Subthreshold Swing Performance Assessment of Nanoscale Junctionless GAA FETs using Machine Learning Approach

In this work, the current-voltage (I-V) characteristics of nanoscale Junctionless Gate-All-Around (JLGAA) Field-Effect Transistor (FET) structures, with and without high-k dielectric materials, were investigated using the ATLAS 2D simulator. Various machine learning (ML) algorithms were employed to analyze and classify the design factors influencing one of the principal electrical parameters of the device, whixh is the subthreshold swing (SS) factor. The SS factor is a critical parameter in evaluating the performance of nanscale Junctionless GAA FE Ts, quantifying how effectively the transistor can switch from the off state (low current) to the on state (high current). Our approach aims to study and classify the impact of oxide thickness, channel radius, channe doping and high-k dielectric materials to better understund and enhance the gate control and achieve optimal performance. It was found that the proposed design can be classified into three categories based on SS values: Good SS ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&lt; 80$</tex> mV/decade), Worst SS (80–500 mV/decade), and Extermly Worst-case SS ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&gt; 500$</tex> mV/dec). This classification of SS values in 20 nm-Junctionless GAA FET provides a framework for understanding the trade-offs between performances and design parameters.