Probing cation intermixing in Li₂SnO₃

Abstract

Li₂MnO₃ holds great promise as a key component for lithium-manganese-rich oxides as high-capacity and high-energy-density cathode materials for lithium-ion batteries. However, its structural complexity remains an unresolved puzzle, hindering the further development of this class of cathode materials. In this work, the structure of Li₂SnO₃ was investigated as a model of Li₂MnO₃. Specifically, the structural evolution of materials during the solid-state synthesis of Li₂SnO₃ was studied using in situ high-energy X-ray diffraction. It was confirmed that Li₂SnO₃ with a C2/c structure was formed using the solid-state process. However, the severe intralayer intermixing between Li and Sn was found to lead to several weakening or vanishing reflection peaks.

---

Introduction

Lithium-manganese-rich transition metal oxides have been long praised for their extremely high specific capacity that is close to the theoretical limit of their stoichiometric counterparts and, potentially, for applications in high energy-density lithium-ion batteries.^1–3 This high reversible specific capacity, up to 250 mA h g⁻¹, can only be achieved by cycling this class of material within a fairly wide potential window, such as between 2.0 V and 4.6 V vs. Li⁺/Li, which subsequently leads to a continuous decrease of average working potential during the continuous charge/discharge process.^4–6 This phenomenon will be referred to as voltage fade in this work. The voltage fade will not only cause the continuous loss of energy density of batteries using these materials, but also challenge the arrayment of the battery management system that needs to calculate the state of charge (SOC) and state of health (SOH) of the battery pack. Therefore, substantial research effort has been devoted to understanding the nature of the voltage fade, as well as effective mitigation strategies.^5,7–13

Previous studies have shown that the voltage fade is a very slow process;^4,14 the stabilization of the average working potential won’t reach within 100 cycles of slow cycling (said C/10), and the surface modification won’t change the behavior of the voltage fade either.^15,16 Both structural studies and ab initio calculations have attributed the voltage fade process to the structural change in the bulk of the oxide materials; and it is believed that the formation of a spinel-like structure is the byproduct of the physical process that leads to the voltage fade.^4,17–21 Tarascon et al.²² believed that the voltage fade was related to the electrochemical activation of Li₂MnO₃ during the initial charge to a high potential, such as 4.6 V vs. Li⁺/Li. They successfully demonstrated that the voltage fade could be suppressed by partially doping Sn into the framework of Li₂RuO₃, which is similar to Li₂MnO₃ in structure, forming Li₂Ru_1−ySn_yO₃. They also indicated that the replacement of Mn by Sn in Li-rich NMC (NMC = Ni, Mn, Co) materials could be a valuable option to improve their cycle life.^9,22 Further improvement was also accomplished by introducing Ru into a Li₂MO₃ (M = transitional metal) framework.^23,24 Therefore, it is widely accepted that the structural instability of Li₂MnO₃ holds the truth of the voltage fade, and substantial effort has been devoted to understanding the structure of Li₂MnO₃ before and after the electrochemical activation, as well as the potential transformation pathway to a spinel-like or spinel structure.

However, determining the structure of Li₂MnO₃ has been proven non-trivial due to its richness in defects.^25,26 For instance, Li₂MnO₃ synthesized in various labs consistently shows a set of poorly-defined fingerprint peaks within a d spacing window between 2.0 Å and 4.5 Å. Hence, Li₂MnO₃ is usually assigned to a monoclinic structure with a space group of C2/m.^27–29 However, controversy about the structure of Li₂MnO₃ exists in the literature. Some research proposed the monoclinic C2/c³⁰ or trigonal P3₁12 (ref. 31 and 32) to describe its structure. The diffraction peaks of an ideal C2/m structure at smaller d spacing values were generally missed from the diffraction patterns of Li₂MnO₃ samples, making it extremely difficult to accurately describe the structure of Li₂MnO₃.^25,29 Researchers sometimes used a layered structure to describe the structure of Li₂MnO₃ by simply ignoring the super reflection peaks of Li₂MnO₃ and missed its unique structure.³³

Li₂SnO₃ and Li₂MnO₃ have similar structures and are made of lithium layers sandwiched between LiM₂ layers forming honeycombs (M surrounding Li).³⁴ The space group C2/c has only a minor structural difference from space group C2/m because of a distortion of the rocksalt oxygen stacking.²³ The structural differences between Li₂MnO₃ and Li₂SnO₃ are subtle and differ only in the stacking sequence of their respective mixed metal layers.³⁵ In this work, Li₂SnO₃ was selected as a model material of Li₂MnO₃, and the structural evolution during the solid-state synthesis of Li₂SnO₃ was followed using in situ high-energy X-ray diffraction.

Experimental section

Preparation of precursor

The precursor was a mixture of SnO₂ and Li₂CO₃ with a molar ratio of 1 : 1.1; 10% excess lithium source was added to minimize the impact of lithium loss during high-temperature treatment. The precursor was first ball-milled overnight to improve the mixing between SnO₂ and Li₂CO₃. The high-energy X-ray diffraction (HEXRD) pattern showed that some Li₂SnO₃ was formed after the ball-milling process, but the reaction is far from being complete. After the ball-milling, the powder was then pressed into a pellet for the in situ HEXRD experiment.

In situ experiment

The in situ HEXRD experiment was carried out at the sector 11-ID-C of Advanced Photon Source (APS) at the Argonne National Laboratory (ANL). The precursor pallet was vertically placed in a Linkam TS1500 programmable furnace, and was heated from room temperature to 1000 °C with a constant heating rate of 1 °C min⁻¹. During the experiment, a high energy X-ray beam (λ = 0.111538 Å) with a beam size of 0.5 mm by 0.5 mm was focused on the center the sample, and a Perkin Elmer 2D X-ray detector was used to collect the diffraction patterns with a collecting rate of 20 seconds per spectrum. A covariance analysis was applied to help identify the change of the crystal structure during the solid-state synthesis. Details of the covariance analysis can be found in our previous paper.³⁶ Rietveld refinement was carried out for some typical HEXRD patterns using the general structure analysis software (GSAS) with a particular focus on the cation intermixing between Li and Sn.

Results and discussion

It is of our great interest to track the structure/defect evolution during the solid-state synthesis of Li₂SnO₃. Hence, it is important to validate that Li₂SnO₃ was formed after the experiment. After the in situ experiment, the sample was cooled down to room temperature and a HEXRD pattern was collected as shown in Fig. 1. A simple pattern matching can quickly assign the diffraction peaks of the final product to a C2/c structure (PDF card # 74-0288). Fig. 1a shows our best effort to fit the experimental data using a C2/c structure without considering the cation intermixing. A good estimate on the cell parameters was obtained by the sample fitting, showing a good match on peak position. However, the fitting to the peak intensity is far from being satisfactory, indicating some major defects haven’t been taken into consideration yet. More specifically, the model without cation intermixing has major difficulty in describing a set of diminished reflection peaks at 1.54° (021), 1.76° (−112), and 2.05° (112). Our previous work suggested that cations with d⁰, d⁵, and d¹⁰ configuration, like Li⁺, Mn²⁺ and Sn⁴⁺, have a higher mobility in the closely packed oxygen framework.³⁷ It implies that a substantial amount of intermixing between Li and Sn can exist in the Li₂SnO₃ sample. Therefore, the optimization of the site occupation of cations was introduced for Rietveld refinement, and a good fit between the model and the experimental HEXRD pattern was obtained as shown in Fig. 1b. Fig. 2 schematically shows the model crystal structure obtained from the Rietveld refinement; the majority of the cation intermixing (between Sn and Li) occurred within the transition metal slab, and the intermixing between Li slab and transition metal slab was negligible. Therefore, it is believed that the diminished peaks are a result of intra-layer intermixing between Li and Sn. In order to validate this statement, a set of model XRD patterns were generated by fixing all variables for the C2/c model except for the degree of intermixing between Sn and Li. Fig. 3 clearly shows the strong dependence of the peak intensity for (021), (−112), and (112) on the degree of cation intermixing. It is shown that these peaks have a moderate intensity for an ideal crystal with C2/c structure, but they almost vanish when about 50% of Li in the transition metal layer (at 4d site) was mixed with their neighboring Sn (at 4e site), primarily due to the loss of long term ordering between Li and Sn (each Li atom is surrounded by 6 Mn atoms).

Fig. 1 Rietveld refinement for HEXRD data (cooled down to room temperature) using a C2/c structure (a) without considering the cation intermixing, (b) considering the cation intermixing.

Fig. 2 The model crystal structure of Li₂SnO₃ obtained from the Rietveld refinement; (a) schematic illustration of the Li₂SnO₃ crystal structure showing the arrangement of layers; (b) top view of figure (a) showing the transition metal slab.

Fig. 3 Model XRD showing the different degree of cation intermixing.

Fig. 4 shows the evolution of HEXRD patterns during the solid-state synthesis of Li₂SnO₃; the starting material is a mixture of nano-sized SnO₂ and Li₂CO₃, whose characteristic peak positions are marked on the bottom of Fig. 4. A small amount of Li₂SnO₃ can also be seen right after the ball-milling. During the heating process, all peaks shifted towards smaller 2θ values, which can be attributed to the thermal expansion of the lattice during the heating process. Fig. 4 also clearly shows the vanishing diffraction peaks corresponding to the raw materials and the emergence of new peaks, which are a strong indication of the ongoing solid-state reaction that consumed the raw materials (SnO₂ and Li₂CO₃) and the generation of the product (Li₂SnO₃) within the temperature window between 550 °C and 820 °C. In order to better identify the key events occurring during the in situ experiment, a covariance analysis was applied to calculate the similarity between the adjunct HEXRD patterns collected during the experiment. The variation of the covariance value as a function of the heating temperature is shown in Fig. 5. A covariance value of 1 is achieved for two spectra that are mathematically identical to each other after a proper scaling. Fig. 5 shows that the covariance value was slightly smaller than 1 from the very beginning of the experiment, this is primarily caused by the thermal expansion of the lattice, leading to the left-shift of peaks and the reduction of similarity between adjunct HEXRD patterns. Fig. 5 also shows several downwards peaks during the course of the in situ experiment; those peaks imply that a major change in the material occurred at that specific temperature window. For instance, Fig. 5 shows two severely overlapped peaks within the window of 500 °C and 800 °C, which can be attributed to the solid-state reaction that generates Li₂SnO₃. In Fig. 4, one can clearly see the evolution a new set of diffraction peaks, attributed to Li₂SnO₃, when the temperature was above 550 °C. At the same time, the diffraction from SnO₂ and Li₂CO₃ started to decrease, and they didn’t completely vanish until about 700 °C. The covariance analysis also implies that this solid-state reaction can be a two-step reaction. Taking a closer look at Fig. 4, one can see that the diffraction peaks for Li₂SnO₃ were broken into different groups. Several major diffraction peaks for Li₂SnO₃, (110) at 1.39°, (−131) at 2.54°, (−133) at 2.90°, and (331) at 4.15°, were observed at about 600 °C. However, another set of peaks, (111) at 1.64°, (−113) at 2.21°, (110) at 1.38°, (−111) at 1.44°, (110) at 1.38°, and (310) at 3.65°, was not visually observed until 722 °C. Moreover, a third group of peaks, (021) at 1.50°, (−112) at 1.73°, and (112) at 2.00° emerged at a temperature as high as 820 °C, above which the intensity of those peaks grew with the heating temperature. This change can be closely related to the downwards peak at about 850 °C shown in Fig. 5. Fig. 5 also mathematically predicts some change at a temperature above 950 °C. However, a physically meaningful change on HEXRD patterns was not identified at this temperature range.

Fig. 4 In situ HEXRD patterns during the solid-state synthesis of Li₂SnO₃.

Fig. 5 Covariance analysis of in situ HEXRD patterns during heating.

Of particular interest is the grouping and step-wise evolution of diffraction peaks of Li₂SnO₃, although they can be consistently indexed using a single C2/c model. Rietveld refinement was then carried out to fit the HEXRD patterns collected at temperatures above 750 °C using the C2/c model. Fig. 6 shows the variation of lattice parameters along the heating temperature, showing an allotropic expansion of lattice parameters. The parameter a showed a continuous increase with the heating temperature with a clear transition between 825 °C and 850 °C while the parameter b remained constant until 850 °C, after which temperature its value increased linearly with the heating temperature. Meanwhile, the parameter c increased linearly with the heating temperature within the whole temperature window, probably dominated by the thermal expansion. Fig. 6d shows that the value of β reached its minimum at about 830 °C, showing good consistency with the observation for a and b. To further elucidate the physical connection between these changes, the c axis can be projected into an orthogonal space as shown in Fig. 7a, the value of c sin(β) represents the spacing between adjacent ab planes while c cos(β) represents the relative displacement between adjunct ab planes. Since the value of β was very close to 90°, there is no surprise to see the linear relationship between c sin(β) and the heating temperature (see Fig. 7b). However, the displacement between adjunct ab panes showed strong dependence on the heating temperature, which peaked at about 830 °C. A more interesting observation is the change of the degree of Li–Sn intermixing, which also peaked at about 830 °C (see Fig. 7d). All these coincidences suggest that the change in HEXRD patterns, as well as the lattice parameters, are strongly connected to the intra-layer intermixing between Li and Sn. Therefore, the value of Li–Sn intermixing was plotted against the value of β, and a strong negative-linear relationship between β and the degree of intermixing can be clearly seen in Fig. 8.

Fig. 6 The variation of lattice parameters (a) a, (b) b, (c) c, (d) β along the heating temperature.

Fig. 7 (a) The schematic plot of geometry, the value of (b) c sin β, (c) c cos β as functions of heating temperature; (d) the exchange of the degree of Li–Sn intermixing along the heating temperature.

Fig. 8 The value of Li–Sn intermixing as a function of the value of β.

Putting the above pieces together, we propose that the structural complexity of Li₂SnO₃ originates from the severe intra-layer intermixing between Li and Sn. On one hand, the Li–Sn intermixing significantly reduces the long-term ordering of LiSn₂ cluster,³¹ and hence reduces the intensity of super-reflection peaks. On other hand, the intermixing between Li and Sn in the transition metal layer causes the reversible displacement of adjunct transition metal layers as seen in Fig. 8.

Conclusion

Li₂SnO₃ was investigated as a model material to gain insight into the structural complexity of Li₂MnO₃. In situ high-energy X-ray diffraction was adopted to trace the structure and defect evolution during the solid-state synthesis of Li₂SnO₃. A limited amount of inter-layer Li–Sn intermixing was observed during the course of solid-state synthesis. However, a significant amount of intra-layer Li–Sn intermixing was observed. It was shown that the intra-layer Li–Sn was the cause of the suppression of characteristic high-order super-reflection peaks for C2/c structure. In addition, the intra-layer Li–Sn intermixing also led to the reversible displacement of adjunct transition metal layers. The results shown here also imply that a severe cation intermixing can also exist in Li₂MnO₃.

Acknowledgements

Research was funded by the National Nature Science Foundation of China (No. 51302017), and the Science and Technology Commission of Beijing (No. Z121100006712002), U.S. Department of Energy (DOE), Vehicle Technologies Office. Support from Tien Duong and Peter Faguy of the U.S. DOE’s Office of Vehicle Technologies Program, is gratefully acknowledged. Argonne National Laboratory operates for the US Department of Energy by U Chicago Argonne, LLC, under contract DE-AC02-06CH11357. Use of the Advanced Photon Source (APS) was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.

References

1. B. Song, H. Liu, Z. Liu, P. Xiao, M. O. Lai and L. Lu, Sci. Rep., 2013, 3, 3094.
2. N. S. Choi, Z. Chen, S. a. Freunberger, X. Ji, Y. K. Sun, K. Amine, G. Yushin, L. F. Nazar, J. Cho and P. G. Bruce, Angew. Chem., Int. Ed., 2012, 51, 9994–10024.
3. J. Ma, Y.-N. Zhou, Y. Gao, Q. Kong, Z. Wang, X.-Q. Yang and L. Chen, Chemistry, 2014, 20, 8723–8730.
4. J. R. Croy, D. Kim, M. Balasubramanian, K. Gallagher, S.-H. Kang and M. M. Thackeray, J. Electrochem. Soc., 2012, 159, A781.
5. M. Bettge, Y. Li, K. Gallagher, Y. Zhu, Q. Wu, W. Lu, I. Bloom and D. P. Abraham, J. Electrochem. Soc., 2013, 160, A2046–A2055.
6. K. G. Gallagher, J. R. Croy, M. Balasubramanian, M. Bettge, D. P. Abraham, A. K. Burrell and M. M. Thackeray, Electrochem. Commun., 2013, 33, 96–98.
7. D. Mohanty, J. Li, D. P. Abraham, E. A. Payzant, D. L. Wood and C. Daniel, Chem. Mater., 2014, 26, 6272–6280.
8. M. Iftekhar, N. E. Drewett, A. R. Armstrong, D. Hesp, F. Braga, S. Ahmed and L. J. Hardwick, J. Electrochem. Soc., 2014, 161, A2109–A2116.
9. F. Dogan, J. R. Croy, M. Balasubramanian, M. D. Slater, H. Iddir, C. S. Johnson, J. T. Vaughey and B. Key, J. Electrochem. Soc., 2014, 162, A235–A243.
10. J. R. Croy, K. G. Gallagher, M. Balasubramanian, B. R. Long and M. M. Thackeray, J. Electrochem. Soc., 2014, 161, A318–A325.
11. J. R. Croy, K. G. Gallagher, M. Balasubramanian, Z. Chen, Y. Ren, D. Kim, S.-H. Kang, D. W. Dees and M. M. Thackeray, J. Phys. Chem. C, 2013, 117, 6525–6536.
12. S. M. Bak, K.-W. Nam, W. Chang, X. Yu, E. Hu, S. Hwang, E. a. Stach, K.-B. Kim, K. Y. Chung and X.-Q. Yang, Chem. Mater., 2013, 25, 337–351.
13. J. Zheng, M. Gu, A. Genc, J. Xiao, P. Xu, X. Chen, Z. Zhu, W. Zhao, L. Pullan, C. Wang and J. Zhang, Nano Lett., 2014, 14, 2628–2635.
14. J. R. Croy, S.-H. Kang, M. Balasubramanian and M. M. Thackeray, Electrochem. Commun., 2011, 13, 1063–1066.
15. I. Bloom, L. Trahey, A. Abouimrane, I. Belharouak, X. Zhang, Q. Wu, W. Lu, D. P. Abraham, M. Bettge, J. W. Elam, X. Meng, A. K. Burrell, C. Ban, R. Tenent, J. Nanda and N. Dudney, J. Power Sources, 2014, 249, 509–514.
16. Q. Xue, J. Li, G. Xu, H. Zhou, X. Wang and F. Kang, J. Mater. Chem. A, 2014, 2, 18613–18623.
17. A. Boulineau, L. Croguennec, C. Delmas and F. Weill, Dalton Trans., 2012, 41, 1574–1581.
18. B. Xu, C. R. Fell, M. Chi and Y. S. Meng, Energy Environ. Sci., 2011, 4, 2223.
19. Y. J. Park, Y. S. Hong, X. L. Wu, M. G. Kim, K. S. Ryu and S. H. Chang, J. Electrochem. Soc., 2004, 151, A720–A727.
20. B. Ammundsen and J. Paulsen, Adv. Mater., 2001, 13, 943–956.
21. Y. S. Meng and M. E. A. Dompablo, Acc. Chem. Res., 2013, 46, 1171–1180.
22. M. Sathiya, G. Rousse, K. Ramesha, C. P. Laisa, H. Vezin, M. T. Sougrati, M.-L. Doublet, D. Foix, D. Gonbeau, W. Walker, a. S. Prakash, M. B. Hassine, L. Dupont and J.-M. Tarascon, Nat. Mater., 2013, 12, 827–835.
23. M. Sathiya, K. Ramesha, G. Rousse, D. Foix, D. Gonbeau, a. S. Prakash, M. L. Doublet, K. Hemalatha and J.-M. Tarascon, Chem. Mater., 2013, 25, 1121–1131.
24. M. Sathiya, A. M. Abakumov, D. Foix, G. Rousse, K. Ramesha, M. Saubanère, M. L. Doublet, H. Vezin, C. P. Laisa, A. S. Prakash, D. Gonbeau, G. VanTendeloo and J.-M. Tarascon, Nat. Mater., 2014, 14, 230–238.
25. A. Boulineau, L. Croguennec, C. Delmas and F. Weill, Solid State Ionics, 2010, 180, 1652–1659.
26. C. H. Lei, J. G. Wen, M. Sardela, J. Bareño, I. Petrov, S.-H. Kang and D. P. Abraham, J. Mater. Sci., 2009, 44, 5579–5587.
27. P. Strobel and B. Lambert-Andron, J. Solid State Chem., 1988, 75, 90–98.
28. V. Massarotti, M. Bini, D. Capsoni, a. Altomare and a. G. G. Moliterni, J. Appl. Crystallogr., 1997, 30, 123–127.
29. J. Bréger, M. Jiang, N. Dupré, Y. S. Meng, Y. Shao-Horn, G. Ceder and C. P. Grey, J. Solid State Chem., 2005, 178, 2575–2585.
30. A. Riou, A. Lecerf, Y. Gerault and Y. Cudennec, Mater. Res. Bull., 1992, 27, 269–275.
31. Y. S. Meng, G. Ceder, C. P. Grey, W.-S. Yoon, M. Jiang, J. Bréger and Y. Shao-Horn, Chem. Mater., 2005, 17, 2386–2394.
32. a. Boulineau, L. Croguennec, C. Delmas and F. Weill, Chem. Mater., 2009, 21, 4216–4222.
33. Y. Kan, Y. Hu, J. Croy, Y. Ren, C. J. Sun, S. M. Heald, J. Bareño, I. Bloom and Z. Chen, J. Power Sources, 2014, 266, 341–346.
34. E. Salager, V. Sarou-Kanian, M. Sathiya, M. Tang, J. Leriche, P. Melin, Z. Wang, H. Vezin, C. Bassada, M. Deschamps and J.-M. Tarascon, Chem. Mater., 2014, 26, 7009–7019.
35. M. J. O. Malley, A Structural, Bonding, And Properties Study of the Ordered Rock Rock Salt Structures, Li₂MO₃ (M = Ru, Ir, Pt), The Ohio State University, 2009.
36. Y. Li, R. Xu, Y. Ren, J. Lu, H. Wu, L. Wang, D. J. Miller, Y. Sun, K. Amine and Z. Chen, Nano Energy, 2015, 19, 522–531.
37. G.-L. Xu, Y. Qin, Y. Ren, L. Cai, K. An, K. Amine and Z. Chen, J. Mater. Chem. A, 2015, 3, 13031–13038.

---

Footnote

† Current affiliation: Qatar Foundation, Qatar Environment & Energy Research Institute, P. O. Box 5825 Doha, Qatar.

---