The technological importance of compound semiconductor GaAs are increasing because of their use in optoelectronic and microelectronic applications. Due to the high conversion efficiency and carrier mobility, GaAs can also be applied in solar cells and the recent study upon GaAs nanowires and their heterostructures has revealed that the conversion efficiency of GaAs nanowire array solar cells conversion is high up to 15.3%. Early the liquid and amorphous properties of GaAs were investigated by employing the first-principles calculations. The emergence of semi-empirical potential and the improvement of computer level have promoted the research and application of molecular dynamics (MD) simulation. MD simulation has now become one of the typical modeling methods at the molecular scale. The simulation is based on the known physical approximation of all particles in the system to solve the equation of motion, and obtain the atomic motion trajectory. Analytical potentials is very important in MD simulation as it is not feasible to solve the Hamiltonian by means of quantum-mechanical methods with huge computational complexity. Abell-Tersoff potential function is a short-ranged bond-order algorithm, which depends on bond lengths and bond angles and hence accesses information about the atomic structure. So it is suitable for simulating covalent bond species. Generally used for the IV elements and compounds like silicon, carbon, and others, but for the III-V compound semiconductor it is not very accurate due to the ionic bonds. Usually the modified tersoff potential, by the addition of Coulomb term, the modified exclusion potential and the truncation parameter, is used to simulate such semiconductor materials. Many studies on the bulk, surface and elastic properties of GaAs by means of MD method, are in good agreement with the experimental results. In this paper Karsten Albe’s Tersoff potential is adopted as it allows one to model a wide range of properties of GaAs compound structure. GaAs has two kinds of tetrahedral crystal structure, namely, Zinc-blende and Wurtzite, the former structure is more stable under normal conditions. But when reduced to a nanoscale scale, Wurtzite structure becomes stable. Different structures have distinct properties, similar to carbon and grapheme. But so far, there is no report on the evolution of the microstructure and the specific crystalline structure of GaAs during crystallization under rapid cooling. In this study, MD simulation was performed for liquid GaAs at the cooling rate 1×1010 K/s. The pair distribution function, the total energy per atom, the bond angle distribution function, the dihedral angle distribution and visualization method were used to analyze the variations of microstructure during the solidification process. Results show that the onset temperature of crystallization of GaAs liquid is 1460 K. The random network is the essential structural feature of liquid. The rapidly cooled crystallization is Zinc-blende based polycrystalline structure, with the grain boundary in a eutectic twin structure is a layer of wurtzite structure. At temperature below 520 K, part of As atoms segregate into simple cubic structure As8.
Read full abstract