Hierarchical carambola-like Li4Ti5O12-TiO2 composites as advanced anode materials for lithium-ion batteries

Hierarchical carambola-like Li4Ti5O12-TiO2 composites as advanced anode materials for lithium-ion batteries

Electrochimica Acta 195 (2016) 124–133 Contents lists available at ScienceDirect Electrochimica Acta journal homepage: www.elsevier.com/locate/elect...

4MB Sizes 1 Downloads 43 Views

Electrochimica Acta 195 (2016) 124–133

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Hierarchical carambola-like Li4Ti5O12-TiO2 composites as advanced anode materials for lithium-ion batteries Yu Zhanga , Yun Zhanga,* , Ling Huanga , Zhongfu Zhoub , Jingfeng Wangb , Heng Liua , Hao Wua,* a b

College of Materials Science and Engineering, Sichuan University, Chengdu 610064, China School of Materials Science and Engineering, Shanghai University, Shanghai 200444, China



Article history: Received 27 January 2016 Accepted 15 February 2016 Available online 18 February 2016

Hierarchically structured Li4Ti5O12-TiO2 (LTO-TiO2) composites are synthesized using a facile hydrothermal approach upon reaction time control. With control over the time of hydrothermal reaction at 18 h, a hierarchical dual-phase LTO-TiO2 composite with appropriate amount of anatase TiO2 can be obtained, and it possesses a uniform carambola-like framework assembled by numerous ultrathin nanosheets, which enable a relatively large specific surface area, along with abundant interlayer channels to favor electrolyte penetration. When used as anode materials for lithium-ion batteries, such carambolalike LTO-TiO2 composite exhibits remarkably improved capacity, high-rate capability, and cycling stability over other LTO-TiO2 samples, which are synthesized at different time of hydrothermal reaction. Specifically, it deliveries a discharge capacity as high as 115.1 and 91.2 mAh g1 at a very high current rate of 20 and 40C, respectively, while a stable reversible capacity of 171.7 mAh g1 can be retained after 200 charge-discharge cycles at 1C, corresponding to 88.6% capacity retention. The excellent electrochemical performances benefit from the unique hierarchical carambola-like structure together with the mutually complementary intrinsic advantages between LTO and TiO2. The robust and porous nanosheets-assembled LTO-TiO2 framework not only offers a shorter transport pathway for electron and Li-ion migration within this composite material, but also is able to alleviate the structure distortion during the fast Li-ion insertion/extraction process. The work described here shows that the hierarchical carambola-like LTO-TiO2 composite is a promising anode material for high-power and long-life lithiumion batteries. ã 2016 Elsevier Ltd. All rights reserved.

Keywords: Hierarchical Carambola-like LTO-TiO2 interlayer channels lithium-ion batteries

1. Introduction Rechargeable lithium ion batteries (LIBs) have been successfully adopted as power sources for portable electronic devices because of their high energy density and long cycle life [1,2], whereas the practical application of LIBs to (hybrid) electric vehicles and stationary electric energy storage largely depends on the electrochemical performance of the electrode materials, such as large specific capacities, high rate capability and good cycling stability [3,4]. Graphite is commonly used as an anode in commercial LIBs, but exhibits a poor rate performance due to its low Li diffusion coefficient and serious safety issues, because of potential solid electrolyte interphase (SEI) film formation. Hence,

* Corresponding authors at: Sichuan University, No. 24, South 1st Section, 1st Ring Road, Chengdu, Sichuan, 610065, P. R. China. Tel.: +86 028 85410272; fax: +86 028 85410272. E-mail addresses: [email protected].cn (Y. Zhang), [email protected] (H. Wu). http://dx.doi.org/10.1016/j.electacta.2016.02.092 0013-4686/ ã 2016 Elsevier Ltd. All rights reserved.

advanced materials with an excellent rate and a better safety capability are critical components for the next generation of LIBs. As an alternative material to a graphite anode, spinel Li4Ti5O12 (LTO) have been deemed as an ideal materials for large-scale LIBs owing to their advantages of low cost, environmental friendliness, improved safety and stability [5,6]. Compared with the conventional carbon/graphite anode material, on the one hand, LTO have a higher intercalation potential (1.55 V vs Li+/Li), which is much safer to avoid the formation of SEI film and lithium dendritic crystal [7]. On the other hand, the spinel LTO is a zero-strain insertion material, and it does not undergo significant structure change during lithiation [8,9], Nevertheless, the low ionic and electrical conductivity which will cause limited capacity and poor cycling performance have seriously hindered the practical electrochemical performance of the LTO anode materials [10–12]. To address the above issues, tremendous strategies have been proposed to improve the conductivity and rate capability of LTO. The first strategy is coating various carbon materials [13,14] or

Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

conductive metal elements [15] or both of carbon and metal to form a core-shell structure [16] to improve electrical conductivity. The other strategy relies on the exploitation of various LTO nanostructure such as sword-shaped [17], sawtooth-like [18], flower-like [19], bowl-type [20] and mesoporous structures [21]. Among them, hierarchically nanostructured LTO material such as hollow microsphere [22] and flower-like particles have received tremendous attention for LIBs. These hierarchical micro-/nanostructures assembled orderly by nanoparticles, nanowire and nanosheets in 1D, 2D and 3D space [23–25], which facilitate the size effect of nanomaterial and structure stability of micromaterial. These advantages have made a significant influence on the kinetics and thermodynamics of electrochemical reaction procedure towards the LTO anode materials. Chou et al. prepared nanoparticles-aggregated LTO microsphere, which showed good cycling stability of 130 mAh g1 up to 200 cycles at the current density of 4C [26]. Some researchers also investigate the effect of raw materials with special morphology on grain size and electrochemical performance of LTO product in detail. Li et al. [27] prepared nanosized LTO by solid state reaction using axiolitic TiO2 as titanium source instead of commercial TiO2. The axiolitic TiO2 which assembled by 10–20 nm nanoparticles was synthesised by solvothermal reaction with acetic and tetrabutyl titanate (TBT) as raw materials. The product spinel LTO showed excellent capacity of 120 mAhg1 at 10C even after 20 cycles. Recently, compositing LTO with other anode materials (e.g. Fe2O3 [28], graphene [29], and TiO2[30]) has been an important and efficient approach towards improving the performance of LTO electrode. TiO2 is a kind of attractive anode material due to its ability of fast lithium insertion/extraction, high insertion potential (2.0 V) and low volume change (3–4%) during the charge/discharge process. Therefore, mixing LTO with TiO2 is deemed to be an advisable way to improve the capacity. Rahman etal. synthesized LTO–TiO2 composites via molten salt method, which yields a high capacity (166 mAh g1at 0.5C) and good rate capability (110 mAh g1 at 10C) [31]. Wang et al. synthesized rutile-terminated dualphase LTO–TiO2 composites without other conductive materials [32]. These composites exhibited much higher capacity than that of pure LTO at various rates. Nevertheless, most of these researches involve a complicated synthesis process, by which the amounts of the LTO and TiO2 phases were not adjustable in the final products. In addition, few attempts have been made so far to elaborately fabricate LTO-TiO2 composite with hierarchical nanostructure and specific morphology, and hence there is still an immediate need for a facile and reliable approach to synthesize structurally welldefined LTO-TiO2 composites with tunable compositions that can be used as anode materials for LIBs. Taking these into consideration, herein, we fabricated a novel hierarchical LTO-TiO2 composite material using a facile and versatile hydrothermal method. Only by adjusting the time of hydrothermal reaction, a hierarchical dual-phase LTO-TiO2 composite with appropriate amount (46.8%) of anatase TiO2 can be obtained, and it possesses a uniform carambola-like framework constructed by numerous ultrathin nanosheets with an ordered microcrystalline structure, wherein a plenty of interlayer channels are presented, thus giving rise to a relatively large specific surface


area and a fast pathway for electron and ion transportation. Benefiting from the unique hierarchical structure, the prepared LTO-TiO2 composites as anode material manifest high reversible capacities, good rate capability and superior cycling performance for lithium-ion batteries. 2. Experimental 2.1. Synthesis of carambola-like LTO-TiO2 composites Hierarchical carambola-like LTO-TiO2 composites were obtained via a simple hydrothermal route, as shown in Fig. 1. Typically, 10 ml H2O2 aqueous solution (30%) was added into 150 mL of 0.5 M LiOH (A. R.) solution. Then, 4.14 g tetrabutyl titanate (TBOT, C. P.) was added into the solution. The H2O2 worked as chelating agent, and can also slow down the hydrolysis of TBOT. After stirring, a light yellow clear liquid can be obtained. Then it was transferred into a PTFE-lined autoclave at 130  C for 18 h. After cooling naturally, the light yellow precipitate was separated by filtration, followed by washing and drying at 80  C. Finally, the dried precipitate was sintered with a heating rate of 3  C/min and maintained at 500  C for 3 h in air. 2.2. Synthesis of urchin-like LTO-TiO2 composites For comparison, urchin-like LTO-TiO2 is synthesized in a similar manner, only by prolonging the hydrothermal reaction time to 24 h. 2.3. Characterizations The crystal structure was characterized by X-ray diffraction (XRD) using a Bruker DX-1000 diffractometer with a CuKaradiation in the 2u angular range 1070 at a scanning rate of 0.06 s1. The surface morphology of LTO-TiO2 was observed by scanning electron microscopy (SEM, Hitachi S-4800). Transmission electron microscope (TEM, JEOL JEM-2100F) and high resolution TEM with an accelerating voltage of 200 kV were used for morphology and structure analysis. The specific surface area and pore size distributions of samples were estimated by N2 adsorption–desorption analysis (Tristar 3020). The atomic ratio of Ti and O in samples were analyzed by using X-ray Fluorescence (XRF, XRF-1800) together with X-ray photoelectron spectroscopy (XPS, SXAM-800), where the binding energy is calibrated with C1s = 285 eV. 2.4. Electrochemical measurements Electrochemical properties were carried out with CR2032 coin cells. The lithium metal was used as counter electrode. The electrolyte was 1 M solution of LiPF6 dissolved in a mixture of ethylene carbonate and dimethyl carbonate (EC/DMC) (1:1). The working electrode mixture was prepared on a copper foil by using a doctor-blade method with a slurry composed of 80 wt% active material (LTO-TiO2), 10 wt% conductivity agent (acetylene black) and 10 wt% binder (polyvinylidene fluoride, PVDF). After being

Fig. 1. Schematic illustration of the synthesis of (a) carambola-like and (b) urchin-like LTO-TiO2 nanocomposites.


Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

dried at 120  C for 10 h in vacuum oven, the foil loaded with active material was cut into little circles with a diameter of 12 mm. The cells were assembled in a glove box filled with argon with both moisture and oxygen contents below 0.1 ppm. Galvanostatic charge and discharge experiment was performed in a potential range 1.0–3.0 V by battery system (Neware, China). Electrochemical impedance spectroscopy (EIS) was performed on a CHI660e electrochemical workstation (Chenhua, China), the frequency ranges from 100 kHz to 10 mHz. Cyclic voltammetry (CV) measurement was carried out and the potential ranges 1.0–3.0 V (vs. Li/Li+) at a scan rate of 0.1 mV s1. 3. Results and discussion By changing the hydrothermal reaction time, the resultant products with different morphology and structures can be obtained. To systematically analyze the influence of hydrothermal reaction time on the morphology and structure change of the subsequent product, a series of experiments with various hydrothermal reaction times were conducted. Fig. 2 displays the scanning electron microscopy (SEM) images of different samples after hydrothermally reacted for 12, 15, 18, and 24 h, respectively. Obvious morphological changes are observed by prolonging the reaction time, from an initial larger and thicker flower-like nanosheets at 12 h (Fig. 2a) to a final smaller urchin-like microflower with intersected thinner nanosheets at 24 h (Fig. 2d), indicating the evolution of the particle growth. In the evolution process, two other morphologies within 15 and 18 h reaction time are also observed and both of which show a nonspherical morphology. When hydrothermally treated for 15 h, as seen from Fig. 2b, the initial nanosheets-stacked spherical particles (12 h, Fig. 2a) undergo an anistropic growth and form a framework

structure composed by longitude, latitude and center axis further. More interestingly, by prolonging the reaction time to 18 h, the resulting nanostructure transforms to a novel unique morphology (Fig. 2c), which has never been reported previously. Through the amplified SEM image in Fig. 2c (inset), it can be seen that plenty of hierarchical ultrathin nanosheets, which grown orderly and symmetrically along the edges of the carambola-like body frame in different directions. The carambola-like product undergoes an obvious morphology transformation, and consequently it converts to a urchin-like particle after 24 h hydrothermal reaction, indicating a possible self-transformation, which may be corresponded to a typical Ostwald ripening process [33]. The average size of a completely carambola-like particle is about 2.0–3.0 mm (Fig. 2c inset). These attractive ordered ultrathin nanosheets are expected to provide extra diffusion path and storage sites for Li-ions. The average diameter of each urchin-like particle is measured to be about 1.0 mm, inserted by lots of ultrasmall nanosheets on its surface disorderly (Fig. 2d inset). Clearly, the size of carambola-like LTO-TiO2 is bigger than that of urchin-like LTO-TiO2, indicating a biggish steric hindrance of the carambola-like LTO-TiO2, which would be reluctant to agglomerate and thus could facilitate the contact between the electrode and electrolyte. Whilst the urchin-like particles with a smaller diameter of about 1 mm is inclined to gather together to form a non-uniform morphology as compared to the carambola-like particles, as shown in Fig. 2d. It should be noticed that the carambola-like morphology is one of the middle states of the particle growth process. We speculate that as no mechanical or magnetic stir was applied during the hydrothermal treatment, the concentration of reactants in the vessel were in some degree nonhomogeneous, resulting in a non-uniform speed of the crystal nucleation and growth process. While the majority grew

Fig. 2. SEM images of the LTO-TiO2 samples obtained at different hydrothermal treatment time: (a) 12, (b) 15, (c) 18, and (d) 24 h.

Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133


Fig. 3. EDS analysis of the LTO-TiO2 samples obtained at hydrothermal treatment time of (a) 15, (b) 18, and (c) 24 h.

homogenously and evenly, and then formed to be uniform particles with carambola-like morphology, after reacted for 18 hours, a small amount of particles might just start to transform from the former morphology (Fig. 2a and b). That is the reason why tiny inhomogeneous morphology in the carambola-like products can be still observed. As the reaction progressed, all particles were eventually grown into urchin-likes LTO-TiO2. During the hydrothermal process, firstly, the TBOT hydrolyze as following: Ti(OC4H9)4 + 4H2O ! Ti(OH)4 + 4C4H9OH

Ti(OH)4 ! TiO2 + 4H2O Then, the TiO2 crystals began to be reacted with LiOH and transformed into LTO. In order to know whether the framework and those sheets grow in small area have the same constituent, energy dispersive spectrometer (EDS) was carried out. As shown in Fig. 3, the titanium and oxygen atomic ratio has a trend of decrease as the particle morphology transformed. It is well known that the titanium and oxygen atomic ratio of TiO2 and LTO is 1:2 and 1:2.4 respectively. It can be speculated qualitatively that the main component of those framework grown firstly is TiO2, while the sheets grow followed in the small area are mainly composed of LTO. As the reaction continuing, the TiO2 crystals of the framework react with LiOH and transform into Li+-containing precursor. In order to understand the reason why the atomic ratio of Ti and O is

smaller than theoretical value, the samples are put in the vacuum oven and keep vacuumizing for 10 hours, then they were characterized by X-ray Fluorescence (XRF). The results are shown in Table 1. It is obvious that after continuing vacuumizing, the content of oxygen decreased and the changing trend of mole ratio is the same as that of EDS analysis. It is speculated reasonablely that the extra oxygen is attributed to the O2 adsorbed on the surface of samples which is produced by H2O2 and existing in the air. X-ray photoelectron spectroscopy (XPS) was used to further confirm the existence of absorbed O2. The XPS spectrum in shown as Fig. 4. The O 1s peak at about 529.5 ev represents the Ti-O bond in LTO or TiO2, the peak at about 531.6 ev assigns to the chemisorbed oxygen on the surface of crystallites. It is obvious that there are chemisorbed oxgen existing on the face. X-ray diffraction was carried out to confirm the component and relative content of each crystal form. The XRD patterns of these samples are shown in Fig. 5. It is obvious that there are two phases, anatase TiO2 (JCPDS, PDF#21-1272) and spinel LTO (JCPDS, PDF#49-0207), existing in the four samples. All of them are well-crystalline. According to the results of the XRD patterns, as Table 1 XRF analysis of samples obtained at hydrothermal treatment time of (a) 15, (b) 18, and (c) 24 h. Samples

15 h

18 h

24 h






Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

Fig. 4. The survey XPS of the LTO-TiO2 samples obtained at hydrothermal treatment time of (a) 15, (b) 18, and (c) 24 h.

the extension of reaction time, the peaks that assigned to anatase TiO2 phase (such as the (101), (004), and (200) facets, etc.) in the composite become weakened gradually. However, the content of spinel LTO phase like (111), (400), and (333) facets, shows an opposite trend of steady increase in peak intensity. The relative content is calculated according to the following equation: WA ¼




K BA ¼



In the equation, IA and IB means the integrated intensity of the strongest peak, RA and RB represents for the K-value of each crystal form. It can be calculated that the content of anatase TiO2

decreased from 46.87% in the carambola-like LTO-TiO2 to 23.51% in the urchin-like LTO-TiO2. This result is consistent with the inference of above EDS analysis. TEM and high resolution TEM (HRTEM) were further performed to examine the morphology and composition of LTO-TiO2. Fig. 6 shows the detailed microstructure of these samples. As displayed in Fig. 6a, the particle consists of uniform nanosheets, with each of them stacking hierarchically and separately, which is in consistent with the SEM image. Compared with the carambola-like LTO-TiO2, as shown in Fig. 6a and 6c, there are thinner and smaller nanosheets intercalating on the surface of the urchin-like LTOTiO2. Fig. 6b and 6d present the HRTEM images of the carambolalike and urchin-like LTO-TiO2, respectively. The clear lattice fringes of about 0.48 and 0.36 nm correspond to the (111) and (101) interplanar spacing of spinel LTO and anatase TiO2, respectively. What should be noticed is that, in the carambola-like LTO-TiO2

Fig. 5. XRD patterns of the LTO-TiO2 samples under different reaction time.

Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133


Fig. 6. TEM images and HRTEM images of (a, b) carambola-like and (c, d) urchin-like LTO-TiO2.

(Fig. 6b), the lattice of anatase TiO2 and spinel LTO grow together, which is not found in the edge of the urchin-like one (Fig. 6d). We attribute this to the distinct growing process of particles. During the reaction, the anatase TiO2 crystals on the surface first contact

with LiOH solution and transform into LTO, which would then surround the crystals inside the particles, preventing them from further contacting and reacting with LiOH. Therefore, the component of nanosheets on the surface of urchin-like LTO-TiO2

Fig. 7. N2 adsorption-desorption curves and pore size distribution of LTO-TiO2 samples:(a) 12, (b) 15, (c) 18, and (d) 24 h.


Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

are more likely to be spinel LTO structure, whilst the internal part includes two kinds of crystal, spinel LTO and anatase TiO2. This is also in correspondence with the XRD results. In addition to the favourable hierarchical structure, this coexisting crystal in carambola-like LTO-TiO2 could also show a synergistic effect in the progress of Li-ion migration [34], by which the ions can diffuse between spinel LTO lattice and anatase TiO2 lattice easily. The nitrogen adsorption-desorption isotherm was then performed to analyze the pore structure of samples. Fig. 7 shows typical type IV curves with H2 hysteresis loop in the area of relative pressure 0.6–1.0 P/P0, demonstrating the co-existence of mesopores and macropores in all four samples. Table 1 summarizes the Brunauer-Emmett-Teller (BET) specific surface area and pore-size distribution for the as-prepared samples. The BET specific surface area of carambola-like LTO-TiO2 is estimated to 72.8 m2/g, and that of urchin-like LTO-TiO2 is 57.2 m2/g. The relatively larger specific area is expected to offer more active insertion site [35]. The Barrett-Joyner-Halenda (BJH) pore-size distribution (inset) indicates that the pores have a relatively wide size distribution between 2.3 and 118 nm, implying that there are mesopores and macropores presented in the carambola-like LTO-TiO2, in which the mesopores are mainly formed by the aggregation of nanosheets, while the macropores arise from the accumulation of particles. The hierarchical pore structure leads to abundant open channels that is favourable to promote electrolyte penetration within the LTO-TiO2 active materials [36]. To gain a further insight into the electrochemical performance of the LTO-TiO2 composite, cyclic voltammogram (CV) and galvanostatic charge/discharge testing are employed to explore the kinetic processes. Fig. 8a is the CV curves for the two electrodes made from the carambola-like LTO-TiO2 and urchin-like LTO-TiO2, respectively. The CV test was carried out between 1.0–3.0 V (vs. Li/ Li+) at a scan rate of 0.1 mVs1. It is clear that there are two pairs of redox peaks in both electrodes. The redox peaks at approximately 1.5/1.7 V is assigned to spinel LTO. The other pair of redox peaks at

1.7/2.0 V can be ascribed to the lithium ion insertion/deintercalation of anatase TiO2. Notably, in comparison with the carambolalike sample, the urchin-like one obviously shows relatively weak redox peaks arising from anatase TiO2, due to its fewer TiO2 content (the inset in Fig. 8a). Moreover, the potential difference between each pair of oxidation/reduction peak of these samples is almost equal, suggesting that rare polarization reaction would happen. In addition, after three circles scanning, the potential of redox peaks almost show no shift, indicating that the electrode materials are highly reversible and stable enough. Fig. 8b shows the initial three discharge-charge profiles of carambola-like LTO-TiO2 at 0.1C (1C = 175 mA g1) between 1.0 and 3.0 V. The potential of two platforms in each charge/discharge curve are almost close to the potential of redox peaks in CV curves. The initial discharge/charge capacites are 281.5 mAh g1/ 200.4 mAhg1. The discharge/charge capacities of second cycle are approximately 227.0 mAhg1/200.1 mAhg1, with the coulombic efficiency improved. After the first cycle, the irreversible loss is 81.1 mAh g1, which may be ascribed to the reduced number of electron transferred per formula unit [37]. Comparing the profiles, it would be found that the irreversible loss occurs in the region between 1.55 V and 1.0 V. This region is assigned to the pseudocapacitive lithium ion storage behavior of LTO, which indicates the different electrochemical reaction characteristics including surface-confined charge-transfer or interfacial storage [38–40]. In comparison, the aforementioned other two electrodes, based on the LTO-TiO2 composite at 12 and 15 h hydrothermal treatment, exhibit a much lower initial discharge capacity. Cyclic performances of series LTO-TiO2 samples are shown in Fig. 8c and d. Different from cathode materials, the coulombic efficiency of anode material is counted as charge capacity divided by discharge capacity. Clearly, the capacity of carambola-like LTO-TiO2 is the highest among all the four samples. For the carambola-like LTOTiO2 electrode, the first charge capacities are 194.9 mAh g1 at 1C, which is superior than that of the urchin-like LTO-TiO2 electrode

Fig. 8. (a) CV curves at a scan rate of 0.1 mV s1 for initial three cycles of carambola-like and urchin-like LTO-TiO2; (b) The initial charge-discharge curves at 0.1C of four samples; (c) Cyclic performance of four samples after 200 cycles at the rate of 1C; (d) Cyclic performance of crambola-like and urchin-like LTO-TiO2 after 200 cycles at the rate of 5C.

Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

(178.6 mAh g1). Even after 200 cycles, a stable capacity of 171.7 mAh g1 are still maintained, corresponding to a capacity retention ratio as high as 88.6% with only 0.06% capacity loss per cycle. In order to discuss the cyclic performance of carambola-like LTO-TiO2 at higher current density, the carambola-like LTO-TiO2 and urchin-like LTO-TiO2 are cycled at rate of 5C. As shown in Fig. 8d, the carambola-like LTO-TiO2 electrode also exhibits a capacity of 134.1 mAh g1 and maintains at 125.3 mAh g1 after 200 cycles, corresponding to a capacity retention ratio as high as 94.4% with only 0.03% capacity loss per cycle, which shows higher capacity and better cycling stability than the urchin-like LTO-TiO2 electrode (119.8 mAh g1 for the first cycle with 100.0 mAh g1 retained after 200 cycles). Fig. 9 shows the SEM images of carambola- and urchin-like LTO-TiO2 composites after 200 cycles at 1C. It can be seen from the images that the carambola-like LTOTiO2 still maintain the framework structure after 200 cycle, the ultra-thin nanosheets are still existing with a little breakage. The urchin-like LTO-TiO2 also keeps its morphology with the nanosheets on the surface blunts. Rate capability is also an important factor for Li-ion batteries in practical high-power application. Therefore, the samples were further tested by charging/discharging at various current rates from 0.2 to 40C and then back to 1C. As shown in Fig. 10a, the carambola-like LTO-TiO2 electrode also delivers a higher specific capacity than the urchin-like LTO-TiO2 electrode, and it still remains a capacity of 119.2, 115.1, 101.9 and 91.2 mAh g1 at a current rate of 10, 20, 30, and 40C, respectively. This rate capability of the carambola-like LTO-TiO2 composite in our work is competitive to the previously reported LTO-TiO2-based anode materials [41–46], as seen in Fig. 10b. When the current density was abruptly switched back to 1C, most of the original capacity was recovered, indicating the excellent stability and reliability of the carambola-like LTO-TiO2 composite structure. However, the urchin-like LTO-TiO2 electrode just maintains a low specific capacity of 72.4 mAh g1 at the rate of 40C. For comparison, the cyclic and rate performance of the other two electrodes, using the dual-phase LTO-TiO2 composites with higher anatase TiO2 contents (54.10% and 54.38%) as the anode materials, are performed, and the results are displayed. Obviously, the specific capacity and rate performance of the two electrodes are much poor than those of the carambola-like and urchin-like electrodes, indicating that higher introduced TiO2 (anatase) would lower the electrochemical activities of LTO. This result is in good agreement with that of other similar researches [30,31]. From the results above, we can see that carambola-like LTOTiO2 electrode has the best electrochemical performances among the four electrodes. The high specific capacity, excellent rate capability, and cycling life of the carambola-like LTO-TiO2 can be attributed to the following reasons. Firstly, the hierarchical


carambola-like structure composed of numerous nanosheets as well as abundant interlayer channels is more open and thus provide more paths for electrolyte infiltration and Li-ion migration. Secondly, benefited from the mutually complementary intrinsic advantages between LTO and TiO2, the transportation of Li-ions and electrons between the spinel LTO and anatase TiO2 could be more easily and rapidly, and therefore improving the capacity and rate capability of the electrode. Moreover, the higher specific surface area of the carambola-like structure, combined with the “zero strain” nature of spinel LTO, could consequently keep the unique carambola-like shape intact, preventing it from collapsing during long-term cycles, thus enabling excellent cycling stability [47,48]. Electrochemical impedance spectroscopy (EIS) measurements were further performed to shed light on the good electrochemical performance of carambola-like LTO-TiO2. The Nyquist plots are shown in Fig. 11a with a insertion of fitting circuit. The impedance curves show a semicircle in the high-frequency region and a line about 45 in the low-frequency region. The intercept at the real axis corresponding to the ohmic resistance (Rs), reflects the electronic conductivity of separator, electrolyte and anodes. The radius of semicircle resulted from the charge transfer resistance (Rct), while the sloping line in low frequency region is associated with the Warburg impedance (W0), which is correlated to the Liion diffusion in active materials. As can be seen from Fig. 10a, it is obviously that the carambola-like LTO-TiO2 shows a significant smaller semicircle than urchin-like LTO-TiO2, indicating a lower charge-transfer resistance (Rct) at the interface between electrode and electrolyte. Table 3 shows the fitted results, the Rs of carambola-like TiO2 is much smaller than that of urchin-like LTO-TiO2, demonstrating a faster electronic conductivity. As shown in Fig. 10a, the slope of line in low-frequency region of carambolalike LTO-TiO2 is smaller than that of urchin-like LTO-TiO2. That means the carambola-like electrode exhibits larger Li-ion diffusion rate, which means its smaller diffusion resistance. The Li-ion diffusion coefficient (DLi) is calculated from the EIS date according to the following equation: DLi ¼

R2 T 2 2A2 F 4 n4 C 2



In the equation, DLi is the diffusion coefficient of Li-ion, R is the gas constant, T is the absolute temperature, A is the surface area of the positive electrode, F is the Faraday’s constant, n is the number of transferred concentration in cathode material, C represents the concentration of lithium-ion and s is the Warburg factor which is obtained from the slope of Z’ vs. v1/2 (Fig. 11b). The results are listed in Table 2. It is obviously that the calculated result for carambola-like LTO-TiO2 is almost 2.78 times of that for urchin-like LTO-TiO2. This result indicates that such a unique hierarchical

Fig. 9. The SEM images of (a) carambola- and (b) urchin-like LTO-TiO2 after 200 cycles at 1C.


Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

Fig. 10. (a) Rate capability of four samples at various rates from 0.2 to 40C and then back to 1C; (b) Comparison of this work with others.

Fig. 11. (a) Electrochemical impedance spectra (Nyquist plots) of electrodes; (b) The relationship between Z0 and v0.5 at low frequency of carambola-like and urchin-like LTO-TiO2 at charge state; (c) Schematic illustration of the transportation of Li-ions and electrons in carambola- and urchin-like LTO-TiO2.

structure of LTO-TiO2 nanocomposites effectively reduces the barrier for Li-ion transfer, and thus increases the rate capability. Fig. 11c simulates the transportation process of electrons and Liions in two kinds of particles, illustrates the possible transportation routes. During the discharge process, the diffusion of Li-ion and electrons in both carambola-like and urchin-like LTO-TiO2 consists of two paths, described as the route 1 and route 2 in Fig. 11c. Route 1 refers to the transportation of Li-ion and electrons

from electrolyte to the surface of anode. It is known that the carambola-like particles has a larger specific surface area, which means a larger contact area with electrolyte. That is a possible cause of the lower Rct of carambola-like LTO-TiO2. Route 2 represents the transportation process of Li-ion and electrons through the active material. As for urchin-like LTO-TiO2, Li-ions and electrons diffuse through the solid core, while for the carambola-like LTO-TiO2, even the center of particles were exposed

Table 2 BET surface area and average pore size of samples synthesized by different reaction time.

Table 3 The fitted results from EIS.

Samples with different reaction time (hours)





BET surface area (m2/g) Average pore width (nm)

26.05 14.68

44.77 17.03

72.82 13.66

57.22 8.49

18 h 24 h




5.6 9.9

25.7 58.8

1.0  1013 3.6  1014

Y. Zhang et al. / Electrochimica Acta 195 (2016) 124–133

outside and contact with electrolyte, shorten the diffusion path. That is one of reasons why carambola-like has a higher DLi. Accordingly, the stable discharge capacities of carambola-like electrode can be reasonably attributed to the relatively lower polarization rate developed within the electrode, whereas the higher charge transfer resistance of urchin-like electrode would lead to capacity decay after long-term cycles at high rates. 4. Conclusion In summary, we have prepared a hierarchical carambola-like LTO-TiO2 composite based on a facile and versatile time-controlled hydrothermal synthesis strategy. The resultant dual-phase LTOTiO2 composite with appropriate amount (46.8%) of anatase TiO2 can be obtained, and it possesses a uniform carambola-like framework assembled by numerous ultrathin nanosheets, which enable a relatively large specific surface area (72.8 m2/g), along with abundant interlayer channels to facilitate electrolyte penetration as well as electron and ion transfer during the charge– discharge process. With this advanced architecture, high specific capacity (281.5 mAh g1 at 0.1C), excellent rate capability (115.1 and 91.2 mAh g1 at 20 and 40C) and good cycling stability (88.6% retention after 200 cycles) can be realized by using the hierarhical carambola-like LTO-TiO2 as anode material for lithium ion batteries. The works described here show a great potential in scalable development of higher rate LTO-based anode materials together with longer life towards next-generation LIBs. Acknowledgements The authors acknowledge the financial support from the National Basic Research Program of China (973 program no. 2013 CB934700), A Foundation for the Author of National Excellent Doctoral Dissertation of P. R. China (no. FANEDD201435), and the Sichuan Province Science and Technology Support Program(no. 2014GZ0093). References [1] H.W. Sisi Huang, Penghui Chen, Yi Guo, Bo Nie, Baojun Chen, Heng Liu, Yun Zhang, J. Mater. Chem. A (2015) 3. [2] O.K. Park, Y. Cho, S. Lee, H.-C. Yoo, H.-K. Song, J. Cho, Energy Environ. Sci. 4 (2011) 1621. [3] P. Gibot, M. Casas-Cabanas, L. Laffont, S. Levasseur, P. Carlach, S. Hamelet, J.M. Tarascon, C. Masquelier, Nat. Mater. 7 (2008) 741–747. [4] M.A.a.J.-M. Tarascon, Building better batteries, Nature 451 (2008) 6. [5] T.F. Yi, S.Y. Yang, Y. Xie, Journal of Materials Chemistry A 3 (2015) 5750–5777. [6] B. Zhao, R. Ran, M. Liu, Z. Shao, Materials Science and Engineering: R: Reports 98 (2015) 1–71. [7] Y. Wang, B. Liu, Q. Li, S. Cartmell, S. Ferrara, Z.D. Deng, J. Xiao, Journal of Power Sources 286 (2015) 330–345. [8] M. Wagemaker, D.R. Simon, E.M. Kelder, J. Schoonman, C. Ringpfeil, U. Haake, D. Lützenkirchen-Hecht, R. Frahm, F.M. Mulder, Adv. Mater. 18 (2006) 3169– 3173.


[9] M. Wagemaker, A. Van Der Ven, D. Morgan, G. Ceder, F.M. Mulder, G.J. Kearley, Chem. Phys. 317 (2005) 130–136. [10] Donghai Wang, Daiwon Choi, Juan Li, Zhenguo Yang, Zimin Nie, Rong Kou, Dehong Hu, Chongmin Wang, Laxmikant V. Saraf, Jiguang Zhang, Ilhan A. Aksay, A.J. Liu, ACS nano 3 (2009) 8. [11] Y.S. Hu, L. Kienle, Y.G. Guo, J. Maier, Adv. Mater. 18 (2006) 1421–1426. [12] A.P.M.K.M. Wagemaker, F.M. Mulder, Nature 418 (2002) 3. [13] M.G.J. Du Hoang Long, Yoon-Sung Lee, Wonchang Choi, Joong Kee Lee, In Hwan Oh, Hun-Gi Jung, ACS Appl. Mater. Interfaces 7 (2015) 10250–10257. [14] H.J.L. Guan Nan Zhu, Ji Hua Zhuang, Cong Xiao Wang, Yong Gang Wang, Yong Yao Xia, Energy Environ. Sci. 4 (2011) 4016–4022. [15] K. Wu, X. Lin, L. Shao, M. Shui, N. Long, Y. Ren, J. Shu, J. Power Sources 259 (2014) 177–182. [16] J.H. Minggui Wang, Huixin Xiong, Rong Guo, Yadong Yin, ACS Appl. Mater. Interfaces 7 (2015) 6909–6918. [17] L. Wu, J. Chen, X. Kong, X. Luo, S. Shi, S. Gou, S. Zhong, Mater. Lett. 143 (2015) 131–134. [18] J. Chen, L. Yang, S. Fang, Y. Tang, Electrochim. Acta 55 (2010) 6596–6600. [19] Z. Zhang, Z. Zhou, S. Nie, H. Wang, H. Peng, G. Li, K. Chen, J. Power Sources 267 (2014) 388–393. [20] H.J. Kim, J.D. Jeon, J.W. Chung, S.Y. Kwak, Microporous and Mesoporous Mater. 198 (2014) 170–174. [21] J. Qiu, C. Lai, Y. Wang, S. Li, S. Zhang, Chem. Eng. J. 256 (2014) 247–254. [22] S. Kim, S. Fang, Z. Zhang, J. Chen, L. Yang, J.E. Penner-Hahn, A. Deb, J. Power Sources 268 (2014) 294–300. [23] X.Y. Jin Xie, Sa Zhou, Dunwei Wang, ACS nano 5 (2010) 77. [24] A.C.B. Jianchao Ye, Y. Morris Wang, Juergen Biener, Monika M. Biener, ACS nano 9 (2015) 9. [25] J.L. Tiejian Zhu, Qingsheng Wu, ACS Appl. Mater. Interfaces 3 (2011) 6. [26] J.Z.W. Shulei Chou, Huakun Liu, Shixue Dou, J. Phys. Chem. C 115 (2011) 8. [27] X. Li, H. Hu, S. Huang, G. Yu, L. Gao, H. Liu, Y. Yu, Electrochimica Acta 112 (2013) 356–363. [28] M. Chen, W. Li, X. Shen, G. Diao, ACS Appl. Mater. Interfaces 6 (2014) 4514– 4523. [29] Y. Tang, F. Huang, W. Zhao, Z. Liu, D. Wan, J. Mater. Chem 22 (2012) 11257. [30] K.M. Kim, K.Y. Kang, S. Kim, Y.G. Lee, Curr. Appl. Phys. 12 (2012) 1199–1206. [31] M. Rahman, J.Z. Wang, M.F. Hassan, D. Wexler, H.K. Liu, Adv. Energy. Mater. 1 (2011) 212–220. [32] Y.Q. Wang, L. Gu, Y.G. Guo, H. Li, X.Q. He, S. Tsukimoto, Y. Ikuhara, L.J. Wan, J. Am. Chem. Soc. 134 (2012) 7874–7879. [33] T.T. Truong, Y. Liu, Y. Ren, L. Trahey, Y. Sun, ACS nano 6 (2012) 8067–8077. [34] Z. Yang, D. Choi, S. Kerisit, K.M. Rosso, D. Wang, J. Zhang, G. Graff, J. Liu, J. Power Sources 192 (2009) 588–598. [35] J.L. Allen, T.R. Jow, J. Wolfenstine, J. Power Sources 159 (2006) 1340–1345. [36] J. Liu, T.E. Conry, X. Song, M.M. Doeff, T.J. Richardson, Energy Environ. Sci. 4 (2011) 885. [37] J.Y. Liao, V. Chabot, M. Gu, C. Wang, X. Xiao, Z. Chen, Nano Energy 9 (2014) 383– 391. [38] C. Lai, Y.Y. Dou, X. Li, X.P. Gao, Journal of Power Sources 195 (2010) 3676–3679. [39] X. Li, C. Lai, C. Xiao, X. Gao, Electrochimica Acta 56 (2011) 9152–9158. [40] Y.M. Jiang, K.X. Wang, X.Y. Wu, H.J. Zhang, B.M. Bartlett, J.S. Chen, ACS applied materials & interfaces 6 (2014) 19791–19796. [41] L. Gao, S. Li, D. Huang, Y. Shen, M. Wang, Journal of Materials Chemistry A 3 (2015) 10107–10113. [42] W. Zhu, H. Yang, W. Zhang, H. Huang, X. Tao, Y. Xia, Y. Gan, X. Guo, RSC Advances 5 (2015) 74774–74782. [43] W. Zhang, Z. Liu, X. Xiao, D. Liu, ChemElectroChem (2015). [44] T.F. Yi, Z.K. Fang, Y. Xie, Y.R. Zhu, S.Y. Yang, ACS applied materials & interfaces 6 (2014) 0205–20213. [45] K.M. Kim, K.-Y. Kang, S. Kim, Y.-G. Lee, Current Applied Physics 12 (2012) 1199– 1206. [46] F. Wu, X. Li, Z. Wang, H. Guo, Nanoscale 5 (2013) 6936–6943. [47] J. Li, Z. Tang, Z. Zhang, Chem. Phys. Lett. 418 (2006) 506–510. [48] J. Wang, J. Polleux, J. Lim, B. Dunn, J. Phys. Chem. C 111 (2007) 14925–14931.