Atomic-scale modelling of interfaces in nanoscale electronics

The evolution of field effect transistors (FETs) has enabled the digital revolution which shapes the everyday lives of people all over the world. Currently, new materials and designs are investigated in pursuit of continuing the improvement of device performance and efficiency. This thesis investigates two-dimensional (2D) materials for future transistor designs.

The research has mainly focused on the transition metal dichalcogenide (TMD) heterophase FET design which uses the metallic phase of the TMD as the source and drain electrodes and the semiconducting phase as the channel. Density functional theory and the non-equilibrium Green’s function method are used to investigate the charge transfer at interfaces between these 2D materials and a method for predicting stable interfaces between two crystals has been developed.

The Schottky barrier between contact and channel in a transistor is perceived as an intrinsic property which predicts how well the device will perform. In this thesis, ab-initio calculations on phase-engineered MoTe2 show that for these 2D Schottky contacts, this is not the case. Interface states and standing waves due to quantum confinement mediate tunneling current and renders the Schottky barrier height a poor descriptor of device performance.

The results also demonstrate that the electrostatic response of the Schottky barrier can’t be predicted by the conventional models which means the Schottky barrier found from the activation energy method isn’t well-defined either. The relevance of using a heterophase MoTe2 design for next-generation transistors is assessed by reviewing the available literature and by comparing with the ab-initio calculations and the performance goals of a 2025 device as defined by the International Roadmap for Devices and Systems.1 The comparison focuses on the ON-current, sub-threshold slope, and power supply voltage and concludes that the heterophase design has the potential to perform according to the goals of the roadmap if a high-κ dielectric is used as the gate oxide.

The largest obstacle for this design to succeed is the lack of a scalable method for doping of 2D semiconductors. If such a method is developed, the heterophase TMD transistors seem to be a viable option for future transistor designs.