Carbon Nanotubes (CNTs) are seen as a viable replacement for silicon technology based on their small dimensions and ballistic transport. CNT based field effect transistors (FETs) outperform silicon based metal-oxide-semiconductor field effect transistors (MOSFETs) in sub-10nm regimes. Within CNTFETs, doped CNTFETs have a three orders better on-to-off current ratio compared to Schottky Barrier CNTFETs, but still, dopant incurred effects in 1D CNTs remain enigmatic due to a lack of reproducible results. Considering doping by an adsorption mechanism, an added dopant in CNTs leads to quantum wave-function interactions, dependent on dopant location, which disrupts the pristine eigen-states. It also causes attractive/ repulsive forces on the π bonds of CNTs, leading to induced strain....[ Read more ]

Carbon Nanotubes (CNTs) are seen as a viable replacement for silicon technology based on their small dimensions and ballistic transport. CNT based field effect transistors (FETs) outperform silicon based metal-oxide-semiconductor field effect transistors (MOSFETs) in sub-10nm regimes. Within CNTFETs, doped CNTFETs have a three orders better on-to-off current ratio compared to Schottky Barrier CNTFETs, but still, dopant incurred effects in 1D CNTs remain enigmatic due to a lack of reproducible results. Considering doping by an adsorption mechanism, an added dopant in CNTs leads to quantum wave-function interactions, dependent on dopant location, which disrupts the pristine eigen-states. It also causes attractive/ repulsive forces on the π bonds of CNTs, leading to induced strain. Electrostatic as well as spatial position and strain effects of dopants are the subject matter of this thesis.

We utilized CNT sparse geometry together with first principle simulations to quantize CNT transport characteristics versus dopant position. A model of the electrostatic dipole variation with the dopant location was used for predicting the conductance degradation in n-type devices due to the dopant proximity to the metal interface. Dopant stationed away from the depletion region gave the least contact resistance, and improvement faded away, though better than the un-doped case, as dopant was placed away from the interface.

Dopant generated the strain effect in the source/drain regions of CNTFETs and its effect on the device current was investigated by employing Density functional theory (DFT) for geometry optimization and numerical simulation for idea qualification. DFT showed that adsorbed dopants provoke strain, and the numerical results discerned ~25% current over-estimation. A surface-potential based model was presented, incorporating the dopant incited strain effect. It agreed well with the numerical results in the on-state, but suffered from band-to-band tunneling (BTBT) in the sub-threshold regime.

BTBT, a ramification of dopant electrostatics at the interface of doped/un-doped CNTs, allowed source confined carriers to tunnel into the channel valence band, degrading the current on-off ratio. A Poisson equation was used to model the interface electrostatics, neglecting the triangular potential approximation based model with constant tunneling length. We then presented a complete compact model for doped CNTFETs, with negligible error at both the on and off-state, for realistic device behavior.