Item

Abstract

In our paper we have proposed a parametrization of the Abell-Tersoff potential for In, Ga, As, InAs and GaAs. This is to report corrections to the results presented there. Specifically, surface energies of non-stoichiometric surfaces, quoted for a specific value of the arsenic chemical potential, were in error due to inconsistent usage of the As chemical potential. The correct surface energies as calculated with the previously published parametrization T1,¹ T2,² T3,³ T4,⁴ T5,⁵ T6,⁶ T7,⁷ and our parametrization (denoted as T9) are reported in Tables I and II, replacing the according entries in Tables VII to X of our paper. The surface energies of stoichiometric surface reconstructions and the relaxation differences (||F₀|| and ‹Δr›) given in these original Tables are not shown here as they are not affected by the inconsistent usage of the As chemical potential. The correct surface energies of our parametrization (T9) deviate from the DFT values since the inconsistent usage of the As chemical potential obstructed the fitting of parameters. In extension to Sec. III.C and Sec. III.D of our paper, we note that the relaxation of surface slabs was limited to 100 iterations and that T5 referred to a modified version of the parameters from Ref. 5 using R^c_ij=3.1 Å and D^c_ij=0.1 Å as cutoff parameters for the As-As interaction. These cutoff parameters effectively define a nearest-neighbour scheme in order to reproduce the results for the GaAs bulk phases presented in the original work.⁵
For potential parameters T8,⁸ we again report all surface energies. Previous results were ambiguous, because the cutoff parameters for the potential T8 had not been provided in Ref. 8. The new results reported in Table IV and V are obtained with the cutoff parameters listed in Table III.