GEOMETRICAL GROWTH-RATE NONUNIFORMITY EFFECTS ON REFLECTION HIGH-ENERGY ELECTRON-DIFFRACTION SIGNAL INTENSITY DECAY

We have discovered a method to use the physical positioning of one or more effusion cells in a molecular-beam epitaxy system to cancel out linear growth rate variations along the area sampled by the reflection high-energy diffraction (RHEED) beam. This cancellation, seen for example by growing with two Ga cells at the correct flux ratio, allows RHEED oscillations to continue for 250 periods where only about 20 oscillation periods are obtained with either source alone. An unrotated growth was performed to measure the nonuniformity from a single source and the results were used in a simple simulation of the RHEED response. The simulated decay and beat pattern agree well with experiment.