First-Principles Theory of Cation and Intercalation Ordering in Li x CoO 2

  • PDF / 879,063 Bytes
  • 12 Pages / 414.72 x 648 pts Page_size
  • 81 Downloads / 193 Views

DOWNLOAD

REPORT


ABSTRACT Several types of cation- and vacancy-ordering exist in the Li•CoO 2 battery material. The ordering patterns are of interest due to the fact that they can control the voltage in rechargeable Li batteries. We present a first-principles total energy theory which can predict both cationand vacancy-ordering patterns at both zero and finite temperatures. Also, by calculating the energetics of the Li intercalation reaction, this theory can provide first-principles predictions of battery voltages of LiCoO2 /Li cells. Our calculations allow us to search the entire configurational space to predict the lowest-energy ground state structures, search for large voltage cathodes, explore metastable low-energy states, and extend our calculations to finite temperatures, thereby searching for order-disorder transitions and states of partial disorder. INTRODUCTION: TYPES OF ORDERING IN LixCoO 2 The LiMO 2 oxides form a series of structures based on an octahedrally-coordinated network with anions (0) on one fcc sublattice and cations (Li and M) on the other. [1] In this paper, we examine the energetics and thermodynamics of ordering tendencies in the Li•CoO 2 oxide. The LiCoO 2 compound is used as a cathode material in rechargeable Li batteries. [2] When Li is de-intercalated from the compound, it creates a vacancy (denoted 0) that can be positioned in different lattice locations. Hence, we will examine three types of ordering problems: (i) Li/Co ordering in LiCoO 2 (x=l) leads to ordered R3m at low temperature and disordered Li/Co (rocksalt) at high temperature. (ii) Similarly, O/Co ordering in the completely deintercalated EJCOO 2 (x=O) is also of interest. (iii) The vacancies left behind by Li extraction can form ordered vacancy compounds in partially deintercalated LiCoO2 , leading to a C/Li ordering problem for intermediate compositions O0.5); However, Gummow et al. [131 have extracted Li electrochemically from the spinel LiCo 2 0 4 , observing a sharp increase in voltage (of --1 V) near x=1/2, in agreement with our calculations. If the capacity of Li extraction in the spinel phase LiCo 2 0 4 could be improved, our calculations provide a prediction that this spinel would make a high voltage (4.78 V) battery cathode. (iii) Because the spinel phase and the D4 are structurally distinct, the D4 Li.CoO 2 system presumably forms a two-phase mixture of spinel+D4 for values of XLi between 1/2 and 1, (as opposed to tolerating a large off-stoichiometry in either of the phases - in other words, when one begins removing Li from D4, one forms small pockets of spinel embedded in the D4 matrix). A two-phase mixture corresponds to straight line (a tie-line) in energy vs. composition (the dashed line in Fig. 3). The voltage is proportional to the slope of the energy vs. composition curve, and hence for a two phase mixture, the voltage is constant. This is consistent with the measured voltage curves for D4 (down to XLi=1/2), which show a nearly constant plateau at 3.6 V. Thus, most of the electrochemical distinctions between the CuPt and D