Authors: U. Monga, T.A. Fjeldly
Affilation: Norwegian University of Science and Technology, and University Graduate Center (UNIK), Norway
Pages: 803 - 805
Keywords: line edge roughness, subthreshold regime, FinFETs, power spectral density
Statistical variability such as line edge roughness (LER) and random dopant effects have become major concerns in the nanoscale regime. Due to the intrinsic body, multigate devices such as FinFETs don’t suffer from random dopant effects, but still LER causes significant fast variations in the electrical parameters. As LER manifests itself as a random process, quite a few forms of power spectral densities (PSD) have been proposed to analyze and model LER. Two most commonly used PSDs are Gaussian and exponential, but it has been pointed out that both of these PSDs give non-zero power at zero frequency (dc).We here propose an alternate solution based on a premise that LER is equivalent to a “modulated” random pulse train. We also study the effect of LER on subthreshold behavior of FinFETs.