1 Introduction

With the rapid increase of energy demand and environmental pollution, it is very important to explore clean and sustainable storage devices equipped with a fascinating energy density and power density [1,2,3,4,5]. Among various energy storage technologies, lithium-ion batteries (LIBs), owing to its fascinating energy density, long cycling lifespan and small memory effect, have evoked extensive attention over past decades [6, 7]. Graphite is widely used as commercial anode materials with a negligible volume expansion during the lithiation/delithiation process. However, the growth of lithium dendrites at low potential platform (~ 0.2 V) and low specific capacity (372 mAh g−1) make it unable to keep up with the increasing demand for new-generation LIBs [8,9,10]. Thus, it is inevitable to exploit advanced alternatives with the high specific capacity to substitute the graphite anode.

Currently, Li3VO4 is deemed as one of the most promising anode materials, attributed to the following reasons: (i) a more considerable capacity (394 mAh g−1) compared with graphite and Li4Ti5O12, which only deliver 372 mAh g−1 and 175 mAh g−1, respectively. Furthermore, the capacity can achieve 592 mAh g−1 when three Li ions reversibly insert in the Li3VO4 lattice [11, 12]; (ii) a high voltage platform (≈ 0.75 V vs. Li/Li+ in average) that can efficiently inhibit the formation of lithium dendrites [13]; (iii) a small volume expansion of only 4% during the cycling process [14]. Unfortunately, the unsatisfying electrochemical properties such as the poor rate performance and large polarization resulted from low intrinsic electronic conductivity severely restrict its practical applications [15,16,17]. A few serious efforts were taken to solve this drawback, for instance: (i) enhancing electrical conductivity by combining with conductive carbonaceous matrix such as organic carbon, graphite, graphene, etc.; Ni et al. [18] prepared Li3VO4/N-doped carbon, which exhibited unprecedented cycling stability (264 mAh g−1 after 2000 cycles at 2 A g−1), due to the improved lithiation degree and electronic conductivity; (ii) modulating the electronic structure through aliovalent ion doping and introducing oxygen defects: Fe-doped Li3VO4, reported by Liu et al. [19], delivered an outstanding rate capability, long lifetime cycling, which was mainly attributed to the formation of oxygen vacancies, reduced bandgap and fast Li+ diffusion; (iii) reducing the particle size or designing the specific morphology to shorten the diffusion path of Li+ and improve reaction activity: Shen et al. [20] synthesized highly homogeneous carbon-coated Li3VO4 spheres by a morphology inheritance route, showing a high capacity, excellent cycle stability and superior rate capability (reversible capacitiy are 287 mAh g−1 at a current rate of 20 C). Among these improved methods, hybridization with carbon is regarded as an easy and low cost method to achieve maximum benefits. In the meantime, LiVO2, as an analog of Li3VO4, which is deemed as the excellent ionic and electronic conductivity materials, becomes the suitable materials for hybridization with Li3VO4 [21,22,23].

Herein, we integrated Li3VO4 with LiVO2 and carbon by calcination at different temperatures in an inert atmosphere. Such a unique design combines some prominent merits: (i) LiVO2, an excellent ionic and electronic conductivity material, plays a critical role in accelerating electron transfer and Li+ diffusion upon charge/discharge; (ii) introduction of the carbonaceous matrix can not only boost electron transfer but also suppress aggregation of particles during cycling. Because of these advantages, LL@C-600 (annealed in an inert atmosphere of Ar at 600 °C) delivers a high specific capacity of 689 mAh g−1 after 100 loops at a current density of 100 mA g−1 and outstanding rate performance of 448 mAh g−1 at the current density of 2 A g−1, which is superior to the previously reported Li3VO4@carbon composites. These results further manifest that surface modified Li3VO4 is one of the most promising electrode materials for new-generation high-power density LIBs.

2 Experimental

2.1 Sample syntheses

2.1.1 Li3VO4 with a conductive surface layer

3 mmol LiOH·H2O and 1 mmol VO(acac)2 were dissolved in 20 ml of distilled water under stirring. Later, the obtained clear solution was freeze-dried for 2 days, and a brownish yellow floc was formed. At last, the above products were grinded, and annealed in an inert atmosphere of Ar at given temperatures (e.g., 550 °C, 600 °C and 650 °C) for 2 h at a heating rate of 5° min−1 to form Li3VO4 with a conductive surface layer. The corresponding samples were named as LL@C-550, LL@C-600 and LL@C-650, respectively.

2.1.2 Reference sample of Li3VO4 without a conductive surface layer

Using the above preparation procedure except for the treatment atmosphere, pure Li3VO4 was obtained by annealing in air at 600 °C for 2 h. The obtained sample was named as LVO.

2.2 Sample characterization

The sample structure was investigated by X-ray diffraction (XRD) on a Rigaku MiniFlex II X-ray diffractometer. The morphologies information of samples were examined by field emission scanning electron microscope (SEM, JEOL, model JSM-6700F) and transmission electron microscopy (TEM) images on a Tecnai G2S-TwinF20 apparatus. Raman spectra of the samples were recorded by a Renishaw Micro-Raman system. The chemical valence analysis of samples was performed using by X-ray photoelectron spectroscopy (XPS) with Al Ka radiation (1486.6 eV) as the X-ray source.

2.3 Electrochemical measurements of samples

The electrochemical performances of the samples were tested by CR2025-type coin cells. The electrodes were constructed in several steps below. The active materials were dried in a vacuum oven at 100 °C for 12 h. Then, active materials, carbon black, and binder at a weight ratio of 7:2:1 were mixed in N-methyl-2-pyrrolidinone (NMP) to form a slurry. The slurry obtained was coated onto a copper foil and dried in vacuum for about 12 h. The cells were assembled in a glove box (the volume fraction of H2O and O2 < 1 × 10−5%). 1 mol L−1 LiPF6 in ethylene carbonate and dimethyl carbonate (1:1 volume ratio) was used as electrolyte and the lithium metal as the reference electrode. The assembled cells were measured by a Neware battery test system. Cyclic voltammetry was obtained by an electrochemical workstation (CHI660C) at a scan rate of 0.2 mV s−1. The electrochemical impedance spectrum (EIS) was recorded in a frequency range of 100 kHz–0.01 Hz.

3 Results and discussion

3.1 Formation of Li3VO4 with a conductive layer LiVO2 via partial surface phase transformation

Partial surface phase transformation of Li3VO4 was highly possible when treated at a high temperature in the inert atmosphere, since under such conditions there may exist a valence reduction such as from V5+ to V3+. To examine this possibility, the preparation conditions were systematically studied by optimizing the treatment temperature in an inert atmosphere. Figure 1 shows the powder XRD patterns of the samples treated at different temperatures in an inert atmosphere of Ar. It is shown that most of the diffraction peaks are identical to those for the orthorhombic structure of Li3VO4 (JCPDS, No. 38-1247), while other three extra diffraction peaks located at two theta of 18°, 48° and 68°, respectively. The extra ones were well attributed to LiVO2 (JCPDS, No. 73-2047), indicating that (i) Li3VO4-based composite was formed by integrating with LiVO2, and (ii) vanadium species were reduced to low valence state likely by organic carbon at high temperatures in the inert atmosphere, as reported elsewhere [24]. Varying the treatment temperature did not change peak positions of both components, but the relative intensities.

Fig. 1
figure 1

Powder XRD patterns of the samples treated at different temperatures (marked as LL@C-550, LL@C-600, LL@C-650) in an inert atmosphere of Ar and at a fixed temperature of 600 °C (marked as LVO) in air. The standard diffraction data for Li3VO4 (JCPDS, No. 38-1247) and LiVO2 (JCPDS, No. 73-2047) were also given for comparison

Full size image

The phase composition for Li3VO4 when treated at a high temperature in inert atmosphere is further confirmed by examining the Raman spectra of LL@C-600, since Raman spectroscopy is sensitive to the lattice variations of materials. As shown in Fig. 2a, there are a D-band and a G-band located at 1348 cm−1 and 1586 cm−1, respectively, which are both typical peaks of carbon. Interestingly, the peak intensity ratio of D band to G band (ID/IG) is about 0.81, manifesting a high graphitization degree of carbon, as reported by others [25, 26]. Such a graphitization appears necessary to promote the electronic conductivity of Li3VO4. The existence of orthorhombic Li3VO4 was indicated by the magnified Raman spectrum in the wavenumber region of 200–1000 cm−1 (inset of Fig. 2a). In addition, the broadened and overlapped Raman signals in the region of 250–450 cm−1 cannot be well distinguished, indicating a mixed phase of Li3VO4 and LiVO2, according to Ref. [21].

Fig. 2
figure 2

Raman spectra of samples treated at 600 °C a in inert atmosphere (LL@C-600) and b in air (LVO). Inset of a is an enlarged view of 200–1000 cm−1 of LL@C-600

Full size image

All sample characterizations mentioned above demonstrate that when treated at a high temperature in the inert atmosphere, Li3VO4 could be achieved by integration with a conductive layer LiVO2 and carbon matrix. It is well established that the appearance of the surface coating could affect the morphology, which can be very important for the electrochemical performance. SEM image of pristine Li3VO4 is shown in Fig. 3a. Particles of the sample LVO exhibit an irregular morphology and a non-uniform particle distribution, which also tend to agglomerate. Figure 3b–d display SEM images of the samples when varying the treatment temperature in the inert atmosphere. All samples are composed of uniform sheet-like morphologies, especially when the treatment temperature is set at 600 °C (Fig. 3c). The uniform particles are believed to play a significance role in promoting the electrochemical properties of Li3VO4, for instance in enhancing the cycling stability to a great extent, which would be described later. Careful data examination showed that the sheet-like morphology is inclined to agglomerate obviously to form chunks as the treatment temperature increases. Figure 3e gives the TEM image of LL@C-600. The sheet-like morphology was composed of many small particles, and a mass of nanopores between particles were also revealed. The high-resolution TEM (HRTEM) image is displayed in Fig. 3f. The lattice spacings of 0.317 nm and 0.544 nm were assigned to the crystal planes (111) and (100) of Li3VO4, respectively. Besides, the interplanar spacings of 0.242 nm and 0.204 nm were associated with the planes (101) and (104) of LiVO2; manifesting that V5+ was reduced to the low valence state of vanadium by carbon matrix during the heat-treatment process in inert atmosphere. A part of Li3VO4 was transformed into LiVO2 phase, which is consistent with the XRD analysis.

Fig. 3
figure 3

SEM images of a the sample treated at a fixed temperature of 600 °C (marked as LVO) in air and b–d samples treated at different temperatures (marked as LL@C-550, LL@C-600, LL@C-650) in an inert atmosphere of Ar; e TEM image and f HRTEM image of LL@C-600

Full size image

To make sure that partial phase transformation occurred at surface, XPS was employed to study the surface chemical states and chemical components. Figure 4a shows the XPS data for Li3VO4, in which the signals located at binding energies of 517.7 and 525.2 eV matched well with those for V 2p3/2 and V 2p1/2 of V5+, as reported in literature [17, 27, 28]. The weak peak at 523.5 eV is in line with an X-ray satellite signal of O 1 s [27,28,29]. Compared to the characteristic XPS signals for LVO, there appeared additional XPS signals at 516.3, 517.0, 523.6, and 524.3 eV when treated in inert atmosphere. The signals at 516.3 and 523.6 eV correspond to the spin–orbit splitting of V 2p3/2 and V 2p1/2 of V3+, while the signals at 517.0 and 524.3 eV are in accordance with the assignments of V4+ as reported earlier [2, 29]. The appearance of the low valence state demonstrates that V5+ was reduced by carbon matrix during the heat-treatment process in inert atmosphere, in accordance with our XRD and HRTEM analyses.

Fig. 4
figure 4

High-resolution V 2p spectra and the corresponding fitting curves for samples treated at 600 °C a in air (marked as LVO) and b in the inert atmosphere (marked as LL@C-600)

Full size image

3.2 Formation of Li3VO4 without a conductive layer

When treated in air at a high temperature, no conductive layer was formed, but the pure phase Li3VO4. This is confirmed by XRD and Raman spectra. As shown by XRD data in Fig. 1, all diffraction peaks of LVO are identical to those of the orthorhombic structure of Li3VO4 (JCPDS 38-1247), while traces of impurity phases were absent. The formation of the pure phase of orthorhombic Li3VO4 is clearly demonstrated by the typical Raman signals appeared in the wavenumber region of 250–900 cm−1 (Fig. 2b), as reported elsewhere [30]. In the following, the sample without a conductive layer would be used as the reference for comparative studies.

3.3 Optimum electrochemical performance of Li3VO4 via a conductive layer

Electrochemical performance of Li3VO4 with and without a conductive layer was comparatively studied. Figure 5a displays the initial discharge–charge curves for samples LVO and LL@C (550 °C, 600 °C, 650 °C) with the potential range of 0.01–3 V at a rate of 100 mA g−1. As indicated by the similar discharge/charge curves, all samples exhibited the same electrochemical processes, while the initial capacities are totally different: the initial discharge/charge capacities of LVO samples are only 421/262 mAh g−1. Comparatively, for LL@C-550, LL@C-600 and LL@C-650 electrodes, the first discharge/charge capacities are much higher, which are 938/561, 1036/678, and 651/405 mAh g−1, respectively. The dramatically increased capacity is mainly ascribed to the synergistic effect of the conductive surface layer LiVO2 and carbon matrix. The 1st, 2nd, 10th, 30th, 50th and 100th charge–discharge voltage curves of the LL@C-600 electrode over a potential range between 0.01 V and 3 V at a current density of 100 mA g−1 are presented in Fig. 5b, in which the galvanostatic charge–discharge curves after the first cycle are overlapped well, implying a favorable reversibility of the electrode material upon charge/discharge [31, 32].

Fig. 5
figure 5

a Initial charge–discharge curves of LVO, LL@C-550, LL@C-600 and LL@C-650 electrodes; b Galvanostatic charge–discharge voltage profiles of the LL@C-600 electrode for different cycles at a current density of 0.1 A g−1

Full size image

The existence of the conductive layer LiVO2 in the LL@C-600 electrode does not change the electrochemical reaction processes, as demonstrated by the similar voltage curves after the first cycle for both LVO and LL@C-600 (Fig. 6). The positions of redox peaks of both samples are, however, slightly different. For Li3VO4 without a conductive layer, two obvious reduction peaks lie in the voltage of 0.496 and 0.805 V in the second cathodic scan, which are mainly related to the lithiation process (Eq. 1) [33, 34].

Fig. 6
figure 6

Cycle voltammograms at a scan rate of 0.2 mV s−1 for the samples a LVO and b LL@C-600 in initial five cycles

Full size image
$${\text{Li}}_{3} {\text{VO}}_{4} \; + \;x{\text{Li}}^{ + } \; + \;xe^{ - } \; \rightleftharpoons \;{\text{Li}}_{3 + x} {\text{VO}}_{4} .$$
(1)

By contrast, for the sample with a conductive layer, both reduction peaks were shifted to 0.518 and 0.821 V, respectively. The higher lithiation voltage of LL@C-600 compared to that of LVO indicates a lower barrier potential, as concluded elsewhere [19, 28, 34]. For the anodic segment, the anodic peak of LVO at 1.386 V is attributed to the delithiation process of Eq. 1. Correspondingly, the oxidation peak of LL@C-600 is centered at 1.364 V. The voltage changes between anodic and reduction peaks after the first cycles are provided in Table 1. The voltage changes of LL@C-600 displayed are obviously smaller than those of LVO, especially as the cycle progresses. This observation suggests that the sample with a conductive layer has lower polarization and faster Li+ diffusion than that without a conductive layer. Similar cases were found elsewhere [26, 33, 34].

Table 1 Voltage differences between anodic peaks and reduction peaks for samples LVO and LL@C-600 after the 1st cycle
Full size table

The cycling properties of all samples were assessed at a current density of 100 mA g−1 over the voltage window from 0.01 to 3 V, as shown in Fig. 7a. The comparison of the cycle property among LVO, LL@C-550, LL@C-600 and LL@C-650 electrodes indicates that LiVO2 and carbon play an important role in improving the electrochemical performance. Compared to the sample without a surface conductive layer (LVO), LL@C-550, LL@C-600 and LL@C-650 showed an obvious increase in the specific capacity, among which LL@C-600 presents the optimum specific capacity. Besides, the initial Coulombic efficiency (CE) of LL@C-600 is 66%, which is higher than 62.2% for LVO, 60.7% for LL@C-550, and 62.8% for LL@C-650. The irreversible capacity fading in the 1st cycle of the LL@C-600 electrode is mainly due to the formation of the irreversible solid electrolyte interface (SEI) film and side reactions related to electrolyte [19, 33]. Notably, after the 100th cycle, the LL@C-600 electrode still showed a high discharge capacity of 689 mAh g−1 that reached 96% of the discharge capacity of the second cycles. The CE retained an average value of 97% among subsequent cycles, manifesting an outstanding reversibility of Li+ in LL@C-600 electrode.

Fig. 7
figure 7

Electrochemical properties of LVO, LL@C-550, LL@C-600, and LL@C-650 electrodes: a cycling performance at a current density of 100 mA g−1; b rate capability at varied current densities; c cycling ability, Coulombic efficiency of LL@C-600 at a current density of 200 mA g−1

Full size image

In sharp contrast, LVO demonstrates a specific capacity of only 273 mAh g−1 after 100 loops. Such an outstanding cycling ability for LL@C-600 is probably attributed to the synergistic effect of LiVO2 and carbon. LiVO2 is an excellent electronic and ionic conductor, which can significantly elevate the electronic conductivity and Li+ ion diffusion coefficient of Li3VO4. In the meantime, the introduction of carbonaceous matrix can not only improve the electronic conductivity and prevent the agglomeration of LVO particles, but also alleviate the volume change. As for LL@C-550 and LL@C-650, the initial discharge/charge capacity were 651 and 395 mAh g−1, 938 and 589 mAh g−1, respectively, which are far larger than those of the LVO anode. This observation also demonstrates that the existence of LiVO2 and carbon matrix would obviously enhance the specific capacity. After 100 loops, the capacity retention for LL@C-550 and LL@C-650 is 105% and 94%, respectively, revealing an excellent cycling stability also.

To satisfy the rising demand of high-power devices, the superior rate capability of batteries is strongly aspired. Rate capability of LVO-based electrodes is displayed in Fig. 7b. The LL@C-600 electrode expresses a more superior rate capability compared with other electrodes. When the current densities are set at current densities of 0.1, 0.2, 0.5, 1 and 2 A g−1, the capacities of the LL@C-600 anode are 634, 600, 573, 529 and 448 mAh g−1, respectively, which are greater than those of other LVO anodes (267, 233, 175, 118, 84 mAh g−1 at rates of 0.1, 0.2, 0.5, 1 and 2 A g−1, respectively). When the current density recovers to 0.1 A g−1, a specific capacity of 690 mAh g−1 is observed, implying a superior reversibility. Besides these advantages, LL@C-650 sample presents a rate performance better than the pristine one, but does not appear as much improvement as the LL@C-550 electrode. The significantly improved rate capability could be the dual consequences of (i) rapid lithium-ion diffusion because of the good ionic conductivity of the conductive layer LiVO2 and (ii) fast electron transfer due to the good electronic conductivity of the carbon matrix.

Figure 7c displays the cycling properties of all samples over a voltage window of 0.01–3 V vs. Li/Li+ at a current density of 200 mA g−1. When at a higher current density, the LL@C-600 sample still shows a high discharge capacity of 522 mAh g−1 with a retention rate of 92% after 200 cycles, and a stable CE of 98 ± 2% after the first three loops. Table 2 compares the electrochemical properties of the LL@C-600 anode with those previously reported for Li3VO4@C based anodes. It appears that LL@C-600, when acts as an anode material for LIBs, exhibits the lithium storage performance superior to those ever reported in literature for LVO-related materials like LVO/C/rGO [11], LVO/C hollow spheres [16], mesoporous Li3VO4/C supported on rGO [39], Li3VO4-BMC [35] and LVO-m/c [29]. The excellent discharge capacity, longtime cycle stability and outstanding rate capability displayed by the LL@C-600 electrode could be attributed to the reasons below: (i) the conductive LiVO2 layer is an ionic conductor that could dramatically improve Li+ ion diffusion during cycling; (ii) Carbon is in situ formed during the calcination process that could be beneficial for fast electron transfer and inhibiting particle aggregation.

Table 2 Comparison of electrochemical performance of LL@C-600 and other LVO@C electrodes
Full size table

To figure out the possible factors responsible for the excellent rate capability, EIS was conducted to explore the electrochemical reaction kinetics. As shown in Fig. 8a, EIS plots of all electrodes consist of depressed semicircles and inclined lines. The impedance data were fit to a corresponding circuit, and the corresponding parameters are listed in Table 3. Apparently, the charge-transfer resistance (Rct) value of LL@C-600 is 67 Ω, which is smaller than 88 Ω for LL@C-650, 275 Ω for LL@C-550, and 350 Ω for LVO. Much smaller Rct value for LL@C-600 is exactly beneficial to the electron transfer and Li+ diffusion during electrochemical processes. Li+ diffusion during cycling was significantly improved by introduction of the conductive layer LiVO2. Lithium diffusion coefficient (D +Li ) could be evaluated according to the following Eq. 2:

Fig. 8
figure 8

a EIS plots and b Z′ vs. ω−1/2 for samples LVO, LL@C-550, LL@C-600 and LL@C-650. Inset of a shows the corresponding equivalent circuit model

Full size image
Table 3 Comparisons of impendence parameters for samples LVO, LL@C-550, LL@C-600, and LL@C-650
Full size table
$$D = R^{2} T^{2} /2A^{2} n^{4} F^{4} C^{2} \sigma^{2}$$
(2)

where R, A, T, n, C and F in the equation above represent the gas constant (R = 8.314 J mol−1 K−1), electrochemical active surface area, absolute temperature (T = 298 K), the number of electrons transferred during electrochemical processes, Faraday constant (F = 96500 C mol−1), molar concentration of Li+ ions (C = 9.8 × 10−3 mol cm−3), respectively, and the Warburg factor (σ) are able to be obtained from Eq. 3,

$$Z_{\text{re}} = R_{\text{ct}} + R_{\text{s}} + \sigma \omega^{ - 1/2}$$
(3)

The linear fitting plots of Zre vs. ω−1/2 are displayed in Fig. 8b. The obtained Li+ diffusion coefficient of LL@C-600 is 4.50 × 10−16, which is around two orders of magnitude higher than that of 7.32 × 10−18 for LVO.

The apparently reduced Rct and enhanced D +Li for LL@C-600 manifest an optimum rate performance relative to other samples, consistent with our electrochemical performance results. Furthermore, from the comparison of Nyquist plots of LVO, LL@C-550, LL@C-600 and LL@C-650 after 100 loops at a current density of 100 mA g−1 in Fig. 9, it is concluded that LL@C-600 has a superior charge-transfer ability during cycles, since after the 100th cycle, Rct of LL@C-600 is 85 Ω, which is smaller than 97 Ω for LL@C-550, 100 Ω for LL@C-650, and 110 Ω for LVO (Table 4).

Fig. 9
figure 9

EIS spectra of electrodes LVO, LL@C-550, LL@C-600, and LL@C-650 after 100 cycles, and the corresponding equivalent circuit model is inserted

Full size image
Table 4 The impendence fitting parameters for samples LVO, LL@C-550, LL@C-600, and LL@C-650 after 100 cycles at 100 mA g−1 (Ω)
Full size table

4 Conclusion

This work reports on the synthesis of Li3VO4 with a conductive layer LiVO2 and carbon matrix via one-step freeze-drying followed by a treatment at high temperatures in the inert atmosphere. The primary results could be summarized as follows: (i) The presence of the conductive layer on Li3VO4 boosted Li+ diffusion during the electrochemical process; (ii) In situ formation of carbon during treatment in the inert atmosphere elevated the electronic conductivity, and inhibited the particle agglomeration; (iii) After treatment at 600 °C, Li3VO4 with a conductive layer LiVO2 showed an optimum lithium storage capability, including the excellent discharge capacity, longtime cycle stability and outstanding rate capability when compared to that without a conductive layer; (iv) Through systematic sample characterizations, the excellent electrochemical performance at the presence of the conductive layer was demonstrated to be the consequence of fast electron transfer and rapid Li+ ion diffusion. The findings reported here may have a broad impact on exploring Li3VO4 and relevant vanadate materials for applications in next-generation lithium-ion batteries.