Atomic scale simulations of quantum transport in semiconductor nanowires

SNIC 2017/4-30


SNAC Small

Principal Investigator:

Sara Kheradsoud


Lunds universitet

Start Date:


End Date:


Primary Classification:

10304: Den kondenserade materiens fysik



Semiconductor nanowires are promising candidates for future high-speed, low-power transistors. However, a well-known problem is that nanowires often suffer from unintended defects, with unclear implications for the quantum transport properties. A particularly important example is the formation of segments with different crystal structures, giving rise to additional scattering at the segment interfaces. Hence, understanding the transport properties of crystal structure junctions would be of strong relevance for transistor research. In addition, investigations of nanowires for topics such as thermoelectric heat engines, spin manipulation and quantum computing would strongly benefit from better understanding of crystal structure interface transport properties. To provide accurate theoretical predictions of the interface transport properties, it is of importance to model the junctions between different crystal structures on the atomic scale. An atomistic model can account for central physical properties as lattice relaxation, impurity doping, electronic band structure formation and bending effects at the crystal structures interfaces. Moreover, to investigate transport properties, including both charge and heat transport, a non-equilibrium model accounting self-consistently for the bias induced charge density and correlation effects, is required. Such an atomistic, non-equilibrium quantum transport model the nanowire is computationally demanding; large computational resources are typically required to reach the number of atoms needed to faithfully capture the physics of the experimentally investigated wires. Since these resources are not available at the Physics Department at Lund University, with our project we apply for supercomputer resources. Our proposed project is part of a theory-experiment collaboration within a single-site European Union research and training network for PhD-students, PhD4Energy, located at Lund University. On the experimental side, at Nanolund, there is a world-leading experimental activity on semiconductor nanowire design, growth and characterization. In particular, by highly controlled growth conditions it is possible to fabricate individual wires with atomically sharp interfaces between different crystal structures. In a number of recent experiments, quantum transport measurements of systems with alternating crystal structures, Wurtzite and Zinc-blende segments of different lengths, have been performed, showing a number of interesting, unexpected physical properties. Typically, wires of InAs but also other material combinations such as GaAs and InP, have been investigated experimentally. However, a realistic, atom-by-atom simulation of the wires has not been performed. To perform such an investigation, on the theory side we collaborate with the transport software company Quantum Wise, benefitting from their long-standing expertise in the field. We perform our simulations using their Atomistix ToolKit (ATK) platform, combining density functional theory and non-equlibrium Green’s functions to describe the quantum transport properties of the nanowires. A detailed specification of the ATK platform, in particular parallelization and scalability for the specific project, as well as the requested computational resources are given in the resource usage section. An accepted application would allow us to carry out our investigation in two steps: first, to interpret the obtained experimental Lund data and second, provide general predictions and schemes for experimental transport characterization of nanowires with different crystal structures.