Three-dimensional plane-wave full-band quantum transport using empirical pseudopotentials

We study theoretically the ballistic performance of future sub-5 nm Field-Effect Transistors (FETs) using an atomistic quantum transport formalism based on empirical pseudopotentials, with armchair Graphene NanoRibbons (aGNRs), Silicon NanoWires (SiNWs) and zigzag Carbon NanoTubes (zCNTs) as channel structures. Due to the heavy computational burden from the plane-wave basis set, we restrict our study to ultrasmall devices, characterized by 5 nm channel lengths and 0.7 nm $\times$ 0.7 nm cross-sectional areas. Band structure calculations show that aGNRs have an oscillating chirality-dependent band gap. AGNRs with dimer lines N=3p+1 have large band gaps and aGNRFETs show promising device performance in terms of high $I_{on}$/$I_{off}$, small drain-induced barrier lowering and limited short channel effects due to their very thin body and associated excellent electrostatics control. N=3p+2 aGNRs have small band gaps and band-to-band tunneling generates a large current at high bias. We also discuss spurious solutions introduced by the envelope function approximation. Device characteristics of SiNWFETs and zCNTFETs are compared to aGNRFETs as well.