Input referred low-frequency noise analysis for single-layer graphene FETs

TL;DR

This paper investigates the bias-dependent input referred low-frequency noise in single-layer graphene FETs using a physics-based model, revealing how different fluctuation effects influence noise behavior across bias points.

Contribution

It introduces a comprehensive compact noise model for graphene FETs that captures the bias-dependent SVG behavior and its underlying physical mechanisms.

Findings

Minimum SVG at maximum transconductance bias point

SVG exhibits a parabolic shape vs. gate voltage

Mobility fluctuation dominates near charge neutrality point

Abstract

The bias-dependence of input referred low-frequency noise (LFN), SVG, is a considerable facet for RF circuit design. SVG was considered constant in silicon transistors but this was contradicted by recent experimental and theoretical studies. In this letter, the behaviour of SVG is investigated for single-layer graphene transistors based on a recently established physics-based complete compact LFN model. A minimum of SVG is recorded at the bias point where maximum transconductance is located which coincides with the peak of the well-known M-shape of the normalized output LFN and the model precisely captures this trend. Mobility fluctuation effect increases SVG towards to lower currents near charge neutrality point (CNP) while carrier number fluctuation and series resistance effects mostly contribute away from CNP; thus, SVG obtains a parabolic shape vs. gate voltage similarly to CMOS devices.