Objective: To identify the influence of strain on epitaxy and surface morphology in the growth and reaction of oxide and metal films on GaAs.

Problem: Heteroepitaxial thin film growth - growth of a thin film on a dissimilar substrate is prevalent in a vast range of technologies including semiconductor electronics, chemical catalysis, optoelectronics, etc. Finding solutions to the problems inherent in heteroepitaxy (chemical reaction, lattice mismatch induced strain, etc.) is nearly always an important issue in any technology attempting to marry the functionality of disparate materials. Success of a given device architecture often hinges on issues of growth morphology governed by surface energy, interfacial reaction, and crystalline compatibility. Strain in a thin film, due to crystalline lattice mismatch, can drastically influence growth energetics and kinetics, producing undersirable surface morphologies and misfit-accommodating dislocations that act, for instance, as device killing recombination sites in semiconductor LED's and lasers. However, strain can also lead to a variety of interesting and potentially useful effects including the formation of coherent one-dimensional and two-dimensional nano-scale ‘dots’ and ‘wires’, and self-organization of these structures into ordered arrays.

Approach: Various aspects of heteroepitaxial growth of oxide and metallic layers on GaAs were investigated using in situ electron diffraction (Reflection High Energy Electron Diffraction) and spectroscopic (Auger spectroscopy) techniques to provide structural and chemical information on interfacial reaction products and growth morphlogy. These techniques were supplemented with Atomic Force Microscopy (AFM) to provide real-space imaging of surface morphology at the nanometer scale. Of particular interest were effects due to strain. 

Results and Future Plans: MgO growth on GaAs is desirable as a potential buffer layer to allow integration of GaAs device capabilities with oxides such as BaTiO₃ and LiNbO₃ with interesting ferroelectric or electro-optical properties. Due to strong reaction, these materials cannot be grown directly on GaAs. In situ growth studies indicated that reactivity issues can be overcome in the case of MgO growth on GaAs, but AFM studies revealed that the large lattice mismatch leads to a rough surface morphology as the growing film relaxes strain by developing a 3-dimensional growth mode and nucleating large, strain-relieving, mis-oriented grains as shown in Fig. 1. This morphology is highly undesirable, particularly for optical applications.

Mg reaction with GaAs was investigated both from the point of view of understanding the role of this reaction in the growth of MgO and also as a potential ohmic contact to GaAs. The reaction produces an epitaxial overlayer with a cubic structure but a lattice constant about 9 % larger than the underlying GaAs. The formation of two-dimensional, wire-like structures was observed by AFM when thin films of Mg were annealed to drive the reaction. Detailed studies of the growth characteristics of these structures again implicated strain in producing this novel growth mode. Strain induces the formation of 3D structures that can relieve some strain elastically. These structures grow to a critical size (~ 40 nm wide and 4 nm high) at which point the accumulated strain triggers anisotropic relaxation processes, leading to subsequent rapid growth along the GaAs [110] direction. Future work may involve systems where interesting magnetic or optical properties will result from this nanoscale structure

Web Contact micro@nist.gov