No description available.

No description available.

Thesis (Ph. D.)--Lehigh University, 2006. / Includes vita. Includes bibliographical references (leaves 115-127).

Thesis (Ph. D.)--State University of New York at Binghamton, Systems Science, 2005. / Includes bibliographical references.

Thesis (M.Phil.)--Hong Kong University of Science and Technology, 2006. / Includes bibliographical references (leaves 97-107). Also available in electronic version.

Today’s thin-film passive components such as capacitors and inductors are limited to low volumetric density and large form-factors that pose as major roadblock to miniaturization of the power modules. These components are also placed far away from the IC’s leading to large interconnect parasitics and lower operating frequencies. Novel thin-film technologies with high densities and small form-factors are, therefore, required to enable miniaturization and performance at high frequencies. Glass- and silicon- based interposer technologies that utilize vertical through-via interconnections have shown way to improve power distribution network (PDN) performance with thin power-ground planes. However, integration of ultra-high density capacitors in such substrates has not yet been demonstrated. This thesis addresses these challenges with tantalum-based, silicon-integrated, ultrathin, high-density capacitors at higher operating frequencies with lower leakage properties (<0.01µA/µF). The anodization kinetics of tantalum pentoxide and the underlying leakage current mechanisms are investigated to provide optimal process guidelines. The thin-film Ta capacitors demonstrated capacitance density of 0.1 µF/mm2 at 1-10 MHz in form-factors of 50 µm, which corresponds to 6X higher volumetric density relative to commercial tantalum capacitors. An innovative approach to address incompatibility of tantalum electrodes with substrates is pursued by prefabricating the electrodes on a free-standing foil, which are then transferred onto the active wafer to form the capacitors on Si. The integration approach is designed to embed these thin tantalum capacitors on alternative substrates such as organic, glass or silicon, with copper via interconnections for lower parasitics. The thesis also explores titanium-based high-density capacitors with high-permittivity titania dielectric as a potential alternate high-density capacitor technology.

This thesis proposes a novel graphical model for inference called the Affinity Network,which displays the closeness between pairs of variables and is an alternative to Bayesian Networks and Dependency Networks. The Affinity Network shares some similarities with Bayesian Networks and Dependency Networks but avoids their heuristic and stochastic graph construction algorithms by using a message passing scheme. A comparison with the above two instances of graphical models is given for sparse discrete and continuous medical data and data taken from the UCI machine learning repository. The experimental study reveals that the Affinity Network graphs tend to be more accurate on the basis of an exhaustive search with the small datasets. Moreover, the graph construction algorithm is faster than the other two methods with huge datasets. The Affinity Network is also applied to data produced by a synchronised system. A detailed analysis and numerical investigation into this dynamical system is provided and it is shown that the Affinity Network can be used to characterise its emergent behaviour even in the presence of noise.

No description available.

Thesis (Ph. D.)--School of Mechanical Engineering, Georgia Institute of Technology, 2004. Directed by Said I. Abdel-Khalik. / Includes bibliographical references (leaves 292-296).

Thesis (M.S.)--Mechanical Engineering, Georgia Institute of Technology, 2004. / Daniel Baldwin, Committee Chair; Suresh Sitaraman, Committee Member; Steven Danyluk, Committee Member. Includes bibliographical references (leaves 238-241).