TITLE:

MULTISCALE MODELING OF DIFFUSION- INDUCED DEFORMATION PROCESSES (No. 20)

DATE:

Friday, September 26th, 2003

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Eugene Olevsky, Department of Mechanical Engineering, San Diego State University

ABSTRACT:

The developed integrated approach of both macroscopic (continuum finite-element), mesoscopic (Monte-Carlo stochastic), and microscopic (Monte-Carlo stochastic) modeling of diffusion-induced deformation processes such as sintering, is employed as a basis of a full-fledged multi-scale virtual reality of the nano-materials processing, which is broadly used in the fabrication of electronic circuitry. This methodology is dedicated to the materials processing system-level analysis, especially the high-temperature fabrication of nano-material systems where hierarchy of the material structure is essential.

Sintering is a high temperature process of bonding together particles at high temperatures. It occurs at temperatures below the melting point, but in some instances involves the formation of a liquid phase. As a result of sintering, density of a powder compact increases and most material properties are improved. For nano-crystalline materials, sintering is a result of matter transport by different diffusion mechanisms driven by the high surface energy of aggregates of fine particles and activated by high temperature.

For modeling of sintering, a special innovative computer procedure for interaction between different material scales is developed. The proposed computational framework at the macroscopic level envelopes the meso-scopic simulators. All constitutive parameters are obtained through numerical experiments in the virtual test environment. Within this approach only one macroscopic element at a time is considered. What is kept in memory for every element is the macroscopic history of loading: the trajectory of the evolution of macroscopic stresses, temperatures and electric fields applied to the element. When a time-step of the calculations is finished for one macroscopic element, the micro- and meso- structures of the next element are restored from the initial state according to the history of loading. A lack of computer memory is compensated by the use of massive parallel computing. The meso-scopic steps for different elements are independent of each other and parallel computers are ideally capable of conducting the calculations.

The developed approach is used for obtaining solutions for problems of co-firing of multi-layer ceramic nano-composites utilized in wireless applications

HOST:

Jose Castillo

DOWNLOAD: