Second-generation charge-optimized many-body potential for Si/SiO2 and amorphous silica

A second-generation dynamic charge transfer, many-body potential function is proposed for crystalline and amorphous silica, and for silicon. The potential is based on the first-generation charge-optimized many-body (COMB) potential for these materials. The materials fidelity of the proposed formalism is demonstrated for several crystalline silica polymorphs and amorphous silica. The correct order of most of the experimentally observed polymorphs of the oxide is obtained and a significant improvement is found for the mechanical properties over the predictions of the first-generation potential. Satisfactory agreement is obtained for predictions of structural properties and defect formation energies compared to experimental and first-principles computational values. This potential can be used in conjunction with recently developed COMB potentials for the Hf/HfO2 systems.