Simulating Vapor–Liquid–Solid Growth of Au-Seeded InGaAs Nanowires

Abstract

Ternary III−V nanowires are commonly grown using the Au-seeded vapor−liquid−solid method, wherein the solid nanowires are grown from nanoscale liquid seed particles, which are supplied with growth species from the surrounding vapor phase. A result of the small size of these seed particles is that their composition can vary significantly during the cyclical layer-by-layer growth, despite experiencing a constant pressure of growth species from the surrounding vapor phase. Variations in the seed particle composition can greatly affect the solid nanowire composition, and these cyclical dynamics are poorly understood for ternary nanowire growth. Here, we present a method for simulating nanowire growth which captures the complex cyclical dynamics using a kinetic Monte Carlo framework. In the framework, a nanowire grows through the attachment or detachment of one III−V pair at the time, with rates that are based on the momentary composition of the seed particle. The composition of the seed evolves through the attachment and detachment of III−V pairs to the solid nanowire and through the impingement or evaporation of single atoms to the surrounding vapor. Here, we implement this framework using the As−Au−Ga−In materials system and use it to simulate the growth of Auseeded InGaAs nanowires with an average solid Ga/III ratio around 0.5. The results show that nucleation preferentially occurs via clusters of InAs and that the compositional hierarchy of the liquid seed (X As < X Ga < X In ) determines much of the dynamics of the system. We see that imposing a constraint on the simulation, that only the most recently attached III−V pair can be detached, resulted in a significant narrowing of the compositional profile of the nanowire. In addition, our results suggest that, for ternary systems where the two binaries are heavily mismatched, the dynamics of the seed particle may result in an oscillating compositional profile.

Theory of MOCVD Growth of III-V Nanowires on Patterned Substrates

“…The surface area per NW equals , where is a geometrical constant (for example, k = 1 for square array and for hexagonal array of NWs). The Au-catalyzed VLS or SAE growth of III-V NWs is controlled by the kinetics of group III species under group V-rich conditions [ 14 , 19 , 20 , 21 , 22 , 23 , 28 ]. Therefore, the central object for the NW growth kinetics is the total volumetric current of group III atoms into the NW (measured in nm 3 /s).…”

“…Therefore, I assume that , with c = kχ s and , is the effective adsorption coefficient. The direct vapor flux into each NW is given by , with being the effective surface area of the NW, including the droplet [ 14 , 18 , 19 , 20 , 21 , 22 , 23 , 24 ]. The current lands on the substrate surface rather than on the NW surface.…”

“…Here, I take into account that the total current cannot be greater than , and can be limited by the desorption of group III species from the NW sidewalls and droplet surface. Therefore, the effective surface areas in Equation (2) are given by [ 14 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 28 ] …”

“…Most models for MOCVD growth of III-V NWs [ 19 , 20 , 21 , 22 , 23 , 24 ] consider surface diffusion of group III adatoms as the governing mechanism of material exchange between the substrate surface and NWs, similar to Ref. [ 25 ] for Si NWs or Refs.…”

Theory of MBE Growth of Nanowires on Adsorbing Substrates: The Role of the Shadowing Effect on the Diffusion Transport

“…[ 14 ] for a review). Most works on this topic [ 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 ], with only a few exceptions [ 30 , 31 , 32 , 33 , 34 , 35 ], addressed the growth kinetics of individual NW without accounting for the shadowing effect [ 30 ], which is very important in molecular beam epitaxy (MBE) and related directional deposition techniques where the material beams are inclined at an angle with respect to the substrate normal.…”