Computational considerations for the study of defects in solids
Department of Electrical and Computer Engineering
A general method for the study of point and extended defects in non-metals has been formulated and a substantial computer program generated to allow the study of such systems in a routine manner. This method requires only a single ansatz; this is the effect of the defect in question is appreciable only in the immediate proximity of the defect. Beyond this region, the influence of the defect may be obtained from a simple response theory, which may be linear but is not required to be so. This response is manifested as a displacement of the ion cores and by the polarization of these atoms. This situation is considered in a mathematically rigorous extension of the local orbitals method of Adams-Gilbert-Kunz, using the approach defined by Kunz-Klein. This approach ultimately defines the system in terms of building blocks for the system, which may be defined in some arbitrary way. These building blocks form a natural point for parallelization of a computer code, and such has been simply accomplished. Each building block in turn is studied using slightly modified quantum chemical techniques at the Hartree-Fock and Moller-Plesset levels. These techniques are also parallelizable and such has been done. Thus a potential two levels of parallelization may be used here, and this makes possible an ultimate use of large-scale MIMD parallelism.
Theoretica Chimica Acta
Computational considerations for the study of defects in solids.
Theoretica Chimica Acta,
Retrieved from: https://digitalcommons.mtu.edu/michigantech-p/4290