Introduction

To date, the preparation of inorganic nanostructures with controlled shape, composition, and desired functions has been explored for the application to biomedicine, catalysts, and energy storages [13]. Li4Ti5O12 has attracted attention as a Li ion battery electrode for energy storage because of its merit as zero-strain lithium insertion hosts and its stable insertion potential at 1.55 V vs. Li [4]. Many researchers have studied the synthesis of Li4Ti5O12 with various nanostructures, including nanowires [57], nanotubes [8], flower-like nanosheets [9], and hollow-spheres [10]. These powders exhibited good electrochemical properties compared to the bulk Li4Ti5O12 powder because of the shorter Li ion diffusion distances. In a wet chemical reaction, the change in the dielectric constant of the solvent can alter the thermodynamics of the reaction system and the nucleation kinetics, which modulate the particle size and size distribution of the resulting particles [11, 12].

In this work, 0D-1D Li4Ti5O12 heterogeneous nanostructures were via a solvothermal reaction using Li(OH)2·H2O and H2Ti3O7 nanowires as the templates in an ethanol/water mixed solvent at 180°C with subsequent heat treatment. The effects of the volume ratio of ethanol to water on the morphologies of the prepared Li4Ti5O12 were examined with respect to the dielectric constant of the mixed solvent. Additionally, the electrochemical performances of the prepared Li4Ti5O12 as an anode for Li ion batteries were evaluated and discussed with regard to the morphology of Li4Ti5O12.

Experimental procedures

H2Ti3O7 (HTO) precursors were synthesized using a hydrothermal treatment in an alkaline environment, which was a well-known solution technique for the preparation of HTO nanowires [13]. First, 1 g of commercial anatase powder (Sigma–Aldrich, USA) was dispersed in 200 ml of a 10 M NaOH aqueous solution. The solution was stirred for 1 h at room temperature to ensure homogeneity, and then the suspension was transferred into a Teflon-lined stainless steel autoclave for the hydrothermal treatment. The autoclave was maintained at 180°C for 24 h. After the hydrothermal treatment, the suspension was centrifuged, and a white sodium titanate (Na2Ti3O7) precipitate was obtained. Then, the precipitate was washed and dried. The prepared precipitate was soaked in a 0. 1 M HCl aqueous solution for 24 h at room temperature, which fully induced cation exchange between the sodium ions and the hydrogen ions. Finally, HTO nanowires were obtained after the suspension was washed with deionized water several times and dried in a freeze dryer for 48 h.

Lithium titanate (Li4Ti5O12) was synthesized using a solvothermal reaction and subsequent calcination. Lithium hydroxide monohydrate (Li(OH)2·H2O) and the HTO nanowires were dissolved in an ethanol/water mixed solvent and autoclaved at 180°C for various reaction times (1–24 h). The volume percentage (vol.%) of ethanol to distilled water was varied from 0 to 100%, and the molar ratio of Li+ to Ti4+ was 1.05. After each reaction was completed, the obtained products were washed with distilled water and ethanol. Then, the powders were dried in a freeze dryer for 48 h and finally calcined at 500 and 700°C for 2 h.

The crystal structures of the powders were analyzed using an X-ray diffractometer (M18XHF-SRA, MAC Science Co., Japan) and Raman spectroscopy (model T64000, Horiba Jovin Yvon, France). The morphologies and microstructures of the as-prepared powders were observed using a scanning electron microscope (SEM, JSM-6330F, JEOL) and a high resolution transmission electron microscope (HRTEM, JEM-3000F, JEOL).

The working electrodes were prepared as pellets with a diameter of 10 mm by pressing and punching the mixture that was composed of 60 wt% active materials, 20 wt% Super P carbon black, and 20 wt% Kynar 2801 binder. The Swagelok-type cells were assembled using the pellets as the positive electrode, a Li metal foil as the negative electrode, and a Celgard 2400 separator [14]. All of the components were immersed in an electrolyte solution of 1 M LiPF6 that was dissolved in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) with the volume ratio of EC/DMC = 1:1.

The electrochemical properties of the assembled cells were measured using a battery cycler (WBCS 3000, WonATech, Korea). The charge/discharge tests were galvanostatically carried out between 2.5 and 1.0 V for 500 cycles at a current rate of 1C, and additionally, the cycling performance was evaluated at various rates of 1C, 5C, 10C, and 20C, in order to determine the highest rate performance.

Results and discussions

Figure 1a shows the XRD patterns of the powders that were calcined at 500°C after the solvothermal reaction in various mixed solvents, including ethanol 20 vol.% (LT-E2), 40 vol.% (LT-E4), 60 vol.% (LT-E6), and 80 vol.% (LT-E8), using the H2Ti3O7 nanowire and Li(OH)2·H2O. When the amount of ethanol was more than 60 vol.%, pure Li4Ti5O12 was well crystallized without a secondary phase and indexed to cubic spinel Li4Ti5O12 (JCPDS Card #: 49-0207). Figure 1b1e show the representative SEM images of the powders that were prepared in various solvents. After the solvothermal reaction and the calcination, the powders retained their wire-like morphology, and interestingly, LT-E6 and LT-E8 exhibited a 0D-1D heterogeneous nanostructures with nanoparticles (0D) on the surface of the wires (1D), which was confirmed in Fig. 2.

Figure 1
figure 1

a XRD patterns of the powders calcined at 500°C after the solvothermal reaction in various mixed solvents. Typical SEM images of b LT-E2, c LT-E4, d LT-E6, and e LT-E8 powders calcined at 500°C

Figure 2
figure 2

TEM images of a LT-E4, b LT-E6, c LT-E8 powders calcined at 500°C, respectively. d and e HRTEM images of A and B regions in the (b), respectively. Insets in d and e show the corresponding FFT and SAED patterns, respectively. f Raman spectra of the LT-E6 and LT-E8 powders

Figure 2a, 2b and 2c represent the TEM images of LT-E4, LT-E6, and LT-E8, respectively. The surface of LT-E4 was smooth without any nanoparticles, whereas LT-E6 and LT-E8 had tetrahedral nanoparticles on the surface. More interestingly, the particle size of LT-E6 was much larger than that of LT-E8. This is attributed to the low dielectric constant (ε) of the mixed solvent (ε water = 80, ε EtOH = 25) [15].

In principle, a solvent with a low dielectric constant is easier to reach higher supersaturation, which can accelerate the nucleation rate. As well, the radius of the nuclei decreases with the reduction of the ε of the solvent from the electrostatic model using the following relationships [16]. Therefore, both the dielectric constant and the primary particle sizes of the powders that were generated in the mixed solvent decreased with increasing ethanol concentration. Figure 2d and 2e represent the HRTEM images of the A and B regions in the Fig. 2b. The measured interplanar distances in both A and B regions closely matched the lattice spacings of Li4Ti5O12. The corresponding fast Fourier transform (FFT) and selected area electron diffraction (SAED) patterns (the insets in Fig. 2d and 2e) also confirm that both the polyhedral particles and the wires are of the same phase Li4Ti5O12. Single phase of Li4Ti5O12 was investigated by Raman spectroscopy. As shown in Fig. 3f, vibration peaks were observed at 225, 272, 340, 420, 514, 674, and 750 cm−1, consistent with the spectra that were reported for the Li4Ti5O12[17]. These observations demonstrate the formation of phase-pure Li4Ti5O12.

Figure 3
figure 3

a Variation of the discharge–charge specific capacities versus the cycle number and b cycling behavior at different C rates for the LT-E6-500, LT-E6-700, LT-E8-500, and LT-E8-700 powders, respectively. Inset in a shows the XRD patterns of the LT-E6-500 and LT-E6-700 powders

Although further experimental observations are needed for the exact mechanism, the formation mechanism of the 0D-1D Li4Ti5O12 in this study can be proposed as follows. In the early stage of the reaction, the Li+ species (or intermediate derivatives of lithium metals) in the solution reacted with the HTO nanowires that were locally dissolved at the interface and in situ precipitated to 0D Li4Ti5O12 nanoparticles. The similar nucleation process was reported in the formation of multiple NaNbO3/Nb2O5 heterostructure nanotubes [18]. This process was possible even at a low temperature, ~180°C, because of the easier supersaturation caused by the lowered dielectric constant of the mixed solvent. The propagation of the reaction was expected to occur from the front of the HTO to the interior until the entire HTO wires were transformed into Li4Ti5O12 through reaction with barium ions that diffused during the growth of the outer Li4Ti5O12 nanoparticles, finally yielding the phase-pure 0D–1D Li4Ti5O12 nanostructures.

Figure 3 represents the charge–discharge behavior of heterogeneous Li4Ti5O12 nanostructures calcined at 500°C (LT-E6-500, LT-E8-500) and 700°C (LT-E6-700, LT-E8-700). LT-E6-700 was also single phase Li4Ti5O12 such as LT-E6-500 (inset in Fig. 3a). Both LT-E6-500 and LT-E6-700 exhibited almost the same capacities (150 mAh/g) in the early stages. After 500 cycles, the capacity of LT-E6-700 was 125 mAh/g, whereas the capacity of LT-E6-500 was 100 mAh/g, which was attributed to the high crystallinity of LT-E6-700 by the calcination at a high temperature [19]. For comparison in our work, TiO2 nanowire (1D) samples were also prepared by the calcinations of H2Ti3O7 nanowires at same temperature, 700°C. Although Li4Ti5O12 and TiO2 had similar theoretical capacities (~170 mAh/g), Li4Ti5O12 nanowires exhibited much higher capacities and long-term cyclabilities than TiO2 nanowires, which might originate from the larger specific surface area (BET surface areas of 40 and 15 m2/g for Li4Ti5O12 and TiO2, respectively) that resulted from the heterogeneous nanostructures. It has been also reported that excellent electrochemical performance could be achieved in high surface area anodes such as V2O5–SnO2 double-shelled nanocapsules and porous Co3O4 nanocapsules due to efficient Li ion and electron transport [20, 21]. However, up until now, this long-life performance of 500 cycles in Li4Ti5O12 has not been reported.

Figure 3b shows the rate capability of LT-E6-500, LT-E8-500, LT-E6-700, and LT-E8-700 while examining rates of up to 20C. The cells were charged and discharged at 1C for the first 10 cycles, and then the rate was increased in stages to 20C. The heterogeneous LT-E6-700 nanowires exhibited a slightly better capacity at high rates than LT-E6-500. A specific discharge capacity of LT-E6-700 after 10 cycles at 1C was 151 mAh/g. This value gradually decreased as the rates increased. However, at a high rate of 20C, the specific discharge capacity was still 114 mAh/g, which was about 75% of the capacity at 1C.

Conclusion

The heterogeneous Li4Ti5O12 nanostructures were synthesized via the solvothermal reaction in an ethanol/water mixed solvent with subsequent calcination at 500–700°C. The heterogeneous nucleation of the Li4Ti5O12 nanocrystals was induced, and the size of the nanoparticles on the surface of the wire was modified by changing the dielectric constant of the solvent during the solvothermal reaction.

The heterogeneous 0D-1D Li4Ti5O12 nanostructures calcined at 700°C exhibited a higher lithium storage capacity of 125 mAh/g after 500 cycles at 1C and a superior cycle performance (114 mAh/g) even at a high rate of 20C. The enhanced reversible capacity and cycling performance were attributed to the large surface area of the 0D-1D heterogeneous nanostructures, which enabled the use of the Li4Ti5O12 nanowires as potential host materials for high-powder Li ion batteries.