Structural and Electrochemical Characterization of Pure LiFePO₄ and Nanocomposite C-LiFePO₄ Cathodes for Lithium Ion Rechargeable Batteries

Abstract

Pure lithium iron phosphate (LiFePO₄) and carbon-coated LiFePO₄ (C-LiFePO₄) cathode materials were synthesized for Li-ion batteries. Structural and electrochemical properties of these materials were compared. X-ray diffraction revealed orthorhombic olivine structure. Micro-Raman scattering analysis indicates amorphous carbon, and TEM micrographs show carbon coating on LiFePO₄ particles. Ex situ Raman spectrum of C-LiFePO₄ at various stages of charging and discharging showed reversibility upon electrochemical cycling. The cyclic voltammograms of LiFePO₄ and C-LiFePO₄ showed only a pair of peaks corresponding to the anodic and cathodic reactions. The first discharge capacities were 63, 43, and 13 mAh/g for C/5, C/3, and C/2, respectively for LiFePO₄ where as in case of C-LiFePO₄ that were 163, 144, 118, and 70 mAh/g for C/5, C/3, C/2, and 1C, respectively. The capacity retention of pure LiFePO₄ was 69% after 25 cycles where as that of C-LiFePO₄ was around 97% after 50 cycles. These results indicate that the capacity and the rate capability improved significantly upon carbon coating.

1. Introduction

Lithium iron phosphate (LiFePO₄) is under intense investigation since its introduction in 1997 as a possible cathode material for Li-ion rechargeable batteries [1]. LiFePO₄ belongs to the olivine type compound exhibiting a theoretical capacity of ~170 mAh/g and a flat charge/discharge profile at ~3.4 V versus Li/Li¹⁺. Additionally, the cost effectiveness, environmental and safety returns (high abuse tolerance), thermal stability at fully charged state, and reasonably good cycleability have made LiFePO₄ as one of the most attractive cathode materials for rechargeable Li-ion batteries. However, for successful commercial acceptance several material related issues, which can be grouped into three categories, should be addressed. (i) Synthesis of phase pure olivine lithium iron phosphate at relatively low processing temperatures. Various synthesis routes have been adopted to synthesize phase pure lithium iron phosphate such as solid-state synthesis [2, 3], sol-gel [4, 5], microemulsion synthesis [6], hydrothermal synthesis [7, 8], and so forth. The reported research [2–8] indicates that it is important to select proper precursor materials as well as optimize process parameters to keep iron in its +2 valence state to suppress the formation of impurity phase such as Li₃PO₄. Use of carbon with the precursor materials and calcinations at inert ambient has been found to be useful to retard the iron oxidation and thereby limiting the formation of impurity phase. (ii) The poor ionic as well as electronic conductivities of LiFePO₄ materials [9]. In several reports, carbon coating improved the Li-ion kinetics in LiFePO₄ cathode materials [2, 5, 7, 10]. In contrast to this popular belief, it is also argued that the effect of carbon coating is marginal [11]. However, LiFePO₄ community now reached a consensus and believes that carbon coating is beneficial in improving the electrochemical performances of LiFePO₄. Efforts have also been undertaken to eliminate carbon coating with other metal dispersion (namely, copper or silver) [12, 13] or conducting organic materials (namely, polypyrole) [14]. Initially, it was believed that the intrinsic electronic conductivity of LiFePO₄ could be improved by aliovalent dopants substitution and is now understood that it was an artifact due to the presence of carbon in the precursor, and also of metallic Fe2P impurities [15–18]. (iii) Control on the particle size in the nanoregime along with a narrow particle size distribution of the synthesized LiFePO₄ cathode materials. Li⁺ has slow diffusivity in LiFePO₄ (𝐷_Li ~10⁻¹⁴–10⁻¹⁶ cm²/s) compared to the widely used layered LiCoO₂ ((𝐷_Li ~10⁻¹²–10⁻¹³ cm²/s) [19–21]. As a result, only ~60%–70% of the capacity could be obtained for the original LiFePO₄ in the early works, and its capacity decreases remarkably at larger current density. Hence, maintaining particle size in nano-regime (to ensure shorter diffusion length for realizing maximum (de)intercalation within the stipulated time) has been reported to improve discharge capacity as well as rate capability [22]. Here again, addition of carbon is beneficial in suppressing the particle growth during high-temperature calcinations. The above discussion on the three material related issues of LiFePO₄ for cathode application clearly suggests that addition of carbon could be beneficial in addressing all three issues and yielding phase-pure olivine, enhancing its electronic conductivity and retarding the particle growth.

Despite aforementioned improvements, the nature of the morphology of C-LiFePO₄ composite is yet to be properly understood. Earlier, it was not clear whether carbon should form a thin coating on LiFePO₄ particles or point contacts between particles to have beneficial effect on the discharge capacity and the rate capability. Research efforts have also been directed to elucidate whether crystalline or amorphous carbon is beneficial for the targeted electrochemical properties. At present it is believed that an amorphous thin coat insures point contact [23]. The process complexities, as outlined above, are reflected in the electrochemical properties of C-LiFePO₄ cathode materials. Hence, a comparative study of pure LiFePO₄ and composite C-LiFePO₄ in terms of structure and electrochemical properties may be of importance to understand the role carbon in LiFePO₄. The details of the electrochemical data from some of the recent literature report are summarized in Table 1. It is seen from Table 1 that the scattering of the data is thought to be stemmed out mostly from one (or more) of the three factors mentioned above. The routes and conditions used to synthesize LiFePO₄ have, in general, profound influence on its electrochemical properties, where highest discharge capacity obtained was around 163 mAh/g with C/10 rate [4]. However, in terms of rate capability best results reported so far are 120 mAh/g with 10 C [24]. In view of the above discussion, there is a compelling need for optimization of the synthesis process to obtain phase pure nanocrystalline C-LiFePO₄ composite cathode materials for its reliable cathode applications in Li-ion batteries. Also, the knowledge about the lattice dynamics is essential for understanding the phase transformation during Li-ion (de)intercalation of C-LiFePO₄ and Raman spectroscopy is a valuable technique for such studies [25–29]. In the present case, a solid-state route under nitrogen ambient has been adopted to synthesize pure LiFePO₄ and C-LiFePO₄ composite cathodes at a relatively low temperature and their structural and electrochemical properties have been studied and compared. Moreover, the Raman spectra of C-LiFePO₄ at various stages of charging and discharging have been taken to study the structural reversibility.

Table 1 

Electrochemical properties of lithium iron phosphate cathode.

2. Experimental Details

Stoichiometric amount of lithium carbonate (Li₂CO₃ (99.999%, Alfa Aeasar)), iron oxalate (FeC₂O₄⋅2H₂O (99.999%, Alfa Aeasar)), and ammonium dihydrogen phosphate (99.995% Alfa Aeaser) were used as precursor materials to prepare LiFePO₄ pure and C-LiFePO₄ composite cathode materials. High-energy balling system was used for the synthesis of LiFePO₄ and C-LiFePO₄ powders. These reagents mixtures were first ball milled for 10–18 hours in acetone media and vacuum dried in the furnace at 150^∘C to predecompose the oxalate and phosphate. The obtained powder was grinded in the glove box under argon atmosphere to suppress the oxidation ferrous iron (Fe²⁺) to ferric iron (Fe³⁺) and subsequently fired at 400^∘C for 5 hours in flowing nitrogen to ensure complete organic removal. After this step, the powder was divided into two halves, and one half was mixed with 8 wt% carbon black and again ball milled without any solvent. Then the two batches of the powders (with and without carbon) were calcined in the temperature range of 500–750^∘C for 4 hours in flowing N₂ atmosphere to obtain pure LiFePO₄ (without carbon) and C-LiFePO₄ (composite) cathode materials.

The phase formation behavior of the synthesized powders was investigated using X-ray diffraction (XRD, Siemens D5000), diffractometer with Cu K𝛼 radiation (𝜆=1.54056Å). The XRD data were collected in the range 10–80^∘ with a step of 0.02^∘ and a count time of 1 second per step. XRD spectra were refined by Reitveld method using the Fullprof package to identify the structural change of pristine samples [31]. The morphology of the synthesized powder was investigated using a scanning electron microscopy (SEM) and transmission electron microscopy (TEM, Carl Zeiss Leo Omega 922 at 200 KeV). The nature of carbon coating on the pristine lithium iron phosphate was characterized using TEM in conjunction with the Raman scattering measurements (T64000 spectrometer equipped with a triple-grating monochromator and a Coherent Innova 90C Ar⁺-laser at 514.5 nm).

In order to evaluate the electrochemical characteristics of the synthesized powder, the working electrode (cathode) was prepared as follows: first a slurry was made by mixing 80 wt% active material (C-LiFePO₄), 12 wt% binder (polyvinylidine fluoride (PVDF, Alfa Aesar)), and 8 wt% carbon black in a solvent N-methyl pyrrolidone. The slurry was coated on aluminum foil current collector and dried in an oven at 60^∘C for 12 hrs. The dried electrode was used as cathode to fabricate coin cell (CR 2032) which was comprised of lithium foil as an anode and LiPF₆ (1 M) (in a mixture of ethylene carbonate and dimethyl carbonate in 1  :  1 ratio) as an electrolyte. The fabrication of coin cell assembly was carried out in Ar atmosphere inside a glove box. A computer controlled potentiostat-galvanostat system (Solatron battery-testing unit 1470E) was utilized for electrochemical measurements. The cyclic voltammograms (CV) were recorded at various voltage scan rate ranging 0.1–0.5 mVs⁻¹ with a cut-off limit 2.5–4.3 V versus Li/Li⁺. The CV data were analyzed to estimate the diffusion coefficient of Li⁺ at room temperature. The charge discharge measurements were performed with various current densities.

3. Results and Discussion

3.1. Structure and Morphology

X-ray diffraction patterns of pure LiFePO₄ and of C-LiFePO₄ materials (calcined at 700^∘C for 4 h in nitrogen ambient) are shown in Figure 1(a). All the major reflections in the XRD pattern were indexed based on orthorhombic olivine structure (space group Pnma). Some additional peaks (marked by arrows, corresponding to the crystalline impurity phases) were also observed in the XRD pattern of LiFePO₄. These peaks at 27^∘, 31^∘, and 33^∘ are probably due to Li₃Fe₂ (PO₄)₃. In some of the reported works it is argued that the deficiency of ferrous iron in oxalate resulted detectable amount of Li₃Fe₂(PO₄)₃ phase [32]. However, XRD pattern of C-LiFePO₄ (8 wt% C) clearly shows the ideal orthorhombic olivine structure (JCPDS card no. 40–1499) without any impurity phase compared to pure LiFePO₄ synthesized without carbon mixing. In Li₃Fe₂ (PO₄)₃ iron present in the Fe³⁺ oxidation state where as in LiFePO₄ it is present in the Fe²⁺ oxidation state. Therefore, we speculate carbon mixing ensured reduction of Fe³⁺ to Fe²⁺ or suppress the formation of trivalent Fe ion in an oxygen deficient atmosphere (like N₂ used in the present case), resulting in the formation of single phase orthorhombic LiFePO₄ [33]. The calculated XRD pattern based on Reitveld fit along with the experimental pattern are also shown in Figure 1(b). Excellent match between the experimental and calculated XRD patterns is clearly seen in Figure 1(b). The calculated lattice parameters were 𝑎=10.328 (1) Å, 𝑏=6.0084 (6), and 𝑐=4.6991 (6) Å with agreement parameters 𝑅_𝑝=2.573% and 𝑅_𝑤𝑝=3.33%. The refined parameters match quite well with the existing literature report in carbon-coated LiFePO₄ particles [34].

(a)

(b)

Figure 1 

XRD patterns of LiFePO4 material (a) with and without carbon and (b) Rietvield refinement of C-LiFePO4 composite material.

Figure 2 shows the scanning electron micrographs (SEMs) of pure LiFePO₄ and C-LiFePO₄ powders (calcined at 700^∘C in nitrogen ambient). In the case of pure LiFePO₄ flake-like particles with large size distribution (200 nm to 1 𝜇m) were observed. However, in the case of C-LiFePO₄, particles were more uniform with narrow size distribution (80 to 200 nm). Hence, the carbon, as an additive, inhibits the particle growth (even when lithium iron phosphate is calcined at relatively higher calcinations temperature) yielding a homogeneous particle size distribution with an average particle size of ~200 nm. To understand the effect of carbon, the C-LiFePO₄ composite powder was characterized by TEM and Figures 3(a)–3(c) show the results. From the TEM micrograph it is apparent that the particles form agglomeration in most of the region (Figure 3(a)); however, in selected region separated particle with size < 200 nm is observed. The energy dispersive spectrometry results of the carbon mapping on the C-LiFePO₄ particles (Figure 3(b)) showed that carbon was evenly distributed on the composite particles.

(a)

(b)

Figure 2 

SEM images of LiFePO4: (a) without carbon; (b) with carbon source.

(a)

(b)

(c)

Figure 3 

Transmission electron micrograph showing (a) particle morphology and size distribution of C-LiFePO4 composite material, (b) C Elemental mapping of LiFePO4, and (c) selected area electro diffraction (SAED) pattern of C-LiFePO4.

Raman scattering is quite sensitive to changes in the local lattice distortions and change in polarizibility arising due to delithiation process in lithium-based rechargeable batteries. The C-LiFePO₄ is characterized by orthorhombic (triphylite) structure with the space group (Pmnb) [27]. The 34 Raman active modes can be classified as follows [27]: Γ=11𝐴_𝑔+7𝐵_1𝑔+11𝐵_2𝑔+7𝐵_3𝑔. The Raman scattering data of C-LiFePO₄ during the electrochemical cycling (half charge(HC), full charge(FC), half discharge(HC), and full discharge(FD)) were obtained by employing normal backscattering geometry and the results are presented in Figure 4(a). We observed intense Raman modes at 216, 278, 390, 441, 950, 992, 1045, and comparatively less intense Raman mode at 429 560, 626 cm⁻¹ in pristine C-LiFePO₄, which are comparable to the earlier reported Raman spectra of C-LiFePO₄ with the same orthorhombic symmetry [27–29]. The Raman mode at 441 cm⁻¹ is the bending mode involving O–P–O symmetry mode and disappeared in full charge Raman spectrum (i.e., believed to be pure FePO₄) during Li deintercalation suggest that it is highly sensitive to the local lithium environment. The Raman mode at 441 cm⁻¹ was not observed in FePO₄; hence it also reveals that the C-LiFePO₄ completely transformed into C-FePO₄ at full charging stage [29]. The Raman modes in the range of 900 to 1150 cm⁻¹ (see Figure 4(b)) are due to the stretching mode of PO₄³⁻ unit and involve symmetric and asymmetric of P–O bonds. These Raman modes show a small red and blue frequency shift during the charging and discharging process could be due to change in bond length and due to this Raman shift the A_g Raman mode at 950 cm⁻¹ appears nicely in full discharge spectrum as shown in inset of Figure 4(a). These Raman modes also show a systemic change in Raman intensity with electrochemical cycling process; that is, Raman intensity of all the PO₄³⁻ generated optical modes decreases during charging process and vice versa during discharging process it reveals that the polarizable derivatives of the C-LiFePO₄ change during Li deintercalation because the vibrational potential energy of the PO₄³⁻ is affected by change in Li/Li⁺ and Fe²⁺/Fe³⁺ ions during the Li deintercalation/intercalation. From the Raman results, structural stability and electrochemical reversibility of C-LiFePO₄ was observed clearly during the charging and discharging process.

(a)

(b)

Figure 4 

(a) Raman spectra of C- LiFePO4 powder at various stages of charge – discharge process (half charge(HC), full charge(FC), half discharge(HC) and full discharge(FD)) in frequency range 150 to 1200 cm−1. (b) Raman spectra of C- LiFePO4 powder in the frequency range 600 to 1800 cm−1 (inset presents resolved Raman spectra of the residual carbon using Gaussian distribution function).

In order to get further insight on the nature of carbon coating, micro-Raman scattering measurements were done on the pristine C-LiFePO₄ (calcined at 700^∘C) and the result is shown in (Figure 4(b)). The weak Raman modes (marked by small arrows) before 850 cm⁻¹ have been discussed previously. The two prominent modes at ~1345 and 1587 cm⁻¹ are the fingerprints of amorphous carbon [35–37]. The mode at ~1587 cm⁻¹ is assigned to sp² graphite like (G band) and the mode at ~1345 cm⁻¹ is assigned to sp³ type amorphous carbonaceous material (D band) [38]. It is apparent from the figure that the intensity ratio between the carbon and PO₄ bands is very high. The large intensity ratio may be due to uniform carbon coating on lithium iron phosphate particles. Interestingly, in a recent report, Nakamura et al. [38] have correlated the measured resistivity of LiFePO₄ with the integrated intensity ratio of carbon bands to PO₄ band. Indeed a marked drop in resistivity was reported with an increase in the intensity ratio of up to 300. To resolve the Raman spectra of the residual carbon, bands were deconvoluted using Gaussian distribution function. As shown in the inset of Figure 4(b), four Raman modes yielded satisfactory fit with minimum fitting error. The mode at 1244 cm⁻¹ may be related to short range vibrations of sp³ coordinated carbons. The mode at 1460 cm⁻¹ might have been arisen from Li₂CO₃ as Li₂CO₃ has Raman active modes at 700, 710, 1100, and 1480 cm⁻¹ [39]. From structural and morphological results it is apparent that carbon addition imparts carbothermal reduction of Fe³⁺ ions and thereby prevents the formation of undesirable iron (II, III) pyrophosphates or phosphate impurity phases. Also, adding carbon to the starting ingredients for the synthesis effectively retards the particle growth [40] and this processing ensured uniform carbon coating on lithium iron phosphate particles.

3.2. Electrochemical Properties

For the potential battery applications of the LiFePO₄ and C-LiFePO₄ cyclic voltammograms (CV) was used to evaluate the electrochemical performance. Figure 5 shows the cyclic voltammograms (CV) of LiFePO₄ and C-LiFePO₄. In both cases, the separation between cathodic (Li⁺ intercalation into the cathode) and anodic (Li⁺ deintercalation from the cathode) peaks and the peak height gradually reduced as the scan rate increased. Various researchers have also reported this kind of observation [41, 42]. If the electron transfer processes were “slow” (relative to the voltage scan rate), this kind of behavior is observed in the electrochemical systems, and low electronic conductivity of LiFePO₄ may be responsible for the slow electron transfer. It is seen from Figure 5(a) that the anodic/cathodic peaks of pure LiFePO₄ are located at ~3.9 V/3.1 V at the scan rate of 0.5 mVs⁻¹ and the 𝐼_𝑝 of the redox peaks is around 8×10⁻⁵A. The big separation between redox peaks (ΔV) of ~0.70 V indicates that the electrochemical behavior is controlled by the diffusion step. From Figure 5(b), it can be seen that the Ip of C-LiFePO₄ composite material increased to 5.5 × 10⁻⁴ A for the same scan rate. Meanwhile, the ΔV between redox peaks were reduced 0.5 V in composite C-LiFePO₄ under the same scan rate. The high current level in composite cathode compared to pure LiFePO₄ for the same scan rate is presumably due to higher electronic conductivity as the result of carbon coating, which might be advantageous for obtaining higher capacity at high C-rate (will be discussed later). Also, for C-LiFePO₄ composite cathode, well-developed CV loop confirms that the kinetics of lithium intercalation and deintercalation is markedly improved by the amorphous carbon coating compared to pure LiFePO₄.

(a)

(b)

Figure 5 

The CV profile of the different sample: (a) without carbon and (b) with carbon at the scan rate of 0.5 mVs−1.

Now it will be interesting to compare the chemical diffusion coefficient of Li into the electrode material as this mainly determines achievable capacity and rapid diffusion of ions, which is of practical importance for fast storage/drainage of energy. The electrochemical methods, like, Impedance Spectroscopy [43, 44], Galvanostatic Intermittent Titration Technique (GITT) [45], and cyclic voltammetry (CV) [46], are widely used to measure the chemical diffusion coefficients. However, calculation of the diffusion coefficient of Li-ion from CV [47–50] is more popular as this technique is straightforward and relatively uncomplicated. In this method, the voltammetric peaks have been used to calculate the chemical diffusion coefficient of Li⁺ as described by the Randles–Sevcik equation (for semi-infinite diffusion): 𝑖_𝑝=2.69×10⁵𝑛^3/2𝐶^𝑏₀𝐴𝐷^1/2_𝐿𝑖𝜈^1/2, (1) where 𝑖_𝑝 the peak current value, 𝑛 is the number of electrons involved in the reaction of the redox couple (for Li¹⁺ it is 1), 𝐶^𝑏₀ is the concentration (0.0228 mol/cm³ in the present case), A is the effective working electrode area (0.423 cm² in the present case), 𝜈 is the rate at which the potential is swept (V/s), and 𝐷_Li is the diffusion coefficient (cm²/s) of Li⁺. In (1), peak current (𝑖_𝑝) is proportional to 𝜐^1/2. Inset of Figures 5(a) and 5(b) shows the variation 𝑖_𝑝 with square root of scan rate, and the observed linearity is consistent with the semi-infinite diffusion-controlled behavior for the range of scan-rate used [48]. From the slope of the linear fit, the calculated Li-ion chemical diffusion coefficients were 1.28 ×10⁻¹⁵ cm²s⁻¹ and 7.13 × 10⁻¹⁴ cm²s⁻¹, respectively, for pure LiFePO₄ and C-LiFePO₄ cathodes. The order of diffusion coefficient matches well with the reported values [19, 21]. Again, the diffusion coefficient of lithium ion increases markedly after mixing with carbon and hence high capacity is expected for the composite cathode under the same charge/discharge conditions. Similar trend was also reported recently [20]. The lower 𝐷_Li in the case of untreated LiFePO₄ may be due to the presence of impurity phases, as diffusion of Li⁺ will be hindered near the region where two phases coexisted due to the phase boundary movement [21].

Coin cells using LiFePO₄ and C-LiFePO₄ cathode and lithium anode were galvanostatically charged and discharged between 2.3 and 4.3 V at room temperature. Figure 6(a) exhibits charge and discharge profiles for 1st and 15th cycles for pure LiFePO₄ at a rate of C/5. The electrode material delivers a first cycle charge and discharge capacity of 70 and 63 mAhg⁻¹, respectively, and has a short plateau near 3.4 V. The discharge capacity reduced to 42 mAhg⁻¹ after 25 cycles (inset of Figure 6(a)), which corresponds to 69% capacity retention. The rate capability of pure LiFePO₄ was also studied and the results are shown in Figure 6(b). The obtained discharge capacities were ~63, 43, and 13 mAh/g for C/5, C/3, and C/2 rates, respectively. The poor charge-discharge characteristics of untreated LiFePO₄ are due to the low electronic conductivity and hindrances to the Li-diffusion into the cathode (lower 𝐷_Li) as mentioned earlier. Figure 7(a) shows charge and discharge profiles for 1st and 50th cycles for C-LiFePO₄ at a rate of C/5. The electrode delivers a discharge capacity of ~163 mAhg⁻¹ in the first cycle, which is significantly higher than the pure LiFePO₄. Recent literature reports (see Table 1) reveal that C-LiFePO₄ material can not only supply large capacity under high-rate but also excellent capacity retention. In the present case, at a rate of C/5, (inset of Figure 7(a)) capacity retention up to ~97% is achieved after 50 charge discharge cycles. The capacity retention is one of the best as compared to C-LiFePO₄ composite cathodes reported by the others. The C-LiFePO₄ cathodes were also characterized in terms of their rate capability and the results are shown in Figure 7(b) for C/5, C/3, C/2, and 1C rates. The obtained discharge capacities were 163, 144, 118, and 70 mAh/g, respectively, for C/5, C/3, C/2 and 1C rates. In the case of carbon-coated LiFePO₄ decreasing D/G intensity ratio is related to the carbon disorder and the width of the Raman line is related to the degree of carbon disorder. In the present case D/G intensity ration was higher than that reported by Julien et al. [35], and hence degree of carbon disorder (amorphous nature) is lower and conductivity may be lower due to this effect. This can be the reason for the reduced capacity at high C-rate; in the present case it may be due to the slow diffusion coefficient of Li ion in LiFePO₄ cathode material and lower conductivity of the conductive carbon coating.

(a)

(b)

Figure 6 

(a) Charge/discharge profile of LiFePO4 sample without carbon source and inset discharge up to 25 cycles; (b) discharge profile of LiFePO4 sample with carbon source at discharge rate of C/5, C/3, and C/2.

(a)

(b)

Figure 7 

(a) Charge / discharge profile of LiFePO4 sample with carbon source and inset discharge up to 50 cycles; (b) discharge profile of LiFePO4 sample with carbon source at discharge rate of C/5, C/3, C/2, and 1C.

Thus, carbon-treated LiFePO₄ proved to be far better compared to untreated LiFePO₄ as it enhanced lithium-ion transport and electronic conductivity. However, the reduction in the discharge capacity (our case) with increase in the C-rate is a concern and mainly depends on the electronic conductivity and Li-ion diffusion. Solid-state diffusion often limits the utilization and rate capability of electrode materials in a lithium-ion battery, especially at high charge/discharge rates. When the fluxes of Li⁺ insertion or extraction exceed the diffusion-limited rate of Li⁺ transport within the bulk phase of an electrode, concentration polarization occurs. Further, large volume changes associated with Li⁺ insertion or extraction could induce stresses in bulk electrodes, potentially leading to mechanical failure. Reducing the particle size of electrodes material ensures high surface to-volume ratio, which would increase the electrochemical reaction surface and suppress the mechanical stress. To get an idea regarding the particle size effect on the rate capability, the relation between diffusion time constant (τ) and characteristic diffusion length (𝐿_Li,max) described by Levi et al. in a recent report [51] can be used: 𝐿_Li,max=𝜏𝐷_Li. (2)

Therefore, the time for intercalation varies as square of the length scale and is faster for smaller particles. In this regard, optimal Li insertion could be achieved when (decided by chosen current density of discharge) 𝐿_Li,max is comparable or larger than the particle size of the cathode material. Alternately, if the τ is higher than charging or discharging time, one can obtain maximum limit of capacity for a given temperature. Hence (2) can be used as a measure to design the particle size for obtaining maximum capacity at high C-rate (fast charging and discharging) for materials with low ionic diffusion as in the case of LiFePO₄. As mentioned earlier, the average diffusion coefficient estimated from the CV data is ~7.13 × 10⁻¹⁴ cm²s⁻¹. As a rough estimation, by taking this value of 𝐷_Li and the discharge time we have calculated 𝐿_Li,max. The characteristic diffusion length decreased with C-rate (see Figure 8) and the values of 𝐷_Li were 160 nm, 226 nm, 277 nm, and 357 nm for 1C, C/2, C/3, and C/5 rate, respectively. It can be seen that for smaller diffusion time (faster charging/discharging) the dependence on the 𝐿_Li,max is stronger. The estimated values are indicative of obtaining better electrochemical performance at higher current rate, with reduced particle size. As envisaged from SEM and TEM analyses discussed before, the average particle size of the C-LiFePO₄ is ~200 nm and hence higher capacity is expected for lower current rate (≤C/2). However in our case, higher capacity was not achieved except for C/5 and this can be attributed to the agglomeration of particles as seen from the TEM results. Hence, by avoiding the agglomeration of the particle and by reducing the particle size, large fraction of cathode can be made accessible for Li-ion (de)intercalation. In short, cathodes with particles in the nanometer scale shorten the Li⁺ diffusion distances and minimize the tortuous transport; hence, less time is needed to achieve full charge or discharge at the same current density.

Figure 8 

Variation of Characteristic diffusion length with C-rate.

4. Conclusions

Pure LiFePO₄ and C-LiFePO₄ composite cathode materials were prepared by solid-state route. Carbon as an additive reduces particle growth and retards the Fe²⁺ to Fe³⁺ oxidation and thereby eliminates the formation of impurity phase during high-temperature calcination. TEM results indicate carbon coating on the surface of LiFePO₄ and micro-Raman analysis suggests carbon in the amorphous form. The structural stability and reversibility upon electrochemical cycling is verified with ex situ Raman scattering. In case of C-LiFePO₄, the Li ion diffusion coefficient was ~7.13 × 10⁻¹⁴  cm²s⁻¹ where as in the case pure LiFePO₄ it was merely ~1.28 × 10¹⁵ cm² s⁻¹, and hence, the Li ion (de)intercalation was better in C-LiFePO₄. Excellent cycleability (~97% retention after 50 cycles) was attained for C-LiFePO₄ compared to pure LiFePO₄ (only 69% retention after 25 cycles). As the C-rate is increased, the discharge capacities of LiFePO₄ and C-LiFePO₄ cathodes are reduced. The poor discharge capacity at high C-rate is thought to be due to lower Li-ion diffusion and hence the rate capability can be improved by reducing the particle size and by avoiding agglomeration in C-LiFePO₄. This comparative study reveals the necessity of conductive coating along with particle size reduction for improving the electrochemical properties of LiFePO₄ at high C-rate.

Acknowledgments

The financial support from NASA-EPSCoR (NNX08AB12A) and NASA-URC (NNX08BA48A) Grants is gratefully acknowledged. Continual support from UPR materials characterization center (MCC) is also acknowledged. One of the authors (SBM) thanks Council of Scientific and Industrial Research, Government of India for partial financial support for this publication.