Impact of the statistical variability on 15nm III–V and Ge MOSFET based SRAM design

The use of III-V and Ge as channel materials is a feasible alternative to Si for the next generation of planar CMOS. We present a set of BSIM4 compact models for III-V and Ge Implant Free Quantum Well (IFQW) transistors which are extracted from well-calibrated TCAD simulations that include the impact of process induced statistical variability. We consider the design of 6T-SRAM using this simulation approach, and investigate the impact of process variability on the memory cell performance. The optimized cell design solutions are presented and discussed.