b shows the different stable potential windows of GMCþ C and GC electrodes, indicates that both CV curves exhibit a nearly rectangular shape demonstrating an ideal capacitive behavior for both electrodes. The stable potential window is between0.9 and 0 V for GC and between 0 and 0.9 V for GMCþ C. The specific ca-pacitances of the GC and GMCþ C are 106 and 521 F g1at a scan rate of 25 mV s1 respectively. In order to find the optimal cell voltage in asymmetric pseudocapacitor, a mass balance between the two electrodes is required to consider. Consequently, the mass ratio of anode electrode (GC) to cathode electrode (GMCþ C) is decided according to the well-known charge balance theory
Fig. 4. (Color online) (a) Schematic illustration of the fabricated asymmetric pseudocapacitor configuration based on GMC þ C as cathode and GC as anode in a neutral aqueous Na2SO4electrolyte, (b) comparative CV curves of GMCþ C and GC nanocomposites electrodes performed in a three-electrode cell in 0.1 M Na2SO4aqueous solution at a scan rate of
25 mV s1, (c) the CV curves in 0.1 M Na2SO4at scan rates of 25, 50, 100, 300 and 500 mV s1(the inset is the summary plot of specific capacitance from CV results), (e) the CC curves
in 0.1 M Na2SO4at current density of 1, 5, 10, 15, and 30 A g1(the inset is the summary plot of specific capacitance from CC results), (e) the cycling stability of the deigned
asymmetric pseudocapacitor at a constant charge/discharge current density of 1 A g1(the inset is the charge/discharge curves of the last 6 cycles), and (f) Nyquist plot of GMCþ C// GC asymmetric pseudocapacitor recorded before and after 10,000 cycles.
(qþ¼ q). In this relationship, the charge stored by each electrode usually depends on the specific capacitance (C), the potential window (
D
E) and the mass of active material (m), as shown in the following equation:q ¼ C
D
E m (4)and in order to get qþ¼ q, the mass balancing is calculated by the following equation:
mþ m ¼
C
D
ECþ
D
Eþ (5)Based on the specific capacitance values and potential windows of GMCþ C and GC electrodes, the asymmetric system voltage can be expressed as the sum of the potential range of GMCþ C and GC. Consequently, it can be extended up to 1.8 V in 0.1 M aqueous Na2SO4solution for the asymmetric pseudocapacitor of GMCþ C// GC with the optimal mass ratio between the two electrodesfixed to mþGMCþC=m
GC ¼ 0:2.
Fig. 4c describes the typical CV curves of optimized asymmetric pseudocapacitor measured at various scan rates between 0 and 1.8 V in 0.1 M Na2SO4aqueous electrolyte. Generally, all the CV curves exhibit a quasi-rectangular shape and the total current in-creases with increasing scan rate. It is worthy to mention, with the increase of scan rates up to 500 mV s1, there are no distortion of CV curve is observed, indicating fast chargeedischarge properties of the asymmetric device. However, it should pay an attention here that the specific capacitance is calculated on the basis of the total mass of active materials on two electrodes in the asymmetric pseudocapacitor, rather than on a single electrode. According to Eq.
(1), the mean calculated specific capacitances are 521, 497, 462, 435, and 400 F g1as a function of different scan rates of 25, 50, 100, 300, and 500 mV s1respectively. It is clearly shown that the specific capacitance gradually decreases with increasing scan rate, as indi-cated in the inset ofFig. 4c, which is attributed to low utilization of active materials at higher scan rates[41]. However, the actual cal-culations based on a single electrode yield much higher values (1.5 times higher than two-electrode configurations); such approach should not be used to predict the asymmetrical configuration’s performance[16,24].
In order to further evaluate the performance of asymmetrical pseudocapacitor, linear charge/discharge testing was also per-formed with different current densities with a voltage window of 1.8 V and the result is shown inFig. 4d. The charge/discharge curves exhibit near symmetric triangular shapes and small equivalent series resistance, indicating good charge storage properties. Ac-cording to Eq.(2), the mean calculated specific capacitances are 676, 603, 537, 444, and 380 F g1as a function of different current density of 1, 5, 10, 15, and 30 A g1, respectively. Even when the discharge current is enhanced 30 times, from 1 to 30 A g1, the specific capacitance still retains approximately 56% of its initial capacitance at a high current density. This result is plotted in the inset ofFig. 4d, corresponding to the behavior of ideal capacitors, which is agreed well with CV results.
Fig. 4e displays the cycling stability of the GMCþ C//GC asym-metric pseudocapacitor at a constant charge/discharge current density of 1 A g1. It is indicated that the capacitance can achieve as high as 602 F g1retaining 89.0% of its initial capacitance after 10,000 continuous charge/discharge cycles, which demonstrates an excellent cycling performance. In addition, the last 6 cycles still exhibit a highly symmetric charge/discharge characteristic further proving the long period durability, as shown in inset ofFig. 4e.
Electrochemical impedance spectroscopy studies were con-ducted within the frequency range of 0.1 Hze100 KHz at an open circuit voltage with AC amplitude of 5 mV.Fig. 4f shows the Nyquist
plots of the EIS spectra for asymmetric pseudocapacitor before and after testing for 10,000 cycles, respectively. Both impedance spectra shown in thisfigure clearly depict the semi-circle type at the end of the Warburg-type line.
Upon long-term cycling, the EIS shows subtle changes. Firstly, the internal resistances after 10,000 cycles increased from 0.8 to 1.0
U
at the high-frequency intercept of the real axis, suggesting that the electron conductivity of the GMCþ C//GC configuration shows only a slight decline tendency. Secondly, the charge transfer resistances are evaluated from the diameter of the semi-circle, increasing from 6.1 to 13.8U
after long-term cycling, which imply that the electrolyte penetration and ion diffusion in the host ma-terials are slightly reduced. In addition, the EIS spectra show almost vertical line at the low frequency region, and no any obvious change after long-term cycling, indicating that it still remain a low ion transport resistance in the pore structure of the nanocomposite. Importantly, even after 10,000 cycles, the internal resistance, charge transfer resistance, and Warburg impedance in the porous structure of the GMC þ C//GC configuration still remain low; demonstrating that our designed asymmetric pseudocapacitors have a high pseudocapacitive performance and excellent cycling stability.Furthermore, power density (P) and energy density (E) both are generally used as important parameters of asymmetric pseudoca-pacitor which evaluate its operational performance/efficiency. On the basis of the charge/discharge test results, the power density and energy density are calculated by using the following equations[42]: E ¼ C
D
V2 2 1000 3600 Energy Density=Wh kg1 (6) P ¼ 4 ESR MD
V2 ac Power Density=kW kg1 (7) where V is the applied voltage, ESR the internal resistance, Macthe total mass of active electrode materials used in asymmetric pseu-docapacitors, and C is the measured total specific capacitance of asymmetric pseudocapacitors.The internal resistance is determined from the voltage drop at the beginning of each discharge
ESR ¼ iR2drop I ðEquivalent Series Resistance=
U
Þ (8) iRdropis the voltage drop between thefirst two points from its top cut-off. I the current applied.These results are summarized as the Ragone plot inFig. 5. It is shown that the energy densities decrease slowly with the increasing of power densities. The maximum energy density of 304.1 Wh kg1 (for a power density of 19.9 kW kg1) and the highest power density of 242.1 kW kg1(for an energy density of 171.2 Wh kg1) are achieved using our designed GMC þ C//GC asymmetric pseudocapacitors at an operation voltage of 1.8 V. This energy density value is indeed much higher than those energy density values of MnO2-based asymmetric pseudocapacitors in aqueous electrolytes reported previously, including graphene/ MnO2//activated carbon nanofibers (51.1 Wh kg1at 198 kW kg1)
[12], graphite oxide/MnO2//graphite oxide (24.3 Wh kg1 at 24.5 kW kg1)[13], MnO2//graphene (23.2 Wh kg1at 1 kW kg1)
[14], graphene/MnO2//graphene/Ag (7.5 Wh kg1at 90.3 kW kg1)
[15], graphene/MnO2/CNTs//activated carbon/CNTs (24 Wh kg1at 7.8 kW kg1) [16], MnO2/CNTs//CNTs (73.6 Wh kg1 at 14.6 kW kg1)[17].
Finally, we also found that the designed GMCþ C//GC asym-metric pseudocapacitors can light a red emitting diode, as shown in
the inset ofFig. 5, demonstrating its practical application in the energy storage/conversion system.
4. Conclusions
In summary, we fabricated novel hybrid nanocomposite thin-film electrodes with high surface area and interconnected pore networks via EPD. The GMCþ C nanocomposite electrode displays excellent electrochemical performance with respect to the maximal specific capacitance of 964 F g1(at the current density of 1 A g1), remarkable rate capability (with the residual capacitance of 529 F g1 at the high scan rate of 500 mV s1), and fast Naþ diffusion (intercalation value: 6.34 107cm2s1, deintercalation value: 8.86 107cm2s1). It indicates that the designed nano-composites can provide a synergistic improvement on electro-chemical performance as well as offering uninterrupted electron-conducting and ion-diffusing pathways. Furthermore, the asym-metric pseudocapacitors assembled using GMCþ C as the cathode, GC as the anode, and 0.1 M Na2SO4 as the aqueous electrolyte, exhibiting a record high energy density of 304 Wh kg1 and retaining 56.2% of its initial specific energy density at the power density up to 242 kW kg1. In addition, the asymmetric cell configuration also possesses excellent cycling stability with 89% of the specific capacitance maintained after 10,000 cycle tests. These results indicate that our designed asymmetric pseudocapacitors are appropriate for practical applications.
Acknowledgments
This work was supported by the National Science Council of ROC under contract No. 102-2221-E-009-044-MY3.
References
[1] M. Yu, T. Zhai, X. Lu, X. Chen, S. Xie, W. Li, C. Liang, W. Zhao, L. Zhang, Y. Tong, J. Power Sourc. 239 (2013) 64.
[2] Y. Gao, Y.S. Zhou, M. Qian, H.M. Li, J. Redepenning, L.S. Fan, X.N. He, W. Xiong, X. Huang, M. Majhouri-Samani, L. Jiangd, Y.F. Lu, RSC Adv. 3 (2013) 20613. [3] H. Jiang, P.S. Lee, C. Li, Energy Environ. Sci. 6 (2013) 41.
[4] I.E. Rauda, V. Augustyn, B. Dunn, S.H. Tolbert, Accounts Chem. Res. 46 (2013) 1113.
[5] X. Lang, A. Hirata, T. Fujita, M. Chen, Nat. Nanotechnol. 6 (2011) 232. [6] R. Gupta, N.K. Sharma, P. Tiwari, A. Gupta, N. Nigam, A. Gupta, Int. J. Eng. Sci.
Technol. 3 (2011) 289.
[7] H. Zhong, F. Xu, Z. Li, R. Fu, D. Wu, Nanoscale 5 (2013) 4678.
[8] H.Q. Wang, Z.S. Li, Y.G. Huang, Q.Y. Li, X.Y. Wang, J. Mater. Chem. 20 (2010) 3883.
[9] Q. Cheng, J. Tang, J. Ma, H. Zhang, N. Shinya, L.C. Qinc, Phys. Chem. Chem. Phys. 13 (2011) 17615.
[10] S. Vaquero, J. Palma, M. Anderson, R. Marcilla, Int. J. Electrochem. Sci. 8 (2013) 10293.
[11] G. Yu, X. Xie, L. Pand, Z. Baod, Y. Cui, Nano Energy 2 (2013) 213.
[12] Z. Fan, J. Yan, T. Wei, L. Zhi, G. Ning, T. Li, F. Wei, Adv. Funct. Mater. 21 (2011) 2366.
[13] X. Zhao, L. Zhang, S. Murali, M.D. Stoller, Q. Zhang, Y. Zhu, R.S. Ruoff, ACS Nano 6 (2012) 5404.
[14] H. Gao, F. Xiao, C.B. Ching, H. Duan, ACS Appl. Mater. Interfaces 4 (2012) 2801. [15] Y. Shao, H. Wang, Q. Zhang, Y. Li, J. Mater. Chem. C 1 (2013) 1245. [16] Y. Cheng, H. Zhang, S. Lu, C.V. Varanasiad, J. Liu, Nanoscale 5 (2013) 1067. [17] J. Shen, A. Liu, Y. Tu, H. Wang, R. Jiang, J. Ouyang, Y. Chen, Electrochim. Acta 78
(2012) 122.
[18] F. Wang, S. Xiao, Y. Hou, C. Hu, L. Liu, Y. Wu, RSC Adv. 3 (2013) 13059. [19] M.R. Lukatskaya, O. Mashtalir, C.E. Ren, Y. Dall’Agnese, P. Rozier, P.L. Taberna,
M. Naguib, P. Simon, M.W. Barsoum, Y. Gogotsi, Science 341 (2013) 1502. [20] H. Wang, Y. Liang, T. Mirfakhrai, Z. Chen, H.S. Casalongue, H. Dai, Nano Res. 4
(2011) 729.
[21] H. Gao, F. Xiao, C.B. Ching, H. Duan, ACS Appl. Mater. Interfaces 4 (2012) 7020. [22] C.J. Hung, J.H. Hung, P. Lin, T.Y. Tseng, J. Electrochem. Soc. 158 (2011) A942. [23] C.J. Hung, P. Lin, T.Y. Tseng, J. Power Sourc. 243 (2013) 594.
[24] R.B. Rakhi, W. Chen, D. Cha, H.N. Alshareef, Adv. Energy Mater. 2 (2012) 381. [25] S. Gurunathan, J.W. Han, V. Eppakayala, A.A. Dayem, D.N. Kwon, J.H. Kim,
Nanoscale Res. Lett. 8 (2013) 393. [26] S. Pei, H.M. Cheng, Carbon 50 (2012) 3210. [27] B.Z. Jang, A. Zhamu, J. Mater. Sci. 43 (2008) 5092.
[28] D.W. Wang, F. Li, Z.S. Wu, W. Ren, H.M. Cheng, Electrochem. Commun. 11 (2009) 1729.
[29] Z. Ni, Y. Wang, T. Yu, Z. Shen, Nano Res. 1 (2008) 273.
[30] H.Q. Wang, G.F. Yang, Q.Y. Li, X.X. Zhong, F.P. Wang, Z.S. Li, Y.H. Li, New J. Chem. 35 (2011) 469.
[31] A.K. Mondal, B. Wang, D. Su, Y. Wang, S. Chen, X. Zhang, G. Wang, Mater. Chem. Phys. 143 (2014) 740.
[32] S.S.J. Aravind, V. Eswaraiah, S. Ramaprabhu, J. Mater. Chem. 21 (2011) 15179. [33] H.Y. Chu, Q.Y. Lai, L. Wang, J.F. Lu, Y. Zhao, Ionics 16 (2010) 233.
[34] O. Ghodbane, J.L. Pascal, F. Favier, ACS Appl. Mater. Interfaces 1 (2009) 1130. [35] J. Liu, J. Jiang, M. Bosmanc, H.J. Fan, J. Mater. Chem. 22 (2012) 2419. [36] J. Zeng, J.R. Nair, C. Francia, S. Bodoardo, N. Penazzi, Int. J. Electrochem. Sci. 8
(2013) 3912.
[37] J. Li, N. Wang, Y. Zhao, Y. Ding, L. Guan, Electrochem. Commun. 13 (2011) 698. [38] M.F. Nazarudin, Z. Zainal, W.T. Tan, I. Hamadneh, E.F. Kadri, Int. J. Electrochem.
Sci. 7 (2012) 2965.
[39] M.N. Patel, X. Wang, B. Wilson, D.A. Ferrer, S. Dai, K.J. Stevenson, K.P. Johnston, J. Mater. Chem. 20 (2010) 390.
[40] M.N. Patel, X. Wang, D.A. Slanac, D.A. Ferrer, S. Dai, K.P. Johnston, K.J. Stevenson, J. Mater. Chem. 22 (2012) 3160.
[41] K. Karthikeyan, S. Amaresh, K.J. Kim, S.H. Kim, K.Y. Chung, B.W. Cho, Y.S. Lee, Nanoscale 5 (2013) 5958.
[42] H. Yu, J. Wu, L. Fan, K. Xu, X. Zhong, Y. Lin, J. Lin, Electrochim. Acta 56 (2011) 6881.
Fig. 5. Ragone plots of GMCþ C//GC asymmetric pseudocapacitors compared with other MnO2-based asymmetric configurations with an aqueous electrolyte in previous
literature (the inset is the photograph of GMCþ C//GC cell lighting up a red emitting diode).
