Lithium titanate microspheres were synthesized by a hydrothermal method. The structure and morphology of samples were characterized by X-ray diffraction, infrared spectroscopy, Raman spectroscopy, scanning electron microscopy, and transmission electron microscopy, respectively. The specific surface area and average pore diameter of samples were studied by N2 adsorption-desorption isotherms. The results indicated that amorphous phase changed to lithium titanium oxide hydrate, accompanying mesopores formed between agglomerated primary particles in hydrothermal reaction. After sintering, mesoporous Li4Ti5O12 microspheres assembled by nanosized particle were obtained and had a diameter of about 400–700 nm. Then, a possible formation process analogous to the Kirkendall effect was proposed. Moreover, the effect of sintering temperature on the electrochemical properties of Li4Ti5O12 microspheres was investigated.

1. Introduction

Lithium titanate has already become one of the most attractive anode materials for lithium-ion batteries (LIBs), due to the zero-strain insertion, high safety, and excellent cycle stability [1, 2]. As is well known, Li4Ti5O12 with a spinel structure is transformed to Li7Ti5O12 with an ordered rock-salt structure during the lithium insertion process through a two-phase equilibrium reaction [3]. The minimal lattice change in the phase transformation is significant for the safety and cycle life of lithium-ion batteries. Herein, LIBs based on Li4Ti5O12 anode material have been expected to be used in hybrid electric vehicles and sustainable energy storage [4, 5]. However, the low electronic conductivity of Li4Ti5O12, only  Scm−1 at room temperature, deteriorates its rate capability [6]. Hence, various works have been reported to improve the electronic conductivity, such as doping with metal ions [7, 8], adding conductive agent [9, 10], and fabrication of nanostructure [11, 12].

However, doping with ions may deteriorate the structural stability and adding conductive agent may reduce the packing density of Li4Ti5O12 electrodes [7, 13]. Without these disadvantages, nanostructure may provide more active surface available for lithium insertion and shorter lithium diffusion length, which may improve the rate capability and reduce electrode polarization [14]. Li4Ti5O12 samples with nanostructure such as spheres [11, 15], 3D-ordered porous [14, 16], nanowires [17], flower-like [18], and hollow spheres [19] have been reported. Especially, the spherical morphology is most favorable due to high tap density and isotropic physical properties. Generally, the spherical Li4Ti5O12 powders are synthesized by spray pyrolysis [20], outer gel [21], and hydrothermal method [15]. Unfortunately, the particle size of precursors or products can be hardly controlled in spray pyrolysis method or outer gel method. The obtained products have a wide size distribution and large secondary particle, which may deteriorate their electrochemical performance. However, the hydrothermal method is a general and effective route for the preparation of micro/nanostructured materials with well-controlled morphology [22]. It is expected that the Li4Ti5O12 samples with a homogeneous particle size distribution and smaller particle size may have good electrochemical properties.

The preparation of Li4Ti5O12 microspheres has been reported by other authors by hydrothermal methods using titanate alkoxide [22], TiCl4 [23], or Ti(SO4)2 [24] as Ti sources. Recently, Zhang et al. [25] reported the preparation of Li4Ti5O12 microsphere with large surface area by a hydrothermal method. However, the hydrolysis reaction of Ti(OC4H9)4 was proceeded using LiOH as precipitant and could not be controlled effectively. As a result, the final product with a large secondary particle about 1–5 μm was obtained. However, the rate capability of Li4Ti5O12 is significantly influenced by its particle sizes [26, 27]. Meanwhile, the formation of Li4Ti5O12 microspheres in hydrothermal synthesis is not well indicated.

In our paper, the precursor was obtained by a controlled hydrolysis reaction and had a homogeneous particle size ranging from 400 to 700 nm. Moreover, the formation process has been thoroughly characterized in order to obtain Li4Ti5O12 samples with a good electrochemical performance.

2. Experimental

2.1. Synthesis

All chemicals were used as received. The preparation of Li4Ti5O12 microspheres was based on the procedure of Tang [22]. In a typical synthesis, 200 mL ethanol was mixed with 1 mL of 0.1 M KCl solution and used as the solvent. Then 8.8 g tetrabutyl titanate was added dropwise into the solvent under continuous sonication for 10 min at ambient temperature. A white precipitate formed immediately. Then the suspension was stirred at a moderate speed for 6 h. After stirring, the precipitate was collected by centrifugation and washed by ethanol and deionized water for serious times. The washed precipitate (denoted as Ta) was dried at 60°C in the oven.

Then, 0.2 g of Ta was mixed with 2.55 mmol LiOH H2O in 25 mL ethanol-water (1 : 1 in volume) mixed solution. The resulting mixture was transferred into a 40 mL teflon-lined stainless steel autoclave and then placed in an oven at 180°C for 36 h. After the autoclave was cooled to room temperature, white powders (denoted as Tb) were collected and washed by ethanol and de-ionized water for several times to remove redundant LiOH. After drying at 60°C for 12 h, the obtained powders were sintering at 600°C, 700°C, and 800°C for 1 h in air, which were denoted as LTO600, LTO700, and LTO800, respectively.

2.2. Characterization

The crystal structure of samples was characterized by X-ray diffraction (XRD, X’Pert Pro, PHLIPS) with Cu-Kα radiation. The lattice parameter of LTO samples was simply refined by the software Jade. The morphology of samples was observed by scanning electron microscope (SEM, JSM-6700F, JEOL) and transmission electron microscopy (TEM, Tecnai G2 20 s-twin, FEI). The N2 adsorption-desorption isotherms at 77 K were measured by a high speed automated surface area and pore size analyzer (NOVA4000e, Quantachrome) after the samples were degassed at 120°C for 2 h. The FT-IR spectrum of the samples was characterized by an IR spectrometer (Nicolet 7000-C, Thermal Scientific) by using KBr as dispersant. The Raman spectra were recorded on a Raman microscope (DXR, Thermal Scientific) with 532 nm excitation length. The Li/Ti ratio of LTO sample was analyzed by ICP-AES spectrometer (IRIS Advantage, Thermal Scientific).

2.3. Electrochemical Characterization

LTO powders, carbon black, and polyvinylidene fluoride binder were homogeneously mixed in N-methyl pyrrolidinone (NMP) at a weight ratio of 85 : 7 : 8. Then the resulting slurry was spread on aluminum current collector and dried at 100°C to remove the redundant NMP solvent. Metallic lithium was used as counter electrode and microporous polypropylene membrane used as separator, respectively. The electrolyte was 1 M LiPF6 in ethylene carbonate and dimethyl carbonate solution (1 : 1 in volume). The cells (CR2016) were assembled in a glove box filled with argon gas. The galvanostatic discharge/charge tests were performed by a LAND cell test system (CT2001A, Wuhan Jinnuo) with potential between 1.0 and 2.5 V (versus Li+/Li) at room temperature.

3. Results and Discussion

3.1. Structure and Morphology

The structure of samples is examined by X-ray diffraction as shown in Figure 1(a). No diffraction peaks in the XRD pattern of Ta are found, indicating the amorphous nature, which is in accordance with the hydrolysis precipitate from TiCl4 or titanium tetraisopropoxide [28, 29]. After hydrothermal reaction, the diffraction peaks of LiTiO2 (JCPDS no. 74-2257) and lithium titanium oxide hydrate (JCPDS no. 47-127) are obviously identified in sample Tb. In the XRD pattern of sample LTO700, all the sharp diffraction peaks can be indexed to cubic spinel structured Li4Ti5O12 (JCPDS no. 26-1198). From Figure 1(b), it is easily observed that the diffraction peaks of spinel Li4Ti5O12 are enhanced with the increased sintering temperature. The refined lattice parameters of samples LTO600, LTO700, and LTO800 are 8.354, 8.357, and 8.360 Å, respectively, which is consistent with the theoretical value [1]. These results demonstrate that the degree of crystallinity of LTO improves when sintering temperature increases. No peaks of anatase or rutile phase TiO2 are detected, indicating the high purity of products synthesized by the hydrothermal method.

The FT-IR spectra for the samples Ta, Tb, and LTO700 are shown in Figure 2(a). There is no significant difference between the absorption peaks of Ta and LTO700. The spectrum of Tb is slightly different from Ta and LTO700. Two absorption peaks at 3443 cm−1 and 1637 cm−1 in all the samples are assigned to hydroxyl groups (–OH) [30]. In the hydrolysis precipitate and precursor, the hydroxyl groups are easy to detect. For Li4Ti5O12 material, it means that a strong absorption of H2O happens in the air [24]. In the spectrum of Tb, the peak at 1503 cm−1 and 1433 cm−1 can be assigned to the antisymmetric stretching vibrations of anions and that at 867 cm−1 can be attributed to the symmetric stretching vibrations of anions, because anions probably formed on the surface of sample Tb by absorbing CO2 molecules [31]. Such absorption of CO2 molecules is not clearly identified in Ta and LTO700, which may ascribe to that sample Tb has a higher specific surface area and pore volume after hydrothermal treatment. The Raman spectra of samples Ta, Tb, and LTO700 are presented in Figure 2(b). As shown in Figure 2(b), the Raman spectrum of Ta does not show any spectral characteristics, which is consistent with the XRD analysis. The weak peaks at 157 and 320 cm−1 in Tb could be assigned to the characteristic peaks of LiTiO2 [31]. The peaks around 234, 270, and 353 cm−1 of sample LTO700 can be assigned to the F2g mode of Li4Ti5O12. The peak at around 672 cm−1 accompanying a shoulder (745 cm−1) and the broad peak at 427 cm−1 can be assigned to the A1g and Eg modes of Li4Ti5O12, respectively [32]. The results above agree well with the XRD analysis of sample LTO.

The scanning electron microscopy in Figure 3 shows the microstructure of precursors and Li4Ti5O12 at different magnification. The obtained amorphous TiO2 microspheres consist of well-defined spherical particles (Figure 3(a)), and the average particle size is about 400–700 nm. In magnification image, it is clear that these spherical powders are assembled by the nanosized particle about 10–50 nm in diameter (Figure 3(b)). Some small primary particles are dropped from these microspheres due to the sonication in preparation. Hence, it is better to wash and filter the precipitation several times to remove small particles. After hydrothermal reaction, the diameter of microspheres does not change significantly (Figures 3(c) and 3(d)). In comparison of Figures 3(a) and 3(c), the surface of particles in Tb became rough after hydrothermal reaction, which may result from the diffusion of ions or molecules. As shown in Figures 3(e) and 3(f), LTO700 also shows a uniform and spherical morphology. Meanwhile, spherical particles in Tb and LTO700 are assembled by nanosized particles.

The morphology of Li4Ti5O12 microspheres is further analyzed by TEM measurements. The TEM image (Figure 4(a)) reveals well-defined spheres with a diameter of 400 nm and rugged surface. Some small particles about 10–40 nm in size could be easily observed on the edge of spheres in Figure 4(b), which is in agreement with SEM observations in Figures 3(e) and 3(f). Meanwhile, microspores are apparently found to be presented among these primary nanoparticles. As shown in Figure 4(c), the calculated interference fringe spacing in HRTEM is about 0.489 nm, typically consistent with the (111) crystal plane of the spinel structured Li4Ti5O12. It indicates that well-crystallized spinel phase is obtained by this hydrothermal method. Only Ti and O elements are detected in the energy dispersive X-ray spectrum in Figure 4(d). The ICP result indicates the mol ratio of Li/Ti is 0.828, which is close to stoichiometry in Li4Ti5O12.

3.2. N2 Adsorption-Desorption Isotherms

The porous structures of samples Ta, Tb, and LTO700 are indicated by N2 adsorption-desorption isotherms and Barrett-Joyner-Halenda (BJH) pore size distribution analysis (inset) as shown in Figure 5. Figure 5(a) exhibits a type I adsorption-desorption isotherm, which is typically characteristic of microporous materials having relatively small internal surfaces [29]. Two distinct capillary condensation steps can clearly be seen in the Figure 5(b). The first hysteresis loop of Tb is at , corresponding to the filling of the framework confined mesoporous formed between intra-agglomerated primary particles. The second hysteresis loop is at , corresponding to the filling of textural meso- and macropores produced by interaggregated secondary particles [29]. However, it is hard to find any loop in the adsorption-desorption curves in Figure 5(c), which indicated that only small amount of these textural micropores preserves in LTO700 because of particles growth during sintering.

The Brunauer-Emmett-Teller (BET) specific surface area, pore volume, and average pore diameter of samples Ta, Tb, and LTO are summarized in Table 1. By comparing the properties of samples Ta, Tb, and LTO, it is observed that the pore volume of sample Tb is the largest, which is ascribed to that a large percentage of mesopores form in hydrothermal treatment. The mean pore diameter of samples Tb is also larger than that of samples Ta and LTO700. These results explain the difference in FT-IR spectra very well. Meanwhile, specific surface area decreases from 22.37 to 10.61 m2/g as the sintering temperature increases from 600°C to 800°C, which may correlate with small primary particle growth.

3.3. The Possible Formation Process

On the basis of the test results and morphology observations aforementioned, a reasonable formation process is proposed for the synthesis of Li4Ti5O12 by a hydrothermal method, as illustrated in Figure 6. Firstly, in the hydrolysis and condensation reactions of tetrabutyl titanate (TBOT), spherical powders of amorphous TiO2 appear after a rapid nucleation and growth process [32]. In hydrothermal reaction, the OH and H2O may enter into the amorphous TiO2 microspheres to form titanium hydroxyl species (probably HTiO3) [22, 29]. Then Li+ may react with HTiO3 to form lithium titanium oxide hydrate. Meanwhile, LiOH may react with TiO2·xH2O to form LiTiO2 and eliminate moles of H2O. The diffusion of the ions or molecules resulted in the vacancy increasing quickly due to the Kirkendall effect and the mesopores forming between intra-agglomerated primary particles [29]. This is why sample Tb has the largest pore volume and mean pore diameter. After sintering, only small percentages of mesopores are preserved, so that no distinct loop is found in the adsorption-desorption curve for sample LTO.

3.4. Electrochemical Performance

Figure 7 shows the initial charge-discharge profiles of Li4Ti5O12 at different rate in the voltage range of 1.0–2.5 V. The charge-discharge curves display very flat plateau at the potential about 1.50 and 1.60 V, which is the characteristic of two-phase reaction based on the Ti4+/Ti3+ redox couple [33]. It can be seen that the coulomb efficiency of the sample is close to 99%, which means the reaction of intercalation/deintercalation in Li4Ti5O12 is highly reversible. The first discharge capacity of the samples is about 119.9, 145.6, and 145.7 mAh/g at 0.1C, when the sintering temperatures change from 600°C to 800°C. And sample LTO800 exhibits the best reversible capacity at all rates. The results demonstrate that the degree of crystallinity of Li4Ti5O12 is crucial to the achievement of a high capacity. Although the capacity is not as high as the samples synthesized by solid-state or sol-gel method, the samples exhibit a good performance at C-rate test. The first charge capacity of sample LTO800 is about 145.7, 141.8, 136.2 and 130.9 mAh/g at the rate of 0.1C, 0.2C, 0.5C, and 1C, respectively. The result of C-rate test is close to the results reported by Lin et al. [34, 35]. As is well known, the lower potential interval indicates the better lithium insertion/deintercalation kinetic [36]. From Figure 7(d), it can be seen that potential interval between discharge and charge plateau at 1C rate is 85, 89, and 142 mV for samples LTO600, LTO700, and LTO800, respectively. It means that the electrode polarization of sample LTO800 is the worst. The reason may be that higher percentage of textural mesopores ensures more active surface and larger electrode-electrolyte contact area to enhance electrochemical kinetic [14]. However, LTO800 has the superior charge-discharge capacity because of the higher degree of crystallinity. Hence, it can be concluded that nanosized particles, high degree of crystallinity, and abundant textural mesopores are crucial for Li4Ti5O12 anode material with good electrochemical performance.

4. Conclusions

In summary, a systematic investigation on the hydrothermal synthesis of Li4Ti5O12 microspheres has been carried out. In the hydrothermal reaction, H+ ions in amorphous TiO2 particles may be exchanged with Li+ ions from LiOH to produce precursors. Meanwhile, mesopores form between agglomerated primary particles due to the Kirkendall effect. The as-synthesized Li4Ti5O12 powders show a high phase purity and degree of crystallinity. The particle size of Li4Ti5O12 powders is about 400–700 nm, and the average pore diameter is about 1.52 nm. Electrochemical test results indicate that high degree of crystallinity results in high charge-discharge capacity, and abundant textural mesopores are in favour of lower electrode polarization. The obtained Li4Ti5O12 microspheres exhibit a good performance at C-rate test.


This work was supported by the Key Basic Research Programs (no. 11JC1414600) and Nano Technology Projects (no. 11nm050050) of Science and Technology Commission of Shanghai Municipality. This work was also supported by the National Science Foundation of China (no. 21373257). The authors thank Lei Zhang, Ph.D., from Tianjin Institute of Urban Construction for assistance with the test of N2 adsorption-desorption isotherms.