Nanowire Transistor Solutions for 5nm and Beyond

Asen Asenov1, Y Wang2, B Cheng1, X Wang3, P Asenov1, T Al-Ameri3, V. P. Georgiev3
1Gold Standard Simulations, 2Peking University, Beijing, 3University of Glasgow


In this paper we present a comprehensive computational study of silicon nanowire transistor (SNT) and a SNM SRAM cell based on advanced design technology co-optimization (DTCO) TCAD tools. Utilizing this methodology, we provide guidelines and solutions for 5 nm and beyond in CMOS technology. At first, drift-diffusion (DD) results are fully calibrated against a Poisson-Schrodinger (PS) solution to calibrate density-gradient quantum corrections, and ensemble Monte Carlo (EMC) simulations to calibrate transport models. The calibrated DD gives us the capability to simulate statistical variability in nanowire transistors of the 5nm node and beyond accurately and efficiently. Various SNT structures are evaluated in terms of device figures of merit, and optimization of SNTs in terms of electrostatics driven performance is carried out. A variability-aware hierarchical compact model approach for SNT is adopted and used for statistical SRAM simulation near the “scaling limit”. The scaling of SNTs beyond the 5 nm is also discussed.