In situ dynamic imaging, using an environmental transmission electron microscope, was employed to evaluate the catalytic activity of Au/SiO(2), Ni/SiO(2), and Au-Ni/SiO(2) nanoparticles for the formation of one-dimensional (1-D) carbon nanostructures such as carbon nanofibers (CNFs) and nanotubes (CNTs). While pure-Au thin-film samples were inactive for carbon deposition at 520 °C in 0.4 Pa of C(2)H(2), multiwalled CNTs formed from Ni thin films samples under these conditions. The number of nanoparticles active for CNF and CNT formation increased for thin films containing 0.1 mol fraction and 0.2 mol fraction of Au but decreased as the overall Au content in thin films was increased above 0.5 mol fraction. Multiwalled CNTs formed with a root growth mechanism for pure Ni samples, while with the addition of 0.1 mol fraction or 0.2 mol fraction of Au, CNFs were formed via a tip growth mechanism at 520 °C. Single-walled CNTs formed at temperatures above 600 °C in samples doped with less than 0.2 mol fraction of Au. Ex situ analysis via high-resolution scanning transmission electron microscopy (STEM) and energy-dispersive X-ray spectroscopy (EDS) revealed that catalytically active particles exhibit a heterogeneous distribution of Au and Ni, where only a small fraction of the overall Au content was found in the portion of each particle actively involved in the nucleation of graphitic layers. Instead, the majority of the Au was found to be segregated to an inactive capping structure at one the end of the particles. Using density-functional theory calculations, we show that the activation energy for bulk diffusion of carbon in Ni reduces from ≈1.62 eV for pure Ni to 0.07 eV with the addition of small amounts (≈0.06 mol fraction) of Au. This suggests that the enhancement of C diffusion through the bulk of the particles may be responsible for improving the number of particles active for nucleating the 1-D carbon nanostructures and thereby the yield.

No abstract

This paper presents a new learning theory (a set of principles for brain-like learning) and a corresponding algorithm for the neural-network field. The learning theory defines computational characteristics that are much more brain-like than that of classical connectionist learning. Robust and reliable learning algorithms would result if these learning principles are followed rigorously when developing neural-network algorithms. This paper also presents a new algorithm for generating radial basis function (RBF) nets for function approximation. The design of the algorithm is based on the proposed set of learning principles. The net generated by this algorithm is not a typical RBF net, but a combination of "truncated" RBF and other types of hidden units. The algorithm uses random clustering and linear programming (LP) to design and train this "mixed" RBF net. Polynomial time complexity of the algorithm is proven and computational results are provided for the well known Mackey-Glass chaotic time series problem, the logistic map prediction problem, various neuro-control problems, and several time series forecasting problems. The algorithm can also be implemented as an online adaptive algorithm.

A Monte Carlo code was developed for simulating the electron cascade in radiation detector materials. The electron differential scattering cross sections were derived from measured electron energy-loss and optical spectra, making the method applicable for a wide range of materials. The detector resolution in a simplified model system shows dependence on the bandgap, the plasmon strength and energy, and the valence band width. In principle, these parameters could be optimized to improve detector performance. The intrinsic energy resolution was calculated for three semiconductors: silicon (Si), gallium arsenide (GaAs), and zinc telluride (ZnTe). Setting the ionization thresholds for electrons and holes is identified as a critical issue, as this strongly affects both the average electron-hole pair energy w and the Fano factor F. Using an ionization threshold from impact ionization calculations as an effective bandgap yields pair energies that are well matched to measured values. Fano factors of 0.091 (Si), 0.100 (GaAs), and 0.075 (ZnTe) were calculated. The Fano factor calculated for silicon using this model was lower than some results from past simulations and experiments. This difference could be attributed to problems in simulating inter-band transitions and the scattering of low-energy electrons.

No abstract