Study on the impact of Fe₂P phase on the electrochemical performance of LiFePO₄

1 Introduction

Recently, lithium iron phosphate (LiFePO₄) has been paid close attention as it could be potentially used in powering batteries of electric vehicles and as an energy storage system because of its high performances and advantages [1–3]. At present, the main problems of lithium iron phosphate; low electrical conductivity and lithium ion conductivity, have obvious improvements [4–8]. For electrical conductivity, it could be improved significantly by adding and/or coating conducting materials. The most common method is carbon coating as it is simple and effective. But there are also other materials being researched, due to their good electrical conductivity, such as Fe₂P. After high-temperature sintering, especially when prepared by the solid state method, LiFePO₄ contains a few impurities, such as Fe₂P, Li₃PO₄ and so on [9, 10]. The impurities will reduce the content of active materials in the cell, resulting in the capacity being decreased. But, Fe₂P and Li₃PO₄ have good electrical conductivity and ion conductivity, respectively, according to reports [11, 12]. Now, Fe₂P or Li₃PO₄ composited with LiFePO₄ to enhance the electrical performance of cathode have been reported by [13–16]. However, there are some problems in previous reports that the Fe₂P phase is obtained by adjusting the reaction condition, and that will change the electrical performance of LiFePO₄, so it is not clear whether the change in electrochemical performance is due to the presence of Fe₂P or the process changes.

In order to confirm the influence of Fe₂P, related work has been studied in this paper. Through the magnet of Fe₂P, the simple magnetic separation method is used to separate the Fe₂P phase. This could divide the LiFePO₄ into a magnet-attracted sample and a retaining sample, which correspond to high and low Fe₂P content samples, respectively. By comparing the character and electrochemical performance of the primary sample, magnet-attracted sample and retaining samples, the effect of Fe₂P could be proven definitely.

2 Materials and methods

The LiFePO₄ sample, provided by Hefei Guoxuan High-tech Power Energy Co., Ltd (Hefei, China), was prepared by the ferrous oxalate process. The magnetic separation experiment process is as follows. Seven grams of LiFePO₄ was dispersed into 150 ml of ethanol by ultrasonic dispersion method. Then a magnet bar, whose magnetic induction intensity is 6000 Gs, was added into the dispersion liquid and mechanically stirred at medium speed for half an hour. Later, the magnet bar was removed and dried; the magnet-attracted sample, named MAS, was collected. The retaining solution was dried to obtain the retaining sample named RS. The primary LiFePO₄ was named PS.

The morphologies of the prepared samples were characterized by field-emission scanning electron microscopy (SEM, Hitachi S-4800, Hitachi, Japan). The crystalline structures of the samples were measured by an X-ray diffractometer (Bruker D8 Advance X-ray diffraction, Bruker-AXS, Germany). The powder resistivity of the samples was characterized by a ST-2258A multifunction digital four-probe meter (Suzhou Jingge Electronic Co., Ltd, Suzhou, China).

For the preparation of electrodes, the active materials (80 wt.%), Super-P (15 wt.%) and PVDF (5 wt.%) were mixed in NMP and stirred for 5 h. The resultant slurry, pasted onto an Al foil, was dried at 80°C under vacuum for 6 h. The loading density of the electrode was about 8 mg/cm². The coin cells (2032) then were assembled in an argon-filled glove box using lithium metal as the anode, Celgard 2600 as the separator, and 1 m LiPF6 (dissolved in ethylene carbonate and dimethyl carbonate with a 1:1 volume ratio) as the electrolyte. Cells were tested at 25°C using a battery tester (Land Co., Ltd, Wuhan, China) within the voltage range of 2.0–4.2 V (vs. Li⁺/Li). Electrochemical characterization was performed using an electrochemical analysis instrument (IviumStat, Holland) at 25°C. The potentials of the electrode for CV were conducted in the voltage range from 2 to 4 V at a scan rate of 0.1 mV/s. Electrochemical impedance (EIS) measurements were performed, the excitation voltage applied to the cell was 10 mV, and the frequency range was between 100 kHz and 0.01 Hz.

3 Results and discussion

Conventional magnetic separation technology is widely used for the removal of tramp iron from a variety of feed materials and for the beneficiation of ferrous ores. In the LiFePO₄ product, the Fe₂P phase is magnetic, and the LiFePO₄ phase is a little magnetic. So in this work, the magnetic separation method was used to separate LiFePO₄, which obtained the different Fe₂P content of the samples. Figure 1 shows the X-ray diffraction (XRD) patterns of LiFePO₄ before and after the magnetic separation and the magnet-attracted substance. The main diffraction peaks for the three samples are indexed to the LiFePO₄ phase, and the second phase, Fe₂P, could be additionally observed in all patterns. The peaks of LiFePO₄ for the three samples are almost entirely consistent with each other comparing the relative intensity. It is due to the samples coming from the same LiFePO₄ samples and proves that the magnetic separation method will not change the crystal structure of LiFePO₄. A comparison of the peak intensity for LiFePO₄ and Fe₂P indicates that the content of Fe₂P could be changed by magnetic separation. The highest content of Fe₂P is for MAS, and the lowest is for RS, and that of PS is in between, when qualitatively analyzed by the XRD result. These results mean that the magnetic separation method could preliminary separate the Fe₂P phase from LiFePO₄. But it could not be hundred-percent separated because some Fe₂P phase is generated in-site and accompanied by LiFePO₄ particles. However, this is enough for this research.

Figure 1:

The XRD patterns of samples (left) and the local zoom (right).

The morphologies of the samples before and after adsorption are observed by SEM and are shown in Figure 2. The morphologies of the three samples are not obviously distinguishable. There are large numbers of particles with irregular shapes and nonuniform size, which is the character of powder prepared by the solid state method. It proves that the Fe₂P phase is not independent particles, but also can be found in LiFePO₄ particles, so that the Fe₂P and LiFePO₄ phase will be attracted to each other.

Figure 2:

The SEM photos of PS (left), RS (middle), and MAS (right).

The cell performances of the samples were measured to evaluate the impact of the Fe₂P content using the 2032 coin-type cell. The first charge/discharge curves of the samples at 0.2 C rates are shown in Figure 3. All samples exhibited typical charge and discharge plateaus around 3.4 V vs. the Li⁺/Li of the LiFePO₄. Owing to the same LiFePO₄ phase and the low discharge rate, the discharge voltages of the three samples are a little different, and the plateaus are all flat. The first columbic efficiency of PS, RS, and MAS are 94.9%, 94.7%, and 93.2%, respectively. The reason for the low columbic efficiency of the MAS is due to too many impurities that contained lithium, which appeared along with the Fe2P phase and may cause a certain amount of irreversible capacity. In terms of specific capacity, there are obvious differences processed by magnetic separation. RS has the best capacity value (150.5 mAh/g), while MAS has the lowest (134.5 mAh/g). PS is combination of another two samples, so the capacity is in the middle (149.4 mAh/g). The distinction of the samples is just the Fe₂P content, shown as follows: MAS has the most, RS has the least, and PS is in between. It is also important to note that the weight of MAS is relatively very small when compared with that of RS using the magnetic separation, resulting in just a little capacity variance between RS and PS. According to their capacities, it could be speculated the Fe₂P phase has none or little contribution to the specific capacity, so that the specific capacity of the samples will be degraded if the Fe₂P phase is in excess.

Figure 3:

The first charge/discharge cycles for samples at 0.2 C rates.

After several cycles at 0.2 C, the cells were fully activated, and then the rate performances and cycling performances were measured as shown in Figure 4. The discharge capacity of RS is about 160.2, 153.1, 143.1, 121.6 mAh/g at the rates of 0.1, 0.5, 1, 2 C, respectively. For MAS, it is about 137.3, 134.2, 127, and 124.1 mAh/g, correspondingly. As the discharge rate increases, the capacity of RS decreases quickly, meaning that the rate performance is bad, and the cell has a great polarization. On the contrary, MAS has good rate performance and can reduce the polarization of the cell. The rate performances of PS, which could be seen as being composed of RS and MAS, show the synergistic effect that it could obtain the preferable rate performance to RS samples and specific capacity to MAS samples (159.3, 154.7, 148.4 and 139.7 mAh/g at the rates of 0.1, 0.5, 1 and 2 C, respectively). Meanwhile, the voltage plateaus of discharge curves are also different. As the discharge rate increased, the voltage plateaus for MAS is still flat and a little reduced. For RS, there is hardly a voltage plateau at the rate of 2 C. For PS, the voltage plateaus decreases slowly and is obvious at the rate of 2 C. The voltage plateau is related to the polarization of the cell system. Therefore, it could be supposed that the Fe₂P phase is good at reducing the polarization; the more Fe₂P, the less polarization.

Figure 4:

Specific capacity at different rates (left) and the cycling performance at 1 C rate (right) of samples.

Figure 4 (right) shows the cycle performance of the samples at 1 C rate. It can be seen that all have a little capacity loss after 100 cycles. Among the samples, PS and MAS have similar capacity retention rates, 98.8% and 98.7%, respectively. But for RS, the capacity retention rate is the lowest, 98.1%. This result is a characteristic of polarization that a small polarization could increase the potential state of the cell system over a shorter period.

Cyclic voltammetry was performed in order to investigate the effect of Fe₂P on the electrochemical properties of LiFePO₄ by using a scanning rate of 0.1 mV/s (Figure 5). The anodic and cathodic peaks correspond to the two-phase charge-discharge reaction of the Fe²⁺/Fe³⁺ redox couple. It can be observed that all the samples exhibited one oxidation peak and one reduction peak for a cycle, which indicates that only one redox reaction proceeds during the insertion and extraction of Li ions. There is no other peak, which proves that the Fe₂P phase has no activity during the redox process in the voltage range from 2 to 4 V. This result agrees with our previous speculation. All the samples exhibit highly symmetrical current peaks, indicating stable electrochemical kinetics due to a good crystallization of the LiFePO₄. The potential difference between oxidation and reduction peaks of MAS is 0.164 V, smaller than that of RS (0.588 V) and PS (0.365 V), implying a better electrode reversibility and less cell polarization, which should be attributed to the more Fe₂P phase.

Figure 5:

CV profiles of samples at 0.1 mV/s scan rates.

The EIS technique has been widely used to investigate charge transfer and Li+ ion diffusion kinetics in various electrode materials. To further understand the electrochemical behavior of MAS, RS, and PS, an EIS measurement was performed. Figure 6 shows the EIS spectra of the electrodes over the frequency range from 0.01 Hz to 100 kHz. The EIS spectrum is composed of a semicircle in the high-to-middle frequency region and an inclined line in the low-frequency region, which are associated with the charge-transfer reaction at the electrolyte-electrode interface (Rct) and lithium ion diffusion in the bulk of the electrode, respectively. By comparing the diameter of the semicircles, it can be seen that the Rct of the MAS electrode is lower. The respective values of Rct for the MAS, RS, and PS electrodes are 55.56, 90.43, and 66.82 Ω. This result agrees with the powder resistivities of MAS, RS, and PS, which are 232, 604, and 385 Ω·m, respectively. From this, it is proven that the Fe₂P phase could enhance the electronic conductivity of cathode, which improved the rate and cycling performance of the cell.

Figure 6:

EIS plot of samples in frequency range from 100 kHz to 0.01 Hz.

4 Conclusion

In this paper, the influence of the Fe₂P phase on LiFePO₄ has been studied. Using the magnetic separation method, the LiFePO₄ samples with different Fe₂P contents were obtained, which has been proved by XRD. By analyzing the cell performances, it could be seen that the Fe₂P phase helps improve the rate and cycling performances, but a too high content will make the specific capacity decrease due to the low content of the active material. This result was explained by electrochemical performances that the Fe₂P phase could improve the electrical conductivity of the material, but it has no electrochemical activity. So this characteristic of Fe₂P could be utilized to obtain different types of LiFePO₄, such as high rate or high capacity type, by controlling the preparation process.