The possibility to expand the range of material combinations in defect-free heterostructures is one of the main motivations for the great interest in semiconductor nanowires. However, most axial nanowire heterostructures suffer from interface compositional gradients and kink formation, as a consequence of nanoparticle-nanowire interactions during the metal-assisted growth. Understanding such interactions and how they affect the growth mode is fundamental to achieve a full control over the morphology and the properties of nanowire heterostructures for device applications. Here we demonstrate that the sole parameter affecting the growth mode (straight or kinked) of InP segments on InAs nanowire stems by the Au-assisted method is the nanoparticle composition. Indeed, straight InAs-InP nanowire heterostructures are obtained only when the In/Au ratio in the nanoparticles is low, typically smaller than 1.5. For higher In content, the InP segments tend to kink. Tailoring the In/Au ratio by the precursor fluxes at a fixed growth temperature enables us to obtain straight and radius-uniform InAs-InP nanowire heterostructures (single and double) with atomically sharp interfaces. We present a model that is capable of describing all the experimentally observed phenomena: straight growth versus kinking, the stationary nanoparticle compositions in pure InAs and InAs-InP nanowires, the crystal phase trends, and the interfacial abruptness. By taking into account different nanowire/nanoparticle interfacial configurations (forming wetting or nonwetting monolayers in vertical or tapered geometry), our generalized model provides the conditions of nanoparticle stability and abrupt heterointerfaces for a rich variety of growth scenarios. Therefore, our results provide a powerful tool for obtaining high quality InAs-InP nanowire heterostructures with well-controlled properties and can be extended to other material combinations based on the group V interchange.
Keywords: III−V nanowires; VLS growth; axial heterostructures; modeling; nanoparticle stability.