Study of a computational-time-saving scheme for quantitative LEED analysis by the matrix inversion method

Abstract

A computational-time-saving scheme for quantitative LEED analysis by the matrix inversion method was studied. Two computational-time-critical parameters, the number of atomic layers (Nl) in the surface slab and the number of ion-core scattering phase shifts (lmax), were tested for energy dependence. Our study on a Cu(210) surface shows that no dependence of Nl on the incident energy exists in the energy range of LEED (50-400 eV), i.e. using energy-dependent Nl is not feasible. The effect of slab thickness reduction on the precision of the Pendry R-factor based best-fit structure search was then studied. This study provides a guide to the selection of surface slab thickness in quantitative LEED analysis by the matrix inversion method. Also, we show that using an energy-dependent lmax can save nearly half of the overall computational time for obtaining a set of I-V curves by the matrix inversion method.