Accurate ionic forces and geometry optimization in linear-scaling density-functional theory with local orbitals

摘要: Linear scaling methods for density-functional theory (DFT) simulations are formulated in terms of localized orbitals real space, rather than the delocalized eigenstates conventional approaches. In local-orbital methods, relative to DFT, desirable properties can be lost some extent, such as translational invariance total energy a system with respect small displacements and smoothness potential-energy surface. This has repercussions calculating accurate ionic forces geometries. this work we present results from onetep, our linear method based on space. The use psinc functions underlying basis set on-the-fly optimization smooth surfaces that consistent calculated using Hellmann-Feynman theorem. enables geometry performed. Results surface reconstructions silicon presented, along three example systems demonstrating performance quasi-Newton algorithm: an organic zwitterion, point defect crystal, semiconductor nanostructure.