With the increasing demand for renewable energy resources, lithium-ion batteries (LIBs) have attracted tremendous attention in recent years (Uddin et al., 2017; Wu et al., 2019; Zheng et al., 2019). As one of the most important parts of LIBs, cathode materials play a critical role in their electrochemical performance (Goodenough and Park, 2013; Nitta et al., 2015; Zheng et al., 2017; Sui et al., 2019a). Recently, various materials have been explored for LIB cathodes, such as olivine-structured materials (LiMPO₄, M = Fe, Mn, Co, and Ni) (Li Y. et al., 2019; Vásquez and Calderón, 2019; Wu et al., 2020), layered oxide materials (LiCoO₂, LiMnO₂, LiNiO₂, etc.) (El-Bana et al., 2017; Zhao et al., 2017; Lin et al., 2020), spinel-structured materials (LiMn₂O₄, LiCo₂O₄, etc.) (Hariprasad et al., 2016; Abou-Rjeily et al., 2019), tavorite-structured materials (LiFeSO₄F, LiVPO₄F, etc.) (Kim and Kang, 2017; Hu et al., 2019), and NASICON-structured materials [Li₃V₂(PO₄)₃, Li₃Fe₂(PO₄)₃, etc.] (Karami and Taala, 2011; Zheng et al., 2016; Jiang, 2018).

Among these materials, LiVPO₄F and Li₃V₂(PO₄)₃ are especially promising and have drawn attention in recent years. LiVPO₄F, which was synthesized by Barker et al. (2003) in 2003 for the first time, is attracting attention due to its high theoretical energy density (655 Wh kg⁻¹) and high working voltage platform (4.2 V vs. Li/Li⁺) (Gover et al., 2006). The synergistic effect between F⁻ and PO43- within the material could enhance the electronegativity and induction property of LiVPO₄F, achieving improved electrochemical performance (Huang et al., 2009). For example, Sui et al. (2019b) have prepared spherical LiVPO₄F/C by spray drying, and the initial discharge capacity of the synthesized LiVPO₄F/C can be as high as 137.9 mAh g⁻¹ at 0.1 C, and discharge capacity remains at 91.4% of the initial capacity after 50 cycles. However, the Li⁺ diffusion coefficient in LiVPO₄F is low, which restricts its application.

Li₃V₂(PO₄)₃, a typical NASICON-structured material, allows the insertion and/or extraction of lithium ions from various pathways, and exhibits an outstanding Li⁺ diffusion coefficient (10⁻⁹-10⁻¹⁰ cm² s⁻¹) (Li R. et al., 2019). In addition, Li₃V₂(PO₄)₃ displays a high average operation voltage (4.0 V vs. Li⁺/Li) and excellent theoretical capacity (197 mAh·g⁻¹) (Zhang et al., 2019). Therefore, it is expected that the combination of Li₃V₂(PO₄)₃ with LiVPO₄F may enhance the ionic conductivity of cathode material and improve electrochemical performance.

Herein, we synthesized xLiVPO₄F·yLi₃V₂(PO₄)₃/C cathode material for LIBs via a feasible sol-gel method. Such a design is demonstrated to combine the advantages of LiVPO₄F and Li₃V₂(PO₄)₃, and it can effectively enhance electronic/ionic conductivity and structural stability. The optimal xLiVPO₄F·yLi₃V₂(PO₄)₃/C shows a high initial discharge capacity, outstanding rate capability, and proper cycling performance, illustrating that it is a promising cathode material for LIBs.

Synthesis of LiVPO₄F/C: firstly, stoichiometric amounts of H₃PO₄, V₂O₅, and citric acid were added into deionized water under stirring, and then ammonia was added into the solution. After stirring at 80°C for 6 h and drying overnight in an oven, the precursor of VPO₄ was obtained. The obtained VPO₄ was pretreated at 300°C for 4 h under the protection of Ar gas and heated at 750°C for 8 h to prepare VPO₄/C. Finally, stoichiometric amounts of the VPO₄-based precursor and LiF were mixed and calcined from 690 to 770°C for 1 h under a flowing Ar atmosphere to obtain LiVPO₄F/C. All chemical reagents used in this work were analytical grade.

Synthesis of Li₃V₂(PO₄)₃/C: stoichiometric amounts of the prepared VPO₄ precursor and Li₃PO₄ were mixed and calcined from 650 to 800°C for 16 h under a flowing Ar atmosphere to obtain Li₃V₂(PO₄)₃/C.

Synthesis of 2LiVPO₄F·Li₃V₂(PO₄)₃/C: Stoichiometric amounts of the prepared VPO₄/C, LiF, and Li₃PO₄ were mixed and calcined at 750°C for 1 h under a flowing Ar atmosphere to obtain 2LiVPO₄F·Li₃V₂(PO₄)₃/C samples.

Crystallographic information was studied by powder X-ray diffractometer (XRD, Rigaku, Ultima VI). The morphology was observed with a Hitachi SU5000 scanning electron microscope (SEM). The electrodes for electrochemical tests were fabricated with 80 wt.% active material and 20 wt.% additives (PVDF/acetylene black, 1:1). CR2025 button cells were assembled in an Ar-filled glove box. The galvanostatic charge/discharge and the cyclic voltammetry (CV) were characterized with a LAND battery test system and a CHI660D electrochemical workstation, respectively.

Figure 1 shows the XRD results of the VPO₄/C precursors synthesized at different temperatures. Obvious peaks indexed to V₂O₃ are observed in the precursor calcined at 650°C. However, these peaks disappear when the calcination temperature is elevated above 700°C. Hence, the precursors synthesized at 700, 750, and 800°C can be confirmed to be amorphous VPO₄/C. Meanwhile, the absence of peaks belong to crystalline carbon suggests the amorphous nature of the carbon coating layer.

The morphology of VPO₄/C precursors synthesized at various temperatures was also investigated, as shown in Figure 2. Figure 2a suggests that the reactants cannot be completely converted to products at 650°C. Moreover, the pores on the samples are uneven. As the temperature rises to 700°C, the product has a more uniform pore diameter. However, many small particles adhere to the surface of the sample. At 750°C, the pores become most uniform, and the small particles disappear. A serious aggregation occurs when the temperature further increases to 800°C. Therefore, the optimal calcination temperature for VPO₄/C precursor is 750°C, which is chosen for the synthesis of xLiVPO₄F·yLi₃V₂(PO₄)₃/C (x:y = 1:0, 2:1, 0:1).

XRD patterns of LiVPO₄F/C (Figure 3A) and Li₃V₂(PO₄)₃/C (Figure 3B) synthesized at different temperatures are shown in Figure 3. As can be seen, the diffraction peaks are stronger and the half-peak width narrower at the higher calcination temperature, indicating that elevated temperature facilitates the crystallization of composites. It can be seen in Figure 3A that the characteristic peaks of Li₃V₂(PO₄)₃ and LiF become weaker as the temperature increases and that they completely disappear at 750°C. However, the impurity of Li₃V₂(PO₄)₃ reappears when the temperature reaches 770°C, suggesting that an excessive temperature has a detrimental effect on the formation of LiVPO₄F. On the contrary, only pure Li₃V₂(PO₄)₃ is observed in the temperature range from 650 to 800°C in Figure 3B.

Figure 4 shows SEM images of LiVPO₄F/C synthesized at different temperatures. Obviously, the samples obtained at 730 and 750°C possess a porous morphology, and the latter has a more uniform pore diameter. In contrast, those calcined at lower temperatures (690 and 710°C) are not porous, since there are lots of unreacted LiF particles left on the surface of the samples. At 770°C, all pores disappear due to aggregation. Figure 5 shows SEM images of Li₃V₂(PO₄)₃/C synthesized at different temperatures. Although the XRD results (Figure 3b) do not show a significant difference between these samples calcined at different temperatures, they can be clearly distinguished from each other in the SEM images, indicating that the porous VPO₄/C and Li₃PO₄ start to react at around 650°C. With an increase in temperature, the particles coarsen since the high temperature can accelerate the growth rate of particles. When calcined at 750°C, the particles display irregular shapes and are uniformly dispersed on the carbon network. As the temperature reaches 800°C, the particles agglomerate seriously. Hence, the optimal calcination temperature for synthesis of LiVPO₄F/C and Li₃V₂(PO₄)₃/C is 750°C.

Based on the above analysis, 2LiVPO₄F·Li₃V₂(PO₄)₃/C was chosen to be synthesized at 750°C. The XRD pattern of the as-prepared 2LiVPO₄F·Li₃V₂(PO₄)₃/C is shown in Figure 6. It can be seen that all diffraction peaks can be indexed to LiVPO₄F and Li₃V₂(PO₄)₃, and no impurity is detected. Moreover, characteristic peaks of crystalline carbon cannot be spotted, suggesting that carbon is amorphous in the composite.

Figure 7 displays SEM images of the as-prepared 2LiVPO₄F·Li₃V₂(PO₄)₃/C. The composite is composed of porous particles with a pore diameter of 0.5–2 μm and nanosheets with a thickness of 200–500 nm. Furthermore, the amorphous carbon can be clearly observed. This uniform morphology of this material could facilitate ion transportation.

The electrochemical performance of the as-prepared materials was tested using galvanostatic charge-discharge in the voltage range of 3.0–4.5 V. The initial discharge curves (Figure 8A) and the cycling performances (Figure 8B) of the LiVPO₄F/C samples were revealed. The one calcined at 750°C delivers the highest initial discharge capacity of 141.1 mAh g⁻¹ at 0.1 C. After 50 cycles, the remaining capacity can be as high as 125.9 mAh g⁻¹, with a capacity retention of 89.22%. The samples calcined at 690, 710, 730, and 770°C display capacities of 90.4, 114.8, 122.3, and 123.9 mAh g⁻¹ at 0.1 C, respectively. The corresponding capacity retention rates after 50 cycles are 37.30, 50.69, 51.43, and 69.57%. The samples calcined in the ranges of 710–770°C show the characteristic discharge plateaus of LiVPO₄F, as well as the characteristic discharge plateaus of Li₃V₂(PO₄)₃ around 4.10, 3.68, and 3.58 V. For the sample calcined at 750°C, no diffraction peaks of Li₃V₂(PO₄)₃ are detected in the XRD pattern, but it indeed shows the characteristic discharge plateaus. This is due to the very tiny amount of Li₃V₂(PO₄)₃ in this sample. Moreover, this sample has a good three-dimensional porous structure, which provides pathways for Li-ion transportation, consequently enhancing the cycling performance.

The initial discharge curves at various discharge rates of the LiVPO₄F/C samples synthesized at different temperatures are illustrated in Figure 9. The one calcined at 750°C shows a better rate capability than the others. For example, the initial discharge capacities at 0.1, 0.2, 0.5, 1, and 2 C are 141.1, 125.2, 106.3, 87.9, and 71.5 mAh g⁻¹, respectively. The discharge plateau around 4.17 V becomes lower and shorter with an increase in discharge rate, indicating that the high current density enlarges the resistance of the electrochemical reactions.

Figure 10 shows the initial discharge curves (Figure 10A) and the cycling performance (Figure 10B) of Li₃V₂(PO₄)₃/C. The samples calcined at 650, 700, 750, and 800°C show initial discharge capacities of 95.3, 128.0, 136.1, and 128.5 mAh g⁻¹, respectively. After 50 cycles, the remaining capacities are 55.5, 92.2, 131.9, and 111.4 mAh g⁻¹, with corresponding capacity retentions of 58.2, 72.0, 96.9, and 86.7%. Compared with the samples calcined at 650, 700, and 800°C, the one calcined at 750°C shows more stable discharge plateaus at 4.07, 3.68, and 3.58 V, suggesting suppressed polarization and a more stable structure.

Figure 11 shows the initial discharge curves at various discharge rates of the Li₃V₂(PO₄)₃/C synthesized at different temperatures. The one calcined at 750°C delivers the highest discharge capacity. When tested under rates of 0.1, 0.2, 0.5, 1, and 2 C, it shows discharge capacities of 136.1, 138.8, 130.2, 124.4, and 115.7 mAh g⁻¹, respectively. Moreover, the discharge plateaus are stable even under higher current density.

Figure 12 shows the electrochemical properties of 2LiVPO₄F·Li₃V₂(PO₄)₃/C. The initial discharge curves at different discharge rates (Figure 12A) show that this composite delivers a significantly improved electrochemical performance in comparison with LiVPO₄F/C and Li₃V₂(PO₄)₃/C. For instance, 2LiVPO₄F·Li₃V₂(PO₄)₃/C has an initial discharge capacity of 143.1 mAh g⁻¹ at 0.1 C (1 C = 156 mA g⁻¹), which is higher than that of the other materials. With an increase in current density, the capacity fading is minor. When tested at 0.2, 0.5, 1, and 2 C, the discharge capacities are 141.6, 133.2, 124.1, and 117.6 mAh g⁻¹, respectively. The discharge curves show four plateaus. The one around 4.2 V is attributed to the discharge of LiVPO₄F, and the others around 4.1, 3.7, and 3.6 V are due to the discharge of Li₃V₂(PO₄)₃. Figure 12B shows the rate capability of the composite. After being successively tested at 0.1 and 0.2 C for 5 cycles and 0.5, 1, and 2 C for 10 cycles (40 cycles in total), the discharge capacity can be almost fully recovered when the discharge rate returns to 0.1 C.

In order to study the lithiation/delithiation processes of 2LiVPO₄F·Li₃V₂(PO₄)₃/C, a CV test was implemented; the results are shown in Figure 13. The anodic peaks at 4.22 and 4.44 V and the cathodic peak at 4.08 V correspond to the reversible redox of V³⁺/V⁴⁺ in LiVPO₄F. The potential difference is 0.36 V. Moreover, the anodic peaks at 3.62, 3.71, and 4.15 V and the corresponding cathodic ones at 3.51, 3.61, and 3.97 V are due to the multi-step Li-ion extraction/insertion in Li₃V₂(PO₄)₃. The potential differences are 0.11, 0.10, and 0.18 V. Apparently, the polarization in Li₃V₂(PO₄)₃ is much smaller than that in LiVPO₄F. Hence, a moderate amount of Li₃V₂(PO₄)₃ could suppress the polarization in the xLiVPO₄F·yLi₃V₂(PO₄)₃/C composite, consequently enhancing the electrochemical performance.

In summary, xLiVPO₄F·yLi₃V₂(PO₄)₃/C (x:y = 1:0, 2:1, 0:1) composites were successfully synthesized via a feasible sol-gel method. Our results indicate that the composites are well-crystallized and composed of porous particles. 2LiVPO₄F·Li₃V₂(PO₄)₃/C outperforms LiVPO₄F/C and Li₃V₂(PO₄)₃/C due to the synergy between the fluorophosphate and phosphate. This study presents a facile approach in synthesizing xLiVPO₄F·yLi₃V₂(PO₄)₃/C for high-performance cathode materials of LIBs.

The datasets generated for this study are available on request to the corresponding author.

SZ, XZ, and JL did the main experiment and wrote the manuscript. JL and YS were involved the discussion of the experiment, revised the manuscript, and made the research plan. XZ assisted in the material synthesis. SZ and YS also provided the financial support.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.