Simulation of intrinsic parameter fluctuations in decananometer and nanometer-scale MOSFETs

TLDR

Intrinsic parameter fluctuations due to charge and matter discreteness will become increasingly important as semiconductor devices shrink to decananometer and nanometer scales, and future challenges to improve simulation accuracy are discussed. The paper reviews analytical and numerical simulation techniques for studying and predicting intrinsic parameter fluctuations. The study examines random discrete dopants, trapped charges, interface and line edge roughness as fluctuation sources, using a Green’s‑function theoretical approach for discrete charges and drift‑diffusion simulations with density‑gradient quantum corrections to model all sources. Results indicate that intrinsic fluctuations in conventional and double‑gate MOSFETs will reach levels that threaten yield and functionality of next‑generation analog and digital circuits unless design changes are implemented.

Abstract

Intrinsic parameter fluctuations introduced by discreteness of charge and matter will play an increasingly important role when semiconductor devices are scaled to decananometer and nanometer dimensions in next-generation integrated circuits and systems. In this paper, we review the analytical and the numerical simulation techniques used to study and predict such intrinsic parameters fluctuations. We consider random discrete dopants, trapped charges, atomic-scale interface roughness, and line edge roughness as sources of intrinsic parameter fluctuations. The presented theoretical approach based on Green's functions is restricted to the case of random discrete charges. The numerical simulation approaches based on the drift diffusion approximation with density gradient quantum corrections covers all of the listed sources of fluctuations. The results show that the intrinsic fluctuations in conventional MOSFETs, and later in double gate architectures, will reach levels that will affect the yield and the functionality of the next generation analog and digital circuits unless appropriate changes to the design are made. The future challenges that have to be addressed in order to improve the accuracy and the predictive power of the intrinsic fluctuation simulations are also discussed.