A group of compounds referred to as metal hydrides, when used as electrode materials, is a less toxic alternative to the cadmium hydroxide electrode found in nickel/cadmium secondary battery systems. For this and other reasons, the nickel/metal hydride battery system is becoming a popular rechargeable battery for electric vehicle and consumer electronics applications. A model of this battery system is presented. Specifically the metal hydride material, LaNi{sub 5}H{sub 6}, is chosen for investigation due to the wealth of information available in the literature on this compound. The model results are compared to experiments found in the literature. Fundamental analyses as well as engineering optimizations are performed from the results of the battery model. In order to examine diffusion limitations in the nickel oxide electrode, a ``pseudo 2-D model`` is developed. This model allows for the theoretical examination of the effects of a diffusion coefficient that is a function of the state of charge of the active material. It is found using present data from the literature that diffusion in the solid phase is usually not an important limitation in the nickel oxide electrode. This finding is contrary to the conclusions reached by other authors. Although diffusion in the nickel oxide …

We investigated the surface bonding of various adsorbates (0, S, C{sub 2}H{sub 3} and NO) along with the resulting relaxation of the Pt(111) surface using low energy electron diffiraction (LEED). LEED experiments have been performed on these ordered overlayers along with theoretical structural analysis using automated tensor LEED (ATLEED). The resulting surface structures of these ordered overlayers exhibit similar adsorbate-induced relaxations. In all cases the adsorbate occupies the fcc hollow site and induces an approximately 0.1 A buckling of the metal surface. The three metal atoms directly bonded to the adsorbate are ``pulled`` out of the surface and the metal atom that is not bound to the adsorbate is `pushed`` inward. In order to understand the reliability of such details, we have carried out a comprehensive study of various non-structural parameters used in a LEED computation. We also studied the adsorption of water on the Pt(lll) surface. We ordered an ultra thin ice film on this surface. The film`s surface is found to be the (0001) face of hexagonal ice. This surface is apparently terminated by a full-bilayer, in which the uppermost water molecules have large vibrational amplitudes even at temperatures as low as 90 K. We examined two other …

The author has applied the Terahertz Time Domain Interferometric (THz-TDI) technique to probe the ultrafast dynamic response of a Superconducting-Insulating-Superconducting (SIS) tunnel junction and a Quantum Point Contact (QPC). The THz-TDI technique involves monitoring changes in the dc current induced by interfering two picosecond electrical pulses on the junction as a function of time delay between them. Measurements of the response of the Nb/AlO{sub x}/Nb SIS tunnel junction from 75--200 GHz are in full agreement with the linear theory for photon-assisted tunneling. Likewise, measurements of the induced current in a QPC as a function of source-drain voltage, gate voltage, frequency, and magnetic field also show strong evidence for photon-assisted transport. These experiments together demonstrate the general applicability of the THz-TDI technique to the characterization of the dynamic response of any micron or nanometer scale device that exhibits a non-linear I-V characteristic. 133 refs., 49 figs.

Atmospheric chemical-radiative-transport (CRT) models are vital in performing research on atmospheric chemical change. Even with the enormous computing capability delivered by massively parallel systems, extended three dimensional CRT simulations are still not computationally feasible. The major obstacle in a CRT model is the nonlinear ODE system describing the chemical kinetics in the model. These ODE systems are usually very stiff and account for anywhere from 75% to 90% of the CPU time required to run a CRT model. In this study, a simple explicit class of time stepping method is developed and demonstrated to be useful in treating chemical ODE systems without the use of a Jacobian matrix. These methods, called preconditioned time differencing methods, are tested on small mathematically idealized problems, box model problems, and full 2-D and 3-D CRT models. The methods are found to be both fast and memory efficient. Studies are performed on both vector and parallel systems. The preconditioned time differencing methods are established as a viable alternative to the more common backward differentiation formulas in terms of CPU speed across architectural platforms.

We describe a spectral approach to the investigation of fluid instability, generalized turbulence, and the interpenetration of fluids across an interface. The Technique also applies to a single fluid with large variations in density. Departures of fluctuating velocity components from the local mean are far subsonic, but the mean Mach number can be large. Validity of the description is demonstrated by comparisons with experiments on turbulent mixing due to the late stages of Rayleigh-Taylor instability, when the dynamics become approximately self-similar in response to a constant body force. Generic forms for anisotropic spectral structure are described and used as a basis for deriving spectrally integrated moment equations that can be incorporated into computer codes for scientific and engineering analyses.

Nuclear magnetic resonance techniques are traditionally used to study molecular level structure and dynamics with a noted exception in medically applied NMR imaging (MRI). In this work, new experimental methods and theory are presented relevant to the study of macroscopic morphology and dynamics using NMR field gradient techniques and solid state two-dimensional exchange NMR. The goal in this work is not to take some particular system and study it in great detail, rather it is to show the utility of a number of new and novel techniques using ideal systems primarily as a proof of principle. By taking advantage of the analogy between NMR imaging and diffraction, one may simplify the experiments necessary for characterizing the statistical properties of the sample morphology. For a sample composed of many small features, e.g. a porous medium, the NMR diffraction techniques take advantage of both the narrow spatial range and spatial isotropy of the sample`s density autocorrelation function to obtain high resolution structural information in considerably less time than that required by conventional NMR imaging approaches. The time savings of the technique indicates that NMR diffraction is capable of finer spatial resolution than conventional NMR imaging techniques. Radio frequency NMR imaging with a …

The ability of organisms to evolve and adapt to the environment has provided mother nature with a rich and diverse set of species. Only organisms well adapted to their environment can survive from one generation to the next, transferring on the traits, that made them successful, to their offspring. Competition for resources and the ever changing environment drives some species to extinction and at the same time others evolve to maintain the delicate balance in nature. In this disertation we present the multi-niche crowding genetic algorithm, a computational metaphor to the survival of species in ecological niches in the face of competition. The multi-niche crowding genetic algorithm maintains stable subpopulations of solutions in multiple niches in multimodal landscapes. The algorithm introduces the concept of crowding selection to promote mating among members with qirnilar traits while allowing many members of the population to participate in mating. The algorithm uses worst among most similar replacement policy to promote competition among members with similar traits while allowing competition among members of different niches as well. We present empirical and theoretical results for the success of the multiniche crowding genetic algorithm for multimodal function optimization. The properties of the algorithm using different parameters are …