Power optimized variation aware dual-threshold SRAM cell design technique

Aminul Islam¹, Mohd Hasan^2
1Department of Electronics and Communication Engineering, Birla Institute of Technology, Mesra, Ranchi, Jharkhand, India; ²Department of Electronics Engineering, Aligarh Muslim University, Aligarh, Uttar Pradesh, India

Abstract: Bulk complementary metal-oxide semiconductor (CMOS) technology is facing enormous challenges at channel lengths below 45 nm, such as gate tunneling, device mismatch, random dopant fluctuations, and mobility degradation. Although multiple gate transistors and strained silicon devices overcome some of the bulk CMOS problems, it is sensible to look for revolutionary new materials and devices to replace silicon. It is obvious that future technology materials should exhibit higher mobility, better channel electrostatics, scalability, and robustness against process variations. Carbon nanotube-based technology is very promising because it has most of these desired features. There is a need to explore the potential of this emerging technology by designing circuits based on this technology and comparing their performance with that of existing bulk CMOS technology. In this paper, we propose a low-power variation-immune dual-threshold voltage carbon nanotube field effect transistor (CNFET)-based seven-transistor (7T) static random access memory (SRAM) cell. The proposed CNFET-based 7T SRAM cell offers ~1.2× improvement in standby power, ~1.3× improvement in read delay, and ~1.1× improvement in write delay. It offers narrower spread in write access time (1.4× at optimum energy point [OEP] and 1.2× at 1 V). It features 56.3% improvement in static noise margin and 40% improvement in read static noise margin. All the simulation measurements are taken at proposed OEP decided by the optimum results obtained after extensive simulation on HSPICE (high-performance simulation program with integrated circuit emphasis) environment.

Keywords: carbon nanotube field effect transistor (CNFET), chirality vector, random dopant fluctuation (RDF), SNM