Noise spectroscopy of silicon grain boundaries

Publisher Website
Google Scholar
Abstract
Noise spectroscopy is a sensitive and versatile tool for the investigation of interface states at semiconductor grain boundaries. We apply a transistorlike model to quantitatively explain observed 1/f noise by trapping of majority carriers in grain boundary states. Our model accounts explicitly for spatial electrostatic potential fluctuations within the interface plane due to the random spatial distribution of grain boundary defects. Such spatial inhomogeneities generally cause deviations from Lorentzian noise spectra towards 1/f behavior; grain boundaries represent a model system for the study of such a transition. We obtain from our quantitative analysis the density of interface states, their capture cross section, and the standard deviation of the spatial potential fluctuations. Our analysis enables us to derive these three characteristic grain boundary parameters for bicrystals as well as for multicrystalline samples.