Brown University
Joint Materials/Solid Mechanics Seminar Series
T.J. Watson Research Center, Yorktown Heights, NY 01598 | |
Abstract: As the scaling of silicon technology becomes more difficult, the search for alternatives has intensified. New schemes will inevitably involve complex materials (e.g. SiGe, silicon-on- insulator, and metal oxides) as well as new atomic-level processing techniques. It is now clear that many of the difficulties in implementing these new technologies can be traced back to a lack of understanding of the fundamental kinetics and thermodynamics of growth. Characterizing complex processes like alloy formation, segragation, and oxidation using "cook and look" microscopy can be difficult due to the ambiguities introduced by ramp-up and cool-down. What is called for is a real-time, in situ microscopy capable of imaging surfaces at elevated temperatures and pressures. One technique that meets many of these requirements is Low-Energy Electron Microscopy, or LEEM.
In this talk I will describe quantitative measurements of thermodynamics and kinetic processes at the Si(111) and Si(001) surfaces. In LEEM a collimated, monochromatic (0-100 eV) electron beam is directed towards a surface and an image is created from the diffracted/reflected electrons. Images are formed in real-time, at arbitrary surface temperature, and with nanometer spatial resolution. Contrast mechanisms are similar to those used in TEM. Three experiments will be discussed that highlight the capabilities of LEEM: (1) determining the rate-limiting kinetic processes during a 2D phase transition, (2) exploiting surface stress to form nanostructures with tunable dimensions, and (3) the discovery of a novel alloying mechanism during the growth of Ge on Si(001).
References:
"Dynamics of the Si(111) Surface Phase Transition," J.B. Hannon, H. Hibino, N.C. Bartlet, B.S. Swartzentruber, T. Ogino, and G.L. Kellogg, Nature, 405 (2000) 299.
"Surface Stress and Thermodynamic Nanoscale Size Selection," J.B. Hannon, J. Tersoff, and R.M. Tromp, Science, 295 (2002) 299.
"Critical Role of Surface Steps in the Alloying of Ge on Si(001)," J.B. Hannon, M. Copel, R. Stumpf, M.C. Reuter, and R.M. Tromp, Phys. Rev. Lett., 92 (2004) 216104 proteins in the airway.
Center for Fluid Mechanics Seminar
Massachusetts Institute of Technology, Cambridge, MA | |
Abstract: Lubrication theory is a powerful technique to investigate flows in which one length scale is considerably 'thinner' than the others. Such thin films occur in many biological systems (e.g. the liquid lining of the lungs, the cornea of the eye), in environmental settings (e.g. agrochemicals, oil recovery from porous rocks), industrial applications (e.g. paints and adhesives, the coating of microcomponents), and material sciences (e.g. foams and colloids).
Traditional lubrication models have been developed primarily for unidirectional films. However, there are many situations which concern liquid films in corner regions, such as the break-up of soap films, the 'fishbone' instabilities of non-Newtonian fluids and flow through capillaries. On a fundamental level, a corner is the simplest prototype of any geometrical inhomogeneity at the scale of the film thickness. To address this important issue we have developed a new method capable of describing flow in such geometries. The method uses a hyperbolic coordinate system, which has enabled substantial analytical progress. We present results for two scenarios: a free film (e.g. foams) and a film over a solid substrate (e.g. microcoating). In addition, we present the results of supporting experiments and discuss the flexibility of the model with respect to the inclusion of new physico-chemical effects (e.g. van der Waals or double layer forces).
Scientific Computing Seminar
Abstract: A fundamental difficulty in the computational analysis of wave propagation problems in the time domain is the accurate truncation of the unbounded physical domain. In recent years new approaches to applying radiation boundary conditions at artificial boundaries have been developed for the wave equation and Maxwell's equations which can provide arbitrary accuracy at reasonable cost. These are based on high-order spatially local and nonlocal complex exponential approximations to convolution kernels. We will give a basic description of the construction, analysis and implementation of these methods, contrasting their relative advantages and disadvantages. We will also discuss ongoing work to find an ultimate method with none of the disadvantages of the current techniques, as well as the prospect for generalizing them to a larger class of wave propagation problems.
Brown University
Graduate School, Dissertation Defense Information
Neural Coding
PDE Seminar
<--- 2004 Index