Bao Zhang,
Xiao-wei Wang and
Jia-feng Zhang*
School of Metallurgy and Environment, Central South University, Changsha, 410083, PR China. E-mail: csuzjf@vip.163.com; Tel: +86-731-88836357
First published on 12th September 2014
A carbon-coated LiMnPO4·Li3V2(PO4)3 composite cathode material is synthesized from a rod-like MnV2O6·4H2O precursor prepared via aqueous precipitation for the first time, followed by chemical reduction and lithiation with oxalic acid as the reducing agent and glucose as the carbon source. XRD results indicate that orthorhombic LiMnPO4 and monoclinic Li3V2(PO4)3 co-exist. SEM results reveal that the thickness of the rod-like MnV2O6·4H2O precursor is about 80 nm and that the LiMnPO4·Li3V2(PO4)3/C composite possesses a micro/nano sphere-like morphology. HRTEM results indicate that the sample is a core–shell structure, where the external shell is amorphous carbon and at the core of the sample are LiMnPO4, Li3V2(PO4)3 and LiMnPO4·Li3V2(PO4)3 unit cells. The initial discharge capacity of the LiMnPO4·Li3V2(PO4)3/C composite is 110 mA h g−1, 104.4 mA h g−1, 100.6 mA h g−1 and 80.4 mA h g−1 at the rate of 0.1 C, 1 C, 3 C and 10 C, respectively. The cell shows excellent cycling stability and good rate capability as a cathode for lithium-ion batteries.
Compared to the commercially successful LiFePO4, LiMnPO4 has an equal theoretical capacity (171 mA h g−1) but can provide nearly 20% higher energy density. However, the low electronic conductivity of LiMnPO4 limits its performance at high currents.8 Similar to LiMnPO4, Li3V2(PO4)3 has been proposed to be a highly promising cathode material for high-power lithium ion batteries because of its open three-dimensional framework, which allows fast transport of lithium ions.9 Many efforts including metal-doping,10–13 coating with electronically conductive materials, such as carbon, metal, and metal oxide,14–17 and optimization of particles with suitable preparation procedures have been made to improve the performance of LiMnPO4 and Li3V2(PO4)3 cathode materials. Recently, Yang et al.18 reported that the discharge capacity at 1/20 C in the voltage range of 2.7–4.8 V was 10 mA h g−1 for LiMn0.95V0.05PO4 and 62 mA h g−1 for pure LiMnPO4. Wang et al.19 reported the performance of a different ratio of LiMnPO4 to Li2V2(PO4)3 in the composite synthesized through spray drying followed by solid-state reaction and demonstrated that the electronic conductivity of the LiMnPO4 could be enhanced by adding some Li3V2(PO4)3. Qin et al.20 reported the relationship between the ratios of LiMnPO4 to Li3V2(PO4)3 and the electrochemical performances of the composite materials, indicating the best ratio to be 0.6:
0.4.
In this work, we report a new way to synthesize LiMnPO4·Li3V2(PO4)3/C composite cathode material. We used NH4VO3 and Mn(CH3COO)·4H2O as starting materials to synthesize a rod-like MnV2O6·4H2O precursor via an aqueous co-precipitation route for the first time; LiMnPO4·Li3V2(PO4)3/C was prepared from the MnV2O6·4H2O precursor by chemical reduction, and the properties were innovationally investigated.
MnV2O6·4H2O + 4LiH2PO4 + 4HOOCCOOH·2H2O → LiMnPO4 + Li3V2(PO4)3 + 8CO2↑ + 12H2O↑ | (1) |
The mole ratio Mn/V of the as-prepared MnV2O6·4H2O was determined by chemical titration. Powder X-ray diffraction (Rint-2000, Rigaku) using Cu K radiation was employed to identify the crystalline structure and purity of the synthesized materials. The surface morphologies and particle information of samples were observed by SEM (JEOL, JSM-5600LV) and transmission electron microscope (TEM) (Tecnai G12). Carbon content of the composite was determined by C–S analysis equipment (Eltar, Germany).
Electrochemical characterizations were carried out using CR2025 coin-type cell. Typical positive electrode loadings were in the range of 2–2.5 mg cm−2, and an electrode diameter of 14 mm was used. For positive electrode fabrication, the prepared powders were mixed with 10% of carbon black and 10% of polyvinylidene fluoride in N-methyl pyrrolidinone until a slurry was obtained. The blended slurries were then pasted onto an aluminum current collector, and the electrode was dried at 120 °C for 4 h in air. The test cell consisted of a positive electrode and a lithium foil as negative electrode separated by a porous polypropylene film, and 1 mol L−1 LiPF6 in EC, EMC and DMC (1:
1
:
1 in volume) was used as the electrolyte. In addition, the assembly of the cells was carried out in a dry Ar-filled glove box. The coin cells were charged at 0.1 C and discharged at 0.1 C, 1 C, 5 C and 10 C over the voltage range of 2.5–4.5 V at ambient temperature using a battery testing system (Neware BTS-2000). Finally, cyclic voltammogram was carried out with a CHI660D electrochemical analyzer.
Fig. 1 shows the XRD patterns of MnV2O6·4H2O. All the peaks are indexed and consistent with the data of JCPDS#86-0720. The crystal structure is identified to be monoclinic, which is the same as that reported by JOEL.21
As illustrated in Fig. 2, the main part of the obtained MnV2O6·4H2O precursor are rod-like with the average width of the particles being around 200 nm and the thickness approaching 100 nm. A small number of rods aggregate and generate rods with the width and thickness approaching 2 μm.
Fig. 3 shows the typical Rietveld refinement XRD patterns of LMP·LVP/C synthesized at 700 °C for 10 h. The refined lattice parameters and phase content are listed in Table 1. The observed and calculated patterns match and the reliability factor (Rw) is acceptable. In addition, the weight ratio of LiMnPO4 (marked as LMP) to Li3V2(PO4)3 (marked as LVP) is 27.75:
72.25, which is consistent with the molar ratio of 1
:
1, indicating the purity of the rod-like MnV2O6·4H2O precursor. It can be seen that the cell volume of LMP in the LMP·LVP/C composites decreases compared with the pure orthorhombic LMP (space group Pnma, 25834-ICSD), which may be attributed to the V doping on the Mn sites because the ionic radius of V3+ (0.74 Å) is smaller than that of Mn2+ (0.80 Å). However, the cell volume of LVP in the LMP·LVP/C composites increases compared with the monoclinic LVP (space group P21/n, 96962-ICSD), indicating that Mn is also doped into the LVP host lattice. The abovementioned results predict that the Mn2+ doping LVP and V3+ doping LMP co-exist in the two-component composites for the utilization of MnV2O6·4H2O. As we know, this two-way behavior improved their electronic conductivity and electrochemical performance.18,22
Sample | a (nm) | b (nm) | c (nm) | Volume (nm3) | Weight ratio (%) | Rw (%) |
---|---|---|---|---|---|---|
LMP in LMP·LVP | 0.60958 | 1.04348 | 0.47418 | 0.30162 | 27.75 | 10.34 |
LMP | 0.610 | 1.046 | 0.4744 | 0.30269 | — | — |
LVP in LMP·LVP | 0.86068 | 0.85922 | 1.20362 | 0.89010 | 72.25 | 10.34 |
LVP | 0.86056 | 0.85917 | 1.20370 | 0.88998 | — | — |
Based on the C–S analysis result, the carbon from the glucose coating on the surface of the LMP·LVP/C composite and existing in the form of a carbon web between the particles add up to about 1.7 wt%. Oxalic acid acted as the reductant and ultimately decomposed into CO2 and H2O and ultimately, was over-contained in the composite. The SEM image of LMP·LVP/C synthesized at 700 °C for 10 h is shown in Fig. 4a. The particles in all the samples have sphere-like morphology. In order to investigate the distribution and the carbon coating information in the LMP·LVP/C composites, TEM was employed. Fig. 4b and c are the TEM images of the LMP·LVP/C synthesized at 700 °C for 10 h. The primary particle size is about 200–400 nm, and a conductive carbon network among the particles can also be observed in Fig. 4b and c. HRTEM of region I was used to further study the microstructure of the as-prepared composites. Two regions of different framework are presented in Fig. 4d. The lattice spacing of 0.23 nm of the right region corresponds to (012) of LMP phase and the lattice spacing of 0.26 nm of the left region corresponds to (113) of LVP phase. Fig. 4d indicates that the sample is a core–shell structure, where the external shell is amorphous carbon and at the core of the sample are LiMnPO4 and Li3V2(PO4)3 unit cells. The thickness of the carbon shell is about 4.8 nm.
The cyclic voltammetry curves of LMP·LVP/C at a scan rate of 0.1 mV s−1 are presented in Fig. 5b. It shows the redox peaks over the voltage range of 3–4.3 V for the composite corresponding to the plateaus of the charging and discharging curves. Three couples of redox peaks can be seen in the curves. The oxidation peaks of 3.64 and 3.73 V correspond to the extraction of the first Li+ in two steps: Li3V23+/3.5+(PO4)3 to Li2.5V23+/3.5+(PO4)3 and Li2.5V23.5+/4+(PO4)3 to Li2V23.5+/4+(PO4)3. The sharp oxidation peaks of 4.14 V corresponds to the extraction of the second Li+:Li2V24+/5+(PO4)3 to Li1V24+/5+(PO4)3. The reduction peaks around 3.53 V, 3.61 V and 3.99 V correspond to the insertion of the two Li+. No characteristic peaks of LMP can be observed in the curves. This phenomenon is unique and may be attributed to the peaks of LMP being covered by the peaks of LVP.
As seen in Fig. 5c, the initial discharge capacity of LMP·LVP/C at a rate of 0.05 C is about 136.5 mA h g−1 (the theoretical specific capacity (CT) of LMP·LVP is calculated by the following equation: CT = (171X1 + 133X2) (mA h g−1), where 171 and 133 are the theoretical capacities of LMP and LVP (<4.3 V), respectively; X1 and X2 are the weight content of LMP and LVP, respectively. Based on our Rietveld refinement results (Table 1 and Fig. 3), the weight ratio of LVP is 72.25%, so the CT for LMP·LVP/C synthesized in this paper is obtained as 143 mA h g−1). The real capacity of the synthesized LiMnPO4 is about 47 mA h g−1, and about 96 mA h g−1 for the Li3V2(PO4)3.
While the initial discharge capacity of LMP·LVP/C at the rate of 0.1 C, 3 C and 10 C can reach 126.2 mA h g−1, 115.7 mA h g−1 and 80.4 mA h g−1, respectively, the discharge capacity is about 113.8 mA h g−1 and 78.2 mA h g−1 after 30 cycles at the rate of 3 C and 10 C, respectively. The cell retains 98.35% and 97.3% of its initial discharge capacity. The composite material performs well at high rate; however, the discharge plateaus of LMP·LVP/C at different rates are obviously fading.
This journal is © The Royal Society of Chemistry 2014 |