A detailed model for defect concentration and dopant activation in GaAs

Abstract

Defects in semi-insulating (SI) GaAs are especially critical in determining the properties of devices in which dopants are introduced by ion-implantation. The defects in GaAs are native to the material and their concentrations are subsequently modified after ion-implantation and annealing. In this work, we have extended the existing models in the literature by incorporating a large set of defects and using the most recent values for formation energies of these defects. The model includes eight types of point defects, the vacancy of Ga and As, their antisites and interstitials of Ga and As on both sub-lattices, along with carbon related defects always present in SI-GaAs. We have also included Si and related defects when this element is implanted as ann-type dopant. All these defects are considered in several charge states allowed by their stability conditions. The model assumes thermodynamic equilibrium between the point defects at an anneal temperature. Then the GaAs wafer is quenched so that the number of defects remain the same as those at the anneal temperature, but redistribution of charges occurs in various charge states. We find that the defect concentrations are extremely sensitive to the crystal stoichiometry, and good agreement with experimental data is shown. However, when we calculate the dopant activation in implanted GaAs, the quantitative agreement with experiments is not adequate. This discrepancy is explained on the basis of available formation energies for the defects.