Structural determination of Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) using a combination of Rietveld analysis and the maximum entropy method

Abstract

The crystal structures and electron density distributions of the layered oxide Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) were studied using a combination of Rietveld analysis of high-resolution synchrotron powder X-ray diffraction data and the maximum entropy method (MEM). Structural analysis revealed that Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) has the lattice parameters a = 4.934 Å, b = 2.852 Å, c = 5.090 Å, β = 108.8° and adopts the space group C2/m. The chemical formula can be expressed as [Ni_0.0815]_2a{Li_0.5Ni_0.0115}_4i[Mn_0.5Ni_0.407□_0.093]_2dO₂. The electron density map obtained using MEM clearly shows that most of the Li ions migrate from the octahedral 2a site to the tetrahedral 4i site during Li de-intercalation.

---

Recent reports have shown that the layered oxides LiNi_0.5Mn_0.5O₂ and LiCo_1/3Ni_1/3Mn_1/3O₂ are promising cathode materials.^1–5 Both LiNi_0.5Mn_0.5O₂ and LiCo_1/3Ni_1/3Mn_1/3O₂ adopt the rhombohedral structure of LiCoO₂ and LiNiO₂ and possess reversible capacities of 150 mAh g⁻¹ in the voltage range 2.5 to 4.3 V and of 200 mAh g⁻¹ in the voltage range 2.5 to 4.6 V, respectively. They show the superior characteristics of larger capacities compared to LiMn₂O₄ and better thermal stability compared to LiNiO₂. We have previously studied the structural change of LiNi_0.5Mn_0.5O₂ taking place during the electrochemical de-intercalation process and we have demonstrated that Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) has a monoclinic lattice.⁶ However, the precise nature of the structural changes taking place in the solid solution Li_1−yNi_0.5Mn_0.5O₂ remained unclear. Detailed information on the phase transition of Li_1−yNi_0.5Mn_0.5O₂ is very important in order to improve the electrochemical properties of not only this solid solution but also of related materials.

It is well known that the maximum entropy method (MEM) can be used to produce an electron density distribution map from a set of X-ray structure factors without the use of any structural model, and that the MEM map is consistent with the observed structural factors and least biased with respect to unobserved structure factors;^7,8 in MEM analysis, any kind of deformation of electron densities is allowed as long as the symmetry requirements are satisfied. Here we report the crystal structures and electron density distributions of the layered oxides Li_1−yNi_0.5Mn_0.5O₂ (y = 0 and 0.5), which we have studied using a combination of Rietveld analysis of high-resolution synchrotron powder X-ray diffraction (XRD) data and the MEM.

LiNi_0.5Mn_0.5O₂ was prepared using appropriate molar ratios of LiOH·H₂O, Mn(NO₃)₂·6H₂O, and Ni(NO₃)₂·6H₂O.⁵ Metal oxalate precipitates were obtained by the coprecipitation method, mixing Mn(NO₃)₂·6H₂O and Ni(NO₃)₂·6H₂O with (NH₄)₂C₂O₄. The dried precipitates were calcined in air at 873 K for 12 h to convert the metal oxalates to oxides. The resulting oxides were weighed with LiOH·H₂O, mixed, pelletized, and fired in air at 1273 K for 12 h. Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) was prepared electrochemically using a coin-type cell comprised of Li/1 M LiPF₆ in EC : DEC (1 ∶ 1) solution/LiNi_0.5Mn_0.5O₂ with a BTS2004 (Nagano Co. Ltd.) apparatus.

High-resolution synchrotron XRD data were collected at room temperature using the powder diffractometer on the BL02B2 beamline at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2002A0370-ND1-np). A large Debye–Scherrer camera with radius 286.5 mm is installed in the experimental hutch and it has an Imaging Plate (IP) on the 2θ–θ arm as a detector. The pixel size of the IP was selected to be 50 µm. The wavelength of the incident synchrotron radiation was fixed at 0.4998 Å. The crystal structure and electron density distributions were studied by Rietveld analysis using the software packages RIETAN2000⁹ and MEED¹⁰/PRIMA,¹¹ respectively.

We have previously reported that the structure of Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) has a monoclinic lattice, as determined by electron diffraction patterns observed using high-resolution TEM.⁶ In the present XRD study, the structural model of Li_0.5NiO₂¹² was used as the starting point for refinements. This material adopts the space group C2/m and Li1/Ni1 is situated at 2a(0, 0, 0), Li2/Ni2/Mn at 2d(0, 1/2, 1/2), and O at 4i(x, 0, z) where x ∼ 0.25 and z ∼ 0.75. Firstly, the structural parameters were refined according to this initial model and the MEM electron density map was calculated for 64 × 64 × 64 pixels using the F_o(Rietveld) structure factors obtained from the refinement results. The occupancies of M (M = Ni, Mn, Li) ions at the 2a and 2d sites were fixed to the corresponding values at the 3a and 3b sites of LiNi_0.5Mn_0.5O₂,⁵ except for the Li1 site where the occupancy was calculated from the lithium content y. The MEM analysis revealed distinct electron density features near the 2a sites, suggesting that a fraction of the Li or Ni ions migrate from the 2a site. Therefore, we added a 4i site for Li3 and Ni3 to the initial structural model, as shown in Fig. 1. The second stage of the procedure was to refine the structural parameters on the basis of this revised model. The occupancy of the 4i Li3/Ni3 site was refined and the sums of the Li1/Li3 and Ni1/Ni3 occupancies were fixed to the values required by the sample stoichiometry. Thirdly, the structural parameters were refined on the basis of the ⁶Li MAS NMR results for Li_1−yNi_0.5Mn_0.5O₂ (y = 0.6),¹³ which suggested that all the Li ions in the transition metal layers are removed on charging. This refinement yielded a significantly lower R value. Finally, the MEM electron density map was recalculated for 64 × 64 × 64 pixels using the obtained F_o(Rietveld) structure factors and whole-pattern fitting was carried out using the calculated F_c(MEM) data during the final Rietveld refinement. The structural refinement results revealed that the chemical formula of Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) can be represented as [Ni_0.0815]_2a{Li_0.5Ni_0.0115}_4i[Mn_0.5Ni_0.407□_0.093]_2dO₂. Fig. 2 shows the observed, calculated and difference diffraction profiles for the Rietveld refinement of Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5), and the structural parameters are summarized in Table 1.

Fig. 1 Schematic representation of the structure of Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5).

Fig. 2 Observed, calculated and difference diffraction profiles obtained by Rietveld refinement of high-resolution synchrotron powder XRD data.

Table 1 Rietveld refinement results and refined structural parameters

Atom	Site	g ^a	x ^a	y ^a	z ^a	B/Å^2^a
a Here g is an occupation factor and x, y, and z are fractional coordinates. B means an isotropic atomic displacement parameter. b a = 4.9342(3) Å, b = 2.85220(14) Å, c = 5.09011(15) Å, β = 108.817(5)°, V = 67.807(6) Å^3 c R wp = 4.17%, Re = 1.67%, RI = 3.34%, RF = 1.22%.
Ni1	2a	0.0815(2)	0	0	0	= B(Mn)
Ni2	2d	0.4069	0	1/2	1/2	= B(Mn)
Mn	2d	0.5	0	1/2	1/2	0.374(7)
Li3	4i	0.25	0.378(12)	0	0.1280(16)	= B(Mn)
Ni3	4i	0.0058(2)	= x(Li3)	0	= z(Li3)	= B(Mn)
O	4i	1	0.2386(15)	0	0.71472(19)	0.789(18)

---

Gummow et al.¹⁴ carried out a neutron diffraction study of the low-temperature-synthesised variant of LiCoO₂ and the Li-extraction process from this material to form Li_0.4CoO₂. They proposed a different structural model for Li_0.4CoO₂, in which most of the Li ions move to the tetrahedral 6c site (0, 0, 3/8) from the octahedral 3a site in the lithium layer. A similar mechanism has been reported when Al and Li ions occupy the 6c interstitial tetrahedral sites in LiAl_0.8Co_0.2O₂ and Li_1−yCo_0.85Fe_0.15O₂^15,16 and when Ir ions occupy the 6c interstitial tetrahedral sites in Li_1.8−yIr_0.6Fe_0.6O₂.¹⁷ From the viewpoint of monoclinic symmetry, the 4i site in the space group C2/m is equivalent to the 6c site in the space group R3̄m. In the highly ordered layered Li_1−yCoO₂ and Li_1−yNiO₂ systems, no migration of the Li ions from the 3a site to the 6c site has been reported, whereas migration of the Li ions from the 2a site to the 4i site is clearly observed in the present study. These results indicate that disorder of the MO₆ (M = transition metal ions) octahedra causes the migration of the Li ions from the octahedral to the tetrahedral site.

Fig. 3 shows the electron density maps for Li_1−yNi_0.5Mn_0.5O₂ (y = 0 and 0.5). No electron density feature near the 6c site was observed for Li_1−yNi_0.5Mn_0.5O₂ (y = 0), while it was unmistakeably present around the 4i site in Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5). Furthermore, bonding electron density between the 2d and 4i sites was clearly apparent. In the Li-containing layered systems, the cationic distribution of Li ions can usually only be probed using neutron diffraction measurements, due to the negative coherent scattering length (b_c(Li) = −0.19 × 10⁻¹² cm) and large scattering factor of Li. However, our results indicate that the electron distribution corresponding mainly to the Li ion is visually observed using the MEM. The detailed structural changes occurring in the Li_1−yNi_0.5Mn_0.5O₂ system will be reported in the near future.

Fig. 3 Three-dimensional electron density distribution of Li_1−yNi_0.5Mn_0.5O₂ (y = 0 (a) and 0.5 (b)) obtained by the maximum entropy method. Iso-surface density level is equal to 0.5 e Å⁻³.

In conclusion, the layered oxide Li_1−yNi_0.5Mn_0.5O₂ (y = 0.5) was synthesized electrochemically and structural analysis revealed that the chemical formula can be expressed as [Ni_0.0815]_2a{Li_0.5Ni_0.0115}_4i[Mn_0.5Ni_0.407□_0.093]_2dO₂. The combination of Rietveld analysis using synchrotron X-ray diffraction data and the MEM is an effective method of investigating the structural changes in this system.

Acknowledgements

The authors would like to thank Mr. Y. Fujii of TOSOH Company for providing the powder samples used in this research. In addition, Fig. 1 and 3 were drawn with the VENUS software developed by Dilanian and Izumi.¹¹

Notes and references

1. T. Ohzuku and Y. Makimura, Chem. Lett., 2001, 744–745.
2. Z. Lu, D. D. MacNeil and J. R. Dahn, Electrochem. Solid State Lett., 2001, 4, A191–A194.
3. L. Zhang, H. Noguchi and M. Yoshio, J. Power Sources, 2002, 110, 57–64.
4. N. Yabuuchi and T. Ohzuku, J. Power Sources, 2003, 119–121, 171–174.
5. H. Kobayashi, H. Sakaebe, H. Kageyama, T. Kuniaki, Y. Arachi and T. Kamiyama, J. Mater. Chem., 2003, 13, 590–595.
6. Y. Arachi, H. Kobayashi, S. Emura, Y. Nakata, M. Tanaka and T. Asai, Chem. Lett., 2003, 32, 60–61.
7. M. Sakata and M. Sato, Acta Crystallogr., 1990, A46, 263–270.
8. M. Takata, B. Umeda, E. Nishibori, M. Sakata, Y. Saito, M. Ohno and H. Shinohara, Nature, 1995, 377, 46–49.
9. F. Izumi, inThe Rietveld Method, ed. R. A. Young, Oxford Univ. Press, Oxford, 1993, ch. 13.
10. S. Kumazawa, Y. Kubota, M. Takata, M. Sakata and Y. Ishibashi, J. Appl. Crystallogr., 1993, 26, 453–457.
11. F. Izumi and R. A. Dilanian, Recent Research Developments in Physics, Vol. 3, Transworld Research Network, Trivandrum, 2002, pp. 699–726 (ISBN 81-7895-046-4).
12. A. Hirano, R. Kanno, Y. Kawamoto, Y. Nitta, K. Okamura, T. Kamiyama and F. Izumi, Solid State Ionics, 1995, 78, 123–131.
13. W.-S. Yoon, Y. Paik, X.-Q. Yang, M. Balasubramanian, J. McBreen and C.-P. Grey, Electrochem. Solid State Lett., 2002, 5, A263–266.
14. R. J. Gummow, M. M. Thackeray, W. I. F. David and S. Hull, Mater. Res. Bull., 1992, 27, 327–337.
15. R. Alcántara, P. Lavela, P. L. Relaño, J. L. Tirado, E. Zhecheva and R. Stoyanova, Inorg. Chem., 1998, 37, 264–269.
16. H. Kobayashi, H. Shigemura, M. Tabuchi, H. Sakaebe, K. Ado, H. Kageyama, A. Hirano, R. Kanno, M. Wakita, S. Morimoto and A. Nasu, J. Electrochem. Soc., 2000, 147, 960–969.
17. H. Kobayashi, Y. Uebou, M. Tabuchi, H. Kageyama, Y. Yamamoto, M. Matsuoka and J. Tamaki, J. Electrochem. Soc., 2003, 150, A1408–A1415.

---