A comparison of carbon supports in MnO₂/C supercapacitors

Abstract

This article presents a comparison of three typical carbon supports, carbon nanotubes (CNTs), reduced graphene oxide (rGO), and acetylene black (AB), with regards to their performance in MnO₂ supercapacitors. To exclude the influence of MnO₂, uniform-sized MnO₂ nanoflakes with a width of ∼50 nm were used to fabricate three MnO₂/C composites. The working electrodes were prepared using a consistent procedure to keep all three types of MnO₂/C electrodes in the same configuration. The influence of these three carbon supports on capacitive performance and cyclability was studied. In particular, the capacitance contribution of carbon supports was subtracted from the overall capacitance. The contribution from MnO₂ was compared among these three types of composites and it was found that rGO could enhance the capacitive performance of MnO₂ at slow scan rates. However, MnO₂/CNTs and MnO₂/AB exhibited better capacitive performance at high rates and better stability.

---

Introduction

The limited fossil fuel resources and increasing environmental problems have stimulated great interest in sustainable energy resources. To relate the sustainable energy resources to our daily life, efficient energy storage and conversion technologies are highly demanded for smart grid, zero-emission transport, and portable devices.¹ Among these technologies, supercapacitors are a class of attractive energy storage devices due to their high power density, low cost, high safety, and long lifetime.² According to the charge storage mechanism, there are two types of supercapacitors. One is the electrical double layer capacitors (EDLCs), which use carbon as electrode materials. EDLCs often give high power density, but low energy density.³ The other is the pseudocapacitors, which use redox materials like metal oxides⁴ and conducting polymers⁵ as active electrode materials. Depending on the active materials, pseudocapacitors can achieve fairly high energy density and power density.^6,7

Manganese dioxide (MnO₂) is a very popular redox capacitor material and the related research has been intensively conducted in the past few years due to the increasing interest in the field of energy storage and conversion.^8–10 MnO₂ is a promising electrode material for low cost, high energy density supercapacitors because of its abundance in the earth and high specific energy capacity.^11,12 MnO₂ is also environmental benign due to its low toxicity. Thus it has a great potential to be widely used in energy devices. The theoretical specific capacitance of MnO₂ is nearly 1400 F g⁻¹ if all manganese can be involved in the redox reaction (one-electron redox).¹⁰ However, to have all manganese participated in the redox reaction is difficult because the redox reaction happens only at the surface region of MnO₂. It therefore stimulated a great interest to innovate nanosized MnO₂ materials to increase the specific surface area. To date, various methods have been reported to synthesize nanosized MnO₂, including co-precipitation,^13,14 sol–gel,¹⁵ electro-deposition,^16,17 hydrothermal,¹⁸ and etc. The use of MnO₂ nanomaterials has given significant improvements in capacitance.¹⁹

However, it should be noted that MnO₂ has frequently been combined with carbon support to be used in supercapacitors. This is because MnO₂ is not conductive. Various carbon support materials have been used in MnO₂ supercapacitors.^20,21 The popular carbon supports include carbon nanotubes (CNTs), graphene, carbon black (CB), etc. These carbon supports are usually mixed with MnO₂ through different approaches, depending on the synthetic methods of MnO₂. For example, some approaches are able to synthesize MnO₂/C composites by one step protocol,¹⁰ while some others use extra steps to make MnO₂/C electrodes.²¹ It has been well-established in literatures that the carbon supports are influential to enhancing the performance of MnO₂ supercapacitors.^19,20 However, the various reported MnO₂/C composites are made from different approaches and those composites are quite different in composition, morphology, MnO₂ materials, mass ratio, and other parameters. As a result, comparisons on the influence of these carbon supports on the MnO₂ supercapacitor performance have been rare.

Here we present a rational comparison of three typical carbon supports, carbon nanotubes, graphene, and carbon black, with regards to their performance in MnO₂ supercapacitors. We employed uniform-sized MnO₂ nanoflakes, which served as a standard active material for three carbon supports. The electrode preparation followed the standard method to keep all three types of MnO₂/C electrodes same in configuration. The influence of these three carbon supports on capacitive performance and cyclability was compared.

Experimental

Synthesis of MnO₂ nanoflakes

MnO₂ nanoflakes were synthesized by modifying a recipe.²² In a typical synthesis, an aqueous solution containing 0.05 M potassium permanganate (KMnO₄) and 0.02 M cetrimonium bromide (CTAB) was heated to boiling in two minutes. The solution turned to dark brown immediately. The solution was kept boiling for 30 minutes. The product was washed with water and acetone by centrifugation at 8000 rpm for three times to remove the surfactant.

Carbon materials

Three carbon materials were selected for this study. Carbon black (acetylene black, Alfa Aesar) and multiwall carbon nanotubes (MWNTs, Sigma-Aldrich) were used as purchased without any treatment. Reduced graphene oxide (rGO) was prepared by a modified Hummer's method.^23,24 All chemicals for rGO synthesis were purchased from Sigma-Aldrich. In a typical synthesis, 1.5 g graphite powder was suspended into 20 ml 98% H₂SO₄ with 1.25 g K₂S₂O₈ and 1.25 g P₂O₅. The suspension was heated to 80 °C and kept at this temperature for 6 hours. The resulting pre-oxidized graphite was washed with deionized water by filtration and dried in vacuum oven. The powder was then suspended in 80 ml 98% H₂SO₄ with 7.5 g KMnO₄ and treated at 40 °C for 2 hours. Subsequently, the solution was cooled down to ∼0 °C using ice bath and then 150 ml of deionized water and 10 ml of H₂O₂ were slowly added into the solution. The yellow solution was washed with 10% HCl. To remove the Cl⁻ ions, the material was washed with the large amount of deionized water until neutral pH was achieved. The powders were finally thermally treated to produce reduced graphene oxide (rGO) at 1050 °C.

Carbon loading

A solution of carbon material (∼0.5 mg L⁻¹) was prepared by sonicating the selected carbon supports (here carbon black, MWCNT, or rGO) in ethanol for 3 hours. Afterwards, a solution of MnO₂ nanoflakes with desired concentration was added drop-wise into the carbon solution and the mixture was further sonicated for 2 hours. The MnO₂/C precipitation was collected by centrifuging for 10 min at 9500 rpm. The MnO₂/C was dried under vacuum for further use. In order to precisely control the mass loading, the concentration of MnO₂ solution used here was determined by ICP. TEM images were obtained on a TEM JEOL 2010 operated at 200 kV.

Electrochemical measurement

The working electrodes were prepared by a conventional method. Typically, as-prepared MnO₂/C composites and polyvinylidene difluoride (PVDF) were mixed with a mass ratio of 90 : 10 and dispersed in N-methyl pyrrolidone (NMP). The resulting slurry was pasted onto nickel foam substrate (1 cm × 1 cm) with a spatula. The total weight of the active material for each sample was controlled at ∼1 mg. After drying at 75 °C for 12 hours, the electrode was pressed at 10 MPa. All electrochemical measurements were carried out on Solartron Electrochemical Interface 1287 and Impedance Analyzer 1260 using typical three-electrode configuration, where a platinum foil, a saturated calomel electrode (SCE), and the prepared composite electrodes served as counter, reference, and working electrodes, respectively. The electrolyte was 1 M Na₂SO₄ aqueous solution. Cyclic voltammetry (CV) was recorded at different scan rates of 5, 10, 20, 50, 100, 200, and 500 mV s⁻¹. The impedance measurements were carried out in the frequency range from 100 kHz to 0.1 Hz at open circuit potential with an AC perturbation of 5 mV. Galvanostatic charge/discharge curves were measured at different current densities of 0.5, 1, 1.5, and 3 A g⁻¹ to evaluate the power density and energy density. The cyclability test was conducted by Galvanostatic charge/discharge approach at the current density of 3 A g⁻¹. The potential window in the range of −0.1 to 0.9 V (vs. SCE) was used in all the measurements.

Results and discussion

Capacitance of carbon supports

Fig. 1a shows the representative cyclic voltammograms of carbon supports, carbon nanotubes (CNTs), graphene (rGO), and carbon black, at the scan rate of 10 mV s⁻¹. The CVs of carbon nanotubes and carbon black are similar in shape, giving a typical feature of double layer capacitor. Graphene gave a larger CV than the other two carbons. It is partially due to the larger specific surface area of graphene. The defects and remaining functional groups also contributed some pseudo-capacitance. The typical pseudo-capacitance features include the deviation from rectangularity and the appearance of polarized peaks at the two sides of the CV, which is consistent with the reported results.^20,25 Fig. 1b shows the specific capacitance of these carbon supports at the different scan rate from 5 to 500 mV s⁻¹. Each data point and its error bar were obtained by three independent measurements. rGO showed the higher specific capacitance than CNTs and AB, especially at the low scan rates. For example, rGO gave ∼54.4 F g⁻¹ at 5 mV s⁻¹, which is nearly three times higher than CNTs (∼18.5 F g⁻¹) and seven times higher than AB (∼8.7 F g⁻¹). Limited by charge transportation at the high rate, the capacitance of these carbons drops with the increase of scan rate. The specific capacitance of rGO dropped about 50% from ∼54.4 F g⁻¹ to ∼26.5 F g⁻¹ as the scan rate increased from 5 to 500 mV s⁻¹. CNTs and AB gave similar drops in their capacitance.

Fig. 1 (a) Representative cyclic voltammograms of carbon supports, carbon black (acetylene black, AB), carbon nanotubes (CNTs), and graphene (rGO), at the scan rate of 10 mV s⁻¹ in 1 M Na₂SO₄ at the room temperature. (b) Specific capacitance of three carbon supports at different scan rates. The inset shows the close view for values at the scan rate 2, 10, 20, and 50 mV s⁻¹. Each data point and its error bar were obtained from three independent measurements.

Capacitive performance of MnO₂/C composites

MnO₂/C composites were synthesized by mixing MnO₂ nanoflakes with AB, CNTs, and rGO. By varying the mass ratio of MnO₂ : C, the composition of each MnO₂/C composite was given at 20, 40, 60, and 80 wt% of MnO₂. Their specific capacitance was recorded using cyclic voltammetry method in 1 M Na₂SO₄ at room temperature. Fig. 2a shows the specific capacitance of MnO₂/C with different MnO₂ mass loading at 5 mV s⁻¹. 0 wt% refers to the pure carbon supports and 100 wt% refers to the electrodes with only MnO₂ on nickel foam. The specific capacitance value in Fig. 2a was normalized by the total mass of MnO₂ and the carbon support. The overall trend for three types of MnO₂/C composites is the same, i.e. the specific capacitance increases with the increase of MnO₂ loading. This is because of the increase of active component MnO₂ in the composite, which mainly contributes the overall capacitance. Among these loading amount, 80 wt% of MnO₂ gave the highest capacitance for all MnO₂/C composites. The specific capacitance of MnO₂/rGO with 80 wt% MnO₂ (MnO₂ : C = 80 : 20) is ∼165.1 F g⁻¹, while the specific capacitance of MnO₂/CNTs and MnO₂/AB with the same composition are ∼134.1 and ∼130.1 F g⁻¹, respectively. In general, MnO₂/rGO showed higher capacitance than MnO₂/CNTs and MnO₂/AB at all compositions. MnO₂/CNTs and MnO₂/AB exhibited similar capacitance at 40, 60, and 80 wt% of MnO₂. At 20 wt%, the specific capacitance of MnO₂/CNTs is slightly higher than MnO₂/AB. The better performance of MnO₂/rGO is probably due to the capacitance contribution of rGO. As shown in Fig. 1b, rGO contributes a capacitance of ∼54 F g⁻¹ at 5 mV s⁻¹.

Fig. 2 Specific capacitance of MnO₂/C composites normalized by the total mass of MnO₂ and carbon at 5 mV s⁻¹ (a), 50 mV s⁻¹ (b), and 500 mV s⁻¹ (c). Specific capacitance of MnO₂ from the contribution of supported MnO₂ at 5 mV s⁻¹ (b), 50 mV s⁻¹ (d), and 500 mV s⁻¹ (f). The mass loading of MnO₂ refers to the composition of MnO₂ in the composite. For example, 20 wt% of MnO₂ refers to a composition of MnO₂ : C = 20 : 80 in mass. Each data point and its error bar were obtained from three independent measurements.

To investigate the influence of rGO and the other two carbon supports, the capacitance contributions of carbon supports were subtracted from the composite capacitance. The specific capacitances at 5 mV s⁻¹ purely from the contribution of MnO₂ are shown in Fig. 2b. It can be seen that in general MnO₂ supported on rGO gave higher capacitance than MnO₂ supported on the other two carbon supports. CNTs and AB showed similar influence on the performance of supported MnO₂. At 20 wt% of MnO₂, supported MnO₂ exhibited highest capacitance. MnO₂ supported on rGO, CNTs, and AB gave specific capacitance of ∼310.2, 292.4, and 274.5 F g⁻¹, respectively. As more MnO₂ was made into the composites, the specific capacitance of supported MnO₂ dropped. From 20 wt% to 80 wt%, the capacitance drop was nearly in a linear relationship with composition. At 80 wt%, the specific capacitance of supported MnO₂ were ∼192.8, ∼163.5, and ∼160.5 F g⁻¹ for MnO₂/rGO, MnO₂/CNTs, and MnO₂/AB, respectively. The capacitance drop for all three types of MnO₂/C composites is nearly 50% from 20 to 80 wt% of MnO₂. The pure MnO₂ supported on nickel foam (100 wt%) gave the lowest specific capacitance. The pure contribution from MnO₂ shown in Fig. 2b indicates that carbon supports are essential and influential to the performance of MnO₂. If more carbon supports and less MnO₂ were made into the composites, higher specific capacitance of supported MnO₂ can be achieved.

The trend of specific capacitance of MnO₂/C composites at the scan rate of 50 mV s⁻¹ is similar to that at 5 mV s⁻¹. Fig. 2c shows the specific capacitance of three types of composites normalized by the total mass of MnO₂ and carbon. Again, MnO₂/rGO showed better performance than the other two MnO₂/C composites. However, this superiority of rGO is not as remarkable as it showed at 5 mV s⁻¹. For example, MnO2/rGO with 20 wt% MnO₂ gave a specific capacitance of ∼71.8 F g⁻¹, which is ∼10 and ∼20 F g⁻¹ more than MnO₂/CNTs and MnO₂/AB. As the weight percentage of MnO₂ increased to 80 wt%, the capacitance of MnO₂/rGO, MnO₂/CNTs, and MnO₂/AB are quite similar. All are around ∼95 F g⁻¹. On the other hand, after subtracting the capacitance contribution of carbons, the specific capacitance of MnO₂ purely from the supported MnO₂ exhibited a change. At 50 mV s⁻¹, the highest capacitance contribution from MnO₂ was given by MnO₂/CNTs with 20 wt% of MnO₂ as shown in Fig. 2d. 20 wt% MnO₂ supported on CNTs gave a specific capacitance of ∼256.1 F g⁻¹. The value for MnO₂ supported on AB and rGO are ∼228.1 and ∼178.2 F g⁻¹, respectively. As more MnO₂ was made into the composites, the specific capacitance of supported MnO₂ decreased. At the mass loading of 40 and 60 wt%, MnO₂ supported on CNTs and AB exhibited similar performance in capacitance, while MnO₂ supported rGO showed lower performance. This indicates that the charge transportation within the MnO₂/rGO probably was not fast as compared to MnO₂/CNTs and MnO₂/AB. The electrochemical impedance spectra of these composites indicated that the resistance of rGO used here was larger than CNTs and AB (Fig. S3†). It should be noted that rGO has higher specific surface area and higher porosity than AB and CNTs (Fig. S4†). However, the double layer capacitance from carbons gave limited contribution and it was also subtracted for comparing the capacitive contributions from MnO₂ supported by three carbons. The high porosity of the rGO electrodes did not show better capacitive behaviour at the high rates. It indicates that the influence of the porosity of carbon supports is very small. The lower conductivity probably is one factor that affected the overall charge transportation efficiency. For all composites, with the increase of the mass loading of MnO₂, the electrode resistance increased gradually. This is consistent with the reported result that the non-conductive MnO₂ usually lowers the overall conductivity of the composite.²¹ It should be noticed that all MnO₂/C composites here are made without any chemical modification at the interface of MnO₂ and carbon supports. Carbon materials here only played the roles of the conducting substrate and the MnO₂ support. Although rGO has a remarkable capacitance contribution, the large size of rGO sheets might hinder the transportation of charge species in the electrolyte as rGO is multiple stacked at the solid state. Another reason is that the resistance of rGO composites is higher than others. This is probably because these composites were mixed physically. Without a chemical bonding or interaction between MnO₂ and carbon, the resistance would be high.^28–30 As the weight percentage of MnO₂ increased to 80 wt%, the specific capacitance of supported MnO₂ in three types of MnO₂/C composites were nearly similar.

At the scan rate of 500 mV s⁻¹, it is interesting to find that the specific capacitance of MnO₂/rGO is lower than that of MnO₂/CNTs and MnO₂/AB no matter how much MnO₂ was loaded (Fig. 2e). At this high rate, MnO₂/CNTs gave higher capacitance at 20 wt% of MnO₂. MnO₂/rGO exhibited a similar specific capacitance as pure rGO. However, the specific capacitance of MnO₂/rGO dropped significantly as the amount of MnO₂ increased more than 20 wt%. Such a trend can be also be found on MnO₂/CNTs and MnO₂/AB. For MnO₂/CNTs, the highest specific capacitance was given by 20 wt%. Then the capacitance dropped as the amount of MnO₂ increased. For MnO₂/AB, the highest specific capacitance was given by 40 wt%, followed by the 20 wt% sample. At the high loading amount of MnO₂ like 60 and 80 wt%, MnO₂/CNTs and MnO₂/AB showed similar specific capacitance. Fig. 2f shows the specific capacitance of supported MnO₂ at 500 mV s⁻¹ after subtracting the contribution of carbons at this scan rate. A remarkable phenomenon can be found on MnO₂/rGO composites. For all mass loading of MnO₂ from 20 to 80 wt%, MnO₂ supported on rGO gave similar capacitance performance, which is nearly same as the MnO₂ on nickel foam. It indicates that at a high scan rate rGO may not be able to enhance the performance of MnO₂. Its function is similar to nickel current collector. Again, this is probably due to the large size of rGO, which hinders the charge transportation within the electrode.²⁰ On the other hand, we found that CNTs and AB were able to play the role of carbon supports to enhance the performance of MnO₂ even at this high scan rate. For example, the capacitance contribution from MnO₂ could reach 160.0 and 140.2 F g⁻¹ as 20 wt% of MnO₂ were supported on CNTs and AB, respectively. Even at 80 wt% of MnO₂, MnO₂ supported on CNTs and AB could still gave a contribution of ∼20 F g⁻¹, while those on rGO gave ∼6.8 F g⁻¹.

The above results show that AB and CNTs gave similar contribution in supporting MnO₂ for enhancing specific capacitance, while rGO can enhance the capacitance much more than AB and CNTs. To demonstrate the capacitance at other scan rates, the average specific capacitance of MnO₂/CNTs and MnO₂/rGO composites at different scan rates are shown in Fig. 3. At all compositions, MnO₂/rGO showed higher capacitance than MnO₂/CNTs at slow scan rates. This is consistent with the results in reported literatures.^19–21,26 But, at high scan rates like 200 and 500 mV s⁻¹, MnO₂/CNTs showed better capacitance. The detailed rate capability analysis is given in the following paragraphs.

Fig. 3 Average specific capacitance of MnO₂/CNTs (a) and MnO₂/rGO (b) composites normalized by the total mass of MnO₂ and carbon at different scan rates. The mass loading of MnO₂ refers to the composition of MnO₂.

The specific capacitances of three types of MnO₂/C composites at different scan rates were plotted in Fig. 4. The capacitance here was normalized by the total mass of MnO₂ and carbon. Fig. 4a shows the three types of MnO₂/C composites with a loading mass of 20 wt% MnO₂. They all exhibited a dropping trend in capacitance as the scan rate increased gradually from 5 to 500 mV s⁻¹. However, MnO₂/rGO showed a remarkable drop as compared to the other two composites. In particular, the specific capacitance of MnO₂/rGO dropped sharply from ∼105.5 to ∼71.8 F g⁻¹ as the scan rate increased from 5 to 50 mV s⁻¹. The capacitance drop is over 30% (see the inset of Fig. 4a). In contrast, the capacitance drops of MnO₂/CNTs and MnO₂/AB from 5 to 50 mV s⁻¹ are only ∼16% and ∼20%, respectively. It should be noticed that the specific capacitance of MnO₂/rGO at the scan rate of 50 mV s⁻¹ is still higher than that of MnO₂/CNTs and MnO₂/AB although MnO₂/rGO gave the bigger drop. As the scan rate was increased further to 100, 200, and 500 mV s⁻¹, the capacitance of MnO₂/rGO dropped to 56%, 43%, and 25% of the highest capacitance obtained at 5 mV s⁻¹. The other two MnO₂/C composites gave smaller drops, i.e. as the scan rate reached to 100, 200, and 500 mV s⁻¹, the capacitance of MnO₂/CNTs dropped to 79%, 71%, and 54%, while that of MnO₂/AB dropped to 75%, 67%, and 49%, respectively. As a result, MnO₂/rGO lost its superiority in the specific capacitance at these high scan rates. At 200 mV s⁻¹, MnO₂/CNTs gave higher capacitance (∼52.1 F g⁻¹) than MnO₂/rGO (∼45.7 F g⁻¹), which in turn is higher than MnO₂/AB (∼41.3 F g⁻¹). At 500 mV s⁻¹, MnO₂/CNTs again gave the highest capacitance (∼39.6 F g⁻¹) and MnO₂/rGO gave the lowest value (∼26.5 F g⁻¹).

Fig. 4 Specific capacitance of MnO₂/C composites normalized by the total mass of MnO₂ and carbon at different scan rates. The four panels refer to the four compositions: 20 wt% (a), 40 wt% (b), 60 wt% (c), and 80 wt% (d) of MnO₂. The inset in each figure shows the capacitance drop in a percentage scale. The highest capacitance achieved at 5 mV s⁻¹ was defined as 100%.

Fig. 4b shows the specific capacitance of MnO₂/C composites with 40 wt% of MnO₂. At the range of 5–50 mV s⁻¹, MnO₂/rGO still exhibited higher capacitance than MnO₂/CNTs and MnO₂/AB. It also gave a faster capacitance drop to 66% of the highest capacitance achieved at 5 mV s⁻¹. Both MnO₂/CNTs and MnO₂/AB dropped their capacitance to ∼83% of the highest value achieved at 5 mV s⁻¹ (the inset of Fig. 4b). At 100 mV s⁻¹, three types of MnO₂/C composites showed similar specific capacitance (∼75 F g⁻¹). As the scan rate went 200 and 500 mV s⁻¹, MnO₂/CNTs and MnO₂/AB gave similar capacitance, however, MnO₂/rGO exhibited its capacitance ∼15 F g⁻¹ lower than MnO₂/CNTs and MnO₂/AB at each of these two scan rates. As compared to the highest capacitance at 5 mV s⁻¹, MnO₂/rGO dropped to 35% at 200 mV s⁻¹ and to 17% at 500 mV s⁻¹, while MnO₂/CNTs and MnO₂/AB could persist ∼62% at 200 mV s⁻¹ and ∼38% at 500 mV s⁻¹.

Fig. 4c and d show the specific capacitance of MnO₂/C composites with 60 and 80 wt% of MnO₂ at different scan rates. At high loading mass of MnO₂, the three types of composites showed similar capacitance value as well as the capacitance drop in percentage. At 60 and 80 wt% of MnO₂, the capacitance of MnO₂/rGO was slightly higher than that of MnO₂/CNTs and MnO₂/AB as the scan rate was below 50 mV s⁻¹. At the higher rates, the capacitance values of three types of composites were quite close with each other. For example, for 60 wt% of MnO₂, the capacitance drop in percentage is similar (the inset of Fig. 4c). They all dropped the capacitance to ∼32% at 200 mV s⁻¹ and ∼15% at 500 mV s⁻¹ of the highest value at 5 mV s⁻¹. For 80 wt% of MnO₂, at 500 mV s⁻¹ MnO₂/rGO could only achieve ∼6% of capacitance obtained at 5 mV s⁻¹, while MnO₂/CNTs and MnO₂/AB could keep ∼14%. The similar and significant drop in capacitance here for all three MnO₂/C composites should be ascribed to the high loading of MnO₂. It lowered the charge transfer resistance within the composites due to non-conductivity of MnO₂.

It is known that the charge storage on MnO₂ includes the charge insertion in the crystal of MnO₂ and the charge absorbed on the surface.²⁷ At a slow rate, the sufficient time allows the charge storage by charge insertion and surface absorption. As the scan rate increases, the insertion is not favoured and the capacitance is limited by the transportation of charge species within the composite electrode. Therefore the capacitance mainly comes from the surface absorption. For these three types of MnO₂/C composites, it is clear that carbon supports have significant influence on the charge storage mechanism. The faster capacitance drop of MnO₂/rGO indicates that the charge transportation in MnO₂/rGO could be slower than that in MnO₂/CNTs and MnO₂/AB. Note that the capacitance drop for pure rGO is similar to CNTs and AB (Fig. 1b) and the faster drop was only observed after combining with MnO₂. It is necessary to investigate the specific capacitance contributed only from MnO₂. Fig. 5 shows the capacitance contributed from MnO₂ by subtracting the capacitance contribution of carbon from that of composite. The four panels correspond to the four compositions. For 20 wt% of MnO₂, MnO₂ supported on rGO, CNTs, and AB gave specific capacitance of ∼310.2, 292.4, and 274.5 F g⁻¹, respectively, at 5 mV s⁻¹. It indicates that at a slow scan rate rGO could enhance the capacitance of supported MnO₂ more than CNTs and AB. However, the capacitance from rGO supported MnO₂ dropped quickly as the scan rate went to 10 mV s⁻¹ and the above. For example, at 10 mV s⁻¹ MnO₂ supported on rGO gave the capacitance of 271.7 F g⁻¹, which is slightly lower than MnO₂ supported on CNTs (274.4 F g⁻¹). At 20 mV s⁻¹, MnO₂ supported on rGO dropped the capacitance to 231.1 F g⁻¹, which is lower than the other two composites. The capacitance drop of rGO supported MnO₂ at 20 mV s⁻¹ was more than 25%, while that of CNTs and AB supported MnO₂ were only ∼10% (the inset of Fig. 5a). Such a drop is very significant for MnO₂/rGO. At 200 mV s⁻¹, MnO₂ supported on rGO gave the capacitance of 85.2 F g⁻¹, which is only 28% of the capacitance achieved at 5 mV s⁻¹. At 500 mV s⁻¹, MnO₂ supported on rGO only gave less than 10% of the capacitance. On the other hand, CNTs and AB exhibited better performance as the scan rate was increased. Their capacitance dropped less than 10% from 5 to 10 mV s⁻¹. Even at 500 mV s⁻¹, CNTs and AB supported MnO₂ still could give more than 50% of the capacitance at 500 mV s⁻¹.

Fig. 5 Specific capacitance of MnO₂ from the contribution of supported MnO₂ at different scan rates. The four panels refer to the four compositions: 20 wt% (a), 40 wt% (b), 60 wt% (c), and 80 wt% (d) of MnO₂. The inset in each figure shows the capacitance drop in a percentage scale. The highest capacitance achieved at 5 mV s⁻¹ was defined as 100%.

For 40 wt% of MnO₂ (Fig. 5b), the capacitance contribution from MnO₂ was similar in trend to 20 wt% of MnO₂. At the slow scan rate, MnO₂ supported on rGO showed higher capacitance contribution than those on CNTs and AB. As the scan rate increased to more than 20 mV s⁻¹, MnO₂ supported on AB contributed more capacitance than CNTs and rGo did. For example, MnO₂ supported on AB gave the capacitance of 199.1 F g⁻¹ at 50 mV s⁻¹, which is higher than CNTs (192.9 F g⁻¹) and rGO (163.1 F g⁻¹). Its capacitance at 100, 200, and 500 mV s⁻¹ are 181.2, 153.1, and 94.9 F g⁻¹, corresponding to 77%, 65%, and 40% of the capacitance at 5 mV s⁻¹, respectively. CNTs supported MnO₂ showed similar capacitance drop value as AB supported MnO₂ at higher rates. However, rGO supported MnO₂ again showed a remarkable drop at higher rates.

As the loading mass of MnO₂ was increased to 60 wt%, MnO₂ supported on AB was able to give a smaller drop as compared to the other two. However, the MnO₂ capacitance drops with the rate increase on all three types of carbons were closer than 20 wt% of MnO₂. At 200 mV s⁻¹, the capacitance contribution of MnO₂ supported on AB, CNTs, and rGO are 81.6, 69.2, and 46.6 F g⁻¹, respectively. From 100 to 500 mV s⁻¹, the dropping slops of the three types of MnO₂/C are similar if the lines are fitted. For 80 wt% of MnO₂ (Fig. 5d), MnO₂ supported on AB and CNTs contributed quite same capacitance at all high scan rates from 5 to 500 mV s⁻¹. MnO₂ supported on rGO exhibited a lower capacitance contribution at the high scan rates from 100 to 500 mV s⁻¹. At the slow rates of 5 and 10 mV s⁻¹, MnO₂ supported on rGO could gave higher contribution than those supported on CNTs and AB. At the scan rates of 20 and 50 mV s⁻¹, MnO₂ supported on the three carbons showed similar capacitance contribution. For all three carbons, supported MnO₂ dropped its capacitance more at high MnO₂ mass loading than those at low MnO₂ mass loading. For example, from 5 to 500 mV s⁻¹, the capacitance drop of 20 wt% MnO₂ supported on AB and CNTs was ∼50%, while that of 80 wt% MnO₂ on same carbons was ∼13%. As for rGO supported MnO₂, this drop in percentage is more significant. For all mass loading ratios, rGO supported MnO₂ dropped its capacitance to less than ∼10% from 5 to 500 mV s⁻¹.

The above results and analysis indicate that rGO is excellent in enhancing the capacitance of MnO₂ at the slow scan rates. This is consistent with the most reported results.^19–26 However, at the higher rates rGO was not able to enhance the performance of MnO₂ more than CNTs and AB. MnO₂ supported on rGO exhibited significant capacitance drop with the increase of scan rates. Note that among these three carbon supports, rGO's size can be over a few micrometres in two dimensions, while the other two carbons are much smaller in the same dimensions. It is probably due to the large size of graphene sheet and which hindered the charge transfer within the composite. As a result, the charge has to be stored through the surface absorption at the higher scan rates and the insertion storage is not facilitated. These results indicate that given same MnO₂, the carbon supports are influential on the capacitance and rate capability of supported MnO₂.

Cyclability of MnO₂/C composites

The cyclability of the three types of MnO₂/C composites with different MnO₂ loading mass was studied by Galvanostatic charge/discharge method at the current density of 3 A g⁻¹. Fig. 6 shows the charge/discharge curves of the three MnO₂/C composites with 20 wt% of MnO₂. The 2nd and 1000th cycles were selected to show the capacitance change. Among these three types of composites, MnO₂/rGO gave the highest capacitance at 2nd and 1000th cycles. However, it can be seen that MnO₂/rGO did not show a better stability after 1000 cycles. The capacitance retention for MnO₂/rGO after 1000 cycles was ∼78%. In contrast, MnO₂/CNTs and MnO₂/AB exhibited ∼97% capacitance retention after same number cycles.

Fig. 6 Galvanostatic charge/discharge curves of MnO₂/C composites with 20 wt% of MnO₂ at the current density of 3 A g⁻¹. Solid lines and dash lines refer to 2nd and 1000th cycles, respectively.

Fig. 7 shows the cyclability performance of the three types of MnO₂/C composites at the current density of 3 A g⁻¹. For 20 wt% of MnO₂ (Fig. 7a), MnO₂/CNTs and MnO₂/AB were stable in 1000 cycles. Their specific capacitance could be maintained at ∼59 and ∼39 F g⁻¹ during these cycles. The capacitance drop for them was about 4%. On the other hand, MnO₂/rGO exhibited a remarkable capacitance drop of >10% within 100 cycles. The drop continued to ∼20% as the electrode was cycled for 300 cycles. Afterwards, the drop became slower. Fig. 7b shows the composites with 40 wt% of MnO₂. MnO₂/CNTs and MnO₂/AB showed better stability during 1000 cycles. Their capacitance retention could reach nearly ∼98% and ∼95%. But MnO₂/rGO showed ∼20% capacitance loss after 1000 cycles. As the mass loading of MnO₂ increased to 60 and 80 wt%, MnO₂/CNTs and MnO₂/AB could achieve over 90% capacitance retention. In contrast, MnO₂/rGO showed a bigger capacitance loss (∼15%). During the cycling measurement, it was found that the electrolyte near the electrode changed to brown color if MnO₂/rGO was the working electrode. It is probably due to the better hydrophility of rGO, which facilitated the composite dissolution. As the mass ratio of rGO decreased, the dissolution of MnO₂/rGO composites was not that significant as compared to those with high mass ratios of rGO : MnO₂. As a result, the rGo composite with the low rGO loading could give slightly higher capacitance retention than those with the high rGO loading.

Fig. 7 Cyclability of MnO₂/C composites tested by Galvanostatic charge/discharge method at the current density of 3 A g⁻¹. The panels show the specific capacitance by total mass of the composite and the capacitance retention (in gray) as the function of cycles for the composites with the MnO₂ mass loading of 20 (a), 40 (b), 60 (c), and 80 wt% (d).

Conclusions

In summary, this article presents a comparison of three types of carbon supports in MnO₂/C supercapacitors in terms of capacitive performance and cyclability using the three-electrode approach. All three types of composites and electrodes were made using the conventional method to keep them same in configuration. The capacitance contribution of carbon supports was subtracted from the overall capacitance. The comparison, in particular, was given to the contribution only from supported MnO₂ among these three types of composites. It was found that rGO could enhance the capacitive performance of MnO₂ at slow scan rates. However, MnO₂/CNTs and MnO₂/AB exhibited better capacitive performance at high rates as well as better stability. The information from this study would be also helpful to the investigation of the influence of carbon supports on supported nanoparticle electrocatalysts.

Acknowledgements

This work is supported by Tier 1 grants of Singapore Ministry of Education (RGT8/13 & RG13/13), the COE Proposal Preparatory Grant 2013, and the Singapore National Research Foundation under its Campus for Research Excellence And Technological Enterprise (CREATE) programme.

Notes and references

1. B. Dunn, H. Kamath and J. M. Tarascon, Science, 2011, 334, 928–935.
2. W. F. Wei, X. W. Cui, W. X. Chen and D. G. Ivey, Chem. Soc. Rev., 2011, 40, 1697–1721.
3. L. L. Zhang and X. S. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531.
4. S. Liu, S. Sun and X.-Z. You, Nanoscale, 2014, 6, 2037–2045.
5. A. Sumboja, C. Y. Foo, J. Yan, C. Yan, R. K. Gupta and P. S. Lee, J. Mater. Chem., 2012, 22, 23921–23928.
6. Y. Cheng, H. Zhang, C. V. Varanasi and J. Liu, Energy Environ. Sci., 2013, 6, 3314–3321.
7. X. Wang, C. Yan, A. Sumboja, J. Yan and P. S. Lee, Adv. Energy Mater., 2013, 4, 1301240.
8. Y. Cheng, S. Lu, H. Zhang, C. Varanasi and J. Liu, Nano Lett., 2012, 12, 4206–4211.
9. A. Sumboja, C. Foo, X. Wang and P. S. Lee, Adv. Mater., 2013, 25, 2809–2815.
10. S. W. Lee, J. H. Kim, S. Chen, P. T. Hammond and Y. Shao-Horn, ACS Nano, 2010, 4, 3889–3896.
11. S. W. Lee, B. M. Gallant, H. R. Byon, P. T. Hammond and Y. Shao-Horn, Energy Environ. Sci., 2011, 4, 1972–1985.
12. A. S. Arico, P. Bruce, B. Scrosati, J. M. Tarascon and W. Van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
13. Q. Qu, P. Zhang, B. Wang, Y. Chen, S. Tian, Y. Wu and R. Holze, J. Phys. Chem. C, 2009, 113, 14020.
14. L. Mao, K. Zhang, H. S. On-Chan and J. Wu, J. Mater. Chem., 2012, 22, 1845.
15. X. Wang, A. Yuan and Y. Wang, J. Power Sources, 2007, 172, 1007.
16. X. Lu, D. Zheng, T. Zhai, Z. Liu, Y. Huang, S. Xie and Y. Tong, Energy Environ. Sci., 2011, 4, 2915.
17. L. Hu, W. Chen, X. Xie, N. Liu, Y. Yang, H. Wu, Y. Yao, M. Pasta, H. N. Alshareef and Y. Cui, ACS Nano, 2011, 5, 8904.
18. W. Xiao, D. Wang and X. W. Lou, J. Phys. Chem. C, 2009, 114, 1694.
19. Y. G. Guo, J. S. Hu and L. J. Wan, Adv. Mater., 2008, 20, 2878–2887.
20. Y. Cheng, H. Zhang, I. Cordova and J. Liu, J. Nano Energy Power Res., 2013, 2, 41–47.
21. J. Yan, A. Sumboja, X. Wang, C. Fu, V. Kumar and P. S. Lee, Small DOI:10.1002/smll.201303553.
22. C. Wei, L. Yu, C. Cui, J. Lin, C. Wei, N. Mathews, F. Huo, T. Sritharan and Z. Xu, Chem. Commun., 2014, 50, 7885–7888.
23. J. Zhu, D. Yang, X. Rui, D. Sim, H. Yu, H. E. Hoster, P. M. Ajayan and Q. Yan, Small, 2013, 9, 3390–3397.
24. W. S. Hummers Jr and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
25. C. Liu, Z. Yu, D. Neff, A. Zhamu and B. Z. Jang, Nano Lett., 2010, 10, 4863–4868.
26. Z. S. Wu, W. Ren, D. W. Wang, F. Li, B. Liu and H.-M. Cheng, ACS Nano, 2010, 4, 5835–5842.
27. M. Toupin, T. Brousse and D. Bélanger, Chem. Mater., 2004, 16, 3184–3190.
28. N. Mahmood, C. Zhang, H. Yin and Y. Hou, J. Mater. Chem. A, 2014, 2, 15–32.
29. N. Mahmood, C. Zhang, F. Liu, J. Zhu and Y. Hou, ACS Nano, 2013, 7, 10307–10318.
30. C. Zhang, N. Mahmood, H. Yin, F. Liu and Y. Hou, Adv. Mater., 2013, 35, 4932–4937.

---

Footnote

† Electronic supplementary information (ESI) available: TEM images of MnO₂ and MnO₂/C composites, cyclic voltammograms and electrochemical impedance spectra of MnO₂/C composites, the specific capacitance of MnO₂/AB at the different scan rates and composition. See DOI: 10.1039/c4ra04914d

---