The use of linear predictive modeling for the analysis of transients from experiments on semiconductor defects

Abstract

Many methods have been employed to analyze data from experiments which measure transient responses to excitations. One such experiment, deep level transient spectroscopy, is widely used in the study of semiconductor defects. The covariance method of linear predictive modeling, which is used in digital signal processing, can be applied to the problem of transient analysis. We describe the covariance method and its limitations, and present the results of its use to analyze the data from capacitance transient experiments.