Improving the properties of Li4Ti5O12 - a promising anode material for lithium-ion batteries

Two materials with the stoichiometric composition Li3.85Ni0.15Ti5O12 and Li3.80Cu0.05Ni0.15Ti5O12 were obtained by solid-state reaction using lithium carbonate Li2CO3, titanium oxide TiO2, nickel oxide NiO and copper oxide CuO. The materials were characterized in terms of phase composition, crystal structure as well as cycle performance. Phase composition and crystal structure parameters were determined using X-ray Panalytical Empyrean XRD diffractometer in the range of 10-110° with CuKa radiation. The results were analyzed using Rietveld refinement which was then implemented in the GSAS computer software. The electrochemical properties of the samples were measured by galvanostatic charge/discharge cycles at different rates over a voltage range of 1.0-2.5 V and 0.2-2.5 V. Cyclic voltammetry measurements were also carried out. It was proved that the addition of both Ni and Cu results in high specific capacity of LTO especially at high current rates (2C and 5C). The sample Li3.80Cu0.05Ni0.15Ti5O12 delivers superior capacity above 200 mAh·g -1 when discharged to 0.2 V.


Introduction
Lithium-ion batteries are considered to be promising energy storage devices and are widely applied in e.g. portable electronic instruments or electric vehicles [1][2][3] because of their high energy and power density as well as long cycle-life [4][5][6]. The properties of electrode materials are crucial for the performance of the battery. Therefore, study of these materials is very important.
Regarding anodes, only materials based on carbon (especially graphite) or Li4Ti5O12 (LTO) are commercialized. Carbonic materials possess excellent electrochemical properties. However, they exhibit volume change during charge/discharge cycles which cause problems with safe usage of the battery. LTO is an attractive alternative to anodes based on carbon [7,8]. LTO has a high working potential (~1.55 V versus Li|Li + ) [9,10]. LTO is considered to be a zero-strain material, because during lithium intercalation and deintercalation the lattice parameter does not change almost at all [11]. Although the high working potential limits significantly the energy density, the operating potential occurs within the thermodynamic stability window of electrolytic solutions, so that it is not necessary to form a solid electrolyte interphase (SEI) layer for proper functioning of the electrode. Anodes based on Li4Ti5O12 exhibit long cycle-life, are resistant to overcharge and can be used in a wide temperature range [10]. Theoretical capacity of non-doped LTO is equal to 175 mAh•g −1 , because of insertion of 3 moles of lithium ion into LTO structure. Li4Ti5O12 has a stable [Li3] 8a [LiTi5] 16d [O12] 32e framework, where all tetrahedral (8a) sites and 1/6 of the 16d sites are taken by Li atoms, while the rest 5/6 of the 16d sites are occupied by Ti atoms. Oxygen atoms reside at 32e sites and the octahedral (16c) sites are empty [12,13]. After intercalation of lithium ions Li4Ti5O12 transforms to rock salt structure of Li7Ti5O12 which framework is presented by [Li6] 16c [LiTi5] 16d O12 [12]. Unfortunately, LTO has one significant disadvantage -it has low electronic conductivity (10 −13 S•cm −1 ) [14]. There are many ways to improve the electronic and ionic conductivity of Li4Ti5O12. It can be done by obtaining nano-materials [7,[14][15][16] doping with metal cations [17][18][19] or composing LTO with carbon [20].
The specific capacity of anodes based on Li4Ti5O12 is higher when the battery is charged in the voltage range of 2.5 to 0.01 V. In this case, the theoretical capacity of LTO is limited by the number of tetravalent titanium ions, but not the octahedral or tetrahedral sites to accommodate lithium ions in the voltage range of 2.5 to 0.01 V, corresponding to 293 mAh•g −1 , but not 175 mAh•g −1 [21].
In this study Li4Ti5O12 was doped with nickel to enhance the specific capacity at higher current rates. Another sample was doped with both nickel and copper for the same purpose. The samples were characterized by x-ray diffraction (XRD). What is more, the electrochemical properties were tested by charge-discharge cycling and cyclic voltammetry.

Synthesis of main electrode materials
The Li3.85Ni0.15Ti5O12 sample was obtained by a solid-state method using Li2CO3 (Aldrich, 99.8%), TiO2-anatase (Acros, 99 + %) and NiO (Aldrich, 99.8%). The starting materials were mixed at the appropriate molar ratio. 5% excess of lithium was provided to avoid the change in the composition because of the evaporation of lithium. The mixed reactants were ball-milled for 30 min with a propanol. After drying, powders were calcined at 800 °C for 10 h in the air. The Li3.80Cu0.05Ni0.15Ti5O12 sample was obtained by using a similar solid-state method mentioned above. In this case, however, the sample was prepared from a mixture of Li2CO3, TiO2-anatase, NiO and CuO (Aldrich, 99.8%).

Preparation of lithium-ion batteries
The electrochemical measurements were done using CR2032 coin-type cells. The working electrodes were prepared by making a black slurry containing 80 wt.% of active material, 10 wt.% of carbon black and 10 wt.% of polyvinylidene fluoride (PVDF) dissolved in N-methyl-2-pyrrolidone. The slurry was spread on an aluminium foil and dried at 70 °C for 1 h in a vacuum drying oven. After that disks with 8 mm diameter were punched out of the foil and roll-pressed. Metal lithium foil served as a counter electrode. 1 M LiPF6 solution in the 1:1 mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) was used as an electrolyte. Batteries were prepared in the glove-box (UNILAB, M. Braun) under argon atmosphere with controlled oxygen and water vapor pressure (<0.1 ppm).

Electrochemical and transport measurements
The electrochemical properties of the samples were measured by galvanostatic charge/discharge cycles at different rates over a voltage range of 1.0-2.5V and 0.2-2.5V. The C rate was calculated based on the weight of the electrode and theoretical capacity of LTO. Cyclic voltammetry measurements were carried out at different scanning rates in the range of 0.1 and 0.5 mV•s −1 and in 1.0-2.5 V voltage range. Cells were tested at a computercontrolled galvanostat (KEST 32k multichannel) and on the electrochemical test instrument (BioLogic). All the electrochemical tests were carried out at room temperature.

Structure
Powders were examined by room-temperature x-ray diffraction in order to detect phase composition in the sintered material (figures 1 and 2). The study shows that Li3.85Ni0.15Ti5O12 and Li3.80Cu0.05Ni0.15Ti5O12 exhibit similar phase composition. Peaks located at 2θ = 18.    It can be clearly seen ( Fig. 1 and Fig. 2) that both samples consist only of cubic Li4Ti5O12 spinel-type phase which allows to claim, that both nickel and copper successfully entered in the structure of Li4Ti5O12 in the lithium lattice. Figure 3 presents discharge capacities of the Li|Li + |Li3.85Ni0.15Ti5O12 and Li|Li + | Li3.80Cu0.05Ni0.15Ti5O12 cells at different current rates and in a voltage range between 1.0 V and 2.5 V. It can be seen, that for the current rate in the range 0.2-2C the sample containing Li3.85Ni0.15Ti5O12 exhibits slightly higher specific capacity than the sample Li3.80Cu0.05Ni0.15Ti5O12. However, for the 5C current rate the capacity of the Li3.80Cu0.05Ni0.15Ti5O12 is about 10 mAh•g −1 higher than for the sample which contains only nickel, which is a significant difference. The possible explanation of this phenomenon is that the addition of copper improves the electron-transport properties while at the same time lowers the maximal capacity of the material.  It can be seen, that both materials exhibit much higher capacity in comparison with the cyclic charging/discharging in the voltage range of 1.0-2.5 V. The major capacity increase could be observed in the 0.2-0.5 V voltage range. The reason of the increase of capacity is the insertion of two additional moles of lithium into the lattice of Li7Ti5O12 resulting in Li9Ti5O12 structure.

Cyclic voltammetry
The samples Li3.85Ni0.15Ti5O12 (LTO-Ni) and Li3.80Cu0.05Ni0.15Ti5O12 (LTO-Ni-Cu) were also tested by cyclic voltammetry at scan rates 0.1 ( fig. 6a) and 0.5 mV•s -1 ( fig. 6b). It can be observed, that Li3.80Cu0.05Ni0.15Ti5O12 sample is characterized by much lower polarization than the sample with stoichiometric composition Li3.85Ni0.15Ti5O12. What is more, in the voltammogram of Li3.80Cu0.05Ni0.15Ti5O12 can be seen another peak around 1.2 V. It is possible that addition of nickel is a reason for higher polarization. The addition of copper helps to reduce this effect. It is probably the explanation why the another peak is observed.

Conclusion
Two materials with the stoichiometric composition Li3.85Ni0.15Ti5O12 and Li3.80Cu0.05Ni0.15Ti5O12 were obtained by solid-state reaction. It was proved that both nickel and copper successfully entered in the lattice of lithium. What is more, doping with nickel, and especially nickel and copper may enhance the specific capacity of LTO at high current rates.
The financial support -the AGH statutory research of Danuta Olszewska