Modeling of the programming window distribution in multinanocrystals memories

Abstract

In this paper, the impact of the Si nanocrystals technological fluctuations on the programming window dispersion of multi nanocrystals memory is thoroughly investigated. Technological dispersions of different nanocrystals populations, directly measured by high-resolution transmission electron microscopy, are used as starting points for the modeling of the device characteristics. Numerical Monte Carlo simulations as well as an original compact modeling, based on the compound distributions (CD) statistics, are here presented. Exact analytical results (CD model), approximated analytical results (CD+Central Limit Theorem model) and numerical results (numerical convolution) are deeply discussed. Finally, the good agreement between our simulations and experimental data of ultrascaled nanocrystal devices, made by conventional UV lithography or by e-beam lithography, definitively confirms the validity of our theoretical approach.