A computational method for partitioning a charge density grid into Bader volumes is presented which is efficient, robust, and scales linearly with the number of grid points. The partitioning algorithm follows the steepest ascent paths along the charge density gradient from grid point to grid point until a charge density maximum is reached. In this paper, we describe how accurate off-lattice ascent paths can be represented with respect to the grid points. This improvement maintains the efficient linear scaling of an earlier version of the algorithm, and eliminates a tendency for the Bader surfaces to be aligned along the grid directions. As the algorithm assigns grid points to charge density maxima, subsequent paths are terminated when they reach previously assigned grid points. It is this grid-based approach which gives the algorithm its efficiency, and allows for the analysis of the large grids generated from plane-wave-based density functional theory calculations.

An improvement to the grid-based algorithm of Henkelman et al. for the calculation of Bader volumes is suggested, which more accurately calculates atomic properties as predicted by the theory of Atoms in Molecules. The CPU time required by the improved algorithm to perform the Bader analysis scales linearly with the number of interatomic surfaces in the system. The new algorithm corrects systematic deviations from the true Bader surface, calculated by the original method and also does not require explicit representation of the interatomic surfaces, resulting in a more robust method of partitioning charge density among atoms in the system. Applications of the method to some small systems are given and it is further demonstrated how the method can be used to define an energy per atom in ab initio calculations.

Zinc selenide, cadmium sulfide, and cadmium selenide clusters were produced by direct laser ablation and analyzed in a time-of-flight mass spectrometer. The positive-ion mass spectra indicated that clusters composed of six and thirteen monomer units were ultrastable in all cases. The geometries and energies of the neutral and positively charged M(n)X(n) clusters up to n = 16 were obtained computationally at the B3LYP level of theory using the SKBJ basis set for the metal atoms and the SKBJ(d,2df) basis set for the chalcogen atoms. Small neutral and positive clusters (n = 1-4) have planar geometries, neutral three-dimensional clusters have the geometry of closed-cage polyhedra, and cationic three-dimensional clusters have structures with a pair of two-coordinated atoms. Physical properties of the clusters as a function of size are reported. The relative stability of the positive stoichiometric clusters provides a thermodynamic explanation for the relative stability observed experimentally from the laser-ablation mass spectrometry.

Zinc sulfide clusters produced by direct laser ablation and analyzed in a time-of-flight mass-spectrometer, showed evidence that clusters composed of 3, 6, and 13 monomer units were ultrastable. The geometry and energies of neutral and positively charged Zn(n)S(n) clusters, up to n = 16, were obtained computationally at the B3LYP/6-311+G level of theory with the assistance of an algorithm to generate all possible structures having predefined constraints. Small neutral and positive clusters were found to have planar geometries, neutral three-dimensional clusters have the geometry of closed-cage polyhedra, and cationic three-dimensional clusters have structures with a pair of two-coordinated atoms. Physical properties of the clusters as a function of size are reported. The relative stability of the positive stoichiometric clusters provides a thermodynamic rationale for the experimental results.