Browsing by Author "Nieuwoudt, Arthur"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Modeling and design of carbon nanotube interconnect for mixed-signal VLSI applications(2008) Nieuwoudt, Arthur; Massoud, YehiaIn future nanoscale integrated circuits, process technology scaling coupled with increasing operating frequencies will exacerbate the resistivity, electromigration, and delay problems that plague interconnect in today's designs. Metallic carbon nanotubes are a promising future replacement for on-chip copper interconnect due to their large conductivity and current carrying capabilities. In this research, we develop modeling and design techniques for carbon nanotube-based interconnect solutions. We create an equivalent RLC circuit model for individual and bundled single-walled and multi-walled carbon nanotubes, which we leverage to determine the optimal design for nanotube-based interconnect solutions. Using the proposed modeling and design techniques, we investigate the performance and reliability of nanotube-based structures in future mixed-signal VLSI applications including digital interconnect and passive components for analog integrated circuits. We also examine the nanotube properties and fabrication requirements necessary for nanotube-based interconnect to be a competitive solution compared to standard copper technology. The results indicate that nanotube-based interconnect solutions will have the potential to revolutionize the next generation of integrated circuits in mixed-signal VLSI applications.Item Modeling, optimization and synthesis for fully integrated spiral inductors(2006) Nieuwoudt, Arthur; Massoud, YehiaAccurate and efficient modeling, optimization, and synthesis of integrated spiral inductors continue to hinder the automated design of mixed-signal circuits in system-on-chip technology. In this thesis, we develop a modeling and automated design methodology for integrated spiral inductors. We have created a wideband inductor model based on closed-form analytical expressions to capture a plethora of resistive, inductive, and capacitive parasitic effects. Leveraging the speed of the inductor model, we have developed a variability-aware automated design methodology that efficiently generates Pareto-optimal inductors based on application requirements. At its core the automated design methodology employs a scalable multi-level single-objective optimization engine that integrates the flexibility of deterministic pattern search optimization with the rapid convergence of local nonlinear convex optimization. The results demonstrate that the inductor modeling, optimization, and synthesis methodology accurately locates and characterizes near-optimal inductor designs with orders of magnitude speed improvement when compared with existing modeling and optimization techniques.