Quantification of substitutional disorder and atomic vibrations by LEED – the role of parameter correlations

The composition of substitutionally disordered surfaces, including the layer and site dependent distribution of the chemical elements involved, has frequently been a target of quantitative low-energy electron diffraction. In the present paper, we demonstrate that, at a certain precision level of the intensity analysis, the structural search for these chemical parameters suffers from their correlations with statistical atomic displacements from the ideal lattice sites. The latter can be due to thermal vibrations and/or induced by the random substitution of atoms. Using the example of two (1 0 0)-oriented Fe1−xAlx alloy surfaces (x=0.15 and 0.47), we show that unusually large error limits for chemical parameters result when such atomic displacements must be determined simultaneously. Even small systematic errors in the intensity analysis can considerably shift the best-fit chemical structure. In order to avoid incorrect conclusions, the parameter correlations should be taken into account explicitly.