Neural Network Modeling of Anion Exchange Using Reflection High-Energy Electron Diffraction Data

Tomas Sarmiento and Gary S. May 
School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA  30322-0150

Arsenide/antimonide heterostructures have important applications in infrared detector, lasers, and high-=speed electronic devices.  The performance, and hence, the manufacture of these devices is compromised by the difficulty of controlling the interface properties as a result of the tendency of group V elements to exchange.  In this paper, a neural-network based model is developed to enable the control of the As-for-Sb exchange. 

Keywords:  Sb Compounds, Molecular Beam Expitaxy (MBE), Reflection High-Energy Electron Diffraction (RHEED), Neural Networks

                                                14.1 pdf                      Return to TOC