Design and Optimization of Novel Shaped FinFET

Abstract

A novel high-performance and miniaturized fin-shaped field effect transistor has been proposed which has been named as rectzoidal (rectz) because of its origin from the existing rectangular (rect) and trapezoidal (trap) structures. The rationale behind proposing this structure is to sustain the integration of millions of transistors on integrated circuits (ICs), further utilizing these scaled transistors in advanced processors of leading semiconductor industries. The work presented here is divided into two phases: first phase presents the proposed transistor design at 20 nm gate length and its comparative simulation analysis with the previous rect and trap transistor structures in terms of short channel effects and other analog and RF parameters like transconductance, output conductance, intrinsic gain, gate capacitance, unity gain frequency etc. using Cogenda three-dimensional Technology Computer-Aided Design (TCAD) tool. In the subsequent phase, i.e., optimization phase, artificial neural network was trained with design parameters of proposed structure and fitness function was formulated using weighted sum approach. Evolutionary and swarm-based optimization algorithms have been applied to obtain optimum design parameters of proposed transistor structure corresponding to minimum fitness function value. Results obtained through these optimizers are in good consistence with TCAD simulation results.