Deep-level-noise spectroscopy of ion-implanted polysilicon thin films

Abstract

Noise measurements represent a new technique for the characterization of traps in polycrystalline silicon. Data of highly-boron-doped polysilicon thin films reveal that the noise can be decomposed into a small number of Lorentz spectra. Each of these Lorentzian curves corresponds to a specific trap in the space-charge regions of the grain boundaries. We therefore introduce deep-level-noise spectroscopy (DLNS) as a new quantitative tool for the analysis of bulk traps in polysilicon. DLNS is based on temperature-dependent noise measurements at different frequencies. Our quantitative analysis yields the energy position, the capture cross section, as well as the concentration of each trap.