Chinese Journal of Chemical Physics  2017, Vol. 30 Issue (1): 103-111

#### The article information

Hai-lin Zou, Xin Liang, Zhong-hui Wang, Sheng Cheng, Hong-fa Xiang

Preparation of Li4Ti5O12 Microspheres with a Pure Cr2O3 Coating Layer and its Effect for Lithium Storage

Chinese Journal of Chemical Physics, 2017, 30(1): 103-111

http://dx.doi.org/10.1063/1674-0068/30/cjcp1607152

### Article history

Accepted on: October 7, 2016
Preparation of Li4Ti5O12 Microspheres with a Pure Cr2O3 Coating Layer and its Effect for Lithium Storage
Hai-lin Zoua, Xin Lianga, Zhong-hui Wanga, Sheng Chengb, Hong-fa Xianga
Dated: Received on July 31, 2016; Accepted on October 7, 2016
a. School of Materials Science and Engineering, Hefei University of Technology, Hefei 230009, China;
b. Instrumental Analysis Center, Hefei University of Technology, Hefei 230009, China
*Author to whom correspondence should be addressed. E-mail:hfxiang@hfut.edu.cn,Tel.:+86-551-62901457,FAX:+86-551-62901362
Abstract: The pure Cr2O3 coated Li4Ti5O12 microspheres were prepared by a facile and cheap solutionbased method with basic chromium(III) nitrate solution (pH=11.9). And their Li-storage properties were investigated as anode materials for lithium rechargeable batteries. The pure Cr2O3 works as an adhesive interface to strengthen the connections between Li4Ti5O12 particles, providing more electric conduction channels, and reduce the inter-particle resistance. Moreover, LixCr2O3, formed by the lithiation of Cr2O3, can further stabilize Li7Ti5O12 with high electric conductivity on the surface of particles. While in the acid chromium solution (pH=3.2) modification, besides Cr2O3, Li2CrO4 and TiO2 phases were also found in the final product. Li2CrO4 is toxic and the presence of TiO2 is not welcome to improve the electrochemical performance of Li4Ti5O12 microspheres. The reversible capacity of 1% Cr2O3-coated sample with the basic chromium solution modification was 180 mAh/g at 0.1 C, and 134 mAh/g at 10 C. Moreover, it was even as high as 127 mAh/g at 5 C after 600 cycles. At-20℃, its reversible specific capacity was still as high as 118 mAh/g.
Key words: Li4Ti5O12    Cr2O3    Lithium-ion battery    Solution-based method
I. INTRODUCTION

Safe, high-power and long-life lithium ion batteries (LIBs) are essential for renewable energy and new energy vehicles [1-5]. Among the novel candidates of anode, spinel Li4Ti5O12 (LTO) is the most promising material for large scale LIBs. It exhibits durability and safety characteristics with zero-strain structural and no lithium dendrite formation during the lithium insertion/extraction process [6-10]. Nevertheless, the low electronic conductivity (merely 10-13 S/cm) of LTO has severely limited its kinetic performance, especially under the rapid charge and discharge conditions [11-13]. Various nanostructures of LTO materials (0 to 3D) [14-21] can increase the electrolyte-electrode contact area and shorten the diffusion path of ions and electrons, resulting in great improvement of performance [22, 23]. Introducing conductive carbon materials, like graphene [24-26], carbon nanotubes (CNTs) [27, 28], can also optimize the electrochemical performance of LTO. However, the above strategies lead to the problem of low tap density.

Still, it remains a great challenge to obtain high tap density and high power density Li4Ti5O12 anode with desirable architectures. In real-world applications, the nanosized LTO primary particles have to be aggregated into secondary microspheres to obtain high tap density [29, 30]. Moreover, the inside active material of the microspheres cannot form electric wiring with the conductive agent or a current collector so that poor electronic transfer between-in the primary particles will further limit the capacity and rate performance of the LTO material [3]. Therefore, to introduce a carbon-free conducting coating layer (i.e. transition metal oxide) on the surface of LTO microspheres is a valuable approach to obtain high tap density as well as high power density by improving conductivity among the particles.

Transition metal oxide as a coating layer to efficiently improve the electrochemical performance of LTO has been found [31, 32], especially for chromium oxide. Feng et al. modified LTO with an aqueous CrO3 solution, which had significant rate capability improvement [33]. For the modified samples with the CrO3 solution, there were Li2CrO4, Cr2O5 and anatase on the surface. The irreversible phase, Li3+xCrO4 and LiyCr2O5, obtained from Li2CrO4 and Cr2O5 with lithium reaction during the first discharge process, can stabilize Li7Ti5O12 with high conductivity and make the modified samples for high rate performance [33]. However, Cr6+$is carcinogenic, which will limit the practical commercial application of this method. In contrast, Cr2O3 is much friendlier to environment relative to high valence state chromium. In previous literatures, the Cr2O3 has been reported to modify LiMn2O4 [34] and LiNi1/3Co1/3Mn1/3O2 [35] cathode materials, which si-gnificantly improves the electrochemical performance of the samples. But to our best knowledge, effects of surface modification with Cr2O3 in the LTO were seldom studied. In this work, a basic chromium(Ⅲ) nitrate solution is used to modify the surface of Li4Ti5O12 microspheres, and obtain pure Cr2O3 coated samples with excellent electrochemical properties. The Cr2O3 layer not only strengthens the electric wiring between the particles, increases capacity, but also reduces the interparticle resistance, and improves the rate performance. Therefore, the Cr2O3 layer works like an adhesive interface. Besides, LixCr2O3, formed by the reaction of Cr2O3 and Li in the first discharge process, can also stabilize Li7Ti5O12 with high electric conductivity. Therefore, this is a simple, economic and efficient strategy to improve the electrochemical performance of Li4Ti5O12 anode. II. EXPERIMENTS A. Materials Cr(NO3)3·9H2O, aqua ammonia, were purchased from Sinopharm Chemical Reagent Co., Ltd. Li4Ti5O12 microspheres were purchased from Yinlong Energy (Zhuhai, Guangdong, China). All the chemicals were used as received without any further purification. B. Material preparation Cr2O3 modified Li4Ti5O12 materials were obtained by modification of basic chromium(Ⅲ) aqueous solution. Firstly, 2.00 g Cr(NO3)3·9H2O was dissolved in 70 mL strong aqua ammonia under magnetic stirring overnight. Then the deionized water was added until the mixture total mass reaches 100 g, and then the basic chromium solution was obtained. Next, the commercial Li4Ti5O12 powder was blended with different quantity of the basic chromium solution by hand grinding and mixing evenly. Afterwards, the mixtures were placed in an electric oven for drying and sintered using a muffle furnace at 370 ℃ for 2 h. The sample named “BA-x” means that Cr2O3 is obtained with xwt% basic chromium solution. For comparison, an acid chromium aqueous solution modified Li4Ti5O12 samples were also prepared and named “AC-x”. The formation of stable basic or acid chromium solution follows reaction (1) and (2), respectively.  (1)  (2) C. Material characterization The crystal structures of these Cr2O3 modified Li4Ti5O12 samples (with basic and acid chromium solution) were characterized by powder X-ray diffraction (XRD) with Cu Kα radiation. Both scanning electron microscopy (SEM, JSM-6390LA, JEOL) and transmission electron microscopy (TEM, FEI Titan 80-300) were used to determine the morphology features and particle size of these samples. The chemical states at the surface of these samples were confirmed by X-ray photoelectron spectroscopy (XPS). D. Electrochemical analysis The electrochemical properties of Cr2O3 modified Li4Ti5O12 samples were performed in half coin cells (CR2032-type). The electrode laminates were formulated with modified Li4Ti5O12, carbon black and polyvinylidene fluoride at 8:1:1 weight ratio, using aluminium foil as a current collector. The electrode laminates were stamped into 14 mm diameter wafer which load about 3 mg active materials. The electrode wafer and high-purity lithium foil were assembled into half coin cell, separated by polypropylene membrane (Celgard 2400) and filled with electrolyte of 1 mol/L LiPF6/ethylene carbonate (EC)+diethyl carbonate (DEC) (1:1, w/w) in an argon-filled glove box (MBraun). Charge-discharge tests were performed using a multi-channel battery cycler (Arbin BT2000) at the voltage range of 1.0-2.5 V (vs. Li/Li+) with different C-rate (1 C was set as 175 mA/g). Electrochemical impedance spectroscopy (EIS) was recorded by a CHI 604 electrochemical workstation (Chenhua Corp.) in the frequency ranged from 10$.5\$ Hz to 0.1 Hz and the tested coin cells were charged to 2.5 V.

III. RESULTS AND DISCUSSION A. Structure and morphology features of the Cr2O3 modified Li4Ti5O12 samples

The powder X-ray diffraction patterns of the Li4Ti5O12 materials after the treatment with different amounts of chromium solution are shown in Fig. 1. The XRD pattern of the pristine Li4Ti5O12 shows that the material is pure spinel phase in Fig. 1(a). And there is no significant change in the XRD patterns for the BA-x (x=0.5, 1, 2, 4) samples relative to the pristine one. Apart from the peaks from Li4Ti5O12, some weak peaks corresponding to Cr2O3 appeared in the BA-8 sample (Fig. 1(a)). This is ascribed to the decomposition of chromic nitrate with the heat treatment at 370 ℃ (Fig.S1 in supplementary materials), and the decomposition reactions can be proposed as follows:

 FIG. 1 XRD patterns of Li4Ti5O12 samples treated with different amounts of Cr solution and LTO. (a) Basic solution and (b) acid solution.
 (3)

As shown in Fig. 1(b), after the treatment with acid chromium solution and sintering, the crystal structure of Li4Ti5O12 does not change, but a lot of impurity peaks appeared. It is found that, in addition to the peaks of Cr2O3, the other weak peaks appeared in the acid solution treated samples, which can be corresponding to Li2CrO4 and TiO2. First of all, the lithium ions of Li4Ti5O12 moved to the aqueous solution due to hydrogen ion exchange function, thus H4Ti5O12-layer is formed on the particle surface. Then after the heat treatment at 370 ℃, H4Ti5O12 can be decomposed into TiO2 anatase. At the same time, trivalent chrome is oxidized to hexavalent chrome by nitrate under high temperature conditions, and combined with lithium ions to form stable Li2CrO4 [36]. The actions can be proposed as follows:

 (4)
 (5)
 (6)

The effects of different chromium solution for Li4Ti5O12 surface are illustrated in Fig. 2. TiO2 is in situ produced from Li4Ti5O12 surface due to the route of acid solution, which connects more closely with Li4Ti5O12 particles and prevents Li4Ti5O12 from contacting with Cr2O3 and Li2CrO4. It is known that the conductivity of TiO2 is very low [33], thus the existence of TiO2-layer will lead to worse electric wiring between the particles of AC-x samples. On contrast, in basic Cr solution, Cr3+ could also be oxidized to high valence state of chromium, but the high valence state chromium oxide is not stable at high temperature, which can be broken down into Cr2O3 [37]. Therefore, the pure Cr2O3 modified Li4Ti5O12 were obtained by the basic chromium solution-based method.

 FIG. 2 Illustration of the formation mechanism of different Cr solution (pH=11.9 and pH=3.2) treated spinel Li4Ti5O12.

The morphologies and particle sizes of the samples before and after the modification with basic Cr solution observed by SEM are shown in Fig.S2 (supplementary materials). The primary particle size distribution of the pristine Li4Ti5O12 sample is uneven with 100-500 nm, and through the granulation process make the particles grow into 4-10 diameter spherical secondary particles (Fig.S2(a) in supplementary materials). After the modification, most of the spherical secondary particles were held, although grinding can make the secondary particle break, and the primary particle size did not change obviously (Fig.S2 (b) and (d) in supplementary materials). This suggests that these samples can keep the tap density of the initial microspheres sample. Furthermore, the particle surface of BA-1 is more smooth and rounded than LTO, which reflects the deposition on the surface of Cr2O3 is very small and uniform.

B. Electrochemical performance

Galvanostatic charge (Li+ extraction)-discharge (Li+ insertion) measurements were used to evaluate the lithium storage performance of Li4Ti5O12 samples treated with varying number of basic (BA-x) Cr solution and the pristine one (LTO) (Fig. 3). As shown in Fig. 3(a), the reversible capacity at 0.1 C reaches 166, 176, 180, 173, and 166 mAh/g for the pristine LTO, BA-0.5, BA-1, BA-2 and BA-4, respectively. Obviously, the capacities of BA-0.5 and BA-1 are significantly improved compared to the pristine sample. This is mainly due to Cr2O3 coated on the surface of the particles, which improved their conductivities, render some LTO particles electrochemical activities. However, the specific capacity decreases as the content of Cr2O3 increases, which is ascribed to the low capacity of Cr2O3 in 1.0-2.5 V (Fig.S3(a) in supplementary materials). It is obviously that the first discharge curves of the modified LTO with basic chromium solution contain two plateaus at about 1.8 and 1.55 V (Fig. 3(a)). To further understand the source of these platforms, the galvanostatic curves of LTO and BA-4 in first cycle are plotted through differential capacity (dQ/dE) at 0.1 C. As shown in Fig. 3(b), most capacity is in the region around 1.55 V, corresponding to the two-phase transformation between Li4Ti5O12 and Li7Ti5O12 [39]. In addition, there are two irreversible peaks on initial reduction around 1.85 and 2.32 V (vs. Li/Li+) for BA-4 sample, which disappear in the subsequent cycles. The peaks corresponding to the plateau of 1.8 V in Fig. 3(a) are assigned to the lithium insertion into Cr2O3 on the surface of Cr2O3-modified LTO particles (the platform of 1.75 V on the first discharge curve of Cr2O3 in Fig.S3(a) in supplementary materials). For the AC-x samples, the peaks of the lithium insertion into Li2CrO4 and TiO2 [40] also can be found from Fig.S3(b), Fig.S4(a) and Fig.S4(b) in supplementary materials). These results are corresponding to the XRD results in Fig. 1.

 FIG. 3 Electrochemical performance of Li4Ti5O12. (a) Voltage profiles of Li4Ti5O12 samples treated with different amounts of basic Cr solution and the pristine one (LTO) at C/10 rate. (b) Incremental capacity plot of the first cycle for LTO and BA-4. (c) Rate performances of Li4Ti5O12 samples treated with different amounts of basic Cr solution at different charge-discharge rates. (d) Cycling performance of Li4Ti5O12 samples treated with different amounts of basic Cr solution and LTO at 5 C.

Figure 3(c) shows the rate capabilities of Li4Ti5O12 samples treated with varying number of basic Cr solution (BA-x) and the pristine one (LTO). The pristine LTO shows low rate capacity, which is about 154 mAh/g at 1 C and only 88 mAh/g at 10 C. The main reason is poor electronic conductivity and large particle size of the pristine LTO. After the modification with the basic chromium solution, the specific capacities of the BA-x samples significantly improved under different rate (Fig. 3(c)). Particularly, the reversible capacities of BA-1 keep high capacity retention with increasing the C-rates. Even at 10 C, it still remains high capacities of 134 mAh/g (75% initial charge capacity at 0.1 C), which is much higher than that of the pristine one. But when the Cr2O3 amount was increased to 4%, the lithium storage performance of modified lithium titanate became worse. This may be because Cr2O3 does not contribute an evident reversible capacity and the excessive deposition will hinder the lithium ion diffusion. Obviously, only a moderate amount of Cr2O3 (less than 2%) can significantly improve the ability of rapid charge and discharge for Li4Ti5O12. As shown in Fig. 3(d), BA-1 exhibits excellent cycling stability compared to other products. After 600 cycles, BA-1 still delivers capacity of 127 mAh/g (retain the initial 84%) at 5 C. The good cycle performance is far better than that of the pristine LTO. Nevertheless, the cycle performance becomes worse when the amount of Cr2O3 increases to 4%, because Cr2O3 has poor cycle performance [41]. Figure S4 (in supplementary materials) displays the electrochemical performance of AC-x, which is worse than that of BA-x. This result proves that TiO2 can decrease the electrochemical performance of Li4Ti5O12 materials. So it is vital for the modification of Li4Ti5O12 materials to obtain pure Cr2O3 coating layer.

C. Further explore the role of Cr2O3 modification

As mentioned above, the BA-1 electrode exhibits a high capacity and good rate performance due to the presence of an appropriate amount of Cr2O3. In order to explore the role of Cr2O3 and find out the reasons for the improvement of electrochemical performance, BA-1 and LTO were chosen to carry out the subsequent study.

To further increase the magnification of SEM, new small grains can be seen in the narrow gap between the Li4Ti5O12 particles of Cr solution treated sample (red circle in Fig. 4(b)). HRTEM images provide a further explanation, as shown in Fig. 4 (c) and (d). It can be observed from the edge of the Li4Ti5O12 particles that the new grains are extremely small, with the size of 6-20 nm. The lattice spaces are counted to be 0.205 nm, with the lattice plane (202) of hexagonal Cr2O3 matching. In order to investigate the distribution of Cr2O3, the EDS element maps of BA-1 are shown in Fig. 4(e) and Fig. 4(f). In Fig. 4(e), the red spots correspond to the presence of the element oxygen, the green spots correspond to the element titanium, and the white spots correspond to the element chromium, in which Ti is the element from LTO and Cr is from Cr2O3. These results show that Cr is distributed uniformly throughout the whole area of the BA-1 composite, which indicates that Cr2O3 uniformly coats on the Li4Ti5O12 particles surface. According to the above account, in addition to the small Cr2O3 particles between the LTO, Cr2O3 also evenly coated on the surface of LTO particles.

 FIG. 4 SEM images of (a) LTO and (b) BA-1. (c) TEM images, (d) HRTEM images, (e) the maps, and (f) EDS of BA-1.

To further study the valences of the elements, the X-ray photoelectron spectroscopy was used to record the Ti2p spectra for the pristine LTO and BA-1 samples before and after the first cycle (Fig. 5). As shown in Fig. 5(a), there are two peaks at 464.9 and 459.2 eV corresponding to the Ti2p1/2 and Ti23/2 binding energies of Ti4+, respectively [42, 43]. It is evident that most of the peak area belongs to Ti4+, only a small peak corresponds to Ti3+ with a content of 4.31%. After the 1st cycle, the content of Ti3+ increased to 19.69% (Fig. 5(b)). the content of Ti3+ increased to 19.69% (Fig. 5(b)). However, the binding energies before cycled BA-1 are 464.2 eV (Ti2p1/2) and 458.6 eV (Ti2p3/2), which are approximately 0.7 and 0.6 eV lower than those in pristine LTO, respectively (Fig. 5(c)). Song et al. reported that Cr3+ ions can provide electrons to Ti4+ ions, to change the charge distribution when Li4Ti5O12 is Cr doped [44]. Therefore, we believe Cr2O3 have a similar effect on the surface of Li4Ti5O12. And then the binding energy shift disappears after cycling. Additionally, the Ti3+ peak at 457.3 eV in Fig. 5(d) can be calculated with the content of 40.13%, much higher relative to the pristine LTO, indicating that the Cr2O3-modified Li4Ti5O12 have higher content of Li7Ti5O12 than the pristine one. According to the results reported by Feng et al., LixCr2O5, which is from the reaction of Cr2O5 and Li, can stabilize a thin layer of Li7Ti5O12 which is in close contact with the LixCr2O5 shell, and will not turn back to Cr2O5 when charged to 2.5 V [33]. Herein, it is proposed that LixCr2O3 can also stabilize a thin layer of Li7Ti5O12, because LixCr2O3 and LixCr2O5 have similar behavior in discharging and charging at the voltage range of 1-2.5 V. The Cr2p XPS spectra of BA-1 sample before and after the first cycle are also shown in Fig.S5 (supplementary materials). In the BA-1 sample, the content of Cr2O3 is only 1%, and x is very small for the LixCr2O3. Therefore, the Cr valence state changes are not obvious in the XPS spectra.

 FIG. 5 Ti2p XPS spectra: (a) LTO sample before cycling. (b) LTO sample after the first cycle and being charged to 2.5 V vs. Li/Li+. (c) BA-1 sample before cycling. (d) BA-1 sample after the first cycle and being charged to 2.5 V vs. Li/Li+.

It is found that Cr2O3 changes into LixCr2O3 from the first discharge curve (Fig. 3(a)). If it is further discharged (cut-off voltage less than 1.0 V), Cr2O3 will change into CrO with poor electronic conductivity [45]. So the evaluation of the impedance change is important. EIS of BA-1, BA-4, and LTO electrode at various potentials from initial to 1.8 V during the first discharge process upon Li-insertion are shown in Fig.S6 (supplementary materials). The Nyquist plots of all samples consist of two parts: the semicircle in the high-to-medium frequency region corresponds to the charge transfer; the line in the low frequency region is Warburg impedance [46, 47]. Obviously, the Cr2O3-modified Li4Ti5O12 always has low charge transfer impedance from initial to 1.8 V, and the charge transfer impedance decreases along with the Cr2O3 content increases. While the impedance of all samples decreases with decreasing voltages, the impedance of BA-4 decreases the fastest, which indicates LixCr2O3 has better conductivity compared with Cr2O3. Here it is clear that the Cr-modified Li4Ti5O12 has excellent electrochemical performance, which benefits from Cr2O3 adding to reduce the charge transfer impedance.

In order to further study the electrochemical behavior of BA-1 and LTO, cyclic voltammetry (CV) was used. At the scan rate of 0.1 mV/s, the two samples have similar redox peaks at 1.5/1.7 V were demonstrated (Fig. 6(a)). While the polarization potential of the BA-1 is lower than that of LTO. Additionally, BA-1 has narrower half peak width. Figure 6(b) and Figure 6(c) show the CV curves at different scan rates for the pristine LTO and BA-1. As seen, the polarization potential increases with the increase of scan rate. However, the peak current of BA-1 is higher than LTO at any sweep rates. These facts indicate better charge transfer kinetics of BA-1 and correspond well with its better rate performance. Moreover, according to previous research results, there is a linear relationship between the peak current and the square roots of scan rates in the CV curve, and the slope can reflect the diffusion of lithium ion [48, 49]. As shown in Fig. 6(d), the diagonal of the BA-1 anode (red line) is much steeper, which suggests that the apparent Li+ diffusion is faster. This also means that the improvement of the apparent lithium ion diffusion of the Cr2O3-modified LTO is attributed to the improved electrical conductivity.

 FIG. 6 (a) CV curves of LTO and BA-1 at 0.1 mV/s. CV curves of LTO (b) and BA-1 (c) at various sweep rates. (d) The relationship between the peak current (Ip) and the square root of scan rate (v1/2) of LTO and BA-1.

In special needs, the battery may be used in harsh environments, such as low temperature. Thus the low temperature performances of the LTO and the BA-1 have been investigated. As shown in Fig. 7, the pristine LTO exhibits discharge capacities of 156, 137, 117, and 89 mAh/g at 25, 0, -10 and -20 ℃, respectively. This is mainly due to both the reaction kinetics and the charge diffusion being lower with the operation temperature decrease. While, the BA-1 sample delivers higher discharge capacities of about 174, 168, 152, and 118 mAh/g at 25, 0, -10, and -20 ℃. Allen et al. proved that the resistance of inter-particle is the main limiting factor at low temperature [50]. So the EIS of LTO and BA-1 (charge to 2.5 V) at various temperatures were performed and shown in Fig.S7 (supplementary materials). At 25 ℃, before the cycling, the comparable electrode resistances are 149 and 140 Ω for LTO and BA-1 (Fig.S6(a) in supplementary materials), respectively. However, after the third cycle, the impedance of BA-1 decreases drastically to about 100 Ω, which is significantly lower than the impedance of LTO (133 Ω, Fig.S7(a) in supplementary materials). This can be related to form a stable Li7Ti5O12 layer for BA-1. With the drop of temperature, the charge transfer resistances increased fast, which indicates that the charge transfer reaction is obviously affected by low temperature. This conclusion agrees well with the finding of Yuan et al. [51]. The pristine LTO showed a larger electrode polarization than BA-1, and the gap is bigger and bigger along with the decreasing of the temperature. For example, they are 3110 and 1740 Ω for the pristine LTO and BA-1 at -20 ℃, respectively. This suggests that the Cr2O3 modification can effectively improve the low temperature reaction kinetics. From 10 ℃ to 20 ℃, the electrolyte conductivity decreased significantly [52], which leads to the sharp drop in capacities of BA-1. So if the electrolyte is optimized, the low temperature performance of BA-1 will be more excellent.

 FIG. 7 Voltage profiles of (a) LTO and (b) BA-1 anode at 0.2 C rate at different discharge temperature with a fixed charge temperature of 25 ℃.

The superior electrochemical performance of 1%Cr2O3-modified Li4Ti5O12 microspheres is attributed to three aspects (as illustrated in Fig. 8). Firstly, Cr2O3 nanoparticles make it easier to achieve electric wiring among the particles inside the Li4Ti5O12 microspheres, render some particles electrochemically activity, and gain higher specific capacities. Secondly, Cr2O3 acts as an adhesive interface to enlarge the contact area among Li4Ti5O12 particles through the indirect way, increase the electronic transmission channels, and improve the rate performance. Thirdly, a stable Li7Ti5O12 layer on the surface of particles after cycling for Cr2O3-modified Li4Ti5O12 microspheres is effective in improving the electric conductivity.

 FIG. 8 Schematic illustration of the effects on Cr2O3 modification in Li4Ti5O12 microspheres.
IV. CONCLUSION

In summary, the Cr2O3-coated commercial Li4Ti5O12 microspheres have been prepared by a facile and cheap solution-based method with chromium(Ⅲ) nitrate solution. After the modification with the acid chromium solution (pH=3.2), the co-existing Li2CrO4, Cr2O3, and TiO2 surface phase were found in the final product. TiO2 blocked electric wiring between particles, which seriously inhibited the electrochemical performance of Li4Ti5O12. While, the pure Cr2O3 phase was obtained in the basic chromium solution modification (pH=11.9). Cr2O3 acts as an adhesive interface to strengthen the connections among particles and provides more electric conduction channels, reducing the inter-particle resistance. LixCr2O3 can stabilize Li7Ti5O12 with high electric conductivity on the surface of particles. Among those samples, 1%Cr2O3 revealed much improved capacities, low-temperature performances and high rate capabilities over pristine Li4Ti5O12, which benefits from the Cr2O3 modification. A reversible capacity of 180 and 134 mAh/g at 0.1 and 10 C, respectively, and even 127 mAh/g was obtained after 600 cycles at 5 C. When the discharge test was at -20 ℃, its specific capacity is up to 118 mAh/g. Therefore, it is believed that Cr2O3 modification is a simple and economic strategy for enhancing the electrochemical properties and maintaining high tap density of LTO anode materials, which shows great potential for large-scale industrialization.

Supplementary materials: Figure S1 shows the XRD pattern of Cr2O3. Figure S2 shows the SEM images of LTO and BA-1. Figure S3 shows the voltage profiles of Cr2O3 and Li2CrO4 in the first cycle process. Figure S4 show the electrochemical performance of AC-x samples. Figure S5 shows the Cr 2p XPS spectra of BA-1 sample before cycling and BA-1 sample after the first cycle. Figure S6 shows the impedance spectra of BA-1, BA-4 and the pristine one at various potentials in first discharge process. Figure S7 shows the impedance spectra of BA-1 and the pristine one (LTO) at various temperature.

V. ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (No.51372060 and No.31501576).

Reference