Fast and Accurate Simulation of Ultrascaled Carbon Nanotube Field-Effect Transistor Using ANN Sub-Modeling Technique

Abstract

In this paper, we have proposed a new modeling methodology based on the artificial neural networks (ANN) to simulate the ultra-scaled carbon nanotube field-effect transistor (CNTFET). The sub-modeling concept has been employed to efficiently simplify the overall modeling process. The developed sub-models have been compared with the mode space nonequilibrium Green's function (MS-NEGF) simulations in terms of the resulted drain current, where a good agreement has been recorded. In addition, simulation tests have shown that the proposed smart models are faster of about two order of magnitude over the standard MS-NEGF simulation. The obtained results indicate that the proposed ANN-based sub-modeling is an accurate and computationally efficient approach, which can be successfully used to simulate, analyze, and optimize the ultra-scaled CNTFETs and the futuristic CNT-based nanoscale integrated circuits.