Nanostructured intercalation compounds as cathode materials for supercapacitors

Two types of SCs

One is the electrical double layer capacitor (EDLC), where the capacitance comes from the pure electrostatic charge accumulated at the electrode/electrolyte interface; therefore it is strongly dependent on the surface area of the electrode materials that is accessible to the electrolyte ions [17]. The thickness of the double layer depends on the concentration of the electrolyte and on the size of the ions and is in the order of 5–10 Å for concentrated electrolytes. The double layer capacitance is about 10–20 μF/cm² for a smooth electrode in concentrated electrolyte solution and can be estimated according to equation eq. (1)

assuming a relative dielectric constant ε_r of 10 for water in the double layer. The thickness and surface area of the double-layer are expressed as d and A, respectively. The corresponding electric field in the electrochemical double layer is very high and up to 10⁶ V/cm [5].

Supercapacitors are constructed much like a battery in that there are two electrodes immersed in an electrolyte, with ion permeable separator located between the electrodes (Fig. 1) [18]. In such a device, each electrode/electrolyte interface represents a capacitor so that the complete cell can be considered as two capacitors in series. For a symmetrical capacitor (similar electrodes), the cell capacitance (C_cell), will therefore be:

Fig. 1

Schematic representation of an electrochemical double layer capacitor in its charged state (modified from ref. [18]).

where C₁and C₂ represent the capacitance of the first and second electrodes, respectively [19]. In the case of EDLCs, carbon materials are used the most often because of their low electrical resistance, easy processability, chemical inertness, stability, and low cost. Typical materials include activated carbon, carbon aero gels, and carbide-derived carbon/ordered mesoporous carbon, carbon nanotubes and graphene [8].

The other one is pseudo-capacitance, which is originated from the redox reaction of the electrode material with the electrolyte [20]. The accumulation of electrons at the electrode is a Faradaic process where the electrons produced by the redox reactions are transferred across the electrolyte/electrode interface (Fig. 2) [11]. The theoretical pseudo-capacitance of metal oxide can be calculated as eq. (3):

Fig. 2

Charge storage mechanism of pseudocapacitive materials showing the surface charge transfer reaction (modified from ref. [11]).

where n is the mean number of the electrons transferred in the redox reaction, F is the Faraday constant, M is the molar mass of the metal oxide and V is the operating voltage window [21]. Redox reactions occur in the electro-active material because of several oxidation states. Often, reactions do not propagate into the bulk material, and occur only at the electrode/electrolyte interface [22]. Typical pseudocapacitive materials include conducting polymers [23] such as polypyrrole, polythiophene and polyaniline, and metal oxides [24] such as RuO₂, MnO₂, LiCoO₂ and LiMn₂O₄. Subsequently, intercalation oxides will be predominantly discussed in this review.

Main parameters of SCs

The key performance parameters of SCs include specific capacitance (normalized by electrode mass, volume, or area), energy density, power density, rate capability (retained capacitance at a high current loading) and cycling stability [25]. The energy density (E) of an SC is expressed as eq. (4):

the maximum power density of an SC is determined by eq. (5):

where C is the direct current capacitance in Farads, V the nominal voltage and R is the equivalent series resistance (ESR) in ohms [26]. In order to increase the energy density and power density of an SC, it is desirable to increase the specific capacitance (C) and the operating voltage window (V) as well as reduce the equivalent series resistance (R).

The capacitance of a SC is largely dependent on the characteristics of the electrode material, specifically, the surface-area and the pore-size distribution. Due to the high porosity, and correspondingly low density of electrode materials, it is generally the volumetric capacitance of each electrode that determines the energy density [18].

The maximum operating voltage window (V_m) is mainly dependent on the electrolyte used, which is limited by the electrochemical window of the electrolyte. Aqueous electrolytes have the advantage of high ionic conductivity (up to about 1 S cm^–1) and low cost. On the other hand, they have the inherent disadvantage of a restricted voltage range with a relatively low decomposition voltage of about 1.23 V [27]. Non-aqueous electrolytes of various types have been developed and the operating voltages of SCs based on them can be up to 3.5 V [28]. The electrical resistivity of non-aqueous electrolytes is, however, at least an order of magnitude higher than that of aqueous electrolytes and therefore the resulting capacitors generally have a higher internal resistance.

A high internal resistance limits the power capability of the SC and its application. In SCs, a number of sources contribute to the internal resistance and are collectively measured and referred to as the equivalent series resistance (ESR) [19]. There are several contributors to the ESR of SCs as follows:

1. electrolyte resistance;

2. ionic resistance of ions moving through the separator;

3. ionic diffusion resistance of ions moving in small pores;

4. electronic resistance of the electrode material; and

5. interfacial resistance between electrode and current collector.

Physical routes

Chemical routes

Other methods

In addition to the synthetic approaches discussed above, some other emerging methods for fast and facile preparation of nanomaterials have been reported in the past few years, such as physical routes including hot-injection method [66] and chemical routes including microemulsions method [67] and chemical vapor deposition [68]. Moreover, a range of other physico-chemical routes to prepare nanomaterials have also been reported. These include sonoelectrochemistry [69] and flame spray pyrolysis [70]. As to these methods, however, there are very rare reports about nanostructure intercalation compounds and they are not discussed here.

LiCoO₂

LiCoO₂ is a hexagonal layered structure belonging to Pnm space group with a=0.2816 nm and c=1.4056 nm. It is widely applied in lithium ion batteries due to its stable structure during charging and discharging process. Due to the good reversibility of intercalation/deintercalation of Li⁺ ions, which is shown in eq. (6), it could be used as positive electrode for aqueous SCs [74].

Its traditional preparation method is solid-state reaction. In order to get better electrochemical performance, nanostructured LiCoO₂ has been prepared by sol-gel methods for application in SCs [75, 76]. CV curves (Fig. 3) of LiCoO₂ in a saturated aqueous Li₂SO₄ solution show that the intercalation and deintercalation of lithium ions are similar to organic electrolyte solutions [77]. There are also three couples of redox peaks for LiCoO₂ in the aqueous solution, located at 0.87/0.71, 0.95/0.90 and 1.06/1.01 V (vs. SCE), which agree well with those in the organic electrolyte solution at 4.08/3.83, 4.13/4.03 and 4.21/4.14 V (vs. Li/Li⁺), respectively. The evolution of oxygen (about 1.7 V vs. SCE) in the aqueous electrolyte occurs at much higher potential, indicating the good electrochemical stability of LiCoO₂ as the positive electrode of the SCs. The diffusion coefficient of the lithium ions is of the same magnitude as that in the organic electrolyte; however, the current response and reversibility of redox behavior in the aqueous solution are better than those in the organic electrolytes due to its higher ionic conductivity [78, 79].

Fig. 3

Cyclic voltammogram of the LiCoO₂-electrode at different scan rates: (a) in the organic electrolyte and (b) in the saturated Li₂SO₄ solution (modified from ref. [77]).

The SC of AC//nano-LiCoO₂ in 0.5 M Li₂SO₄ solution at 1 A g^–1 between 0 and 1.8 V shows high capacitance and efficiency. The efficiency of this system increases to nearly 100 % after the initial cycle and the capacity does not change much after 40 cycles [80]. Due to the high cost of cobalt, there is not much work on its redox behavior in aqueous solutions.

LiMn₂O₄

LiMn₂O₄, whose reaction mechanism is shown in eq. (7), is found to be a good positive electrode for aqueous SCs due to its low cost compared to LiCoO₂ [81]. It mainly exists in a spinel structure. Manganese cations occupy half of the octahedral interstitial sites and Li⁺ ions occupy one eighth of tetrahedral sites. The Mn₂O₄ framework provides 3D interstitial space for Li⁺ ion transport, maintaining its structure over the compositional range Li_xMnO₄ (0 < x < 1) by changing the average Mn oxidation state between 3.5 and 4.0 [82]. As early as 1994, it was reported for the first time that LiMn₂O₄ can deintercalate and intercalate Li⁺ ions in aqueous electrolyte [83].

The LiMn₂O₄ spinel can be prepared by solid-state reaction, sol-gel method and hydrothermal methods [76]. Among them, one-dimensional (1D) nanostructures including nanowires, nanotubes and nanorods have attracted special attention [84, 85]. To improve the structural stability of LiMn₂O₄ from solid-state reaction, it can also be doped by heteroatoms such as Al, Cr, Cr–Fe and Ni like in organic electrolytes [86–88]. As it can be seen from Fig. 4a, a LiMn₂O₄ nanohybrid (LMO-NH) consisting of nanotubes, nanorods and nanoparticles has been synthesized using α-MnO₂ nanotubes from hydrothermal reaction as a precursor. This LiMn₂O₄ nanohybrid exhibits a high specific discharge capacitance of 415 F g^–1 at 0.5 A g^–1 in 0.5 M Li₂SO₄ aqueous solution. Even at 10 A g^–1, it still has a specific discharge capacitance of 208 F g^–1. The Ragone plots of the asymmetric SC based on AC/0.5 M Li₂SO₄/LMO-NH and the symmetric SC based on AC/0.5 M Li₂SO₄/AC are shown in Fig. 4b. The asymmetric SC presents an energy density of 29.8 Wh kg^–1 at the power density of 90 W kg^–1, much higher than that of the symmetric AC//AC capacitor. Moreover, it keeps an energy density of 11.2 Wh kg^–1 at power density of 1500 W kg^–1. Moreover, the capacity retention is 91 % after 1000 cycles [89].

Fig. 4

(a) The formation process of LiMn₂O₄ nanohybrid and (b) Ragone plots of the SC AC//LMO-NH and AC//AC (modified from ref. [89]).

LiMn₂O₄ nanochains exhibit a high reversible capacity of 110 mA h g^–1 at 4.5 C and 95 mAh g^–1 even at 91 C in 0.5 M Li₂SO₄ aqueous electrolyte. When charged at 136 C, 84 % capacity could be obtained [90]. As for the nanostructured LiMn₂O₄, when they are assembled into SCs, power density and cycling behavior are greatly improved. For example, the SC of AC// LiMn₂O₄ nanorod presents very high power density in 0.5 M Li₂SO₄, up to 14.5 kW kg^–1, and there is still 94 % capacity retention after 1200 full cycles [84]. The SC of AC//porous LiMn₂O₄ in 0.5 M Li₂SO₄ aqueous solution shows excellent cycling performance at the rate of 9C (1000 mA g^–1), with no more than 7 % capacity loss after 10 000 cycles [85]. The cycling behavior of AC//porous LiMn₂O₄ is the best amongst the reported LiMn₂O₄. The nanograins in the porous LiMn₂O₄ could restrain the Mn³⁺ on the surface from dissolution, which can also accommodate effectively the strain caused by Jahn–Teller distortion during the charge and discharge process and favor the morphological and structural stability [85]. Nanoporous LiMn₂O₄ spinel with a pore size of about 40–50 nm exhibits a high specific capacitance of 189 F g^–1 at 0.3 A g^–1. Even at 12 A g^–1, it still has a capacitance of 166 F g^–1. After 1500 cycles, there is no evident capacity fading [91].

Doping the LiMn₂O₄ with several cations is also a good way to effectively reduce the Jahn–Teller distortion and further improve the cycling stability [92]. In the CV curve of LiCr_0.15Mn_1.85O₄ in the saturated aqueous LiNO₃ (9 mol l^–1) solution, faster ‘‘CV response’’ of LiCr_0.15Mn_1.85O₄ is in correlation with higher capacity retention in comparison to undoped LiMn₂O₄ [86]. The Ni-doped LiMn₂O₄ presents much better rate capability and cycling behavior [87].

Li[Ni_1/3Co_1/3Mn_1/3]O₂

Li[Ni_1/3Co_1/3Mn_1/3]O₂ has a rhombohedral structure belonging to the R3m space group of a hexagonal α-NaFeO₂ structure. The lattice is formed by oxygen atoms in ABC stacking with alternating layers containing mixtures of nickel (+2), cobalt (+3), and manganese (+4) atoms. Its reaction mechanism as a positive electrode for SCs is as shown in eq. (8). During the de-intercalation of Li⁺ ions, the valence of Ni is changed from +2 to +3, and that of Co from +3 to +4. In the meanwhile, Mn stays at +4. Li[Ni_1/3Co_1/3Mn_1/3]O₂ powders are traditionally synthesized from heat-treating the coprecipitated spherical metal hydroxide with lithium salt or hydroxide [93].

The electrochemical stability of Li[Ni_1/3Co_1/3Mn_1/3]O₂ in a Li⁺-containing aqueous solution is critically dependent on the pH value [94]. As shown in Fig. 5, one couple of its redox peaks in 2 M Li₂SO₄ solution are situated at 0.48 and 0.68 V (vs. SCE), corresponding to the intercalation and de-intercalation of lithium ions in Li[Ni_1/3Co_1/3Mn_1/3]O₂, respectively. The oxygen evolution potential shifts to 1.29 V (vs. SCE), which is much lower than those for LiCoO₂ and LiMn₂O₄ systems. It shows that it is possible to extract lithium ions from the host before the evolution of oxygen [95]. The species of aqueous electrolyte, current density, scan rate and potential limit have influence on the capacitive property of Li[Ni_1/3Co_1/3Mn_1/3]O₂. The initial discharge specific capacitance of 298 F g⁻¹ was obtained in 1 M Li₂SO₄ solution within potential range 0–1.4 V at the current density of 100 mA g⁻¹ and was cut down <0.058 F g⁻¹ per cycling period in 1000 cycles [96]. In addition, it is curious that Li[Ni_1/3Co_1/3Mn_1/3]O₂ synthesized by the sol-gel method can also exhibit a good cycling performance in a 2 M LiNO₃ aqueous solution at different charge/discharge rates from 80 to 200 C in spite of low capacity [97].

Fig. 5

CV curve of Li[Ni_1/3Co_1/3Mn_1/3]O₂ in 2 M Li₂SO₄ aqueous solution (modified from ref. [95]).

The electrochemical performance of Li[Ni_1/3Co_1/3Mn_1/3]O₂ in aqueous solution can be modified by mixing with PPy. The redox reversibility and capacity retention are much improved compared to those of the pristine Li[Ni_1/3Co_1/3Mn_1/3]O₂. The main reason is that the added PPy could enhance the electronic conductivity of Li[Ni_1/3Co_1/3Mn_1/3]O₂ electrode [98].

Li_1+xV₃O₈

Li_1+xV₃O₈can be used as a lithium insertion host, and possess many advantages such as the large discharge capacity, the high rate capability and the good cycle performance. These are caused by the uptake of more than three moles of lithium per formula unit, the fast diffusion of lithium in the compound, and the structural stability against lithium insertion and extraction, respectively [99]. This vanadate belongs to a monoclinic system with the space group P2₁/m and has a layered structure where adjacent layers with the nominal composition of V₃O₈^- are held together by Li ions at the octahedral sites in the interlayer. Excess lithium corresponding to the amount x is accommodated at the tetrahedral sites between the layers [100]. Lithium insertion reaction of Li_1+xV₃O₈ is described as three steps composed of a single-phase reaction for the range 0 < x < 2.0 in Li_1+xV₃O₈, a two-phase reaction for 2.0 < x < 3.2, and a single-phase reaction for 3.2 < x < 4.0. Phase transformation occurs at x=2.0 from the original LiV₃O₈ phase to the second Li₄V₃O₈ one [101].

It is well known that the preparation method for Li_1+xV₃O₈ strongly influences its electrochemical properties. Traditional synthesis method of Li_1+xV₃O₈ was that Li₂CO₃ reacted with V₂O₅ at 680 °C [102]. To improve the electrochemical properties of this material, much attention has been paid to the new synthetic methods and post treatments, mainly the sol–gel method [103], hydrothermal reaction [104], microwave-assisted synthesis [105], and ultrasonic treatment [106]. To be more precise, single-crystalline LiV₃O₈ nanobelts have been synthesized by using V₂O₅ nanobelts as the vanadium resource via a facile solid-state reaction method. The LiV₃O₈ nanobelts with thickness of <20 nm, width of 100–300 nm, and length up to several micrometers can grow along the [–201] direction [107]. In addition, an amorphous wrapped [98] orientated nanorod structure has been fabricated in a LiV₃O₈ thin film by adjusting the oxygen partial pressure in the deposition process using radio frequency (RF) magnetron sputtering [108].

LiV₃O₈ nanobelts exhibit stable lithium-ion intercalation/deintercalation reversibility and deliver initial specific discharge capacities of around 356.2 mA h/g at 0.02 A/g and 234 mAh g^–1 at 0.1 A/g, respectively. After 30 cycles, the specific discharge capacities are lowered to 298.6 mAh g^–1 at 0.02 A/g and 195.5 mAh g^–1 at 0.1 A/g, respectively [107]. In addition, LiV₃O₈ nanorods with a (100) preferred orientation can provide more entrances and shorter diffusion pathways for insertion of Li ions, which facilitate high capacity and excellent rate capability. However, the lithiation planes are perpendicular to the Li ion transport pathways when the (100) planes are parallel to the substrate for the film electrode, which blocks the flow of Li ions, as shown in Fig. 6a. Li ions therefore have to move through the close-packed oxygen layer, which leads to poor rate properties for a film electrode. After introducing the amorphous wrapping layer, the migration of Li ions proceeds firstly in a three-dimensional manner along the outside of the nanorods and then across the interface towards the (100) plane. This amorphous wrapping layer can avoid the close-packed oxygen layer and remove the anisotropy of the nanorods, which therefore improves the sluggish kinetics of Li ion insertion, as shown schematically in Fig. 6a. Moreover, the amorphous layer is not an inert phase but the same as the bulk material, so that it can provide Li ions in various sites and be convenient for delivery of Li ions to the (100) planes where it can be inserted [108].

Fig. 6

(a) Illustration of the diffusion pathway for Li ions in LiV₃O₈ nanorods with and without amorphous wrapping and (b) cyclic voltammograms of nanostructured LiV₃O₈ thin film recorded in the first 5 cycles (modified from ref. [108]).

The cyclic voltammogram (Fig. 6b) exhibits two reduction peaks, at 2.56 V and 2.81 V, and oxidation peaks at 2.52 V and 2.78 V, corresponding to the extraction and insertion of Li ions in monoclinic LiV₃O₈, respectively. The sharp peaks at 2.81 and 2.52 V indicate the depth of the phase transformation during the charge and discharge processes. The curves obtained for five cycles remain identical to each other, indicating a stable and good reversible insertion/extraction reaction [108]. As for the Co_0.58Ni_0.42 oxide coated LiV₃O₈ nanomaterials, 5.0 wt % Co_0.58Ni_0.42 oxide coated LiV₃O₈ shows the best rate capability and cyclability. The initial discharge capacity of 5.0 wt % Co_0.58Ni_0.42 oxide coated LiV₃O₈ is 243.3 mAh g^–1 at 1 C, which is nearly the same as that at 0.5 C. After 60 cycles, the discharge capacity of 5.0 wt % Co_0.58Ni_0.42 oxide coated LiV₃O₈ remains at 203.9 mAh g^–1 [109].

Na_xMnO₂

The known phases of Na_xMnO₂ (x=0.2, 0.40, 0.44, 0.70, 1) have been summarized [110]. There are two phases for NaMnO₂. Low-temperature α-NaMnO₂has an O3 layered structure with a monoclinic structural distortion due to the Jahn-Teller distortion of the Mn³⁺ ion. At high temperature, the orthorhombic β-NaMnO₂ forms in a different layered structure containing MnO₂ sheets consisting of a double stack of edge-sharing MnO₆ octahedra. Na occupies the octahedral sites between two neighboring sheets [111]. First principles computations indicate that the monoclinic NaMnO₂ is energetically more stable than other competing phases. This is in contrast to LiMnO₂, which prefers an orthorhombic structure [112, 113]. Both α- and β-NaMnO₂ were electrochemically tested as a positive electrode material in 1985. Their results showed that only 0.22 and 0.15 Na could be reversibly extracted and reintercalated in α- and β-NaMnO₂ respectively. Besides NaMnO₂, Na_0.4MnO₂ and Na_0.6MnO₂ have also been studied as a positive electrode [113].

Among the small number of oxide materials of potential interest identified, Na_0.44MnO₂ is particularly attractive because of its adequate crystal structure forming suitable large-size tunnels for sodium incorporation [114]. Moreover, in previous reports, the phases prepared by solid-state synthesis contained Mn₂O₃-bixbyite impurities. To avoid the presence of such parasitic phases, acidic treatment using HCl is used for their dissolution. However, this procedure concomitantly induces sodium leaching, leading to an isostructurally deficient sodium phase of approximate composition Na_0.2MnO₂ [114]. Recently, some researchers pointed out the possibility of preparing pure Na_0.44MnO₂ powders by accurately adjusting the solid-state synthesis parameters [115]. In addition, NaMnO₂ was synthesized by ball milling mixtures of Na₂CO₃ and MnO₂ at a molar ratio of 1:2 for 12 h followed by heating at 870 °C for 10 h [116]. Na_0.35MnO₂ nanowire is prepared by a simple and low energy consumption hydrothermal method [117].

The capacitance of the nanowire Na_0.35MnO₂ (157 F g^–1) is much higher than that of the rod-like Na_0.95MnO₂ (92 F g^–1). It presents excellent cycling performance even the oxygen in the aqueous electrolyte is not removed, no evident capacitance fading after 5000 cycles. The Na_0.35MnO₂ nanowire delivers an energy density 42.6 Wh kg^–1 at a power density of 129.8 W kg^–1 when testing using activated carbon as the anode, higher than that of the rod-like Na_0.95MnO₂, 27.3 Wh kg^–1 at a power density of 74.8 W kg^–1 [117]. Fig. 7a shows the cyclic voltammogramms (CV) of NaMnO₂ electrode in 0.5 M Na₂SO₄ aqueous solution at a scan rate of 5 mV s⁻¹. The current collector (Ni-mesh) is very stable in the range −1.0 < E_SCE< 1.2 V in Na₂SO₄ aqueous solution. Hydrogen and oxygen evolutions occur at E_SCE < −1.0 V and > 1.2 V, respectively, because of significant overpotentials consistent with results previously reported for aqueous Li₂SO₄ solution [118]. In the potential range of 0–1.1 V the behavior of the NaMnO₂ electrode slightly deviates from the ideal rectangular shape with two small redox couples, indicative of the pseudocapacitance properties of NaMnO₂ cathode. These redox peaks are similar to that of Na_xMnO₂·H₂O, which can be definitely ascribed to the intercalation/deintercalation of Na⁺ into and from the solid lattice [119]. The SC exhibits a sloping voltage–time curve in the entire voltage region of 0–1.9 V and delivers an energy density of 19.5 Wh kg⁻¹ at a power density of 130 W kg⁻¹ based on the total mass of the active electrode materials. It can be seen from Fig. 7b that also shows excellent cycling behavior with not more than 3 % capacitance loss after 10 000 cycles at a current rate of 10C [116].

Fig. 7

(a) CV curves of NaMnO₂ in 0.5 M Na₂SO₄ aqueous solution and (b) the cycling behavior of the asymmetric AC//NaMnO₂ SC (modified from ref. [116]).

K_xMnO₂

Lithium and sodium intercalation compounds can be a promising cathode for SCs as mentioned above, potassium intercalation compounds can also be used as positive electrodes. The lamellar structure of LiMnO₂ has been demonstrated to be unstable during cycling and it transforms into the spinel form easily since the lamellar compound has a cubic-close-packed arrangement of oxide ions which is identical to that of a spinel [120], therefore introduction of large alkali ions and H₂O molecules into the interlayer space of manganese oxide could stabilize the lamellar structure [121]. For example, lamellar K_xMnO₂, with the large K⁺ ions as pillars, shows better cycling performance than Li_xMnO₂ during the Li⁺ intercalation/deintercalation in organic electrolytes. This can be ascribed to the expansion of the interlayer space. At the same time, large K⁺ ions make manganese cations diffusion into the interlayer region to form spinel structure less favorable [122]. When K_xMnO₂·yH₂O is used as electrode material in an aqueous capacitor, the existence of these interlayer H₂O molecules is not supposed to influence the performance of the capacitor, unlike commercial lithium-ion batteries using organic electrolytes, which require anhydrous environment for the assembly of battery. In addition, lamellar K_xMnO₂·yH₂O lattice possesses a large interlayer distance of about 0.7 nm that could be intercalated by large quaternary ammonium cations and other alkaline cations [123–125].

There are many scientific reports about the synthetic routes to prepare K_xMnO₂. For instance, sol–gel synthesis usually involves various sugars and other organic polyalcohols as well as further heating to develop good crystallinity and remove the organic additives [126]. Precipitation routes usually involve not only the oxidation of aqueous Mn²⁺ cations but also the oxidation of amorphous solid precursors via the redox reaction between Mn²⁺ and MnO₄⁻ [127]. A soft-template method for the synthesis of layered manganate (K_xMnO₂, birnessite) compounds with a relatively good control at the nanoscale regime on the number of octahedral layers is described [128]. K_xMnO₂ precursor is prepared by ball milling the mixture of K₂CO₃ and MnO₂ in a molar ratio of 1:2 for 12 h, followed by calcination at 550 °C for 8 h. The precursor is further washed several times with water to remove residual K₂CO₃ [129].

A schematic diagram is shown in Fig. 8a to illustrate the different impedance behaviors of K_xMnO₂·nH₂O electrodes in Li₂SO₄, Na₂SO₄ and K₂SO₄ aqueous electrolytes. The molar ionic conductivity of the three types of alkaline cations in aqueous solutions (K: 73.5 S cm²mol^–1, Na: 50.1 S cm² mol^–1, Li: 38.6 S cm²mol^–1) [130], namely their migration rate, is supposed to account for the different electrolyte resistance. Since the hydrated ionic radius of the three alkaline cations are similar (K⁺: 3.31 Å, Na⁺: 3.58 Å, and Li⁺: 3.82 Å) [130], the different capacitive features should be mainly affected by their different solvation interactions. K⁺ possesses the weakest solvation interaction with H₂O due to its smallest charge density, making it dehydrate readily in the interior of the porous electrode. As a result, the charge transfer process occurs facilely and the double-layer capacitance is formed rapidly on the electrode/electrolyte interface. All the above results explain the best power ability of K_xMnO₂·nH₂O in the K₂SO₄ electrolyte [131].

Fig. 8

(a) Schematic illustration of the hydrated ionic radius and migration rate of three alkaline cations and (b) CV curves of K_xMnO₂·nH₂O electrodes in the three aqueous electrolytes at the scan rate of 5 mV s^–1 (modified from ref. [131]).

CV curves of K_xMnO₂·nH₂O electrodes in the three aqueous electrolytes at the scan rate of 5 mV s^–1 are presented in Fig. 8b. A couple of reversible redox peaks can be observed distinctly for all the three electrolytes, suggesting the Faradic pseudocapacitive nature of K_xMnO₂·nH₂O material. Previous work on the crystalline structure and chemical composition of K_xMnO₂·nH₂O during electrochemical cycling in K₂SO₄ electrolyte indicates that this couple of redox peaks is concerned with the reversible intercalation/deintercalation of K⁺ into/from crystalline K_xMnO₂·nH₂O lattice [132]. In this case, the three electrolytes lead to slightly different potentials of these redox peaks. In addition, the redox peaks of K_xMnO₂·nH₂O electrodes are considerably more distinct than those of MnO₂ reported in literatures, which can be ascribed to the high crystallinity of K_xMnO₂·nH₂O material [131]. Studies of the electrochemical behavior of K_0.27MnO₂·0.6H₂O in K₂SO₄ show an energy density of 25.3 Wh kg⁻¹ at power density of 140 W kg⁻¹ based on the total mass of the active electrode materials. It also shows excellent cycling behavior with no more than 2 % capacitance loss after 10 000 cycles at a current rate of 25C [129].