Multiphysics Simulation of Nanostructured Semiconductor Devices

Thesis (Ph.D.)--University of Washington, 2013

This dissertation describes how to utilize multiphysics simulation techniques to develop physical (optics and device) models of nanostructured semiconductor devices. A hierarchy of approaches (ab-initio, continuum, transfer-matrix, finite-difference time-domain) is used to bridge the gaps in computational efficiency and physical accuracy. This modeling approach is demonstrated by investigating several challenging applications: organic solar cells and atomic force microscope direct writing. In the first application, optics and device simulation techniques are coupled together with the aim of improving organic solar cell efficiency. Transfer matrix method (TMM) is used to understand the photon distribution and absorption in photovoltaics with layered structure. The optimized thickness of organic light-absorbing layer can be matched accurately with experiments. We also investigate the light absorption enhancement due to surface-plasmonic effect in presence of silver nano-prisms. Finite-difference time-domain (FDTD) solutions of Maxwell's equations suggest benefits of utilizing silver nano-prisms only for ultra-thin organic light-absorbing layer (thickness < 50 nm). Device models involve the Poisson equation, charge continuity equations, and organic-specific singlet exciton generation/dissociation equation. By considering the effect from deep trap states, we calibrate our model via matching the quantum efficiency and current-voltage curves with experimental data. We can also control the current-voltage response by tuning the work-function of metal nano-prisms, due to tunneling current through ultra-thin dielectric layer between silver nano-prism and surrounding organic light-absorbing material, which leads to charge accumulation. Finally, by combining optics and device simulation, we identify a trade-off between plasmon-mediated optical enhancement and optical and electronic losses associated with low aspect-ratio silver nanoprisms in plasmonic organic solar cells. We predict up to 26% enhancement of power conversion efficiency (PCE) for active layer thicknesses less than 50 nm. The second application addresses the effect of temporary-negative-ion (TNI) generation and dissociation in the liquid precursor located between the tip of atomic force microscope and silicon wafer coated with SiO2. Ab-initio calculations were used to extract the density of states of the TNI's at their energy-favorable states. Device simulation illustrates nonlocalized Fowler-Nordheim type of tunneling current, in which the current density is a function of applied electric field. Based on the device simulation models considering the TNI's tunneling effect, we can match the current-voltage curve of direct AFM writing with experiments under various applied voltages. We also developed a methodology to extract the shape and deposition rate of silicon/germanium nanostructures under extremely high external electric field at the tip of atomic force microscopy. Finally, we derived an analytical model that can predict the deposited shape of silicon/germanium nanostructures as a function of AFM writing voltage and speed.