Introduction

In recent years, lithium-ion battery has been widely used due to its high voltage and high specific and volumetric energy densities, and easy transportability. Currently, a key factor in the lithium-ion battery is the research of high-voltage cathode materials that has mainly focused on layered transition metal oxides LiMO2 (M = Co, Ni) and spinel structures LiM2O4 (M = Mn, Ni) [1–7]. The widely used LiCoO2-based batteries have problems regarding safety, cost, and environment, which are difficult to overcome. Searching for cheaper, more secure, stable and environmentally friendly cathode materials has become a top priority. LiNiO2, which has a similar structure as LiCoO2, has an advantage with respect to cost and raw materials. However, the difficulties in the synthetic process, the unstable structure, and the poor chemical stability become shortcomings of LiNiO2 to limit its application. Finally, the spinel LiMnO4 has advantages in rich resources of raw materials, low cost, increased safety, environmental friendliness, and simple synthetic process. But further research, development, and application of the synthetic process are hindered by the Jahn–Teller effect in the process of charging and discharging.

In 1997, the Goodenough group [8] first reported that olivine structure lithium iron phosphate (LiFePO4) can display the intercalation/de-intercalation of lithium ions reversibly. Since, LiFePO4 has been considered to be the most promising cathode material for the lithium-ion battery due to its nontoxicity, low cost, and high thermal stability characteristics [9–17]. Unlike other Li electrodes, LiFePO4 exhibits almost theoretical capacity of about 170 mAh g−1 [2], and Li4Ti5O12 also almost shows theoretical capacity (Fig. 1); however, it must be mentioned that the practical capacity depends on C-rate.

Fig. 1
figure 1

A range of lithium-ion battery electrode available or currently under development and comparison of their theoretical capacity and practical capacity (NCA: LiNi0.8Co0.15Al0.05O2, NCM: LiNi1/3Co1/3Mn1/3O2)

Full size image

The iron sources for preparation of lithium iron phosphate can be divided into trivalent iron sources [18–24] and divalent iron sources [25, 26]. Divalent iron sources are expensive and easily oxidized. Even in an inert atmosphere, it is still difficult to avoid the appearance of Fe3+ impurity. FePO4, as a trivalent iron source, which has a low cost and a high chemical stability, is an ideal material for the synthesis of lithium iron phosphate. LiFePO4 particles can be obtained by controlling the size and morphology of FePO4 with similar characteristics, so as to achieve the target of high bulk density, high discharge rate, and high specific capacity, overcoming the limitation in ionic conductivity of LiFePO4 to a certain extent. In this paper, we review the crystal structure, physical and chemical properties, and charging/discharging mechanism of Li iron phosphate and focus on the aspect of synthesizing iron phosphate as a precursor for lithium-ion batteries.

Crystal structure, and charging/discharging mechanism of iron phosphate

FePO4 [27–30] with amorphous and different crystalline phases has already been synthesized, including heterosite, α-quartz phase, monoclinic system, and orthorhombic system. In this paper, the amorphous state and various crystalline phases of iron phosphate are discussed. However, discussion is limited to phosphate ions in the form of independent PO4 3−, with no expansion to the aggregate state, meaning that the n in (P n O3n + 1)(n + 2)− can only take the value 1, but not 2 or 3. Pressure and temperature can affect the structure of FePO4. At normal pressure, FePO4 exists as α-quartz phase, the same structure as AlPO4. Each phosphorus atom and iron atom are connected to four oxygen atoms. Under a higher pressure, it changes to a tetragonal system similar to CrPO4 with the cell parameters a = 5.227 Å, b = 7.770 Å, and c = 6.332 Å. Then, under a pressure of 2.5 GPa, FePO4 changes to a structure between VCrO4 phase and amorphous state. Recent experimental observations suggest that LiFePO4/FePO4 interfaces are the juxtapositions of the two end members (FePO4 and LiFePO4) instead of solid solutions [31]. However, a solid solution might exist under particular conditions, e.g., at high temperatures [32] or in nanosized particles [33]. FePO4 changes into a trigonal system at 650 °C. When the temperature rises to 705 °C, an α–β phase transition from a P6222 to a P31121 space group occurs [34]. Compared to the tetragonal system of LiFePO4, FeO6 octahedron is replaced by FeO4 tetrahedron in the trigonal system [35] (Fig. 2).

Fig. 2
figure 2

Idealized crystal structures of orthorhombic and trigonal FePO4

Full size image

The intercalation/de-intercalation process of lithium ions between FePO4 and LiFePO4 can be expressed as LiFePO4 ⇔ xFePO4 + (1 − x)LiFePO4 + xLi+ + xe− (at 3.45 V/Li) over a large composition range, with a theoretical capacity of 170 mAh g−1. To further study the intercalation process of lithium ions, Padhi et al. [8] proposed a radial model (Fig. 3a). It was suggested that this process is carried out through a two-phase interface with FePO4 and LiFePO4 looking like a coaxial core–shell. During charging, along with lithium-ion inserting, the interface migrates to the particle center and the interfacial area decreases, and when a critical value of area is reached, lithium-ion migration through the interface cannot support the current, and the electrochemical behavior becomes limited by the rate of diffusion (Fig. 3).

Fig. 3
figure 3

Schematics of three growth models for the LiFePO4–FePO4 phase transition. a Isotropic shrinking core model, b anisotropic growth mode, and c anisotropic growth model. Reprinted from Ming et al. [38] copyright (2010), with permission from Annual Reviews

Full size image

Anderson and Thomas [36] proposed a different “Mosaic model.” It is believed that the de-intercalation process of lithium ions can occur over the entire surface area rather than only at the two-phase interface. With the FePO4 phase of the de-lithiation region increasing gradually, the LiFePO4 that remained un-reacted is coated by FePO4 layers in the process, becoming a source of capacity loss. Chen et al. [37] proposed that the FePO4 region and the LiFePO4 region are separated by a dislocation line along the c-axis with anisotropy. In the de-intercalation process of lithium ions, the two-phase interface located at the b–c plane advances along the a-axis horizontally (for details, see Fig. 3b). Laffont et al. [31] proposed that in both the intercalation and de-intercalation processes of lithium ions, the FePO4 region is always in the center of the plane, while the LiFePO4 region can be found close to the edge of area (Fig. 3c). Li-ion migration paths in a unit cell of LiFePO4 are represented in Fig. 4, as summarized by the above three models.

Fig. 4
figure 4

Li-ion migration paths in a unit cell of LiFePO4 [39]. Mechanism A, [010] direction; mechanism B, [001] direction; mechanism C, [101] direction. Redrawn from Saiful et al. [39] copyright (2005), with permission from the American Chemical Society

Full size image

The FePO4 obtained from LiFePO4 with a de-intercalation process of lithium ions, referred to as heterosite structure, belongs to the same Pnma space group with LiFePO4 with a similar structure, exhibiting a small contraction in a and b parameters and a small increase in the c parameter [40] (Table 1).

Table 1 Lattice parameters of LiFePO4 and FePO4 [8]
Full size table

Synthesis of FePO4 by wet chemical methods

When the synthetic process involves water or is carried out in aqueous solution, the FePO4 obtained often contains crystal water, namely, FePO4·xH2O. Figure 5 shows the thermogravimetry–differential scanning calorimetry (TG-DSC) diagram of FePO4·xH2O obtained by Qian et al. [41]. In the range of 18–500 °C, an obvious endothermic peak in the DSC curve and a well-defined weight loss in the TG curve were observed, corresponding to the loss of crystal water of the precursor. According to the weight loss in this process, the x in FePO4·xH2O can be estimated. The exothermic peaks at 614.46 and 690.63 °C without obvious weight loss can be assigned to structural transformation from amorphous to hexagonal FePO4 and the α–β transition, respectively. Of course, TG-DSC curves vary with different synthetic methods [41, 42] but are generally similar with the previous example.

Fig. 5
figure 5

TG-DSC curves of the amorphous FePO4. Reprinted from Qian et al. [41] copyright (2012), with permission from Elsevier

Full size image

Xu et al. [40] proposed that Fe2P2O7 appeared at 380 °C and transformed to FePO4 at 460 °C during the heating process. Zhang et al. [43] divided the dehydration process of FePO4·4H2O into a multistep reaction by an iso-conversional rate method and analyzed it by a multivariate nonlinear regression method. They concluded that no branching reaction existed, and the most probable model for the dehydration process is a two-step consecutive reaction with kinetic parameters, such as activation energy and pre-exponential calculated out, for each of the two steps. The most probable kinetic model was estimated with a multivariate nonlinear regression method assuming a two-step consecutive reaction: D4 → Fn. The activation energy E and ln(A/s−1) of D4 were 79.62 kJ mol−1 and 19.35. Those of Fn were 103.04 kJ mol−1 and 25.38 [43].

Boonchom and Danvirutai [44] studied the thermal decomposition kinetics of FePO4·3H2O in air. The FePO4·3H2O decomposed in two steps after 50 °C: the first and second decomposition steps are the loss of one and two molecules of water in crystallization, respectively. The results indicated that the kinetic model, which better describes the second reaction of dehydration for FePO4·3H2O, was the Fn model as a simple nth-order reaction (n = 2.50). The calculated kinetic parameters of the Coats–Redfern method were E a = 77.95 ± 1.18 kJ mol−1 and ln(A/s−1) = 14.26 ± 0.87. Additionally, Boonchom and Puttawong [45] also reported the dehydration reaction of FePO4·2H2O in dynamical air atmosphere. FePO4·2H2O decomposed in one step, and the possible conversion function was based on the three-dimensional diffusion mechanism (D4 model), with the correlated kinetic parameters E a = 65.94 ± 2.83 kJ mol−1 and ln(A/s−1) = 8.32 ± 1.14.

Okawa et al. [46] reported that amorphous FePO4 can be obtained at 350 °C and transformed into a crystalline state at 700 °C, with the specific capacity dropped down to less than a half of the former. This can be attributed to the low activity for FePO4 of the trigonal system and the formation of a glassy phase, Fe3P5O7, covering the surface of FePO4 at 580 °C. This is consistent with the study by Prosini et al. [47]. Chang et al. [48] proposed that a hexagonal layered structure of FePO4 is beneficial to the diffusion of lithium ions to the center of grains, the formation of olivine LiFePO4 along the c-axis, and the carbon coating in the reduction process, as well as the metal ion doping.

Synthetic methods of iron phosphate

FePO4 is not only used as a precursor for LiFePO4, but can also be used as the cathode material directly, however, with a poor cycle performance. In addition, FePO4 has also been used widely as absorbents, catalysts, antirust pigments, and additives. In the electrochemical field, it is imperative that FePO4 contains no undesirable impurities. Wang et al. [49] proposed that the residual moisture in FePO4 can react with Li+ to produce LiOH (Fig. 6a, b), resulting in a higher first capacity and an irreversible degradation in the subsequent cycles. Song et al. [50] proposed that a lower activity for FePO4 sintered at a temperature of 700 °C was caused by a glass phase on the surface (Fig. 6c). Secondly, it requires a high specific surface area and a high bulk density. This can be achieved by producing particles with a small size, a uniform distribution, or a spherical morphology. The electrochemical performance of LiFePO4 is the most important outcome of the synthetic method. Finally, we hope that the synthetic method is simple, convenient, feasible, low cost, fast, environmentally friendly, and easy to achieve industrialization. At present, the synthetic methods of iron phosphate mainly include coordinate precipitation, sonochemical methods, hydrothermal methods, sol–gel methods, homogeneous precipitation, fluoride system methods, surfactant templates, and biologic templates.

Fig. 6
figure 6

SEM images of the a amorphous FePO4·2H2O, b crystalline FePO4, and c discharge curves of amorphous FePO4 obtained at different temperatures. Reprinted from Wang et al. and Song et al. [49, 50], copyright (2012 and 2002), with permission from Elsevier

Full size image

Co-precipitation method

The co-precipitation method, also known as liquid phase precipitation method, has a comparable convenient procedure, consumes little energy, requires simple equipment, and can produce particles with small size and uniform distribution. However, this method requires similar precipitation conditions for the raw materials, which may restrict the range of choice for the starting materials. Over the past few years, there have been a number of reports regarding this method. The co-precipitation method can be done via two approaches: the liquid phase oxidation precipitation method and liquid phase non-oxidation precipitation method. The former uses divalent iron source and H2O2 as the oxidant, while the latter uses trivalent iron source, which is cost effective.

Generally, a stoichiometric Fe/P ratio of 1 is used. However, Hu et al. [51] believed that the electrochemical performance can be improved by increasing the H3PO4 proportion. FeSO4 was mainly selected as the iron source for the liquid phase oxidation precipitation method, while Fe(NO3)3 and Fe2(SO4)3 were chosen to be the iron sources for the liquid phase non-oxidation precipitation method. Chang et al. [48] used several kinds of trivalent iron sources and reached the conclusion that a maximum bulk density can be gained by using Fe(NO3)3, which was about double of the bulk density, when Fe2(SO4)3 and FeCl3 were used. H3PO4 or (NH4)2HPO4 is mainly chosen as a phosphorus source. Jiang et al. [52] proposed that a H3PO4/(NH4)2HPO4 ratios of 3:1 can reduce the difficulty of adjusting the pH value.

A too large concentration will result in a large number of lattice defects, while a too low concentration will lead to coarse grain. In addition, the concentration of solution can affect the amount of crystal water, meaning that a higher concentration causes a smaller value for x in FePO4·xH2O. Both the adjustment and the final value of pH value in the synthesis are of great importance. When pH >2.2, a red-brown precipitate of Fe(OH)3 appears, while a too low pH value will result in a incomplete precipitation of Fe3+ and waste. Ma et al. [53] proposed that a low temperature (70 °C or less) was not favorable for the Fe(OH)3 to FePO4·xH2O conversion, while a high temperature was promoting the crystallization of FePO4·xH2O. The whole synthesis process also needed stirring. When the stirring speed increased, the raw materials would react entirely. However, a too high speed can cause a whirlpool in the liquid center, causing uneven mixing away from the center. The FePO4·xH2O amorphous precipitation was collected by filtration, washed with distilled water, dried, grinded, and heated. Chang et al. [48] proposed that the aggregation of FePO4 can be reduced by drying two times (drying under vapor after natural drying), and the bulk density of FePO4 can be increased by nearly 50 % to 2.14 g cm−3 simultaneously. Xia et al. [54] reported that a small amount of Fe2P with a good dispersion after calcination at 800 °C.

In addition, when the flow rate was controlled strictly by a constant flow pump [48], a metering pump [55], or a peristaltic pump [56], the solution was added uniformly rather than mixed directly, and a high-density FePO4 was obtained. Cao et al. [56] reported that the FePO4 particles of large size can get an excellent bulk density (2.03 g cm−3), and the FePO4 particles of middle size delivered a capacity density of 230.4 mAh cm−3 at a rate of 0.1C due to the relatively high bulk density and a lower electrochemical polarization. Zhu et al. [57] reported a facile co-precipitation route through which amorphous iron phosphate particles were synthesized, and the FePO4 obtained featured narrow size distribution, abundant porous structure, and large specific surface.

The homogeneous precipitation method, in which a precipitant is added to achieve a uniform precipitation, can be classified in the liquid phase precipitation method. Carbamide is a commonly used precipitant. Gong et al. [58] reported that the wafer-like FePO4 of hexagonal structure with uniform particle size was obtained by adding a carbamide precipitant and a shape control agent. In the sonochemical method, proposed by Okawa et al. [46], an ultrasonic vibration measure was employed to initiate a series of reactions, mainly generating H2 and H2O2, and achieving oxidation of FeSO4. Although a different name was used, in essence, it was still in the range of the co-precipitation method. This method had the following advantages: no use of oxidation agents, reduction in reaction time, and the ability to control particle size. However, we believe that it may not be suitable for large-scale industrialization due to repeated filtration and washing. The common synthetic process is shown in Fig. 7.

Fig. 7
figure 7

Synthetic process of coordinate precipitation method [46, 48, 51–54, 56, 57]

Full size image

Hydrothermal method

Mal et al. [59] reported that a novel organic–inorganic mesoporous FePO4 had been synthesized by using sodium dodecyl sulfate (SDS) molecules as surfactant, adding FeCl3 and C6H5PO(OH)2 as raw materials under stirring, followed by a heat treatment at 453 K for 15 h in a Teflon-lined stainless steel autoclave (Fig. 8). The hydrothermal method had the following advantages: small size and uniform distribution of particles obtained, simple reaction condition, and completeness of grain growth. The use of autoclave meant that raw materials should be added all at once, that the process was extremely difficult to observe and control, and that industrial production was limited due to the small autoclave container size.

Fig. 8
figure 8

Synthetic process of hydrothermal method [59]

Full size image

Sol–gel method

Guo [60] reported that the trigonal FePO4 with a particle size of 100 nm was synthesized by using citric acid as ligand after adjusting the pH value and the ligand molar ratio. Lu et al. [61] reported that FePO4 with uniform distribution was synthesized by thermal decomposition of FeOPO(OC6H4COOH)2·0.5H2O precursor prepared from a HOPO(OC6H4COOH)2 ligand dissolved in distilled water (70 %) and ethanol (30 %) with the addition of Fe(NO3)3 drop-wise under continuous stirring (Fig. 9). It was suggested that the existence of a little residual carbon after thermal decomposition can improve the conductivity of the cathode material. The advantages of the sol–gel method were low temperature and high uniformity. But a large number of micropores in the gel may cause contraction during the drying process. The best discharge capacity of the synthesized samples around 380 °C was 146 mAh g−1.

Fig. 9
figure 9

Synthetic process of sol–gel method [61]

Full size image

Fluoride system method

Guo [60] reported that the hexagonal mesoporous FePO4 with a particle size of 2.6 nm was synthesized by adding HF as mineralizer and sodium dodecyl sulfate as surfactant into FePO4 (prepared by mixing Na2HPO4 with Fe(NO3)3 suspension and reacting at 60 °C for 2.5 h) (Fig. 10). However, the removal of impurities was cumbersome. It was proposed that neither NaF nor NH4F can work well, while HF with a low dissociation degree could form a hydrogen bonding between [FeHPO4]+–F− ion pairs. Also, the F− anions had an effect on the electrostatic interaction between ion pairs and anionic surfactant micelles. They also enhanced the formation of the hexagonal mesostructured phase.

Fig. 10
figure 10

Synthetic process of fluoride system method [60]

Full size image

Surfactant template method

In this method, mesoporous materials are synthesized under the guide function of templates. Zhu et al. [62] reported that highly ordered hexagonal mesoporous FePO4 was synthesized by adding cetyltrimethylammonium chloride (CTMACl) into Fe2(SO4)3 and H3PO4 raw materials. However, it was pointed out that only 50 % of CTMACl can be removed effectively. Similarly, Shi et al. [63] took a surfactant (EO20–PO20–EO20, Plurinic P123) as the template. They proposed that a little residual carbon would be left in FePO4 after heating and a collapse of the mesoporous structure would happen in the temperature range of 500–600 °C. The mesoporous FePO4 delivered enhanced specific capacity of 160 mAh g−1 at the first discharge process and 135 mAh g−1 in the following cycles at 0.1C rate. Wang et al. [64] reported that cetyltrimethylammonium bromide (CTAB) was also suitable to be a template (Fig. 11). The LiFePO4/C nanoparticles lithiated from the obtained FePO4·2H2O nanoplates delivered discharge capacities of more than 150, 120, 110, 100, and 75 mAh g−1 at rates of 5C, 10C, 15C, 20C, and 30C, respectively.

Fig. 11
figure 11

Synthetic process of surfactant template method [64]

Full size image

Biologic template method

In the so-called biologic template method, organisms provided nucleation sites to regulate the grain growth. Zhou et al. [65] reported that iron phosphate nanopowders with flake morphology had been synthesized with yeast cells as a biologic template. In the synthetic process, the biologic template adsorbed Fe3+ cations firstly, which led to the formation of FePO4. They pointed that the organisms could be removed completely after heating at 500 °C and only little conductive carbon was left. Similarly, Cao and Li [66] reported that iron phosphate hollow microspheres were obtained with pollen grains as biologic template (Fig. 12).

Fig. 12
figure 12

Synthetic process of biologic template method [66]

Full size image

Other synthetic methods

Chen et al. [67] reported that crystalline FePO4 had been synthesized through a reaction between iron powder, H3PO4, and NH4H2PO4 in aqueous solution and mixed with conductive agent after grinding. This synthetic method was simple and low cost; however, the electrochemical performance was poor. Wu et al. [68] prepared nano FePO4·xH2O powders by controlled crystallization method, using Fe(III) compound as the iron source. Then, the olivine nano LiFePO4/C composites were obtained through carbothermal reduction process at different temperatures. The results show that the nano LiFePO4/C composite calcined at 700 °C for 10 h has fine particle sizes of about 40–100 nm. The nano LiFePO4/C composite cathode material can deliver an initial discharge capacity of 156.5, 134.9, 105.8, 90.3, and 80.9 mAh g−1 in the voltage range of 2.5–4.2 V, at a rate of 0.1C, 1C, 5C, 10C, and 15C, respectively, which exhibits good rate performance.

Okada et al. [35] reported that an increase of more than 35 % in specific capacity can be obtained by mixing P2O5 with Fe, ball milling with water for 24 h, and subsequent heating and ball milling without water for 24 h. Delacourt et al. [69] reported that several different hydrated iron phosphates were prepared under refluxing conditions. The low temperature solid phase method [70] took high energy consumption in the synthesis process, and the uniformity of particle size distribution was relatively poor. Qian et al. [41] reported that amorphous FePO4 had been synthesized by an electrochemical synthetic method with H3PO4 as the electrolyte, iron as the anode, and H2O2 as the oxidant. This method may be a new direction to prepare FePO4.

Conclusion and prospection

In this paper, the structure, properties, and various synthetic methods of iron phosphate were discussed, with a focus on the synthetic methods. One common issue that needs to be addressed is that although some researchers had synthesized FePO4 of a large surface area or small particle size, the electrochemical performances of the ultimate product LiFePO4 were still unknown. On one hand, to obtain an appropriate FePO4 precursor, researchers should continue to increase bulk density, reduce cost, and optimize details to adapt to industrialization. On the other hand, the coordination between synthetic methods for FePO4 precursor and the subsequent lithium insertion operation should be taken into consideration. In general, the method to prepare LiFePO4 was mixing the FePO4 precursor with a carbon source and a lithium source in a tube furnace by carbothermal reduction method, which inevitably led to a reduction of bulk density and an enlargement of particle size. At present, some researchers also take a rheological phase method [71] as well as other methods.