Problems and prospects in the ab initio treatment of pure and defective crystals

Abstract

Accurate Hartree-Fock LCAO calculations for moderately complex crystalline systems are now feasible; a number of important applications may be envisaged in the areas of material science and technology. Some critical aspects of the corresponding computer schemes are discussed which are of fundamental importance in determining the cost of the computation. Data are provided concerning actual computations which are indicative of the kind of periodic systems that can (or cannot) be treated at present. The result of a perfect-crystal ab initio HF study can be used as an input for treating with the same approximation local-defect problems, by use of suitable embedding techniques. A scheme of this kind is presented, and its computational implications are discussed: due to the intrinsic complexity of this problem, it may be foreseen that the study of defects in crystals will be a typical application of supercomputers in the area of quantum chemistry.