Saving Moore's Law Down To 1nm Channels With Anisotropic Effective Mass

TL;DR

This paper demonstrates that using anisotropic effective mass materials enables the scaling of transistors down to 1nm, maintaining excellent switching performance in MOSFETs and TFETs, thus advancing Moore's Law.

Contribution

It introduces a novel design leveraging anisotropic effective mass materials for ultra-scaled transistors, supported by atomistic quantum transport simulations.

Findings

Channel scaling to 1nm is feasible with anisotropic mass materials.

The proposed TFET design improves ON/OFF performance.

Simulations confirm excellent switching at 1nm scale.

Abstract

Scaling transistors' dimensions has been the thrust for the semiconductor industry in the last 4 decades. However, scaling channel lengths beyond 10 nm has become exceptionally challenging due to the direct tunneling between source and drain which degrades gate control, switching functionality, and worsens power dissipation. Fortunately, the emergence of novel classes of materials with exotic properties in recent times has opened up new avenues in device design. Here, we show that by using channel materials with an anisotropic effective mass, the channel can be scaled down to 1nm and still provide an excellent switching performance in both MOSFETs and TFETs. In the case of TFETs, a novel design has been proposed to take advantage of anisotropic mass in both ON- and OFF-state of the TFETs. Full-band atomistic quantum transport simulations of phosphorene nanoribbon MOSFETs and TFETs based on the new design have been performed as a proof.