First-Principles Study of Stoichiometric and Nonstoichiometric LiCoO2 (104) Surfaces

Monday, 6 October 2014: 17:40
Sunrise, 2nd Floor, Galactic Ballroom 2 (Moon Palace Resort)
Y. Koyama, H. Arai (Office of Society-Academia Collaboration for Innovation, Kyoto University), I. Tanaka (Department of Materials Science and Engineering, Kyoto University), Y. Uchimoto (Graduate School of Human and Environmental Studies, Kyoto University), and Z. Ogumi (Office of Society-Academia Collaboration for Innovation, Kyoto University)

Electrode/electrolyte interfaces have a great influence on performance of lithium-ion batteries, e.g., rate capability and cycle durability. However, structures of the electrode surfaces are not yet clear on the atomic scale. Recently, we have found using a surface sensitive XAS technique and thin-film model electrodes that Co ions at the surface of LiCoO2are reduced just by immersion into organic electrolyte [1]. This suggests that the electrode surfaces are likely to be modified in the real batteries by chemical reactions with organic electrolyte components. Therefore, it is necessary to consider stability of the surface structures including changes in compositions and arrangements under the environment in the real batteries.

In this study, first-principles calculations have been carried out on several models of LiCoO2(104) surface. The stoichiometric (104) surface is reported to be one of the low energy surfaces [2]. The (104) planes have the stoichiometric composition as in the bulk, and thus the naturally cleaved (104) surface is in stoichiometry. In addition to the stoichiometric surface model, surface models with modification on the topmost ions, i.e., addition, removal, and substitution, have been examined. Most of the modified surface models have nonstoichiometric compositions, and thus the stability of the surface models is discussed considering the environment of the real batteries.

Calculation method

The surface models used in this study are slabs consisting of 15–19 layers of the (104) planes with sufficiently thick vacuum regions. Inversion symmetry is assumed to avoid artificial dipole effects. The plane-wave basis PAW method is used for the first-principles calculations with the GGA+U exchange correlation functional.

Surface energy is estimated as a function of the environment, i.e., chemical potentials, μi, as

Es = (E(slab) – Σi Ni μi) / S,

where E(slab) is the energy of the slab model, Ni is the number of atoms of species i in the slab model, and Sis the surface area, respectively. The chemical potentials have a constraint of

μLi + μCo + 2 μO = E(LiCoO2) ,

where E(LiCoO2) is the energy of LiCoO2 bulk. Assuming that LiCoO2 coexists with Li2O, i.e., 2 μLi + μO = E(Li2O), the surface energy is estimated as a function of μO.

Results and discussion

The stoichiometric surface model has the topmost Co ions whose oxidation state is +3 with the intermediate spin configuration (dup4, ddown2), whereas the spin configuration of the other Co ions is the low spin configuration (dup3, ddown3). The electronic states of the stoichiometric (104) surface are consistent with the literature [3].

Here, we show one of the nonstoichiometric surface models, in which Co ions are substituted for the topmost Li sites. The oxidation state of the topmost Co ions is +2 with the high spin configuration (dup5, ddown2), whereas the oxidation state of the other Co ions is +3 with the low spin configuration. Thus, this Co substitution surface model has a similar structure to rocksalt CoO.

Surface energy of the stoichiometric model is independent from μO. On the other hand, surface energy of the Co substitution model decreases as mO decreases. It becomes smaller than the surface energy of the stoichiometric model at μO < -1 eV versus the standard state of O2 gas at 300 K. This suggests that the surface of LiCoO2 is modified at low mO in the organic electrolytes by the chemical reaction with an electrolyte component E

LiCoO2 + E → CoLi + Li2O + EO.

The suggested Co reduction at the surface is consistent with the experimental observation by the surface sensitive XAS measurements [1].


This work was supported by RISING battery project from NEDO, Japan.


[1] D. Takamatsu, et al., Angew. Chem.-Int. Edit., 51, 11597 (2012).

[2] D. Kramer and G. Ceder, Chem. Mater., 21, 3799 (2009).

[3] D. N. Qian, et al., J. Am. Chem. Soc., 134, 6096 (2012).