Electrochemical properties of non-stoichiometric nanocrystalline Li₄Mn₅O₁₂ for hybrid capacitors

Nanocrystalline Li₄Mn₅O₁₂ was synthesized using the sol–gel method with the precursors Li(CH₃COO).2H₂O (Sigma Aldrich, 99.9%) and Mn(CH₃COO)₂ (Sigma Aldrich, 99.9%). The first step in the sol–gel process was to dissolve and agitate the precursors (with a molar ratio of Li:Mn = 4:5) into a mixture of ethylene glycol, ethanol and distilled water (with a volume ratio 6:3:1). Then, the citric acid solution (as a chelated agent) was added to the mixed solution under stirring, the ratio of citric acid to total metal ions was used in 1:1. The pH of solution was then adjusted using NH₃ solution (pH = 8–11). The 'sol' solution was heated and stirred at 100 °C until gel formation was observed. Finally, the gel was heated at 300 °C for 5 h in air.

The structures of the materials were identified using x-ray diffraction (XRD) using a D8-Advance (Bruker) diffractometer with a copper anode (λ_Kα = 1.5408 Å), XRD patterns were collected in the range 10–70° (0.029° s⁻¹). Lattice parameters were calculated using the software Celref. The morphology and the grain size distribution were determined using field emission scanning electron microscopy (FE-SEM; S4800 Hitachi, Japan) and transmission electron microscopy (TEM; JEOL JEM 1400).

The electrode paste was prepared by mixing Li₄Mn₅O₁₂ or active carbon (AC) powder with acetylene black, graphite and polytetraflouroethylene at a weight ratio of 80:15:5. The paste was laminated to 0.1 mm thickness, cut into pellets with a diameter of 10 mm and dried at 130 °C under a vacuum.

The electrochemical properties of nanocrystalline Li₄Mn₅O₁₂ were evaluated using cyclic voltammetry (CV) and the charge–discharge test at a high rate. The CV was performed in the 1 M Li₂SO₄ aqueous solution using a three-electrode cell where the working electrode is platinum coated with Li₄Mn₅O₁₂, the counter electrode is platinum and the reference electrode is Ag/AgCl (KCl saturated), respectively. The Li/Li₄Mn₅O₁₂ half-cell test was accomplished using Swagelok cells. The electrode Li₄Mn₅O₁₂ and Li foil were used as the positive and negative materials in the half-cell and a solution of 1 M LiPF₆ in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) at a ratio of EC:DCM = 2:1 was used as the electrolyte.

The asymmetric hybrid capacitors AC/Li₄Mn₅O₁₂ in the Swagelok cell consisted of Li₄Mn₅O₁₂ as the cathode and AC as the anode. The high rate charge–discharge test was performed in the potential window 0–1.4 V for the aqueous solution 1 M Li₂SO₄ and 0–3 V for the non-aqueous solution 1 M LiPF₆/EC:DMC (2:1). All electrochemical experiments were carried out using a MPG2 apparatus (Biologic, France).

The Li₄Mn₅O₁₂ phase is normally crystallized in the spinel phase with a cubic face center structure (space group: Fd3m): lithium ions are located simultaneously at the tetrahedral 8a and the octahedral 16d sites, manganese ions are distributed at the octahedral 16d sites and the oxygen ions occupy the 32e sites. As shown in figure 1, the XRD patterns of all the samples could be identified to a pure phase of spinel structure Li₄Mn₅O₁₂ (JCPDS: No. 46-0810) without the emergent second phase. According to the XRD results, the lattice parameters of the samples were calculated to be 8.1904 Å, 8.1990 Å, 8.1795 Å and 8.1893 Å, respectively (table 1). According to previous reports, the literature value of the lattice parameter is 8.161 Å [14, 17], thus the synthesized samples were non-stoichiometric. The oxidation value of manganese and the ratio Li:Mn were determined by oxidation–reduction titration and atomic absorption spectroscopy. As can be seen in table 1, all samples exhibiting manganese oxidation have values less than 4, the chemical formula was thus identified. It was apparent that the preparation of stoichiometric Li₄Mn₅O₁₂ was still not realized, in particular at temperatures lower than 600 °C [15]. The non-stoichiometric compounds are able to intercalate lithium reversible not only in the 3 V range but also in 4 V range at the expense of the reduction of Mn⁴⁺ and Mn³⁺.

Zoom In Zoom Out Reset image size

The XRD patterns of the low-temperature Li₄Mn₅O₁₂ were quite large (figure 1); it was considered that the samples crystallized in small particle sizes. The small particles and high surface area played important roles in improving the electrochemical double layer capacitance (the non-Faradic reaction).

SEM images of the Li₄Mn₅O₁₂ samples are shown in figure 2. The morphology of the Li₄Mn₅O₁₂ particles was affected by the pH value and the particles appear to be more regular and smaller with low pH values (pH = 8, 9). As can be seen in the figure 3(b), the sample S09 had a homogenous grain distribution, and the particle size fell into the nanometric dimension scale. The agglomeration of particles was clearly observed for other samples at pH values = 8, 10, 11. The TEM images of sample S09 confirmed that this sample had a polycrystalline structure and the particles size varied from 50–100 nm (figure 3).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The CV was performed by a three-electrode cell in the range 0.2–1.2 V (versus Ag/AgCl) for the aqueous 1 M Li₂SO₄. Figure 4(a) shows the CV curve of sample S09 at various scan rates (v) of 0.1, 1.0, 2.0, 5.0 and 10.0 mV s⁻¹. Two-redox reversible peaks were observed in all the CV curves, which confirms the reversible intercalation of Li⁺ ions into the host Li₄Mn₅O₁₂, corresponding to the redox reaction Mn(III) ⇄ Mn(IV). At low scan rates, two broad redox peaks could be observed: two anode peaks occurring at 0.91 and 1.04 V and two cathode peaks appearing at 0.83 and 0.93 V. Following the increase of the scan rate, it was observed that the peak positions has shifted around 300 mV and the redox peaks gradually became inconspicuous.

Zoom In Zoom Out Reset image size

The CV characterization was also performed in non-aqueous 1 M LiPF₆/EC:DMC (2:1) solution in the range 2.5–5.0 V. Figure 5(a) shows two oxidation–reduction peaks, similarly to the CV in aqueous solution. These peaks are also related to the redox reaction Mn(III) ⇄ Mn(IV). Increasing the scan rate results in shifted potentials of the oxidation–reduction peaks as well as the increase of the peak current.

Zoom In Zoom Out Reset image size

The specific capacitance (C_sp) is calculated by using half the integrated area of the CV curve to obtain the charge (Q), and subsequently dividing the charge by the mass of the electrode (m) and the potential window width (ΔV)

$$\begin{eqnarray*}&&{C}_{{\rm{s}}{\rm{p}}}=\displaystyle \frac{Q}{m{\rm{\Delta }}V}\cdot \end{eqnarray*}$$

The C_sp of samples at different scan rates in aqueous and non-aqueous electrolyte are presented in table 2. The values for C_sp in the aqueous solution are higher than the values in the non-aqueous solution due to the high conductivity of the 1 M Li₂SO₄ solution. In the aqueous solution, the samples reached high specific capacitances of 200, 190, 185 and 203 F g⁻¹ at the lowest scan rate (0.1 mV s⁻¹), and C_sp decreased with increasing of scan rate. At a scan rate of 10 mV s⁻¹, the C_sp of the samples was found to be around 50 F g⁻¹ for both the aqueous and non-aqueous solution. The CV results indicate that the charge storage of Li₄Mn₅O₁₂ relies on the mechanism of insertion/extraction of Li⁺ ions and the intercalation of Li⁺ occurring simultaneously on the surface and in the internal structure of the material.

Figure 6(a) shows the charge–discharge profile of samples at a C/5 rate in a half-cell. The charge–discharge curves showed two potential plateaus at 4.1–4.2 V and 4.1–3.9 V, corresponding to two oxidation–reduction peaks in the CV curve. All the samples showed stable performance after 50 cycles and sample S09 exhibited the best discharge capacity of 72 mAh g⁻¹. Kim et al [18]  reported that the Li₄Mn₅O₁₂ phase, prepared at 400 °C for 5 h, exhibited the discharge capacity of 46 mAh g⁻¹ at a C/5 rate.

Zoom In Zoom Out Reset image size

The charge–discharge performances of S09 and S10 were evaluated under different rates: C/5, C/2, 1C and 2C (figure 7). At the low rates (C/5, C/2), S09 and S10 reached a specific capacity around 70 mAh g⁻¹, while they presented a value of 12 mAh.g⁻¹ at the high rate 2C. The sample S10 exhibited a lower discharge capacity than S09 due to the high content of Mn⁴⁺ in the chemical formula.

Zoom In Zoom Out Reset image size

To investigate the performance of samples in a full-cell, the charge–discharged test of the hybrid capacitor AC/Li₄Mn₅O₁₂ was performed at rate of 4 A g⁻¹ in the aqueous and non-aqueous solutions. In the aqueous electrolyte (figure 8), the AC/Li₄Mn₅O₁₂ hybrid capacitor shows a relative high specific discharge capacitance and excellent cycling stability of capacitance during 200 cycles. The C_sp obtained for S08, S09, S10 and S11 were 14 F g⁻¹, 21 F g⁻¹, 6 F g⁻¹ and 12 F g⁻¹, respectively. The aqueous full-cell AC/Li₄Mn₅O₁₂ (S09) hybrid capacitor exhibited a specific energy of 31.1 Wh kg⁻¹. The highest C_sp of S09 could be attributed to the homogenous size distribution and nanometric particles compared to other samples presenting the agglomeration phenomena.

Zoom In Zoom Out Reset image size

Figure 9 shows the charge–discharge curve and cycle performance of a non-aqueous hybrid capacitor with the samples. S09 shows a good C_sp of 33 F g⁻¹, corresponding to a specific energy of 100 Wh kg⁻¹. However, the C_sp decreased below 15 F g⁻¹ after 100 cycles.

Zoom In Zoom Out Reset image size