(398b) An Investigation into Gallium Arsenide Thin Film Growth: Molecular Dynamics Simulation

Quantum dots are known as superatoms for good reason. The atoms in these three dimensional structures act, in concert, as one atom to confine electrons to discrete energy levels. Their low threshold current increases efficiency in laser applications and the reduced density of electronic states opens the door for quantum computing. In order to take advantage of these properties, the dots should have uniform size, shape, composition and spacing. Quantum dots self-assemble due to Stranski-Krastinov growth in heteroepitaxial systems with a lattice mismatch above approximately 2%, for example in the deposition of InAs on GaAs (001). However, much is left unknown about the fundamentals of this growth such as the role of strain and composition in the formation of the wetting layer. Simulation methods such as molecular dynamics (MD) can provide insight into deposition at the atomic scale.

Until recently, all of the empirical potentials used to study GaAs/InAs systems were fit to bulk properties alone. A new parameterization has recently been developed which was fit to surface geometries and energies of the GaAs (001) and InAs (001) surfaces, including the β2(2x4) and α(2x4) reconstructions, as well as bulk lattice constants and cohesive energies¹. We have performed a number of studies aimed at validating this potential. Using NPT MD, we studied the thermal properties and melting of bulk GaAs. We determined the bulk melting point, the isothermal compressibility, and the coefficient of thermal expansion. The potential closely matches experimental densities as the temperature increases, but it underestimates the bulk melting point and overestimates the coefficient of thermal expansion as well as the isothermal compressibility. We also utilized MD to probe the stability of the GaAs(001)β2(2x4) reconstruction against melting.

To probe diffusion and the preferred Ga-atom binding sites, we calculated the minimum potential-energy surface (PES), using conjugate-gradient energy minimization for a gallium atom on the GaAs (001) β2(2x4) reconstruction. We also evaluated various diffusion pathways and energy barriers using the nudged elastic band method. The empirical potential captures the deepest minima and the diffusion barriers from the map compare favorably to first-principles DFT results. 1. T. Hammerschmidt, P. Kratzer, and M. Scheffler, to be published.