Performance prediction of graphene-nanoribbon and carbon nanotube transistors

Abstract

Technology exploration is carried out through the modeling of zigzag carbon nanotube field-effect-transistors (z-CNTFETs) and armchair graphene nanoribbon field-effect-transistors (a-GNRFETs) with top gate design. The devices are simulated using a top-of-the-barrier model [1] where the energy dispersion for CNTs and GNRs is based on the tight-binding approximation [2]. The structure of these transistors is shown in Fig. 1. In armchair GNRs, two Dirac points (K and K′) are merged into one valley (gv=1), whereas for CNTs two discrete valleys (gv=2) are included [3]. Unlike gapless two-dimensional (2D) graphene, nanometer-wide GNRs can have semiconducting characteristics due to quantum confinement by tailoring its width as illustrated Fig. 2. Table I shows the contact, channel and quantum resistance for a GNR and a CNT computated using Ron (L) = h/(2gvq²) × (L/ℓ) + h/(2gvq²) + Rnc where ℓ is the electron mean free path (MFP) given as ℓ=(1/λAP+1/λOP+1/λEDGE(GNR))⁻¹, Rnc is the non-transparent resistance, Rc=RQ+ Rnc is the contact resistance and RQ is the quantum resistance given by h/(2gvq²) [4]. In addition, the MFP of optical phonon, acoustic phonon and edge scattering are as follows; λOP,300 ≈15d, λAP,300 ≈ 280d, λEDGE= 15nm where d is diameter [5–6].