In a typical synthesis, reagent-grade Li₂SO₄ (0.15 mol), P₂O₅ (0.05 mol), Fe₂O₃ (0.05 mol) and ascorbic acid (1 g) were added into 40 mL deionized water and stirred for 30 min at room temperature, then the mixture was placed in a PTFE-lined stainless steel autoclave and maintained at 200℃ for 24 h. After being cooled naturally to room temperature, the deep green precipitates were collected by centrifugation without further washing and then dried at 80℃ in vacuum overnight before heat-treatment at 600℃ in Ar/H₂ (95 and 5 mL/min ) atmosphere for 10 h. To investigate the effects of ascorbic acid on the morphology and phase purity of products, the amount of the ascorbic acid were changed to 0 or 0.5 g, while keeping other parameters constant. The synthesized samples are named as LFP-x, and the value of x is based on the amount of the ascorbic acid used.

Fourier transform infrared (FT-IR) spectra (4000- 400 cm^-1) were obtained from KBr pellets using a Nicolet FT-IR Impact 400 system. Raman spectrum was taken under ambient condition by using a RM 2000 microscope confocal Raman spectrometer in the range of 400- 2000 cm^-1. Powder X-ray diffraction (XRD) was carried out through a Bruker AXS D8 diffractometer using Cu Kα radiation. The carbon content was measured in a Thermogravimetric analyzer (TGA) with a heating rate of 5 ℃/min to 600℃ in air. The structures were analyzed using a JSM-5610LV scanning electron microscope (SEM, JEOL) and a JEM 2010-FEF transmission electron microscope (TEM, JEOL).

Electrochemical performances of LiFePO₄/C nanocomposites were characterized using CR2016 coin-type cell. A composite electrode was fabricated with the active material, super P carbon back and polyvinylidene fluoride (PVDF) with a weight ratio of 75:15:10 using N-methylpyrolidinone (NMP) as solvent. The slurry was then pasted onto an Al foil and the NMP was evaporated at 120℃ for 12 h in vacuum. Each individual electrode has an area of 0.65 cm² and contains 4-6 mg/cm² of active materials. A disk cut from Al foil coated with LiFePO₄/C nanocomposites was used as the tested electrode. The electrolyte was 1 mol/L LiPF₆ dissolved in ethyl carbonate, diethyl carbonate and dimethyl carbonate (volume ratio, 1:1:1). The coin cells were composed of LiFePO₄/C as the cathode, lithium metal as the anode and a microporous polypropylene sheet as the separator. The cells were charge/discharged between 4.2 V and 2.5 V at room temperature with a rate of 0.1 C-10 C using a charge/discharge apparatus.

Figure 1(a) presents the morphology of the pristine Fe₂O₃ with an average size of 500 nm. When there is no ascorbic acid in the reaction system, the prepared product exhibits a cone-shaped structure with the edge length of about 1.5 μm (Fig. 1(b)). When the amount of the ascorbic acid is 0.5 g, there are cone-shaped particles accompanied with some smaller irregular particles in the product. However, with 1 g ascorbic acid added into the reaction, the product is composed of nanoparticles with the size in the range of 200-300 nm, as shown in Fig. 1(c). Compared with Fe₂O₃, the sample synthesized without ascorbic acid has a larger size, while the sample synthesized with 1 g ascorbic acid has a smaller size. The results indicate that Fe₂O₃ may experience a dissolution and recrystallization process, and the addition of the ascorbic acid into the reaction medium may restrict the growth of the product. This phenomenon can be explained by the space steric effect, which increased the diffusion activation energy of the reactants and slowed down the growth rate of grains^{[ 14]}. It can be found from the high resolution TEM image (Fig. 1(e) and (f)) that both of the primary crystallite of the sample LFP-0.5 and LFP-1 are coated by a fine carbon layer. But, LFP-1 shows a much higher carbon coating with a thickness of 2 nm. The selected area electron diffraction of LFP-1 reveals that the single particle is well crystallized. The fine carbon layer is produced from the decomposition of the ascorbic acid during heat treatment process, through which Li⁺ can easily intercalate into the framework of the synthesized crystal. Although the presence of agglomerated particles usually decreases tap density, the present heat treated samples show relatively high tap density of 1.4 and 1.1 g/cm³for LFP-0.5 and LFP-1, respectively.

Fig. 1

	Figure Option ViewDownloadNew Window
	Fig. 1 SEM (a-d) and TEM (e, f) images of the samples(a) pristine Fe2O3; (b) synthesized without ascorbic acid; (c) and (e) LFP-0.5; (d) and (f) LFP-1

XRD patterns of the products synthesized with different conditions are displayed in Fig. 2. The peaks of the pristine material synthesized without ascorbic acid can be indexed as a mixture of Fe₂O₃ and Fe₅(PO₄)₄(OH)•2H₂O phase (Fig. 2(a)). As for the pristine material prepared with 1 g ascorbic acid (Fig. 2(b)), the Fe₂O₃ phase is vanished, whereas the peaks of the Fe₃(PO₄)₂(OH)₂ phase turn up. After annealed at 600℃ in H₂/Ar atmosphere, the phases of pure LiFePO₄ are obtained and the sharp diffraction peaks illustrate the highly crystalline nature (Fig. 2(c)). As indicated by Fig. 2(d), when adding 0.5 g ascorbic acid the products after heat treatment also show a pure LiFePO₄ phase. However, after heat-treated at 600℃, the product prepared without ascorbic acid exhibits a mixture phase of Fe and LiFePO₄ (Fig. 2(e)), in which Fe should be ascribed to the reduction of Fe₂O₃ by H₂. It should be noted that the precipitates collected after the hydrothermal reaction need not be washed to obtain a pure LiFePO₄ phase. Together with the images shown in Fig. 1, ascorbic acid in the system can be considered not only as structure director, but also as an anti-oxidation source to favor the formation of pure LiFePO₄. And the peaks corresponding to carbon are not visible, owing to its amorphous structures.

Fig. 2

	Figure Option ViewDownloadNew Window
	Fig. 2 XRD patterns of the samples prepared under different conditionsPristine sample synthesized with 0 (a) and 1 g (b) ascorbic acid; the heat treated sample synthesized with 1 g (c), 0.5 g (d) and 0 (e) ascorbic acid, respectively

The FT-IR absorption spectra of the pristine and heat-treated samples synthesized with 1 g ascorbic acid are shown in Fig. 3(a). Except for the similar bands located in the region of 1200-700 cm^-1and 630-540 cm^-1assigned to the PO₄^3- groups^{[ 15, 16]}, several additional absorption peaks exist in the pristine sample without heat treatment. The bands at 3400 and 1634 cm^-1 are attributed to the stretching vibrations of hydroxyls and bending modes of absorbed water, respectively. The band at 2340 cm^-1 originates from a un-deprotonated P-OH group^{[ 17]}. The absorption peak at 1680 cm^-1is attributed to the stretching vibration of C=C in ascorbic acid^{[ 18]}, which confirms that the amorphous carbon layer around the LiFePO₄ crystal lattice seen in the TEM image (Fig. 1(f)) possibly due to carbon resulting from pyrolyzed ascorbic acid. A typical Raman spectrum of the carbon coated LiFePO₄ is shown in Fig. 3(b). The carbon layer makes it difficult to see the details of the spectrum of the olivine structure of LiFePO₄ due to the attenuation of the signal and the overlapping of the spectral bands^{[ 19]}. Only a very weak signal at 945 cm^-1 related to PO₄^3- anion symmetric stretching (the strongest peak in the olivine LiFePO₄ spectrum) is observed. Two peaks at 1600 and 1358 cm^-1correspond to the D and G bands of carbon, respectively. The D band arises because of the disorder induced in sp²-bonded carbon, whereas the G band arises from the in-plane vibration of sp² carbon atoms. The results further reveal that the synthesized sample is composed of LiFePO₄ and amorphous carbon.

Fig. 3

	Figure Option ViewDownloadNew Window
	Fig. 3 FT-IR (a) and Raman (b) spectra of the LiFePO4/C nanocomposites prepared with 1 g ascorbic acid

Thermogravimetric (TG) measurement was used to estimate the carbon content in the final product. Figure 4 gives the TG curve of the sample in air. As can be seen from Fig. 4, the prepared nanocomposite has a high thermal stability below 260℃. Olive LiFePO₄ can be oxidized into Li₃Fe₂(PO₄)₃ and Fe₂O₃ in the temperature range of 250-500℃, corresponding to a theoretical weight gain of 5.07wt%^{[ 20]}. When the sample is heated above 350℃, the carbon in the mixture is oxidized into CO₂ gas. Finally, taking into account the theoretical weight gain (5.07wt%) of pure LiFePO₄ during TGA measurement in air, the amount of coated carbon for the synthesized sample is calculated to be 2.06wt%.

Fig. 4

	Figure Option ViewDownloadNew Window
	Fig. 4 TGA curve of the LiFePO4/C sample prepared with 1 g ascorbic acid

Figure 5(a) shows the cyclic voltammetry curves of the LFP-1. As can be seen, no peak appears at 2.63 V (characteristic of Fe³⁺ in Fe₂O₃), indicating that all the iron atoms in the LiFePO₄/C nanocomposites are Fe^{2+[ 21]}. The couple of redox peaks are observed between 3.2 and 3.6 V ( vs Li⁺/Li), which should be attributed to the Fe²⁺/Fe³⁺ redox couple reaction, corresponding to lithium extraction and insertion in LiFePO₄ crystal structure. Figure 5(b) shows that a flat discharge potential plateau was observed for LiFePO₄/C sample at around 3.4 V, suggesting that the two-phase redox reaction proceeds via a first-order transition between FePO₄ and LiFePO₄.

Fig. 5

	Figure Option ViewDownloadNew Window
	Fig. 5 Cyclic voltammograms (a), cyclability (b, c) and rate performance (d) of LiFePO4/C nanocomposites

Long-term cyclic discharge performances of the samples are shown in Fig. 5(b) and (c). After 500 cycles at 0.1 C, it demonstrates high cycling stability of LFP-1 by delivering the discharge capacity of 159 mAh/g with a capacity retention of 97.5%, while LFP-0.5 exhibit a specific capacity of 129 mAh/g with capacity retention of 92.8%. And the charge/discharge curves inserted in the Fig. 5(b) and (c) indicated that the reversible capacities are relative stable at low rate. Although Fe₂O₃ is used in the hydrothermal system, the typical discharge plateau at 0.75 V corresponding to Fe₂O₃^{[ 22, 23]} does not appear, which could further indicate the phase purity of the prepared materials.

To study the rate capabilities of the fabricated electrodes, the cells were charged and discharged at various current rates. As shown in Fig. 5(d), charge/discharge measurement was firstly conducted at a rate of 0.1 C, and then increased to 10 C stepwise and finally returned back to 0.1 C. There is a decrease in the discharge capacity of the prepared sample with the current rates increasing. The initial discharge capacities of LFP-1 at a rate of 0.1 C are 163 mAh/g and the columbic efficiency is 97%, while that of LFP-0.5 is 136 mAh/g and 96%, respectively. However, the discharge capacity for the LFP-1 at 10th cycle decreases to 144 and 140 mAh/g at current rates of 1 C and 2 C, respectively. Although the specific capacity of electrode decreases drastically when the sample is cycled at higher rates from 1 C to 10 C, the initial capacity of the composite is retained if returning to the low current rate (0.1 C). As exhibited in Fig. 5(d), there is a striking difference between the prepared samples when discharged at high rates. There is almost no capacity retention when LFP-0.5 discharged over 2 C rate. But the initial specific capacity can be retained after the current change to 0.1 C rate. These results indicate that the capacity loss at high charge/discharge rate is mainly caused by the low diffusion rate of lithium ions within the electrode materials due to the larger particle size of sample LFP-0.5^{[ 21, 24]}.

A hydrothermal reaction-calcination process for the synthesis of LiFePO₄/C nanocomposites in the water medium was developed using inexpensive and stable Fe₂O₃ as the iron source, ascorbic acid as the antioxidant and carbon source. It is found that the amount of ascorbic acid has an effect on the structure, phase and carbon amount of the final product. With 1 g ascorbic acid used in the reaction, the particle size of synthesized LiFePO₄/C nanocomposites is in a range of 220-280 nm. When used as the cathode materials for the lithium-ion batteries, the as-prepared material shows high capacity and good cycle stability (159 mAh/g at 0.1 C rate over 500 cycles), as well as good rate capability. Its excellent cycling behavior make it a promising cathode material for the lithium-ion batteries.