Estimating ID Current of 45nm FinFET by Artificial Neural Networks

Abstract

Fin field-effect transistor (FinFET) is non-planar 3D transistor. The silicon wing that forms the main body of the transistor distinguishes it from the others. Traditional MOSFETs and CMOS have many problems, such as large gate-to-channel leakage currents and increased computational power. Increased computational power means increased computational intensity. In addition, there is confusion in dimensioning the fin structure, which is three-dimensional compared to the planar Mosfet. All these problems require more transistors and make it very difficult to determine the ID current obtained against the channel width and channel length parameters of the FinFET element in circuits designed with FinFET element. To solve these problems, channel widths and lengths are estimated with Artificial Neural Networks. In this study, the closeness of the ID current obtained in response to the changing channel width and channel length parameters of the FinFET element to the theoretical ID current obtained from the SPICE program is estimated using Artificial Neural Networks. This estimated closeness is indicated by the error rate, and the obtained ID current versus the channel width and channel length parameters provided the circuit designer with the opportunity to determine the physical properties of the element without loss of time. © 2023 IEEE.