Statistical constraints in nanocrystal memory scaling

This paper presents a detailed investigation of the statistical constraints that may limit nanocrystal memory scaling to the deca-nanometer scale-size. We adopted a Monte Carlo simulation approach to evaluate the probability distribution of the threshold voltage shift after program and of the retention time in presence of stress-induced leakage current. These distributions were used to extract a program and a retention fail probability. Both of them increase with cell dimensions scaling, strongly reducing the benefits offered by the nanocrystal technology in future microelectronics nodes.