Formation mechanism of spinel LiTi₂O₄ prepared by carbon thermal reduction reaction

Abstract

The formation mechanism of LiTi₂O₄, prepared by a carbon thermal reduction reaction using Li₂CO₃ and TiO₂ (anatase) as starting materials and acetylene black as a reducing agent, was investigated by in situ variable temperature X-ray diffraction and thermal gravimetric analysis/differential scanning calorimetry system. It was found that the cooling rate significantly impacts on obtaining pure phase LiTi₂O₄ sample after forming LiTi₂O₄ product during the carbon thermal reduction reaction. LiTi₂O₄ has excellent cycling stability, remaining a specific capacity of 126.6/111.9 mA h g⁻¹ with a capacity fade of 5.1%/3.1% at 0.5C/1C rate after 200 cycles.

---

In recent years, great efforts have been made to develop spinel-structured Li₄Ti₅O₁₂ as an alternative replacement for currently used graphite anode material in lithium-ion batteries (LIBs) due to its superior advantages such as high power capability, unique safety performance (1.55 V vs. Li/Li⁺), and excellent cycling stability (zero-strain insertion material).^1,2 But Li₄Ti₅O₁₂ has a poor electrical conductivity (<10⁻¹³ S cm⁻¹) which seriously limits its high-rate performances for practical applications.³ Spinel LiTi₂O₄ has a similar crystal structure and lithium-ion migration path with Li₄Ti₅O₁₂ ⁴ and especially has a significantly higher electrical conductivity of 5.56 × 10⁶ S cm^−1 5–7 compared with Li₄Ti₅O₁₂. These features make LiTi₂O₄ a more promising anode material as a replacement of Li₄Ti₅O₁₂ in lithium-ion batteries.

Various approaches have been explored to synthesize LiTi₂O₄, such as solid state reaction,^8,9 sol–gel method,^10,11 molten salts electrolysis,^12,13 hydrothermal reaction.¹⁴ Among of these methods, solid-state reaction usually uses titanium⁵ or low-valence titanium oxides¹⁵ as reductants or other reducing elements such as lithium or hydrogen.¹⁶ However, most synthesis reported requires rigorous conditions, costly starting materials or dangerous hydrogen gas. Moreover, the synthesis of spinel LiTi₂O₄ with pure phase is difficult compared to Li₄Ti₅O₁₂, because of the existence of mixed valence state titanium ions (Ti³⁺ and Ti⁴⁺) in LiTi₂O₄. In 2010, we reported an improved one-step carbon thermal reduction method to prepare LiTi₂O₄ using Li₂CO₃ and anatase as starting materials and acetylene black as a reducing agent.¹⁷ Nonetheless, we have reproducible problem to prepare spinel LiTi₂O₄ with pure phase due to the unclear underlying mechanism during the solid state reaction.

In this work, we investigate the formation mechanism of LiTi₂O₄ in our one-step carbon thermal reduction reaction using in situ variable temperature X-ray diffraction (VT-XRD) and thermal gravimetric analysis/differential scanning calorimetry system (TGA-DSC). The electrochemical performances of pure spinel LiTi₂O₄ are also reported.

In situ VT-XRD was carried out on a PANalytical X'Pert Powder diffractometer with Cu Kα radiation (λ = 1.5405 Å), equipped with an Anton Parr HTK 1200N high temperature attachment. The samples were heated in N₂ atmosphere from room temperature to 900 °C with a heating rate of 10 °C min⁻¹, and then stabilized for 30 min at each integer point of temperature. SDTQ600 thermal gravimetric analysis/differential scanning calorimetry system was used to identify the phase transition temperature in N₂ atmosphere. The morphologies of the LiTi₂O₄ samples at different cooling rate were observed by SEM. ⁷Li MAS NMR spectrum of LiTi₂O₄ was acquired on an AV 400 Bruker spectrometer under magic angle sinning at 5 kHz using 4 mm zirconium rotors. CV was conducted in cells at 0.2 mV s⁻¹ from 0.8 V to 2.5 V. EIS was measured by applying an alternating voltage of 5 mV over the frequency ranging from 10² to 10⁵ Hz.

The in situ VT-XRD measurement was used to analyze the formation mechanism of LiTi₂O₄ during the carbon thermal reduction. The XRD patterns over heating from room temperature to 900 °C in a 2θ range of 10°–70° are shown in Fig. 1a. The XRD patterns of the starting materials before heating are well consisted with the corresponding crystal structure in the database. The peaks at 20.3°, 35.8° and 43.4° corresponding to (0 2 0), (−1 3 1) and (2 0 2) plane of Li₂TiO₃ intermediate started to appear at 600 °C. At 700 °C, the peaks at 35.5° and 40.3° corresponding to (3 1 1) and (4 0 0) plane of LiTi₂O₄ appeared, suggesting that LiTi₂O₄ product was generated since 700 °C. The peak at 18° corresponding to (1 1 1) plane of LiTi₂O₄ and Li₂TiO₃ was not distinguishable due to overlapping of these two phases. After heating to 800 °C and kept for 30 min, anatase disappeared completely. The intensity of rutile increased rapidly because of anatase turned into rutile from 700 °C to 800 °C, and then decreased after 800 °C due to its continuous consumption by reacting with Li₂TiO₃ to generate LiTi₂O₄. In our experiments, rutile completely disappeared when the temperature was heated to 900 °C and kept for 12 h.

Fig. 1 (a) The in situ VT-XRD patterns of LiTi₂O₄ during heating stage from room temperature to 900 °C; (b) in situ VT-XRD patterns of LiTi₂O₄ during cooling stage from 900 °C to room temperature; (c) XRD patterns of LiTi₂O₄ during different cooling rate.

It worth to point out that the cooling rate has a significant impact on obtaining pure phase LiTi₂O₄ sample.¹⁸ Fig. 1b shows the in situ VT-XRD patterns of LiTi₂O₄ product cooling from 900 °C to room temperature with a rate of 10 °C min⁻¹. As the temperature decreased after 850 °C, the diffraction peaks of LiTi₂O₄ product shifted to larger angles, indicating that the impurity was generated in the product (Fig. 1b). A comparative experiment was conducted by cooling the LiTi₂O₄ product at different rate of 180 °C min⁻¹ and 10 °C min⁻¹, respectively. At a fast cooling rate of 180 °C min⁻¹, the pure phase LiTi₂O₄ was obtained as a deep-blue powder with well-defined and sharp XRD diffraction peaks which is consistent well with the reference LiTi₂O₄ (JCPDS no. 82-2318, Fig. 1c). As comparison, impurity was observed after cooling at 10 °C min⁻¹ as shown in Fig. 1c.

SEM observation shows that LiTi₂O₄ sample cooled at 180 °C min⁻¹ has a size range of 400 to 500 nm with perfect cubic shape (Fig. S1a†). As comparison, LiTi₂O₄ sample cooled at 10 °C min⁻¹ has similar size, but consists of a lot of shapeless broken particles (Fig. S1b†). The ⁷Li MAS NMR spectrum of pure LiTi₂O₄ cooled at 180 °C min⁻¹ displays only one chemical shift at −0.053 ppm with three symmetric spinning sidebands at 42.957/−43.082, 85.727/−85.775, 128.510/−128.531 ppm (Fig. S1c†), due to the average oxidation state of Ti³⁺ and Ti⁴⁺ ions in LiTi₂O₄.^18,19

Based on the above results, the formation mechanism of LiTi₂O₄ was proposed as following, and schematically illustrated in Scheme 1:

Scheme 1 Schematic formation mechanism of LiTi₂O₄.

TGA-DSC experiment was carried out from room temperature to 1000 °C to further confirm the formation mechanism of LiTi₂O₄ during the solid state reaction (Fig. 2). The TGA curve of the reaction mixture (Li₂CO₃, TiO₂ and acetylene black) for synthesizing LiTi₂O₄ showed a weight loss of 0.83% at a temperature range of room temperature to 500 °C due to the dehydration of the reactants which is correlated with the broad endothermic peak at the same temperature range in DSC curve. The second weight loss of 18.3% at 500 °C–720 °C in TGA curve corresponds to the exothermic peak at the same temperature range in DSC curve which composes of two relatively sharp exothermic peaks at 580 °C–680 °C and 680 °C–720 °C, respectively. These two sharp exothermic peaks correspond to the eqn (1) reaction with a weight loss of 11.61 wt% (fitting well with the theoretical value of 10.83 wt%) and the eqn (2) reaction between Li₂TiO₃ and anatase with a weight loss of 6.69 wt% (fitting well with the theoretical value of 7.01 wt%), respectively.

Fig. 2 TGA-DSC curve of LiTi₂O₄ measured from room temperature to 1000 °C.

To correlate with the previous XRD data, the intensities of in situ VT-XRD peaks for reactants, intermediate and product are normalized as shown in Fig. 3. The small sharp exothermic peak observed in Fig. 2 from 720 °C to 750 °C corresponds to the phase transformation of anatase to thermodynamically stable rutile in eqn (3).²⁰ Anatase disappeared completely at 800 °C, evidenced by that the normalized intensity of anatase decreased to zero (800-2C in Fig. 3a). It is worth to note that Li₂TiO₃ continued to react with rutile to generate LiTi₂O₄ product after 800 °C (eqn (4)) which correlates well with the intensity change of rutile, Li₂TiO₃ and LiTi₂O₄ product (Fig. 3b–d). There is no obvious weight loss in TGA curve, mainly because of the very little amount of the rutile and the reaction is slow. Therefore, it is necessary to keep at 900 °C for 12 h in experiments to obtain the LiTi₂O₄ with pure phase.

Fig. 3 The normalized intensities of in situ VT-XRD peaks for (a) anatase, Li₂CO₃ (b) intermediate Li₂TiO₃ and (c) LiTi₂O₄ product and (d) rutile at the heating and cooling temperature.

CV and EIS measurements were used to characterize and compare the electrochemical kinetics of LiTi₂O₄ electrode at different cooling rates. Both samples display a pair of reversible oxidation and reduction peaks at 1.63/1.44 V and 1.86/1.33 V for LiTi₂O₄ cooled at 180/10 °C min⁻¹, respectively (Fig. S2a†). But LiTi₂O₄ cooled at 180 °C min⁻¹ shows relatively sharper peaks and smaller potential difference (ϕ_a − ϕ_b) of 0.19 V between anodic and cathodic peaks as compared with 0.53 V for LiTi₂O₄ cooled at 10 °C min⁻¹, suggesting a lower polarization of the electrode. The Nyquist plot for LiTi₂O₄ electrode at different cooling rate both consists of a semicircle and a linear part (Fig. S2b†); the fitted parameters including the electrolyte resistance (R_s), the charge-transfer resistance (R_ct), the double layer capacitance and passivation film capacitance (CPE), the exchange current density (i⁰ = RT/nFR_ct) are collected in Table S1.†²¹ These data, together with the above CV results, imply that LiTi₂O₄ sample cooled at 180 °C min⁻¹ has an improved electrical conductivity and better electrochemical kinetics as compared with LiTi₂O₄ sample cooled at 10 °C min⁻¹.

The voltage–capacity curve for the initial charge/discharge cycle at 0.2C is shown as the inset in Fig. 4a. The potential profile of LiTi₂O₄ electrode exhibited a typical long and flat potential plateau at around 1.55 V (vs. Li/Li⁺), corresponding to the lithium insertion/extraction reaction between LiTi₂O₄ and Li₂Ti₂O₄ phase. At 0.2C rate, the electrode showed a reversible discharge capacity of 144.5 mA h g⁻¹ with a coulombic efficiency of 98% between 1.0 and 3.0 V. The long-term cycling performances of LiTi₂O₄ electrode at 0.5C and 1C rate are displayed in Fig. 4a. The LiTi₂O₄ electrode showed an initial charge capacity of 133.4/115.1 mA h g⁻¹ at 0.5C/1C rate, respectively. Over 200 cycles, LiTi₂O₄ electrode remarkably retained a stable cycling capacity of 126.6/111.9 mA h g⁻¹ with corresponding capacity retention of 94.9%/96.9% at 0.5C/1C rate, respectively. These data are better, in terms of capacity and cycling stability, than those of spinel LiTi₂O₄ samples recently reported in literature,^22,23 even including that doped with metal ions (Fe and V).^24,25 Fig. 4b shows the rate capability of the LiTi₂O₄ electrode. The current density was increased gradually from 0.2C to 10C and then returned back to 0.2C. When the rate increased from 0.2C to 5C, the LiTi₂O₄ electrode maintained an average reversible capacity of 54.3 mA h g⁻¹. When the current density was returned back to 0.2C, a stable reversible capacity of 137.5 mA h g⁻¹ was obtained, which is about 95.2% of the initial charge capacity at 0.2C, indicating that the integrity of LiTi₂O₄ electrode was maintained even after high rate charge and discharge tests.

Fig. 4 (a) The cycling performance of LiTi₂O₄ during 180 °C min⁻¹ cooling at 0.5C/1C rate and (b) rate capability.

Conclusions

The formation mechanism of LiTi₂O₄ in the one-step solid-state reaction using Li₂CO₃ and TiO₂ (anatase) as starting materials and acetylene black as reductant is studied by in situ VT-XRD measurements and TGA-DSC analysis. The formation of spinel LiTi₂O₄ consists of three steps. Firstly, Li₂CO₃ reacts with TiO₂ (anatase) to form intermediate Li₂TiO₃. Secondly, Li₂TiO₃ reacts with TiO₂ (anatase) in the presence of acetylene black to form LiTi₂O₄ at 700 °C. In the same time, anatase starts to transform into rutile at 700 °C. Thirdly, Li₂TiO₃ continues to react with TiO₂ (anatase and rutile) and acetylene black to generate LiTi₂O₄ product. Interestingly, the cooling rate significantly impacts on obtaining pure phase LiTi₂O₄ sample. Pure LiTi₂O₄ sample can be reproducibly prepared as a deep-blue powder at a fast cooling rate of 180 °C min⁻¹. Comparative CV and EIS study shows that LiTi₂O₄ cooled at 180 °C min⁻¹ has improved electrical conductivity and favorable electrochemical kinetics as compared with the sample cooled at 10 °C min⁻¹. The electrochemical investigation shows that LiTi₂O₄ exhibits excellent cycling stability with a capacity retention of 96.9% at 1C after 200 cycles. We believe that this work will construct a solid foundation for optimizing the synthesis and further developing LiTi₂O₄ as anode material for lithium-ion batteries.

Acknowledgements

This work was supported by National Natural Science Foundation of China (51164006/21573239), CAS Key Laboratory of Renewable Energy (Y407K41001), Science & Technology Program of Guangxi Province (1298025-9), Special Support Program of Guangdong Province for High-Level Talents (2014TX01N014), Collaboration Project of CAS-Guangdong Province (2013B09 1300017) and Guangzhou Municipal Project for Science & Technology (201423/2014Y2-00219).

Notes and references

1. T. Ohzuku, A. Ueda and N. Yamamoto, J. Electrochem. Soc., 1995, 142, 1431–1435.
2. G. Q. Zhang, B. Y. Xia, C. Xiao, L. Yu, X. Wang, Y. Xie and X. W. Lou, Angew. Chem., Int. Ed., 2013, 52, 8643–8647.
3. Q. Y. Zhang, H. S. Lu, H. X. Zhong, X. D. Yan, C. Y. Ouyang and L. Z. Zhang, J. Mater. Chem. A, 2015, 3, 13706–13716.
4. T. F. Yi, Y. Xie and Y. R. Zhu, J. Power Sources, 2012, 198, 318–321.
5. K. M. Colbow, J. R. Dahn and R. R. Haering, J. Power Sources, 1989, 26, 397–402.
6. A. Kuhn, C. Baehtz and F. Garcia-Alvarado, J. Power Sources, 2007, 174, 421–427.
7. H. X. Geng, A. F. Dong, G. C. Che, W. W. Huang, S. L. Jia and Z. X. Zhao, Physica C, 2005, 432, 53–58.
8. M. Manickam and M. Takata, J. Power Sources, 2003, 114, 298–302.
9. J. A. Mergos and C. T. Dervos, Mater. Charact., 2009, 60, 848–857.
10. L. Persi, F. Croce and B. Scrosati, Electrochem. Commun., 2002, 4, 92–95.
11. M. Łapiński, B. Kościelska and W. Sadowski, J. Phys. Chem. Solids, 2013, 74, 575–578.
12. N. Kumada, M. H. K. Rubel, A. Miura and T. Takei, J. Ceram. Soc., 2014, 122, 307–309.
13. J. Akimoto, Y. Gotoh, K. Kawaguchi and Y. Oosawa, J. Solid State Chem., 1992, 96, 446–450.
14. D. Fattakhova, V. Petrykin, J. Brus, T. Kostlánová, J. Dědeček and P. Krtil, Solid State Ionics, 2005, 176, 1877–1885.
15. A. Kuhn, M. Martín and F. García-Alvarado, J. Solid State Chem., 2010, 183, 20–26.
16. T. Kanno, J. Awaka, F. Kariya, S. Ebisu and S. Nagata, Phys. B, 2006, 381, 30–33.
17. J. W. Yang, J. Zhao, Y. Z. Chen and Y. Li, Ionics, 2010, 16, 425–429.
18. J. P. Kartha, D. P. Tunstall and J. T. S. Irvine, J. Solid State Chem., 2000, 152, 397–402.
19. C. P. Grey and N. Dupre, Chem. Rev., 2004, 104, 4493–4512.
20. K. J. D. Mackenzie and P. J. Melling, Trans. J. Br. Ceram. Soc., 1974, 73, 179–183.
21. L. Yue, S. Q. Wang, X. Y. Zhao and L. Z. Zhang, J. Mater. Chem., 2012, 22, 1094–1099.
22. C. Q. Feng, L. Li and Z. P. Guo, J. Alloys Compd., 2009, 478, 767–770.
23. M. J. Pan, Y. X. Chen and H. B. Liu, Ionics, 2015, 21, 2417–2422.
24. S. Chakrabarti, A. K. Thakur and K. Biswas, Solid State Ionics, 2014, 262, 49–55.
25. J. Liu, C. Q. Du, Y. Q. Lin, Z. Y. Tang and X. H. Zhang, J. Alloys Compd., 2015, 622, 250–253.

---

Footnote

† Electronic supplementary information (ESI) available: Details of the synthesis, characterization, and electrochemical measurements. See DOI: 10.1039/c5ra19621c

---