A novel high-density dual threshold GNRFET SRAM design with improved stability

We propose a novel single-ended static random access memory (SRAM) design with nine graphene nanoribbon FETs (9-GNRFET) in this paper. Single-ended has an impact on density, delays, static noise margin (SNM) and power consumption. The proposed model is implemented in HSPICE as a library for 16 nm GN...

Full description

Saved in:
Bibliographic Details
Published in:Microprocessors and microsystems 2020-03, Vol.73, p.102956, Article 102956
Main Authors: Patel, Pramod Kumar, Malik, M.M., Gupta, Tarun K.
Format: Article
Language:English
Subjects:
Circuit design
Computer architecture
Computer simulation
Density
Electric potential
Energy conservation
GNRFET
Graphene
Liability
Model accuracy
Power consumption
Random access memory
Simulation
SRAM
Static noise margin
Static random access memory
Ultra-low power
Voltage
Write voltage margin
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a novel single-ended static random access memory (SRAM) design with nine graphene nanoribbon FETs (9-GNRFET) in this paper. Single-ended has an impact on density, delays, static noise margin (SNM) and power consumption. The proposed model is implemented in HSPICE as a library for 16 nm GNRFET technology. This HSPICE-compatible compact model provides accuracy while maintain compactness, and make possible efficient circuit level simulations of futuristic GNRFET-based SRAM cells design. Simulations at low supply voltage of 0.325 V have shown that proposed cell provides power saving 4.8 × as compared to a supply voltage of 0.7 V. The half-select free technique provides bit interleaving architecture, consisting of error-free operations with VDD down to 325 mV. The proposed architecture implemented in 16 nm low leakage GNRFET technology presents the scalability of these cells near threshed voltage region, which can significantly reduce power consumptions with 0.21µW. The proposed SRAM cell design is based on simulations and results are verified on GNRFET HSPICE-compact model. The proposed cell verified under process variation, and is demonstrated with write-assist, the impact of geometrical liability and adaptive supply voltage scaling.
ISSN:0141-9331
1872-9436
DOI:10.1016/j.micpro.2019.102956