Introduction

Rechargeable lithium-ion batteries (LIBs) have been used in hybrid electric vehicles (HEVs), electric vehicles (EVs) and portable electronic equipment owing to their high voltage, high energy density and excellent cycling life [1,2,3]. Recently, lithium transition metal phosphates such as LiFePO4 [4, 5], LiMnPO4 [6, 7], LiCoPO4 [8] and Li3V2(PO4)3 [9,10,11,12] have attracted a great interest as potential cathode materials for lithium-ion batteries. The NASICON-structured Li3V2(PO4)3 has been synthesized in two different crystallographic structures: the monoclinic Li3V2(PO4)3 (M-LVP) and rhombohedral Li3V2(PO4)3 (R-LVP) [13, 14]. For the M-LVP, a three-dimensional structure is composed of slightly distorted PO4 tetrahedral and VO6 octahedral by sharing the oxygen vertexes [13]. The M-LVP has several advantages, including high operating voltage, ionic mobility and theoretical specific capacity (197 mAh g−1). As a cathode material, the M-LVP displays three voltage plateaus in charge–discharge curves corresponding to extraction/insertion of two lithium ions in a potential range of 3.0–4.3 V [15]. In contrast to the M-LVP, the R-LVP shows only a single voltage platform at 3.75 V, corresponding to the V3+/V4+ redox couple [14, 16]. The theoretical capacity of the R-LVP is 133 mAh g−1, corresponding to extraction/insertion of two lithium ions. The R-LVP with a single operation potential and considerable capacity is very attractive because it can supply a stable output power for electronic devices and electric vehicles to run well. Compared with the M-LVP, however, it is very difficult to directly synthesize the R-LVP due to its poor stability. [17,18,19]. Up to now, several feasible methods have been proposed to obtain rhombohedral compounds [16, 18, 20]. For example, Nazar’s and Goodenough’s groups have synthesized the R-LVP via chemical ion-exchange method [16, 18]. The partial substitution of metal Na+ ions for Li+ ions in pristine LVP can obtain R-Li3-xNa x V2(PO4)3 (R-LNVP) by stabilizing the rhombohedral structure [20]. Nevertheless, the electrochemical performance of the R-LVP or R-LNVP as cathode materials is usually poor, which restricts their practical application in lithium-ion batteries.

Recently, a few strategies have been applied to improve the electrochemical performance of R-LVP or R-LNVP, such as carbon coating [14] and nanosized particles [19, 21]. Yao’s group has synthesized carbon-coated R-LVP by ion-exchange reaction as a cathode material for lithium-ion batteries, which shows the reversible capacities of 97.7 and 76 mAh g−1 at 2 C and 5 C, respectively, and a capacity retention of 77% after 300 cycles at a rate of 1 C [14]. Zhang et al. [19] have prepared porous nanosheet-structured rhombohedral Li2NaV2(PO4)3/C using a facile sol–gel method. This Li2NaV2(PO4)3/C cathode displays a remarkable high rate capability (80 mAh g−1 at 5 C and 68 mAh g−1 at 10 C) and long-term cyclability. (About 93% of the initial capacity is retained over 500 cycles at 2 C.) However, the electrochemical performance, especially high-rate performance, cannot meet the requirement of their practical applications in lithium-ion batteries and needs to be further enhanced.

In this paper, rhombohedral Li2.4Na0.6V2(PO4)3@C (R-LNVP) nanoplates were successfully synthesized through a simple sol–gel method. The synthesized rhombohedral Li2.4Na0.6V2(PO4)3@C nanoplates as cathode materials have a remarkable electrochemical performance with excellent rate capability and long cycle life, delivering an initial discharge capacity of 121.6 mAh g−1 and a capacity retention of 95.6% after 200 cycles at 1 C rate in a potential range of 3.0–4.3 V. Even at a high rate of 10 C, the initial discharge capacity is 115.3 mAh g−1 and keeps 76.7% after 500 cycles.

Experimental

Preparation of the materials

Rhombohedral Li2.4Na0.6V2(PO4)3@C (R-LNVP) nanoplates were synthesized via a simple sol–gel method. In a typical procedure, stoichiometric amount of NH4VO3, C2H2O4·H2O, LiOH·H2O, NH4H2PO4, Na2CO3 and glucose was dissolved in 50 mL of deionized water under vigorous stirring at 70 °C, and the mixed solution was stirred continually till the green precursor gel was obtained. The obtained precursor gel was dried in a vacuum oven at 120 °C for 12 h to achieve a dry gel. Afterward, the dried precursor was pre-annealed at 350 °C for 4 h and then calcined at 750 °C for 10 h in an atmosphere of Ar gas containing 10% H2 to yield black R-LNVP nanoplates. Hybrid-phase Li2.4Na0.6V2(PO4)3@C (H-LNVP) nano/microspheres were prepared by a solvothermal method using NH4VO3, oxalic acid and polyvinylpyrrolidone (PVP) as starting materials. First, PVP was dissolved in 80 mL of ethylene glycol, and 0.01 mol NH4VO3 and 0.015 mol C2H2O4·H2O were added to the mixed solution under vigorous stirring at 70 °C. Stoichiometric LiOH·H2O, NH4H2PO4 and Na2CO3 dissolved completely into the mixed solution. The amount of PVP was based on 10 wt% residual carbon in Li2.4Na0.6V2(PO4)3@C. Then, the mixed solution was transformed into a 100-mL Teflon-lined stainless steel autoclave and heated at 180 °C for 2 h. After the solution cooled to room temperature, the solution was dried at 80 °C until the ethylene glycol was completely vaporized and dried at 120 °C for 12 h in a vacuum oven to achieve the precursor powders. Here, PVP not only served as a template, but also acted as a carbon source to enhance the electronic conductivity of Li2.4Na0.6V2(PO4)3. The sintering process was same as the synthesis of R-LNVP nanoplates.

Characterization of the materials

The crystal structure of the as-synthesized samples was identified using a PIXcel3D X-ray diffractometer with Cu Kα source. A field emission scanning electron microscope (FE-SEM, S-4800, 15 kV) was used to observe the morphologies of the samples. The microstructure of the samples was investigated by using a Tecnai G2 F20 H-800 transmission electron microscope with an accelerating voltage of 200 kV. The carbon content of the samples was determined by thermogravimetric (TG) analysis. X-ray photoelectron spectroscopy of the samples was collected using an ESCALAB-250Xi spectrometer. The Raman spectra were obtained using a HORIBA LabRAM HR800 equipped with a 514-nm Ar-ion laser.

Electrochemical measurements

The electrochemical properties of the samples were performed using CR2032 coin cells. The electrode was prepared by mixing active materials, acetylene black and polyvinylidene fluoride (PVDF) in a weight ratio of 80:10:10 in N-methyl-2-pyrrolidone (NMP). The slurry of mixture was coated onto aluminum foils using a film-coating machine and dried at 110 °C for 12 h in vacuum condition. The dried film was pressed in a fixed pressure, and then, the electrode sheet was punched into disks with a diameter of 14 mm. The electrolyte was 1 M LiPF6 in a mixture solution of ethylene carbonate (EC), dimethyl carbonate (DMC) and ethyl methyl carbonate (EMC) in a volume ratio of 1:1:1. The prepared electrode was used as the working electrode. The pure lithium plate served as the anode, and a polypropylene membrane (Celgard 2400) was used as the separator. The coin cells were assembled in an argon-filled glove box. Galvanostatic charge/discharge measurements were made on an Arbin battery testing system. Electrochemical impedance spectroscopy (EIS) testing was conducted using a Solartron electrochemical workstation with an AC signal of 5 mV and a frequency range from 10 mHz to 100 kHz. The cyclic voltammetry (CV) tests were also carried out using the Solartron electrochemical workstation at various scanning rates in the potential range of 3.0–4.3 V.

Results and discussion

Material characterization

XRD patterns of H-LNVP and R-LNVP samples are shown in Fig. 1a, b. All the diffraction peaks of the H-LNVP sample (Fig. 1a) could be indexed to a two-phase mixture of R-LVP and M-LVP. According to Rietveld refinement, the mole ratio of M-LVP to R-LVP is 58:42. The lattice parameters of the M-LVP are a = 8.664 Å, b = 12.212 Å, c = 0.8603 Å, and those of the R-LVP are a = b=8.339 Å, c = 22.454 Å, which are in good agreement with those of the previous studies [16, 18]. The obtained parameters are Rexp = 3.03%, Rwp = 6.03% and Rp = 4.62%, which represent the satisfactory quality of the Rietveld refinement. The R-LNVP sample (Fig. 1b) is nearly pure R-LVP phase with a space group of \( R\overline{3} C \) analyzed by Rietveld refinement with satisfactory reliability factors (Rexp = 2.07%, Rwp = 4.98% and Rp = 3.5%). The unit cell parameters are a = b=8.341 Å and c = 22.458 Å, and the atomic positions of the R-LNVP sample are listed in Table 1. The absence of rhombohedral Na3V2(PO4)3 in these two samples suggests that small amount Na can replace completely Li of pristine LVP structure. No diffraction peaks of carbon phase are observed in the XRD patterns of these two samples, indicating that the residual carbon is amorphous. XPS measurements were made to study the content and oxidation states of the elements of the prepared samples, as displayed in Fig. 1(c). The peaks at the binding energy of 55.1, 133.6, 190.8, 285.4, 531.2 and 1071.2 eV corresponds to Li1s, P2p, P2s, C1s, O1s and Na1s, respectively. The high-resolution V2p3/2 XPS spectrum in Fig. 1d shows a peak at 516.9 eV, which represents the oxidation state of V3+ and is in good agreement with the previously reported papers [22, 23]. The content of Na in H-LNVP and R-LNVP determined by XPS is approximately identical, with a molar ratio value about 0.55 and 0.56, respectively. In order to identify the structural characteristics of the carbon coated on the surface of LNVP particles, Raman spectroscopy measurements were made, as shown in Fig. 1e. Two intense broad peaks located at about 1347 and 1586 cm−1 in the Raman spectrum of the R-LNVP sample can be assigned to the disorder-induced phonon mode (D-band) and graphite band (G-band), respectively. The ratio of ID/IG is 0.97, indicating that a high graphitization degree of carbon existed in the carbon matrices [24]. The carbon content is determined to be around 6.5 and 8.9 wt% for the H-LNVP and R-LNVP samples (Fig. 4f), respectively, using the TG analysis technique.

Figure 1
figure 1

a, b XRD patterns of the H-LNVP and R-LNVP samples. c, d The whole XPS spectra and the V2p3/2 XPS spectrum of the R-LNVP sample. e Raman spectra of the R-LNVP sample. f TG curves for the H-LNVP and R-LNVP samples

Full size image
Table 1 Atomic positions of the R-LNVP sample deducted from Rietveld refinement of the XRD data
Full size table

Figure 2 shows SEM and TEM images of the samples. As illustrated in Fig. 2a, the H-LNVP sample consists of nano/micro-sized spheres. From Fig. 2b, the R-LNVP sample exhibits nanoplate-like particles with a thickness of about 20 nm and a width of around 500 nm. The morphology of two samples is confirmed by TEM (Fig. 2c, d) to be nano/microspheres and nanoplates, respectively. High-resolution TEM (HRTEM) images (Fig. 2e, f) reveal an amorphous carbon layer with a thickness of about 4.8 nm and 5.8 nm covering on the surface of the H-LNVP and R-LNVP particles, which is expected to provide a high electronic conductivity to the materials. The characteristic of two-phase coexistence in the H-LNVP sample is verified in Fig. 2e. The lattice fringes with a spacing of 0.440 and 0.373 nm correspond to the (012) planes of M-LVP and the (113) planes of R-LVP, respectively. In addition, the two-phase boundary is clearly observed in the HRTEM image of the H-LNVP sample. In contrast, a typical well-crystalline R-LVP is clearly identified in the R-LNVP sample. As shown in Fig. 2f, a lattice fringe spacing of 0.433 nm in the single-phase region is well matched to that of the (110) planes of R-LVP.

Figure 2
figure 2

SEM, TEM and HRTEM images of a, c, e the H-LNVP and b, d, f R-LNVP samples

Full size image

Electrochemical properties of the samples

CV curves of two samples were obtained at a scanning rate of 0.1 mV s−1 in the potential range of 3.0–4.3 V, as shown in Fig. 3a. Three anodic peaks at around 3.61, 3.69 and 4.1 V for the H-LNVP sample correspond to the extraction of lithium ion in the monoclinic phase, while the anodic peak at 3.82 V is attributed to the extraction reaction of lithium ions in the rhombohedral phase [14, 25]. Four corresponding cathodic peaks appear around 3.55, 3.63, 3.639 and 4.03 V, respectively. CV curve demonstrates the coexistence of monoclinic and rhombohedral phases in the H-LNVP sample. In contrast to the H-LNVP sample, the R-LNVP sample shows a pair of sharp redox peaks at around 3.81/3.69 V relating to the insertion/extraction of lithium ions in the rhombohedral phase and three very small auxiliary redox peaks caused by the insertion/extraction of lithium ions in the monoclinic phase, indicating that the R-LNVP sample consists of rhombohedral LVP and a negligible amount of monoclinic LVP. According to the CV curve, the mole ratio of R-LVP to M-LVP is estimated to be 95:5.

Figure 3
figure 3

a Cyclic voltammograms of the H-LNVP and R-LNVP samples at a scanning rate of 0.1 mV s−1 in the potential range of 3.0–4.3 V. b, c Cyclic voltammograms of two samples at various scanning rates. d A linear relationship between the oxidation peak current and the square root of scanning rate for two samples

Full size image

Figure 3b, c displays the CV curves of the H-LNVP and R-LNVP samples at various scanning rates of 0.1, 0.2, 0.5, 1 and 2 mV s−1, respectively. Even at a relatively high scanning rate (2 mV s−1), the well-defined and symmetrical redox peaks can be observed, implying the good lithium intercalation reversibility and rate performance of the samples. As shown in Fig. 3d, the highest oxidation peak current has a linear relationship with the square root of scanning rate, which is indicative of electrode reaction controlled by diffusion [26]. The lithium-ion diffusion coefficient D (cm2 s−1) can be calculated by the following Randles–Sevcik Eq. (1) [27]:

$$ i_{\text{p}} = \left( {2.69 \times 10^{5} } \right)n^{3/2} AD^{1/2} C_{\text{Li}}^{*} \upsilon^{1/2} $$
(1)

where ip is the peak current (A), n is the number of electrons involved during the redox process, A is the surface area of electrode (cm2), \( C_{\text{Li}}^{*} \) is the concentration of lithium ions in the electrode (mol cm−3), υ is the scan rate (V s−1), and D is the diffusion coefficient of lithium ions (cm2 s−1). Based on Eq. (1) and the slope of ip versus υ1/2 plots in Fig. 3d, the chemical diffusion coefficient of lithium ion in the H-LNVP and R-LNVP samples is calculated to be 5.06 × 10−10 and 1.01 × 10−9 cm2 s−1, respectively. The R-LNVP sample has a higher Li-ion chemical diffusion coefficient than the H-LNVP sample and LVP nanocrystals (10−10 cm2 s−1) [10], indicating better rate capability. It is resulted from a larger free volume of the interstitial space in rhombohedral LVP, supplying a faster lithium-ion movement [16].

The electrochemical performance of two samples was firstly evaluated by galvanostatic charge/discharge tests at 1 C (1 C = 133 mA g−1) in the potential range of 3.0–4.3 V. Figure 4a, b presents the charge/discharge curves of the H-LNVP and R-LNVP samples. The H-LNVP sample exhibits four pairs of charge/discharge plateaus, which agrees well with the redox peaks in the CV curves. The plateaus around 3.6, 3.68 and 4.10 V correspond to the reversible reaction of the lithium ions in M-LVP, and the plateau near 3.76 V is assigned to lithium-ion insertion/extraction in R-LVP. In comparison, the R-LNVP sample shows clearly a pair of flat charge/discharge platforms at 3.77/3.72 V, which is attributed to the V3+/V4+ redox couple in R-LVP [14, 16, 19]. The corresponding cycle performance of the two samples is presented in Fig. 4c. Both samples deliver high capacities and excellent cycling stability. An initial discharge capacity of the H-LNVP and R-LNVP samples is 115.9 and 121.6 mAh g−1, and a capacity retention is 100 and 95.6% after 200 cycles, respectively. Except the first several cycles, the coulombic efficiencies of two samples are close to 100% in the whole cycling processes. In order to evaluate objectively the superiority of the R-LNVP sample, a comparison of cycle performance of the H-LNVP and R-LNVP samples with same carbon content (8.9 wt%) is carried out, as illustrated in Fig. 4c. The H-LNVP sample (8.9 wt%C) was prepared by adding 2.4 wt% conductive carbon (acetylene black) to the H-LNVP sample (6.5 wt%C) and mixing uniformly. The capacities and cycle performance were investigated by charge–discharge testing at 1 C. The H-LNVP sample (8.9 wt%C) shows an initial discharge capacity of 117.5 mAh g−1, which is inferior to that of the R-LNVP sample (121.6 mAh g−1). In addition, the capacity retention of the H-LNVP sample (8.9 wt%C) is a little lower than that of the H-LNVP sample (6.5 wt%C). The results indicate that the effect of the increasing small carbon content (2.4 wt%) on the electrochemical performance of the samples is ignorable.

Figure 4
figure 4

The electrochemical performance of the H-LNVP and R-LNVP samples. a, b The charge–discharge profiles of at a rate of 1 C in the voltage of 3.0–4.3 V. c Cycle performances at 1 C rate. d Rate capability. e The long-term cycle stability at 10 C

Full size image

To evaluate the rate capability, the H-LNVP and R-LNVP samples are also tested at various current rates of 0.1 C, 0.5 C, 1 C, 5 C, 10 C and 20 C in the potential range of 3.0-4.3 V. The rate capability of two samples is displayed in Fig. 4d. The discharge capacities of the H-LNVP sample are 114.1, 115.6, 113.8, 105.4, 97.4 and 80.1 mAh g−1 at 0.1 C, 0.5 C, 1 C, 5 C, 10 C and 20 C, respectively. Compared with the H-LNVP sample, the R-LNVP sample exhibits higher discharge capacities of 121.6, 119.3, 115.6, 108.4, 102.9 and 94.7 mAh g−1 under the same testing procedures, indicating an outstanding rate capability. After 55 cycles, the current density recovers to 0.5 C; the discharge capacity of the R-LNVP sample can remain 116.7 mAh g−1, as high as 96.0% of the first capacity (121.6 mAh g−1). The gravimetric energy density of the R-LNVP sample (451.5 Wh kg−1) is higher than that of the H-LNVP sample (435.1 Wh kg−1) and LiMn2O4 (~ 430 Wh kg−1) [28]. The remarkable rate capability of the R-LNVP sample may be attributed to the synergetic effect of nanoplate-like structure and stable rhombohedral LVP. Compared with spherical H-LNVP, the R-LNVP nanoplates can provide a higher surface area and a shorter lithium-ion diffusion path, in favor of improving the rate capability. The stabilized NASICON framework of rhombohedral LVP could provide a better lithium-ion transport in LVP, leading to enhance the rate capability and cycle stability [16]. Furthermore, the uniform and thick carbon coating layer on the surface of the R-LNVP particles effectively increases its electronic conductivity and improves the electrochemical performance.

The high-rate long-term cycle stability of the H-LNVP and R-LNVP samples at 10 C is shown in Fig. 4e. The R-LNVP sample exhibits higher cycle stability than the H-LNVP sample, with an initial discharge capacity of 115.3 mAh g−1 and a capacity retention of 76.7% after 500 cycles. The excellent long-term cycle performance of the R-LNVP sample is due to the synergetic effect of nanoplate-like structure, stable rhombohedral LVP and uniform carbon coating layer. The electrochemical performance of the R-LNVP sample is comparable to previous reported plate-like V-based phosphate cathodes (Table 2), confirming the superiority of the R-LNVP sample developed in this work.

Table 2 A comparison of the electrochemical performance of R-LNVP@C nanoplates in this work and previous reported plate-like V-based phosphate cathodes
Full size table

To further understand lithium-ion insertion/extraction kinetics of the samples, EIS spectra of two samples were tested in fully discharge state of the cells after different cycles at 1 C. The AC voltage amplitude of 5 mV was employed in a frequency range of 10 mHz–100 kHz. From Fig. 5, all EIS spectra consist of a depressed semicircle at the high-frequency region and a sloping line at the low-frequency region. All the EIS spectra were simulated by an equivalent circuit model using the Zview program [32]. The semicircle at the high-frequency region is related to the charge transfer resistance (Rct). The sloping line in the low-frequency region represents the diffusion impedance of lithium ions in active materials. The fitting results are listed in Table 3. The Rct value of the R-LNVP fresh cell is 24 Ω and then increases to 54, 52 and 50 Ω after 1, 5 and 200 cycles, respectively. The large increase in Rct value in the first cycle can be attributed to the formation of solid electrolyte interface (SEI) film [11], then slightly decreases after 5 cycles owing to the perfect carbon coating layer and keeps almost unchanged over 200 cycles due to the stable rhombohedral structure. It can be found that the Rct value of the R-LNVP sample (50 Ω) is lower than that of the H-LNVP sample (70 Ω), suggesting the nanoplate-structured R-LNVP has fast charged species transport and enhanced rate capability.

Figure 5
figure 5

Impedance spectra of two samples. a Nyquist plots of the R-LNVP sample under different cycles at 1 C. b Nyquist plots of two samples after the 200th cycle at 1 C

Full size image
Table 3 The fitting results of EIS spectra for the H-LNVP and R-LNVP samples
Full size table

Conclusions

In summary, rhombohedral Li2.4Na0.6V2(PO4)3@C nanoplates have been successfully synthesized by a simple sol–gel method. For comparison, hybrid-phase Li2.4Na0.6V2(PO4)3@C nano/microspheres have also been prepared by a solvothermal method. As cathode materials for lithium-ion batteries, the rhombohedral Li2.4Na0.6V2(PO4)3@C nanoplates show superior electrochemical properties than the hybrid-phase Li2.4Na0.6V2(PO4)3@C nano/microspheres. Rhombohedral Li2.4Na0.6V2(PO4)3@C nanoplates deliver a high initial discharge capacity of 121.6 mAh g−1 and an excellent capacity retention of 95.6% after 200 cycles at 1 C in the potential range of 3.0–4.3 V. Even at a high rate of 10 C, they also exhibit an initial discharge capacity of 115.3 mAh g−1 and a capacity retention of 76.7% after 500 cycles. The outstanding electrochemical performance is ascribed to the synergetic effect of nanoplate structure, stable rhombohedral LVP phase and uniform carbon coating layer.