Surface Effects on the Atomic and Electronic Structure of Unpassivated GaAs Nanowires

On the basis of accurate ab initio calculations, we propose a model for predicting the stability of III-V nanowires (NW) having different side walls and ridge configurations. The model allows us to obtain the NW formation energies by performing calculations only on relatively small systems, small diameter NWs and flat surfaces, to extract the contributions to the stability of each structural motif. Despite the idea illustrated here for the case of hexagonally shaped GaAs NWs grown along the [111]/[0001] direction, the method can also be applied generally to other differently shaped and oriented III-V NWs. The model shows that NW surfaces (side walls plus ridges) mainly determine the NW stability, so the changes to the surface structure (e.g., induced by defects or different growth conditions) would modify the final NW structure in a remarkable way. We find that wurtzite and zinc blende nanowires have similar energies over a wide diameter range, thus explaining the observed polytypism. Furthermore, new more stable ridge reconstructions are proposed for zinc blende nanowires. The surface-related structural motifs also have clear fingerprints on the NW electronic structure. We find that the more stable nanowires are all semiconducting. The band gaps are ruled by surface states and do not follow the trend mandated by the quantum confinement effect. Small diameter wurtzite nanowires have an indirect band gap, but for some of them, an indirect to direct transition can be foreseen to occur at larger diameters. Surface states have a larger impact on the zinc blende NW band gaps than on the wurtzite NW ones. Zinc blende nanowire band gaps reduce significantly with increasing nanowire radius, reaching the bulk value at a diameter of about 30 A. The surface structure and the high surface related DOS below the conduction band are going to affect the nanowire dopant incorporation and efficiency when doping is carried out during the NW growth.