Impact of Randomly Distributed Dopants on $\Omega$ -Gate Junctionless Silicon Nanowire Transistors

Abstract

This paper presents experimental and simulation analysis of an -shaped silicon junctionless nanowire field-effect transistor (JL-NWT) with gate lengths of 150 nm and diameter of the Si channel of 8 nm. Our experimental measurements reveal that the ON-currents up to 1.15 mA/μm for 1.0 V and 2.52 mA/μm for the 1.8-V gate overdrive with an OFF-current set at 100 nA/μm. Also, the experiment data reveal more than eight orders of magnitude ON-current to OFF-current ratios and an excellent subthreshold slope of 66 mV/dec recorded at room temperature. The obtained experimental current-voltage characteristics are used as a reference point to calibrate the simulations models used in this paper. Our simulation data show good agreement with the experimental results. All simulations are based on driftdiffusion formalism with activated density gradient quantum corrections. Once the simulations methodology is established, the simulations are calibrated to the experimental data. After this, we have performed statistical numerical experiments of a set of 500 different JL-NWTs. Each device has a unique random distribution of the discrete dopants within the silicon body. From those statistical simulations, we extracted important figures of merit, such as OFF-current and ON-current, subthreshold slope, and voltage threshold. The performed statistical analysis, on samples of those 500 JL-NWTs, shows that the mean I D -V Gs characteristic is in excellent agreement with the experimental measurements. Moreover, the mean I D -V Gs characteristic reproduces better the subthreshold slope data obtained from the experiment in comparison to the continuous model simulation. Finally, performance predictions for the JL transistor with shorter gate lengths and thinner oxide regions are carried out. Among the simulated JL transistors, the configuration with 25-nm gate length and 2-nm oxide thickness shows the most promising characteristics offering scalable designs.Index Terms-Drift diffusion (DD), electronic transport, junctionless (JL) nanowire field-effect transistors (FETs) (JL-NWTs), random dopants, scattering mechanisms, silicon nanowire, simulations, variability.

Nanowire Electronics: From Nanoscale to Macroscale

“…When the channel or gate lengths of a transistor are greatly larger than the mean-free-path of charge carriers, mobility is a meaningful index for devices in the drift-diffusion regime. Compared to a bulk material, the cross-sectional dimensions of a nanowire are reduced, which leads to different scattering effects to alter mobility. − For instance, for the nanowires with diameters between strong quantum confinement and pure classical regimes, Das et al found that mobility fluctuates with increasing diameter ( d NW ) of nanowires, which originates from appeared new channels for interband scattering . Due to the distinct band splitting and modifying of nanowires with different crystal directions, anisotropic effects on mobility take place in nanowires.…”

Reduction of Random Dopant Fluctuation-induced Variation in Junctionless FinFETs via Negative Capacitance Effect

“…To prevent the disturbance of unphysical charge trapping and initiate further efforts to specify the nano-scale electronic devices with the short channel effect [29], the quantum mechanical effect was considered during the simulation of the NCJL-FinFET. The mobility model was incorporated with the high-field saturation and Shockley Read Hall physical model to accomplish the recombination and generation [30]. As a consequence of the time consumption of each randomization device, 200 randomization samples comprising varying doping profiles were reproduced.…”

Machine Learning Approach for Predicting the Effect of Statistical Variability in Si Junctionless Nanowire Transistors

“…In our previous work, we have shown an extensive comparison between simulations and experimental results for JL-NWTs with Ω-gated region and with a channel length of 150 nm [6]. In another study, we have discussed statistical simulation results based on an ensemble of 500 JL-NWTs, where each device is atomistically unique with random distribution of discreet dopants in the channels [7]. Results obtained from those previously reported works have allowed us to suggest an improvement of the device design, predict the device performance and to extract (FoM), such as OFF-current (IOFF) and ON-current (ION), subthreshold slope (SS) and voltage threshold (VTH).…”

“…In this letter, we would like to extend our previous works [6] [7] further and explore the opportunity to include multi-layer neural network (NN) in our analysis in order to predict device characteristics without running numerical TCAD device simulations. To the best of our knowledge this is the first paper that establishes a direct link between NN and TCAD simulations.…”