1 Introduction

Rechargeable lithium-ion batteries (LIBs) have been highlighted for their large scale application of electronic portable devices and electric vehicles [1,2,3]. High capacity, high energy and power densities, low cost and lifespan are the key factors in practical applications of LIBs [4, 5]. Recently, extensive attention has been focused on lithium iron orthosilicate (Li2FeSiO4) because of their distinct advantages such as high security performance, obvious environmental benignity and low cost owing to the abundant resources of Fe and Si [6]. Additionally, Li2FeSiO4 shows high theoretical capacity (> 330 mAh g−1) as a result of extracting two Li+ ions per formula unit [7, 8]. However, like most polyanion-based cathode materials, Li2FeSiO4 also suffers from very low electronic conductivity (~ 6 × 10−14 S cm−1) [8, 9] together with slow lithium ion diffusion rate (~ 1 × 10−14 cm2 S−1) [10], thus leading to extremely poor rate capability.

To solve these problems, several attempts have been made. Metal ion doping has been reported to be a suitable way to enhance the electrochemical properties of Li2FeSiO4 for the improved electronic conductivity and accelerated lithium ion diffusion caused by lattice defects [11,12,13,14]. Yi et al. [11] reported that Li2FeSiO4/C doped with 5% Mn showed excellent electrochemical performance with highest discharge capacity (248.5 mAh g−1 at 0.1 C) in the first cycle, outstanding high-rate performance and high value of capacity retention. Morphology control has also been illustrated to exert positive effects on the electrochemical properties of the material [15,16,17,18,19]. Varieties of shapes of Li2FeSiO4 have been investigated, such as shuttle-like [15], flower-like [16] and nanoworm-like shapes [17] as well as hollow spheres [18] and three-dimensionally ordered macroporous structure [19]. Moreover, Li2FeSiO4 coated with carbon has been confirmed to achieve larger discharge capacity and better cycling performance [20,21,22]. It is worth noting that the as-synthesized composites coated with a thin carbon layer can not only increase the electronic conductivity, but also control the particle size of material during the modification process [23, 24].

Nowadays, coating with fast ion conductor has been a widely used modified method. Various active materials such as Li2SiO3 [25], Fe7SiO10 [26] and AlPO4 [27] have been coated on the surface of Li2FeSiO4, which greatly improve the rate performance, especially the high rate performance in comparison with the original material. For instance, Li et al. [26] synthesized Li2FeSiO4/Fe7SiO10/C via a sol–gel method by adjusting the molar ratio of Fe and Si. Compared with Li2FeSiO4/C, the resulting composite showed much better high-rate performance with discharge capacities of 128 and 112 mAh g−1, respectively, at 5 and 10 C. As a fast lithium ion conductor, Li3PO4 exhibits unique advantages with ionic conductivity about 10−6~10−8 S cm−1 [28, 29]. By the virtue of high ionic conductivity and strong P=O bond of Li3PO4, the migration rate of Li+ between electrode and electrolyte is expected to be increased in a certain extent [30]. Importantly, Li3PO4 provides the possibility to reduce the corrosion of electrolyte to electrode materials when used as a coating material [31]. Hence, Li3PO4 has been commonly used to enhance the cyclability and rate performance of some cathodes such as LiCoO2 [32], Li(Ni0.4Co0.3Mn0.3)O2 [31], LiNi0.5Mn1.5O4 [33], LiFePO4 [34], LiMn2O4 [35] and Li-rich oxide [36,37,38,39].

In the present work, Li3PO4 has been introduced into Li2FeSiO4/C as a property improvement additive. The effects of different amounts of Li3PO4 on the structure, morphology and electrochemical performances of cathodes have been systematically investigated. By introducing Li3PO4, the rate performance of Li2FeSiO4/C can be significantly improved with nearly no capacity decay.

2 Experimental

2.1 Synthesis of Li3PO4

The Li3PO4 powder was precipitated by adding H3PO4 (85% solution) dropwise into 0.4 M LiOH aqueous solution with a molar ratio of Li: P equal to 3:1. The resulting white precipitates were washed with distilled water and dried at 120 °C. Finally, the Li3PO4 was obtained by ball-milling the dried products for 3 h.

2.2 Synthesis of Li2FeSiO4/C composite

Li2FeSiO4/C composite was prepared by a modified sol–gel method according to Ref. [40]. Firstly, 2 g poly (ethylene oxide)-b-poly (propylene oxide)-b-poly (ethylene oxide) triblock copolymer P123 (EO20PO70EO20) was dissolved in 40 mL absolute alcohol, and the stoichiometric ratio of CH3COOLi·2H2O and Fe(NO3)3·9H2O were added into alcohol solution in order under vigorous magnetic stir. The solution was stirred and dissolved to get a uniform and transparent red-brown solution. Then the corresponding amount of TEOS was added to the mixture. The red gel was obtained under vigorous magnetic stir at 50 °C and dried at 100 °C for 12 h. Finally, the dry gel was annealed at 600 °C for 10 h under argon atmosphere. After cooling to room temperature, the products were obtained and marked as LFS.

2.3 Synthesis of Li2FeSiO4/Li3PO4/C composites

The synthesized Li3PO4 was added into the solution containing P123, CH3COOLi·2H2O and Fe(NO3)3·9H2O according to the given molar ratio, and the stoichiometric ratio of TEOS was dissolved in the above solution. The subsequent processing was the same as the process of Li2FeSiO4/C. For comparison, the molar percent of Li3PO4 to Li2FeSiO4 were 4%, 6%, 8% and 10%, respectively. The samples were marked as LFS-4, LFS-6, LFS-8 and LFS-10.

2.4 Characterization

The phase structure of as-synthesized samples was conducted by powder X-ray diffraction (Bruker D8 Advance, Germany) using Cu Kα. The morphologies and microstructures of samples were observed by scanning electron microscope (Zeiss SIGMA FESEM) and Transmission electron microscope (JEM-2100F, JEOL Ltd.). The carbon contents in the Li2FeSiO4/Li3PO4/C composites and Li2FeSiO4/C were characterized by VARIOELIII elemental analyzer (Elementar Analysen System GmbH, Germany).

2.5 Electrochemical tests

The Li2FeSiO4/Li3PO4/C composites and Li2FeSiO4/C, acetylene black and polytetrafluoroethylene (PTFE) binder were mixed in proportion at a ratio of 75:20:5 (in mass, wt%), and a proper amount of isopropanol was added as the dispersant. The mixture was made into paste in the mortar and transferred to the rolling machine to make film. The film was cut into disks about 1 cm in diameter and coated on stainless steel, after vacuum drying at 60 °C. In an Ar-filled glovebox, batteries were assembled using lithium foil, Celgard 2300 microporous films, 1 M LiPF6 in the ethylene carbonate (EC) and dimethyl carbonate (DMC) (1:1 v/v) as anodes, separators and electrolytes, respectively. The active materials were assembled into CR2016 coin-type cell for electrochemical measurements. Galvanostatically charge–discharge tests were done between 1.5 V and 4.8 V using battery test system (Neware, China). And all the specific capacities were calculated with the mass of Li2FeSiO4/Li3PO4/C composites and Li2FeSiO4/C, not including carbon content. Electrochemical impedance spectroscopies (EIS) were tested with an AC amplitude of 5 mV in the applied frequency of 0.01 Hz to 100 kHz via CHI760C electrochemistry workstation.

3 Results and discussion

Table 1 lists the carbon contents in Li2FeSiO4/Li3PO4/C composites and Li2FeSiO4/C, which were characterized by elemental analyzer. Obviously, carbon contents in all the composites are very close to each other.

Table 1 The carbon contents in all the composites
Full size table

The XRD pattern and SEM image of the prepared Li3PO4 are presented in Fig. 1. Figure 1a displays that all the diffraction peaks match well with the orthogonal phase Li3PO4 (JCPDS NO. 71-1528). There are no impurities can be found, demonstrating that the obtained sample is pure phase. Moreover, the peak type of the sample is relatively sharp and the base is flat, suggesting the high crystallinity of the synthesized sample. From SEM image (Fig. 1b), it is clear that the particles of Li3PO4 after ball milling are uniform and the average diameter is about 40 nm.

Fig. 1
figure 1

The XRD pattern (a) and SEM image (b) of the prepared Li3PO4

Full size image

Figure 2 illustrates the XRD pattern and SEM image of Li2FeSiO4/C. As can be seen in Fig. 2a, the reflection peaks can be assigned to the monoclinic structure Li2FeSiO4 with a space group of P21/n [41]. No impurity peaks can be detected in the XRD pattern, suggesting the high purity of prepared sample. Meanwhile, no reflection peaks associated with carbon can be seen, which may be related to the amorphous property of carbon. From SEM image (In Fig. 2b), we can see that the nanoparticles are uniformly distributed and the average diameter is around 30 nm.

Fig. 2
figure 2

The XRD pattern (a) and SEM image (b) of LFS

Full size image

The XRD patterns of the samples with different contents of Li3PO4 (LFS-4, LFS-6, LFS-8 and LFS-10) are displayed in Fig. 3. The main reflection peaks of the samples with Li3PO4 are the same as those of Li2FeSiO4/C expect the Li3PO4 peaks, indicating that the crystal structure of Li2FeSiO4 has not been affected by Li3PO4. When the molar ratio of Li3PO4 is up to 6%, the star-labeled orthogonal phase Li3PO4 (JCPDS NO. 71-1528) appears in the samples. The diffraction peaks of Li3PO4 exhibit an increasing trend with enhanced amounts of Li3PO4. However, there are no characteristic peaks of Li3PO4 for LFS-4, which may be attributed to a small amount of introduced Li3PO4. There are also no diffraction peaks from the carbon phase shown in the patterns.

Fig. 3
figure 3

The XRD patterns of the samples with different contents of Li3PO4 (LFS-4, LFS-6, LFS-8 and LFS-10)

Full size image

The morphologies of the samples with different contents of Li3PO4 were identified by SEM and the images are displayed in Fig. 4. From that, we can see the average size of nanoparticles is around 20 nm, and they are agglomerated together. The samples with different contents of Li3PO4 also preserve the morphology of Li2FeSiO4/C (Fig. 2b) with slightly decreased diameter, suggesting that Li3PO4 added has no effect on the morphology of Li2FeSiO4/C composite. To study the microstructure of LFS-6, TEM and HRTEM of material were characterized. TEM image (Fig. 5a) shows that the particles of LFS-6 are coated with amorphous carbon. From HRTEM image shown in Fig. 5b, we can clearly see the lattice fringes of Li2FeSiO4 and Li3PO4, respectively. Moreover, the interplanar spacing of 0.330、0.235 and 0.283 nm can be well assigned to the (201), (212) and (112) facets of monoclinic Li2FeSiO4, while the interplanar spacing of 0.231 nm correlated well with the (021) plane of orthogonal phase Li3PO4. The above results suggest that Li3PO4 and Li2FeSiO4/C form nanocomposite in LFS-6.

Fig. 4
figure 4

SEM images of LFS-4 (a), LFS-6 (b), LFS-8 (c) and LFS-10 (d)

Full size image
Fig. 5
figure 5

TEM (a) and HRTEM (b) images of LFS-6

Full size image

To investigate the influences of Li3PO4 content on the rate performance and high-rate cyclability of Li2FeSiO4/C, charge–discharge experiments were performed at voltage window of 1.5–4.8 V. Figure 6a exhibits the rate performance of LFS, LFS-4, LFS-6, LFS-8 and LFS-10 at different rates from 1 C to 10 C (1 C = 166 mA g−1) for 50 cycles each. It is apparently observed that Li2FeSiO4/C with different contents of Li3PO4 exhibit much better performance than LFS, except that LFS-10 is almost the same as that of LFS at 1 C and 2 C rates. Among the samples after adding fast ion conductor Li3PO4, LFS-6 possesses the best rate performance. This indicates that the introduction of Li3PO4 has a significant effect on improving the rate property of Li2FeSiO4/C, and the advantage is much more obvious on the high rate. Figure 6b presents the cycling performance of LFS, LFS-4, LFS-6, LFS-8 and LFS-10 at 10 C for 100 cycles. It is evident from the figure that the performance of LFS-6 is the best. The discharge capacity is 114.1 mAh g−1, and there is no capacity decay after 100 cycles. In addition, the cyclic stability of Li3PO4-Li2FeSiO4/C composites are higher than that of LFS without Li3PO4.

Fig. 6
figure 6

Rate performance (a) and the cycling performance at 10 C (b) of LFS, LFS-4, LFS-6, LFS-8 and LFS-10

Full size image

LFS and LFS-6 were selected to further study the electrochemical performance differences between pristine sample and sample with Li3PO4 additive at 0.1 C at voltage window of 1.5–4.8 V. Figure 7 compares the charge–discharge curves of the two electrodes for the initial 5 cycles. Obviously, the charge–discharge curves of LFS-6 are similar to the LFS, which suggests the intrinsic properties of Li2FeSiO4 are not affected by Li3PO4. For both composites, the initial charge curves show two voltage plateaus at around 3.3 V and 4.3 V, which should be related to the Fe2+/Fe3+ and Fe3+/Fe4+ redox couples, respectively. However, in the subsequent charge–discharge process, the charge curves are quite different from those of the first cycle, which is attributed to structural arrangement [42]. From Fig. 7a, it can be seen the discharge capacities of LFS are 209.3, 199.4, 193.1, 186.6 and 184.8 mAh g−1 for the initial 5 cycles, respectively, and the capacity retention is 88.3%, while the discharge specific capacities of LFS-6 cathode material (Fig. 7b) are 195.5, 183.2, 180, 6, 179.8 and 179 mAh−1, respectively, with a capacity retention of 91.8%. The LFS shows higher capacities than LFS-6 composite, however, the value of capacity retention is slightly lower than that of LFS-6. This is mainly due to the fact that Li3PO4 has no electrochemical activity. Although it has a negative effect on the specific capacity of Li2FeSiO4, it can enhance the cyclic stability of material to a certain degree.

Fig. 7
figure 7

The galvanostatic charge–discharge curves of LFS (a) and LFS-6 (b) at 0.1 C rate in the voltage of 1.5–4.8 V for the initial 5 cycles

Full size image

In order to understand the influence of Li3PO4 additive on the rate performance and cyclic stability of Li2FeSiO4 more comprehensively, we test LFS and LFS-6 at high rate, and results are plotted in Fig. 8. For LFS-6 (Fig. 8c and d), the discharge capacities are 188.8, 172.8, 154.2 and 114.1 mAh g−1, respectively, at rates of 1 C, 2 C, 5 C and 10 C. With the increasing current rate, the discharge capacities of the composite decrease gradually. After 100 cycles, the capacities are 164.9, 153.1, 152.4 and 118.9 mAh g−1, respectively, with 87.3%, 88.6%, 98.8% and 104.2% of capacity attainable. In Fig. 8a and b, the discharge capacities of LFS are 177.5, 159.2, 133.0 and 89.0 mAh g−1, respectively, under the same rates. After 100 cycles, the capacities declined to 129.9, 123.5, 109.1 and 75.9 mAh g−1, respectively, and their corresponding capacity retentions are 73.2%, 77.6%, 82.0% and 85.3%. Additionally, our Li2FeSiO4 with 6 mmol% Li3PO4 added shows superior performances with an excellent cyclability and an outstanding rate performance than those of the previous reported Li2FeSiO4 summarized in Table 2.

Fig. 8
figure 8

The galvanostatic charge–discharge curves (a and c) and the corresponding cycling performances (b and d) of LFS and LFS-6 at rates of 1C, 2C, 5C and 10C

Full size image
Table 2 Comparisons of the electrochemical performances of LFS-6 and the previous reported Li2FeSiO4 composites
Full size table

Cyclic voltammetry (CV) was conducted to analyze the kinetic characteristics of LFS and LFS-6 composites in the second cycle, and the curves are presented in Fig. 9. From Fig. 9, it is clearly seen that the CV curves of LFS-6 are similar to that of LFS, demonstrating that Li3PO4 adding has no effect on the electrochemical behavior. All the electrodes exhibit a major redox couple of Fe2+/Fe3+ at a scan rate of 0.1 mV s−1 in the voltage range of 1.5–4.8 V. The anodic peak at 3.04 V in LFS corresponds to the oxidation of Fe2+ to Fe3+, and the cathodic peak at 2.45 V involving with the reduction of Fe3+ to Fe2+, with a potential interval of 0.59 V between these two redox peaks. Compared with LFS, LFS-6 show little narrower separations of the redox peaks (a potential interval of 0.42 V), suggesting that LFS-6 has better electrochemical kinetics.

Fig. 9
figure 9

Cyclic voltammetry of the composites of LFS and LFS-6 at scanning rate of 0.1 mV s−1

Full size image

EIS measurements of LFS and LFS-6 were carried out to further study the kinetic properties of the materials. Figure 10a illustrates the Nyquist plots of the two composites and the equivalent circuit is shown inset. All the EIS spectra are divided into three characteristic regions. The semicircle at the high to medium frequency region that is assigned to the charge-transfer resistance (Rct) through the solid-electrolyte interfaces, and an inclined line in the low frequency range which represents the lithium ion diffusion in the cathode material. The small intercepts by the semicircle on the real axis (Zre) in the high frequency which usually correspond to the ohmic resistance of the cell (Re), including the resistance of electrolyte, separator and electrodes. The diffusion coefficient of lithium ion (DLi+) can be obtained based on the relationship between Zre and the reciprocal square root of the frequency (ω−1/2) in the low frequency region, according to the equations mentioned [48]. From the impedance parameters summarized in Table 3, it is evident that the Rct value of LFS-6 (112.09 Ω) is much smaller than that of LFS (281.67 Ω) before cycling. After 100 cycles, the Rct value of LFS increased considerably in comparison with that of LFS-6. The Rct of LFS increased to 472.99 Ω, but the Rct of LFS-6 was 183.43 Ω. Meanwhile, the lithium ion diffusion coefficients (D +Li ) of LFS and LFS-6 before cycling can be calculated as 2.50 × 10−15 and 2.04 × 10−14 cm2 s−1, respectively. After 100 cycles, the DLi+ value of LFS-6 is still higher than that of LFS. The results above clearly imply that LFS-6 shows a smaller value of charge transfer resistance with a faster diffusion coefficient. Obviously, the Li3PO4 adding suppress the rising of charge transfer resistance upon cycling and enhance the kinetics of lithium-ion diffusion (Fig. 10).

Fig. 10
figure 10

The Nyquist plots and equivalent circuit of LFS and LFS-6 at room temperature (a, c) and the relationship between Zre and ω−1/2 in the frequency region (b, d) before and after 100 cycles at 2 C rate

Full size image
Table 3 Impedance parameters of LFS and LFS-6
Full size table

4 Conclusion

In summary, we synthesized and characterized Li2FeSiO4/Li3PO4/C composites, and compared their electrochemical performance with pristine one as LIB cathodes. The Li2FeSiO4/Li3PO4/C composites exhibited superior rate capability and largely improved cycling performance, especially with the amount as high as 6 mmol%. The specific discharge capacities of LFS-6 can be achieved 188.8, 172.8, 154.2 and 114.1 mAh g−1 at 1 C, 2 C, 5 C and 10 C, respectively, and retained 164.9, 153.1, 152.4 and 118.9 mAh g−1 over 100 cycles. The overall improved electrochemical performances are mainly attributed to the improved lithium ion diffusion, promoted charge transfer process and retarded side reactions by introducing lithium ion conductor Li3PO4. Accordingly, adding Li3PO4 into Li2FeSiO4 can be a useful way to optimize the electrochemical performance, which can also be extended to other cathodes for LIBs.