• lithium-oxide cluster;
  • genetic algorithm;
  • density functional theory;
  • geometric optimization;
  • charge transformation


We apply genetic algorithm combining directly with density functional method to search the potential energy surface of lithium-oxide clusters (Li2O)n up to n = 8. In (Li2O)n (n = 1–8) clusters, the planar structures are found to be global minimum up to n = 2, and the global minimum structures are all three-dimensional at n ≥ 3. At n ≥ 4, the tetrahedral unit (TU) is found in most of the stable structures. In the TU, the central Li is bonded with four O atoms in sp3 interactions, which leads to unusual charge transformation, and the probability of the central Li participating in the bonding is higher by adaptive natural density partitioning analysis, so the central Li is in particularly low positive charge. At large cluster size, distortion of structures is viewed, which breaks the symmetry and may make energy higher. The global minimum structures of (Li2O)2, (Li2O)6, and (Li2O)7 clusters are the most stable magic numbers, where the first one is planar and the later both have stable structural units of tetrahedral and C4v. © 2012 Wiley Periodicals, Inc.