Genetic Algorithm to Optimize Performance of Tunneling Field-Effect Transistor

In In this paper we investigate a double-gated tunneling FET, DGTFET, with two channel lengths (50 and 20 nm) and the effect of device design variation on its DC and RF collective performance. The studied design parameters are drain doping abruptness (or spread) and its shift relative to the edge of gate electrode. Additionally, the gate work function, the dielectric material and its thickness are investigated. The studied performance parameters are ON/OFF ratio, maximum cutoff frequency, subthreshold swing and ambipolar current. Hence, the device's figure-of-merit, FOM, is expressed as a weighted-sum objective function to be optimized by a genetic algorithm, GA. Afterward, the truthfulness of GA results is validated by multifactorial experimentation and studying the effect of changing every parameter on device's FOM. In this work, the genetic optimization is shown to enhance device performance when it has 20 nm channel to be comparable to the 50-nm-long-channel device. The GA is tested multiple times with aid of parallel processing, MATLAB class and a TCAD model to validate its efficiency in optimizing electronic devices when a TCAD model is built without much need to find a closed-form mathematical model.