Development of a Massively Parallel Nanoelectronic Modeling Tool and Its Application to Quantum Computing Devices

Lee, Sun Hee Ph.D., Purdue University, December 2011. Development of a Massively Parallel Nanoelectronic Modeling Tool and Its Application to Quantum Computing Devices . Major Professor: Gerhard Klimeck. The rapid progress in nanofabrication technologies has led to the possibility of realizing scalable solid-state quantum computers (QC) which have the potential to outperform conventional microprocessors. Using STM patterning followed by low temperature MBE, phosphorus doping in silicon can be controlled with atomic precision to fabricate Si:P delta doping layers for semi-metallic contacts and leads as well as for few-donor quantum dots (QDs). To model Si:P quantum dot systems in realistic domains, we have developed a nanoelectronic modeling tool (NEMO3D-peta) to expand existing NEMO3D capabilities with advanced parallelization schemes and to provide more flexibility to the code through an object-oriented design approach. Benchmark studies are performed on various aspects and NEMO3D-peta is proven to scale up to 32,000 processors. Furthermore, a variety of applications including a charge-potential self-consistent module are implemented. In the first step of QD system modeling with NEMO3D-peta, the electronic structure of delta doped contacts under equilibrium condition is computed self-consistently based on an atomistic sp3d5s∗ tight-binding Hamiltonian. Bandstructure results are compared against previous ab-initio studies and shown to be in good agreement in terms of valley minima and Fermi level positions. In addition, the effect of random dopant disorder in the delta layers is investigated in extended domains. Although fluctuations in the density of states (DOS) in a disordered supercell are expected, a high DOS will exhibit little effects on the conduction properties of Si:P delta doped layers. Finally, work is in progress on predicting valley splitting (VS) and excited states in electrostatically