The adsorption properties of sodium adatom on silicon (Si) surfaces are investigated using ah initio molecular orbital theory. Twenty-three bare silicon clusters are studied at the unrestricted Hartree-Fock (UHF) many-body perturbation theory (MP4) levels. Twenty-seven alkali atom-silicon (AM-Sin) adsorbate systems are studied with adsorption sites classified as on-top, open, and bridge sites. Adsorption distances are found by optimizing the distance between the alkali atom and Sin clusters at the UHF level and then chemisorption energies are calculated at the MP4 level. Electron correlation is found to have significant effects on cluster stability and chemisorption energy. The chemisorption energies are smaller at the MP4 level than at the SCF level for most systems, and usually increase with increasing number of atoms in the cluster. AM-Si bonds appear to be of mixed type. preferred adsorption position for K atoms. In addition, the bonding was found to be mixed, neither purely covalent nor purely ionic. In yet another theoretical investigation, Ramirez [ l l ] used a slab calculation in a SCF-HF-INDO formalism to conclude also that the cave site is the most favorable for K adsorption, but with a bond length of 0.265 nm and the Si metallization at a half monolayer (0.5 ML) K coverage.The evidence from experimental work is also ambiguous. While the angle-resolved electron-energy loss spectroscopy (AREELS) studies of Tochihara [12] and Aruga et al. [13] found the metallization of Si substrate on K adsorption, the experimental measurements of clean and exposed K/Si surfaces by Pervan et al. [14] revealed the metallic character of K overlayers for coverages greater than or equal to 0.5 ML. Furthermore, the angle-resolved UPS study of Enta et al. 115 to 171 showed a semiconducting surface for Si at K and Cs saturation coverages close to 1.0 ML, which contrasts with the metallic picture of the Si substrate obtained from the experiments of Tochihara and Aruga et al. just mentioned.Compared to the theoretical and experimental results on the K adsorption on the Si surfaces published in the literature, there has been much less work on the adsorption of Na and Li. For Na adsorption on Si (loo), Glander and Webb 118,191 found a well-ordered 4 x 1 structure at low coverage and a 2 x 1 structure at increased coverage in their experiments. Soukiassian et al. [20] performed photoemission spectroscopic (PES) studies of the adsorption of Na and Cs on the Si (100) 2 x 1 surface. Some of the important conclusions from their work include the characterization of the alkali-silicon bonding as weak, polarized and covalent, the metallization of the adsorbate, the hybridization between Na s and Si p valence orbitals and the existence of an alkali-induced electronic surface state near the Fermi level. On the theoretical side, Batra [21] investigated Na adsorption on the Si (100) 2 x 1 surface. Some important conclusions from his work are the ionic nature of the Na-Si bonding, the preference for the pedestal (or hollow) site at 0.5 ML co...