Enhanced fast charge–discharge performance of Li₄Ti₅O₁₂ as anode materials for lithium-ion batteries by Ce and CeO₂ modification using a facile method

Abstract

A facile solid-state method to improve the fast charge–discharge and kinetic performance of Li₄Ti₅O₁₂ in lithium-ion batteries by Ce and CeO₂ in situ modification is presented in this work. XRD shows that the Ce doping and CeO₂ modification do not change the spinel structure of Li₄Ti₅O₁₂. Little Ce doping (Ti/Ce = 4.9 : 0.1 and Ti/Ce = 4.85 : 0.15) reduces the lattice parameter of doped Li₄Ti₅O₁₂, but more Ce⁴⁺ doping (Ti/Ce = 4.8 : 0.2) increases the lattice parameter due to the large ionic radius of Ce⁴⁺. Raman spectra reveal that CeO₂ is not completely incorporated into the host structure and leads to the formation of a uniform coating on the surface of Li₄Ti₅O₁₂. The doping of Ce⁴⁺ and the combination with in situ generated CeO₂ in Li₄Ti₅O₁₂ are favorable for reducing the electrode polarization and charge-transfer resistance and improve the lithium insertion/extraction kinetics of Li₄Ti₅O₁₂, resulting in its relatively higher capacity at a high charge–discharge rate. The Ce-doped Li₄Ti₅O₁₂–CeO₂ composites show a much improved rate capability and cycling stability compared with pristine Li₄Ti₅O₁₂ at a 10 C charge–discharge rate in a broad voltage window (0–2.5 V). The introduction of Ce and CeO₂ enhances not only the electric conductivity of Li₄Ti₅O₁₂, but also the lithium ion diffusivity in Li₄Ti₅O₁₂, resulting in a significantly improved high-rate capability, cycling stability, and fast charge–discharge performance of Li₄Ti₅O₁₂.

---

1. Introduction

Lithium-ion batteries have been considered as one of the most promising power batteries for electric vehicles (EVs) and hybrid electric vehicles (HEVs) due to their high energy density. However, safety is the essential criterion to evaluate if an LIB can be applied to these vehicles.¹ At fast discharge rates, the anode potential drops to such a degree that dendritic lithium may deposit over the anode surface, which then causes safety issues. However, the batteries used in EVs and HEVs are required to be charged–discharged continually at high rates, which then causes serious safety issues. Spinel lithium titanate (Li₄Ti₅O₁₂) has received considerable attention as a potential candidate for lithium-ion batteries due to its high lithium ion insertion potential at around 1.55 V (vs. Li/Li⁺), and excellent structural and thermodynamic stability during cycling.² These indicate that Li₄Ti₅O₁₂ used as an anode in high power lithium ion batteries is safe, long-living and reliable. Unfortunately, the low electronic conductivity (ca. 10⁻¹³ S cm⁻¹) and lithium diffusion coefficient (ca. 10⁻⁹ to 10⁻¹³ cm² s⁻¹) result in the poor performance of Li₄Ti₅O₁₂ materials, preventing them from being widely used.³ In order to overcome the kinetic problems of Li₄Ti₅O₁₂, several methods have been developed. These include the preparation of nano-Li₄Ti₅O₁₂ particles,^4,5 doping Li₄Ti₅O₁₂ with metal and non-metal ions (Na⁺,⁶ Mg²⁺,⁷ Zn²⁺,⁸ Ca²⁺,⁹ Al³⁺,¹⁰ La³⁺,¹¹ Zr⁴⁺,¹² Ru⁴⁺,¹³ V⁵⁺,¹⁴ Ta⁵⁺,¹⁵ W⁶⁺,¹⁶ F⁻,¹⁷ and Br⁻ (ref. 18)) in the Li, Ti or O sites, and surface modification via coating with a conductive second phase including Ag,¹⁹ Au,²⁰ carbon,²¹ graphene,²² polyacene (PAS),²³ TiN,²⁴ ZrO₂,²⁵ and TiO₂ (ref. 26) etc. However, nanomaterial or carbon coating frequently results in a low volumetric energy density of the cell. Surface coating with a conductive second phase can increase the electronic conductivity of Li₄Ti₅O₁₂, but the ionic conductivity is difficult to improve. It has been reported that ceria (CeO₂) can produce a good electrical contact between oxides which facilitates the electron transfer between CeO₂ and the supported metal oxide.²⁷ CeO₂ with good electrical conductivity has already been used as a coating material for improving the electrochemical performance of cathode materials, such as in LiMn₂O₄,²⁸ LiFePO₄,²⁹ LiCo_1/3Ni_1/3Mn_1/3O₂,³⁰ and Li-rich electrodes³¹ due to the direct and fast transformation between Ce(III) and Ce(IV). Yang et al.³² reported the electrochemical performance of CeO₂-coated Li₄Ti₅O₁₂ between 1 and 3 V synthesized by a one-pot co-precipitation method using tetrabutyl titanate as the raw material. However, this is difficult for commercial applications because of its relatively high synthetic cost. From a commercial viewpoint, the solid-state synthesis of Li₄Ti₅O₁₂ powders exhibits a potential commercial application due to the simple synthesis route and low synthesis cost. In addition, a uniform surface coating around the whole Li₄Ti₅O₁₂ particles is also obviously difficult to achieve. Compared with traditional structures, this unique inlaid architecture of Ce-doped Li₄Ti₅O₁₂–CeO₂ can enhance the lithium ion diffusion ability and conductivity of composites, and inherits the advantages of CeO₂ coating and doping. It is well known that a high energy density of a battery results from a high voltage or high capacity. According to a previous report,³³ Li₄Ti₅O₁₂ can be lithiated to the state Li_8.5Ti₅O₁₂ (discharged to 0 V), and can then provide a theoretical capacity of 260 mA h g⁻¹ by means of first-principles calculations, which is about 1.5 times higher than that of the compound lithiated to Li₇Ti₅O₁₂. In addition, considering the safety because of the risk of explosion from a possible inner short circuit of the lithium-ion battery, it is important to study the over-discharge behaviors of anode materials at a lower voltage.³⁴ To the best of our knowledge, however, there have been no reports on the fast charge–discharge performance of Ce-doped Li₄Ti₅O₁₂–CeO₂ composites in a broad voltage window. In this paper, Ce-doped Li₄Ti₅O₁₂–CeO₂ composites as anode materials were synthesized by a facile solid-state reaction, and the composites show significantly improved fast charge–discharged performances.

2. Experimental

2.1 Materials preparation

Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites were synthesized by a solid-state method. Li₂CO₃, anatase-phase TiO₂ and Ce(NO₃)₃·6H₂O were used as the starting materials at a Li : (Ti + Ce) molar ratio of 4.3 : 5. Excessive Li (5 wt%) was provided to compensate for the volatilization of Li during the synthesis. The molar ratios of Ti : Ce are 4.9 : 0.1, 4.85 : 0.15, and 4.8 : 0.2, respectively. The overall fabrication procedures of all samples are schematically illustrated in Fig. 1. All raw materials are mixed and ball-milled. The high-energy ball-milling processes were carried out at a constant speed of 400 rpm and a processing time of 3 h. Following the ball-milling process, the remaining alcohol medium was evaporated and the obtained powders were subsequently dried at 120 °C for 8 hours in an air atmosphere. The mixed raw materials were calcined at 450 °C for 6 h in a flowing air atmosphere to obtain the precursor. Then, the precursor was ball-milled with alcohol as the medium for about 3 hours. Finally, the precursors were calcined at 850 °C for 20 h in a flowing air atmosphere to obtain Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites. After cooling, all samples were ball-milled with alcohol as the medium for about 30 minutes to form the final products.

Fig. 1 Fabrication procedures of Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples in this study.

2.2 Battery preparation

A CR2025 coin-cell assembly was used for the electrochemical characterization. A slurry was formed by mixing the active material (80%), super P conductive carbon (10%), and the binder (10 wt% polyvinylidene fluoride, dissolved in N-methyl-2-pyrrolidone). After coating onto the Cu foil, the film was dried in a vacuum oven at 110 °C for 10 h and then cut into discs with a radius of 7 mm. The half-cells were assembled with the active material and a metallic lithium anode separated by a porous polypropylene film (Celgard 2300) filled with a 1 M LiPF₆ in ethylene carbonate/dimethyl carbonate (1 : 1 by vol) solution.

2.3 Materials characterization and electrochemical tests

X-ray diffractometry (XRD) measurements were performed on a Rigaku instrument with Cu Kα radiation. Particle sizes, morphologies and microstructures were examined using scanning electron microscopy (SEM, SU8000). Raman measurements were performed on a SPEX-1403 Raman spectrometer. The laser light source was the 488 nm line of an Art laser excited at 400 mW. Cyclic voltammetry (CV) tests were carried out on a CHI 1000C electrochemical workstation with a voltage between 0 and 2.5 V at a scanning rate of 0.2 mV s⁻¹. Electrochemical impedance spectroscopy (EIS) is measured by a Princeton P4000 electrochemical working station over a frequency range from 0.01 Hz to 10 kHz at a potentiostatic signal amplitude of 5 mV. The charge–discharge measurements were recorded on a multichannel Land Battery Test System (Wuhan Jinnuo, China) at room temperature in the 0.0–2.5 V (vs. Li/Li⁺) range carried out at different charge–discharge rates.

3. Results and discussion

Fig. 2 shows X-ray diffraction patterns of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites. No impurities were observed for pristine Li₄Ti₅O₁₂. All the sharp diffraction peaks can be attributed to the cubic spinel structure of Li₄Ti₅O₁₂. This indicates that the Ce doping and CeO₂ modification do not change the spinel structure of Li₄Ti₅O₁₂. However, a few impurity peaks can be observed in the XRD patterns of modified Li₄Ti₅O₁₂, as shown in Fig. 2(b–d), and these can be ascribed to CeO₂. In addition, with the increase in CeO₂ additive, the peak intensity of CeO₂ is enhanced correspondingly. Interestingly, there is a slight shift of the diffraction to higher diffraction angles for the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites with Ti/Ce = 4.9 : 0.1 and Ti/Ce = 4.85 : 0.15, revealing that a few cerium ions can substitute titanium ions in the octahedron 16d site, and then reduce the lattice parameter of doped Li₄Ti₅O₁₂. The reason may be that the bond strength of Ce–O is larger than that of Ti–O, so it has higher octahedral site preference energies (OPE) which results in the diminution of the bond length. However, with the increase in CeO₂ additive, a slight shift of the diffraction to lower diffraction angles can be found for the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites with Ti/Ce = 4.8 : 0.2. The reason may be due to the difference in the ionic radius between Ce⁴⁺ (0.87 Å)³⁵ and Ti⁴⁺(0.605 Å), indicating that more cerium ions enter the crystal lattice. In this work, the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites are prepared at 850 °C for 20 h in a flowing air atmosphere. Ce(NO₃)₃·6H₂O can be decomposed into CeO₂ at such a high temperature (850 °C) and long sintering time (20 h) in an air atmosphere. The decomposition process is described in eqn (1) and (2):

Fig. 2 XRD patterns of the as-prepared of Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Hence, it can be confirmed that the main valence state of Ce is +4 in the Ti-based and Ce-based oxides. However, it is well known that CeO₂ is a good oxygen storage material based on a reversible redox reaction as follows:

CeO₂ ⇌ CeO_2−x + 0.5xO₂ (3)

Hence, it can be confirmed that the valence states of Ce are +4 and +3 in the Ti-based and Ce-based oxides. This phenomenon may be compared to a CeO₂-coated Li(Li_0.17Ni_0.2Co_0.05Mn_0.58)O₂ cathode.³¹

There are three possible doping sites in a Ce-doped Li₄Ti₅O₁₂ crystal (Li site, Ti site, or both Li and Ti sites). The determination of the locations of the Li⁺ ions in transition metal oxides is usually very difficult to study by XRD alone because the atomic scattering factor of the lithium ion is very small.³⁶ However, the location of the lithium ions at the tetrahedral (8a) or octahedral (16c) sites reflects upon the XRD intensities in the pattern especially for the (3 1 1) and (4 0 0) lines.³⁷ As we know, the structural cation distribution of Li₄Ti₅O₁₂ can be shown as [Li₃]_8a[LiTi₄]_16d[O₁₂]_32e. The Li ions sit in the tetrahedral and octahedral sites, and Ti only takes up octahedral sites. According to Ohzuku et al.,³⁸ the integrated intensity ratios of the (3 1 1)/(4 0 0) and (2 2 0)/(3 1 1) peaks are indices of the extent of occupancy of the substituent ions in the 8a lithium sites. The (220) (at 2θ ≈ 30°) peaks cannot be found in the XRD patterns of Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites, indicating that there are no heavy cations (Ti⁴⁺ or Ce⁴⁺) in the tetrahedral 8a sites of the spinel-type structure.³⁹ The relative intensity ratios between the (3 1 1) and (4 0 0) peaks of Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites with Ti/Ce = 4.9 : 0.1, Ti/Ce = 4.85 : 0.15 and Ti/Ce = 4.8 : 0.2 are 0.674, 0.720, 0.713 and 0.656, respectively. This indicates that Ce doping leads to a confused degree of ion locations, and the confused degree of ion locations decreases with increasing Ce content. The increased confused degree means that the locations of the lithium ions are changed because of the doped Ce, indicating that Li and Ti of the 16d sites may be substituted by Ce. This means that the partial charge compensation can be accomplished by a Li vacancy. Hence, it can be expected that the electronic conductivity of Li₄Ti₅O₁₂ can be improved by Ce doping. The analysis described above indicates that the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites can be successfully prepared by the solid-state method.

Fig. 3 shows the Raman spectra of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites. As shown in Fig. 3, it can be found that there are five main vibration peaks at about 238, 347, 427, 670 and 755 cm⁻¹ representing the features of the spinel structure (A1g + Eg + 3F2u).⁴⁰ No obvious differences of the Raman signals between pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ powders were found. However, the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites show low Raman signals. The reason may be that there is a lower homogeneity in the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites because the low dose of Ce⁴⁺ ion doping may result in an increased confused degree of ion locations during the long time and high temperature sintering. The two higher frequency bands (670 and 755 cm⁻¹) can be assigned to the vibrations of the Ti–O bonds in TiO₆ octahedra. The middle frequency bands (350 and 426 cm⁻¹) can be assigned to the stretching vibrations of the Li–O bonds in LiO₄ and LiO₆ polyhedra, respectively. The lower frequency bands (238 cm⁻¹) are assigned to the bending vibrations of the O–Ti–O bonds.⁴¹ In addition, a weak feature at 267 cm⁻¹ is assigned to the F2g modes in Fig. 3, which is well consistent with the results reported in the literature.^42,43 As shown in Fig. 4, the CeO₂ crystal possesses a cubic structure. There exists a symmetrical stretching mode of the Ce–O vibrational unit in CeO₂. From Fig. 3, a sharp and symmetric peak at 465 cm⁻¹ can be found, and it is associated with the F2g Raman active mode of the polycrystalline cubic structure of CeO₂.⁴⁴ These are the Raman characteristics of the spinel structure of Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂, which are consistent with the XRD characterization mentioned above.

Fig. 3 Raman spectra of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Fig. 4 (a) Side view and (b) top view of the CeO₂ bulk structure.

The scanning electron microscopy (SEM) images of Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ materials are shown in Fig. 5. As shown in Fig. 5, it is apparent that the morphologies of all samples are similar, and all samples show small and uniform particle dimensions in the range of 1–2 μm. Hence, it can be concluded that the effect of particle size on the electrochemical properties of Li₄Ti₅O₁₂ can be ruled out, and the dissimilarity of performance can be mainly attributed to the Ce and CeO₂ modification. The surface morphology of Li₄Ti₅O₁₂ is highly smooth. However, some very small particles can be found to be highly dispersed on the Li₄Ti₅O₁₂ particles, as shown in Fig. 5(b–d). The denser the small particles accumulate, the higher the coated CeO₂ content. From a comparison of these four powder surface morphologies, it can be speculated that the surface of the prepared Li₄Ti₅O₁₂ is covered with small CeO₂ particles. This indicates that the surface modification leads to the formation of a uniform coating.

Fig. 5 SEM images of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Typical steady-state cyclic voltammetry (CV) plots of Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites obtained at a slow scan rate of 0.2 mV s⁻¹ are presented in Fig. 6. A pair of redox peaks of all the samples appears at about 1.70/1.47 V, which can be attributed to the phase transition between spinel (Li₄Ti₅O₁₂) and rock-salt (Li₇Ti₅O₁₂). The pair of redox peaks below 0.6 V exists also due to further reduction of Ti⁴⁺ to Ti³⁺, indicating that the electrochemical lithium extraction/insertion reaction in these composite electrodes takes place in at least two stages during the discharge. The broad peak with envelope feature below 0.5 V indicates the formation of an amorphous phase. Different electrochemical polarization degrees can be found in different samples even at such a low scanning rate (0.2 mV s⁻¹). The potential difference of the CV peaks can reflect the polarization degree of the electrodes. Table 1 gives the potential differences between the anodic and cathodic peaks of all samples. It is obvious that the potential differences of the Ce-doped Li₄Ti₅O₁₂–CeO₂ electrodes are lower than that of pristine Li₄Ti₅O₁₂. The narrow potential differences of the Ce-doped Li₄Ti₅O₁₂–CeO₂ electrodes indicate the reduced polarization and enhanced electrochemical kinetics.²⁵ The low potential interval demonstrates that the lithium insertion into the Ce-doped Li₄Ti₅O₁₂–CeO₂ electrodes can be treated as a quasi-reversible system with improved dynamic behaviors. Among all samples, the Ce-doped Li₄Ti₅O₁₂–CeO₂ composite with Ti/Ce = 4.9 : 0.1 exhibits the smallest potential difference, indicating the lowest polarization and excellent kinetics.

Fig. 6 Cyclic voltammetry (CV) plots of Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Table 1 Potential differences between the anodic and cathodic peaks and lithium ion diffusion coefficients (D_Li) of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce

Samples with different molar ratios	φpa (V)	φpc (V)	Δφp (mV)	DLi (cm^2 s^−1)
Ti/Ce = 5 : 0	1.816	1.359	457	3.03 × 10^−16
Ti/Ce = 4.9 : 0.1	1.722	1.472	250	6.78 × 10^−16
Ti/Ce = 4.85 : 0.15	1.728	1.453	275	8.68 × 10^−16
Ti/Ce = 4.8 : 0.2	1.722	1.366	356	4.11 × 10^−15

---

To clarify the effects of the Ce and CeO₂ modification on the kinetic properties of Li₄Ti₅O₁₂, EIS (Electrochemical impedance spectroscopy) were further measured as shown in Fig. 7. Models used to fit the EIS of Li₄Ti₅O₁₂ have also been reported in many references, such as Li₄Ti₅O₁₂,⁴⁵ Au-coated Li₄Ti₅O₁₂,⁴⁶ and TiO₂-coated Li₄Ti₅O₁₂.⁴⁷ The inset shows the selected equivalent circuit used to fit the EIS and the enlarged Nyquist plots of all anode materials. In the Niquist plot, the high frequency intercept at the real axis corresponding to the ohmic resistance of the cell is caused by the electrolyte. The capacitive loop is mainly contributed by the charge transfer resistance (R_ct) at the electrode–electrolyte interfaces. The linear Warburg (Z_W) region in the low-frequency region is associated with the lithium-ion diffusion process in Li₄Ti₅O₁₂.^45–47 From the enlarged Nyquist plots, it can be seen that the pristine Li₄Ti₅O₁₂ electrode exhibits a higher charge transfer resistance than those of the Ce-doped Li₄Ti₅O₁₂–CeO₂ composite electrodes. This indicates that the Ce and CeO₂ modification can enhance the conductivity of the active materials, and then decrease the charge transfer resistance at the electrode–electrolyte interface, suggesting that the Ce-doped Li₄Ti₅O₁₂–CeO₂ composites may have a higher electrochemical activity than that of pristine Li₄Ti₅O₁₂ during cycling. The reason may be that CeO₂ with a high electrical conductivity enhances the total electrical conductivities of the Ce-doped Li₄Ti₅O₁₂–CeO₂ composite electrodes. The lithium ion diffusion coefficient (D_Li) can be calculated from the plots in the low-frequency region according to the following eqn (4) and (5) (similar approaches have also been applied to various electrode materials like LiFePO₄,⁴⁸ LiNi_0.5Mn_1.5O₄,⁴⁹ LiMn₂SiO₄,⁵⁰ and Li₄Ti₅O₁₂⁵¹ electrode materials):^48–51

where T is the absolute temperature, R is the gas constant, A is the surface area of the electrode, F is the Faraday constant, n is the number of electrons per molecule during oxidation, C_Li is the concentration of lithium ions, and σ is the Warburg factor which is related to Z_re obtained from the slope of the lines in Fig. 8.

Fig. 7 Electrochemical impedance spectra of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2. Insets: Equivalent circuit and enlarged Niquist plots.

Fig. 8 Graph of Z_re plotted against ω^−1/2 in the low-frequency region for the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

The D_Li parameters of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ electrodes obtained from the EIS are recorded in Table 1. The calculated D_Li values of the samples are 3.03 × 10⁻¹⁶, 6.78 × 10⁻¹⁶, 8.68 × 10⁻¹⁶ and 4.11 × 10⁻¹⁵ cm² s⁻¹. It can be found that the lithium diffusion coefficient values of the Ce-doped Li₄Ti₅O₁₂–CeO₂ electrodes are higher than that of pristine Li₄Ti₅O₁₂. The enhancement of the diffusion coefficient may be mainly attributed to the Ce and CeO₂ modification. The substitution of Ti⁴⁺ by Ce⁴⁺ can increase the proportion between Ti³⁺ and Ti⁴⁺, and then improve the lithium diffusion coefficient.

Fig. 9 shows the initial charge–discharge profiles of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ anode materials between 0.0–2.5 V at an 0.1 C rate. All discharge profiles present two flat plateaus near 1.55 and 0.7 V, and an inclined curve from 0.6 to 0.0 V, indicating a new electrochemical insertion processes below 1.0 V. It may be a two-phase reaction between Li₇Ti₅O₁₂ and Li_8.5Ti₅O₁₂. All samples deliver specific capacities of 280–300 mA h g⁻¹ at an 0.1 C rate, and exceed the theoretical capacity (260 mA h g⁻¹) of Li₄Ti₅O₁₂.³³ The extra capacity is from the formation of a solid electrolyte interface (SEI) film and the discharge of carbon black.⁵² As we know, the insertion/extraction of lithium ions is rather sufficient even in bulk materials during the low charge–discharge current density.⁵³ However, the potential difference (ΔE) between the charge and discharge plateaus of all samples is different, as shown in Fig. 10, suggesting different polarization for all electrodes. It can be found that the Ce-doped Li₄Ti₅O₁₂–CeO₂ anodes show a smaller potential difference than that of pristine Li₄Ti₅O₁₂. In addition, the Ce-doped Li₄Ti₅O₁₂–CeO₂ anodes show a higher plateau voltage than that of pristine Li₄Ti₅O₁₂, suggesting increased energy densities for the modified electrodes. Fig. 11 shows the differential capacity curves of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ samples in the first cycle between 0 and 2.5 V. It can be found that all the peaks exhibit similar shapes. The peak heights decrease sequentially in the order of pristine Li₄Ti₅O₁₂ and the Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with increasing Ce content, indicating a sequential decrease in initial capacity. The reason may be due to the decrease of active Ti⁴⁺. However, the potential differences between the anode and cathode peaks also decrease in the same order. This reveals that the Ce-doped Li₄Ti₅O₁₂–CeO₂ samples possess a lower polarization of the charge transfer reaction and higher diffusivity of lithium ions inside the electrode than those of pristine Li₄Ti₅O₁₂.

Fig. 9 Initial galvanostatic charge–discharge curves of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce at a rate of 0.1 C: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Fig. 10 Enlarged potential plateaus of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples from Fig. 9 with different molar ratios between Ti and Ce at a rate of 0.1 C: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Fig. 11 Differential capacity vs. voltage plots of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce at a rate of 0.1 C: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Cycling performance curves of all samples at a charge–discharge rate of 0.1 C are given in Fig. 12. Compared with the initial discharge capacity, all samples show a big irreversible capacity loss for the second cycle. This can be ascribed to the formation of a solid electrolyte interface (SEI) film below 0.7 V (vs. Li/Li⁺), and irreversible reactions at the surface with e.g. organic lithium alkylcarbonates. It is well known that it is difficult to obtain the theoretical capacity of a bulk or micrometer sized particles. All electrode materials synthesized by the solid-state method have a lower homogeneity than that of materials prepared by a soft chemical method.⁵⁴ However, the Ce-doped Li₄Ti₅O₁₂–CeO₂ samples also show high discharge capacities. This indicates that moderate Ce and CeO₂ modification can improve the discharge capacity of Li₄Ti₅O₁₂ at a low rate.

Fig. 12 Cycling performances of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce at a rate of 0.1 C: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15 and (d) Ti/Ce = 4.8 : 0.2.

As we know, a high rate performance is important for applications in which a fast charge and discharge are needed, such as in HEV and PHEV applications. Simultaneously, a high cyclability with a fast charge–discharge process contributes to the commercial success of lithium-ion batteries as power sources. It has been reported that fast charging often induces severe underpotential deposition, resulting in hazardous Li plating and subsequent deterioration of the cell performance.⁵³ To provide more information about the electrochemical performance difference of Ce- and CeO₂-modified Li₄Ti₅O₁₂, fast charge–discharge measurements carried out at a 10 C charge–discharge rate, are presented in Fig. 13. Compared with the discharge capacity shown in Fig. 12, the discharge capacity of Li₄Ti₅O₁₂ decreases quickly with increasing discharge rate due to the large polarization at the high rate, whereas those of Ce- and CeO₂-modified Li₄Ti₅O₁₂ decrease slowly at the same rate. Although the synthesis method of the Ce- and CeO₂-modified Li₄Ti₅O₁₂ samples is a facile solid-state method, they show high discharge capacities at a 10 C charge–discharge rate. After 50 cycles, the discharge capacities of Ce- and CeO₂-modified Li₄Ti₅O₁₂ with Ti/Ce = 4.9 : 0.1, Ti/Ce = 4.85 : 0.15, and Ti/Ce = 4.8 : 0.2 are 91, 161 and 113 mA h g⁻¹, respectively, and it can still retain 100% of the initial capacity, exhibiting an excellent cycling stability at a high charge–discharge rate. However, pristine Li₄Ti₅O₁₂ only delivers a capacity of 32 mA h g⁻¹ after 50 cycles, revealing a poor rate capability at a high charge–discharge rate. The higher discharge capacity of the cell based on the Ce- and CeO₂-modified Li₄Ti₅O₁₂ anode can be attributed to the promotion of charge transfer reactions at the electrode–electrolyte interface, decreasing the electrode polarization and improving the lithium diffusion ability.

Fig. 13 Cycling performance of the as-prepared Li₄Ti₅O₁₂ and Ce-doped Li₄Ti₅O₁₂–CeO₂ samples with different molar ratios between Ti and Ce at a 10 C charge–discharged rate: (a) Ti/Ce = 5 : 0, (b) Ti/Ce = 4.9 : 0.1, (c) Ti/Ce = 4.85 : 0.15, and (d) Ti/Ce = 4.8 : 0.2.

Fig. 14(a) gives the structural transformation model of Li₄Ti₅O₁₂ during the charge–discharge process. In Li₄Ti₅O₁₂, oxygen and lithium occupy the 32e and 8a sites, respectively, while the 16d sites are occupied by titanium and lithium with a ratio of 5 : 1. During discharge of the Li₄Ti₅O₁₂/Li cell, the lithium atoms at the 8a sites move to the 16c sites, and the inserted Li⁺ ions also occupy 16c sites via 8a sites. During charging of the Li₄Ti₅O₁₂/Li cell, the lithium atoms are extracted out from the 16c sites via 8a sites, and the other lithium atoms move back to the 8a sites from 16c sites. Hence, it can be concluded that the occupation of Li⁺ ions at 16c sites can impede the Li⁺ insertion and extraction. According to our synthesis process, a model of the Li₄Ti₅O₁₂–CeO₂ composites is shown in Fig. 14(b). In the Li₄Ti₅O₁₂–CeO₂ composites, in situ generated CeO₂ is tightly combined with Li₄Ti₅O₁₂, and then many Li₄Ti₅O₁₂–CeO₂ phase interfaces can be formed. These interfaces can store electrolyte and provide more places for the reactions of lithium ion insertion/extraction,⁵⁵ leading to the improved reaction kinetics in polycrystalline materials or the liquid electrolyte at the solid/liquid interface, and then reduce the electrochemical polarization during the charge–discharge process. In addition, it has been reported that an internal adsorption of ions at the CeO₂ surface can cause space-charge effects.⁵⁶ The increased cation vacancy concentration at the interface can form highly conducting interfacial layers between Li₄Ti₅O₁₂ and CeO₂. Hence, these additional defects are responsible for the decreased charge transfer resistance and enhanced conductivity (see Fig. 7). Hence, it can be concluded that the CeO₂ modification enhances the conductivity of Li₄Ti₅O₁₂ and produces a good electrical contact between the oxides which facilitates the electron transfer between cerium oxide and the supported Li₄Ti₅O₁₂. In addition, it has been reported that there is an interfacial reaction between Li₄Ti₅O₁₂ and the surrounding alkyl carbonate solvents, which results in the formation of anatase TiO₂.⁵⁷ Hence, the CeO₂ can also prevent the direct contact between the active material and electrolyte, and improve the rate capability of Li₄Ti₅O₁₂ at a high charge–discharge rate. This is why the Ce- and CeO₂-modified Li₄Ti₅O₁₂ composites show a higher rate capacity and cycling stability than those of pristine Li₄Ti₅O₁₂ during the fast charge–discharge process (see Fig. 13). Hence, the Ce and CeO₂ in situ modification is an effective way to improve the electrochemical performance of Li₄Ti₅O₁₂.

Fig. 14 (a) Structural transformation model of Li₄Ti₅O₁₂ during the charge–discharge process, and (b) model of the Li₄Ti₅O₁₂–CeO₂ composites.

4. Conclusions

Ce- and CeO₂-modified Li₄Ti₅O₁₂ were successfully prepared by a solid-state method. Ce and CeO₂ modification does not affect the structure and particle size of the Li₄Ti₅O₁₂ powder. However, the Ce and CeO₂-modified Li₄Ti₅O₁₂ electrodes exhibit larger discharge capacities and cycling stabilities during the fast charge–discharge process. In comparison, the modified sample with Ti/Ce = 4.85 : 0.15 shows the optimal cycle stability and high-rate discharge capability. CV measurements provide strong evidence that the Ce and CeO₂ modification can improve the reversibility of lithium ion intercalation and de-intercalation, and then decrease the polarization degrees of the electrodes. Correspondingly, EIS tests prove that the Ce and CeO₂ modification can decrease the charge transfer resistance, and increase the lithium diffusion coefficient of the electrodes. Ce and CeO₂ modification is an effective way to improve the electrochemical properties of the Li₄Ti₅O₁₂ anode material by producing a good electrical contact between the oxides and Li₄Ti₅O₁₂, resulting in an improved conductivity, lithium diffusion coefficient and fast charge–discharge performance.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (nos 51274002 and 51404002), Anhui Provincial Natural Science Foundation (no. 1508085MB25), the Postdoctoral Science-research Developmental Foundation of Heilongjiang Province (no. LBH-Q13138), and the Program for Innovative Research Team in Anhui University of Technology (no. TD201202).

References

1. Y. Li, J. Song and J. Yang, Renewable Sustainable Energy Rev., 2014, 37, 627–633.
2. K. Zaghib, M. Dontigny, P. Perret, A. Guerfi, M. Ramanathan, J. Prakash, A. Mauger and C. M. Julien, J. Power Sources, 2014, 248, 1050–1057.
3. Y. Q. Wang, L. Gu, Y. G. Guo, H. Li, X. Q. He, S. Tsukimoto, Y. Ikuhara and L. J. Wan, J. Am. Chem. Soc., 2012, 134, 7874–7879.
4. H. Kageyama, Y. Oaki and H. Imai, RSC Adv., 2014, 4, 44124–44129.
5. C. Cheng, H. Liu, X. Xue, S. Cao, H. Cao and L. Shi, RSC Adv., 2014, 4, 63105–63109.
6. C. W. Xiao, Y. Ding, J. T. Zhang, X. Q. Su, G. R. Li, X. P. Gao and P. W. Shen, J. Power Sources, 2014, 248, 323–329.
7. W. Wang, B. Jiang, W. Xiong, Z. Wang and S. Jiao, Electrochim. Acta, 2013, 114, 198–204.
8. T. F. Yi, H. P. Liu, Y. R. Zhu, L. J. Jiang, Y. Xie and R. S. Zhu, J. Power Sources, 2012, 215, 258–265.
9. Q. Y. Zhang, C. L. Zhang, B. Li, S. F. Kang, X. Li and Y. G. Wang, Electrochim. Acta, 2013, 98, 146–152.
10. H. Zhao, Y. Li, Z. Zhu, J. Lin, Z. Tian and R. Wang, Electrochim. Acta, 2008, 53, 7079–7083.
11. D. Wang, C. M. Zhang, Y. Y. Zhang, J. Wang and D. N. He, Ceram. Int., 2013, 39, 5145–5149.
12. X. Li, S. H. Tang, M. Z. Qu, P. X. Huang, W. Li and Z. L. Yu, J. Alloys Compd., 2014, 58, 817–824.
13. W. Wang, H. Wang, S. Wang, Y. Hu, Q. Tian and S. Jiao, J. Power Sources, 2013, 228, 244–249.
14. P. Kubiak, A. Garcia, M. Womes, L. Aldon, J. O. Fourcade, P. E. Lippens and J. C. Jumas, J. Power Sources, 2003, 119, 626–630.
15. J. Wolfenstine and J. L. Allen, J. Power Sources, 2008, 180, 582–585.
16. Q. Y. Zhang, C. L. Zhang, B. Li, D. D. Jiang, S. F. Kang, X. Li and Y. G. Wang, Electrochim. Acta, 2013, 107, 139–146.
17. X. Han, Z. Zhao, Y. Xu, D. Liu, H. Zhang and C. Zhao, RSC Adv., 2014, 4, 41968–41975.
18. J. Q. Wang, Z. Z. Yang, W. H. Li, X. W. Zhong, L. Gu and Y. Yu, J. Power Sources, 2014, 266, 323–331.
19. Z. M. Liu, N. Q. Zhang, Z. J. Wang and K. N. Sun, J. Power Sources, 2012, 205, 479–482.
20. C. C. Li, Q. H. Li, L. B. Chen and T. H. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 1233–1238.
21. J. Shu, L. Hou, R. Ma, M. Shui, L. Shao, D. Wang, Y. Ren and W. Zheng, RSC Adv., 2012, 2, 10306–10309.
22. Y. Ding, G. R. Li, C. W. Xiao and X. P. Gao, Electrochim. Acta, 2013, 102, 282–289.
23. H. Yu, X. Zhang, A. F. Jalbout, X. Yan, X. Pan, H. Xie and R. Wang, Electrochim. Acta, 2008, 53, 4200–4204.
24. H. Park, T. Song, H. Han and U. Paik, J. Power Sources, 2013, 244, 726–730.
25. J. Liu, X. Li, M. Cai, R. Li and X. Sun, Electrochim. Acta, 2013, 93, 195–201.
26. X.-P. Li and J. Mao, Ceram. Int., 2014, 40, 13553–13558.
27. L. Daza, C. M. Rangel, J. Baranda, M. T. Casais, M. J. Martinez and J. A. Alonso, J. Power Sources, 2000, 86, 329.
28. D. Arumugam and G. P. Kalaignan, Electrochim. Acta, 2010, 55, 8709–8716.
29. J. W. Yao, F. Wu, X. P. Qiu, N. Li and Y. F. Su, Electrochim. Acta, 2011, 56, 5587–5592.
30. F. Wu, M. Wang, Y. F. Su, L. Y. Bao and S. Chen, Electrochim. Acta, 2009, 54, 6803–6807.
31. W. Yuan, H. Z. Zhang, Q. Liu, G. R. Li and X. P. Gao, Electrochim. Acta, 2014, 135, 199–207.
32. X. J. Yang, Y. D. Huang, X. C. Wang, D. Z. Jia, W. K. Pang, Z. P. Guo and X. C. Tang, J. Power Sources, 2014, 257, 280–285.
33. Z. Zhong, C. Ouyang, S. Shi and M. Lei, ChemPhysChem, 2008, 9, 2104–2108.
34. J. Shu, Electrochem. Solid-State Lett., 2008, 11, A238–A240.
35. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
36. T.-F. Yi, J. Shu, Y.-R. Zhu, X.-D. Zhu, C.-B. Yue, A.-N. Zhou and R.-S. Zhu, Electrochim. Acta, 2009, 54, 7464–7470.
37. T. Ohzuku, A. Ueda and N. Yamamoto, J. Electrochem. Soc., 1995, 142, 1431–1435.
38. T. Ohzuku, K. Ariyoshi, S. Takeda and Y. Sakai, Electrochim. Acta, 2001, 46, 2327–2336.
39. T.-F. Yi, C.-Y. Li, Y.-R. Zhu, J. Shu and R.-S. Zhu, J. Solid State Electrochem., 2009, 13, 913–919.
40. C. K. Lan, S. I. Chuang, Q. Bao, Y. T. Liao and J. G. Duh, J. Power Sources, 2015, 275, 660–667.
41. R. B. Hadjean and J. P. P. Ramos, Chem. Rev., 2010, 110, 1278–1319.
42. G. Yan, H. Fang, H. Zhao, G. Li, Y. Yang and L. Li, J. Alloys Compd., 2009, 470, 544–547.
43. L. Aldon, P. Kubiak, M. Womes, J. C. Jumas, J. O. Fourcade, J. L. Tirado, J. I. Corredor and C. P. Vicente, Chem. Mater., 2004, 16, 5721–5725.
44. G. Balakrishnan, C. M. Raghavan, C. Ghosh, R. Divakar, E. Mohandas, J. I. Song, S. I. Bae and T. G. Kim, Ceram. Int., 2013, 39, 8327–8333.
45. W. Fang, P. J. Zuo, Y. L. Ma, X. Q. Cheng, L. X. Liao and G. P. Yin, Electrochim. Acta, 2013, 94, 294–299.
46. W. Wang, Y. Y. Guo, L. X. Liu, S. X. Wang, X. G. Yang and H. Guo, J. Power Sources, 2014, 245, 624–629.
47. J. Y. Liao, X. C. Xiao, D. Higgins and D. G. Lee, Electrochim. Acta, 2013, 108, 104–111.
48. H. Liu, Q. Cao, L. J. Fu, C. Li, Y. P. Wu and H. Q. Wu, Electrochem. Commun., 2006, 8, 1553–1557.
49. J. Wang, W. Lin, B. Wu and J. Zhao, Electrochim. Acta, 2014, 145, 245–253.
50. S. Liu, J. Xu, D. Li, Y. Hu, X. Liu and K. Xie, J. Power Sources, 2013, 232, 258–263.
51. B. Li, C. Han, Y. B. He, C. Yang, H. Du, Q. H. Yang and F. Kang, Energy Environ. Sci., 2012, 5, 9595–9602.
52. J. Shu, J. Solid State Electrochem., 2009, 13, 1535–1539.
53. G. G. Amatucci, F. Badway, A. D. Pasquier and T. Zheng, J. Electrochem. Soc., 2001, 148, A930–A939.
54. C. Zhang, Y. Zhang, J. Wang, D. Wang, D. He and Y. Xia, J. Power Sources, 2013, 236, 118–125.
55. M. M. Rahman, J. Z. Wang, M. F. Hassan, D. Wexler and H. K. Liu, Adv. Energy Mater., 2011, 1, 212–220.
56. M. M. E. Jacob, S. Rajendran, R. Gangadharan, M. S. Michael and S. R. S. Prabaharan, Solid State Ionics, 1996, 86, 595–602.
57. Y. B. He, B. H. Li, M. Liu, C. Zhang, W. Lv, C. Yang, J. Li, H. D. Du, B. Zhang, Q. H. Yang, J. K. Kim and F. Y. Kang, Sci. Rep., 2012, 2, 9.

---