Growing GaAs nanowires with well-defined crystal structures is a challenging task, but may be required for the fabrication of future devices. In terms of crystal phase selection, the connection between theory and experiment is limited, leaving experimentalists with a trial and error approach to achieve the desired crystal structures. In this work, we present a modeling approach designed to provide the missing connection, combining classical nucleation theory, stochastic simulation, and mass transport through the seed particle. The main input parameters for the model are the flows of the growth species and the temperature of the process, giving the simulations the same flexibility as experimental growth. The output of the model can also be directly compared to experimental observables, such as crystal structure of each bilayer throughout the length of the nanowire and the composition of the seed particle. The model thus enables for observed experimental trends to be directly explored theoretically. Here, we use the model to simulate nanowire growth with varying As flows, and our results match experimental trends with a good agreement. By analyzing the data from our simulation, we find theoretical explanations for these experimental results, providing new insights into how the crystal structure is affected by the experimental parameters available for growth.
Keywords: GaAs; Wurtzite; nanowire; simulation; zinc blende.