Atomistic modeling of electronic structure and transport in disordered nanostructures

Neerav Kharche, Purdue University

Abstract

As the Si-CMOS technology approaches the end of the International Technology Roadmap for Semiconductors (ITRS), the semiconductor industry faces a formidable challenge to continue the transistor scaling according to Moore’s law. To continue the scaling of classical devices, alternative channel materials such as SiGe, carbon nanotubes, nanowires, and III-V based materials are being investigated along with novel 3D device geometries. Researchers are also investigating radically new quantum computing devices, which are expected to perform calculations faster than the existing classical Si-CMOS based structures. Atomic scale disorders such as interface roughness, alloy randomness, non-uniform strain, and dopant fluctuations are routinely present in the experimental realization of such devices. These disorders now play an increasingly important role in determining the electronic structure and transport properties as device sizes enter the nanometer regime. This work employs the atomistic tight-binding technique, which is ideally suited for modeling systems with local disorders on an atomic scale. High-precision multi-million atom electronic structure calculations of (111) Si surface quantum wells and (100) SiGe/Si/SiGe heterostructure quantum wells are performed to investigate the modulation of valley splitting induced by atomic scale disorders. The calculations presented here resolve the existing discrepancies between theoretically predicted and experimentally measured valley splitting, which is an important design parameter in quantum computing devices. Supercell calculations and the zone-unfolding method are used to compute the bandstructures of inhomogeneous nanowires made of AlGaAs and SiGe and their connection with the transmission coefficients computed using non-equilibrium Green’s function method is established. A unified picture of alloy nanowires emerges, in which the nanodevice (transmission) and nanomaterials (bandstructure) viewpoints complement each other and illuminate the interesting physics of these disordered nanostructures that otherwise can not be explained using the traditional averaging methods such as the virtual crystal approximation. Finally, a multiscale modeling approach is employed, which combines the atomistic tight-binding method to compute the electronic structure and the real-space effective mass based quantum transport model including gate leakage to simulate the three terminal characteristics of III-V quantum well field effect transistors (QWFETs). The simulation methodology has been benchmarked against experimental data and it is then applied to investigate the logic performance of ultra-scaled III-V QWFETs with high mobility InAs channels.

Degree

Ph.D.

Advisors

Boykin, Purdue University.

Subject Area

Electrical engineering|Condensed matter physics|Nanotechnology

COinS