Multiscale Design of Advanced
Materials based on Hybrid Ab-Initio
and Quasicontinuum Methods
EMSL Project ID
18590
Abstract
A multiscale method based on hydrid ab initio and quasicontinuum methods will be developed on a rigorous mathematical, physical, and chemical foundation. Rather than being based on classical interatomic potentials, this method will utilize quantum mechanics-based potentials capable of realistically describing the complex chemical bonding required to meet the design needs of advanced materials. Longer time scales will be achieved by a new accelerated ab initio molecular dynamics algorithm. Mathematical analysis of the quasicontinuum method rigorously validating the method and ensuring its accuracy will be given. This will provide error indicators to be used with modern adaptive mesh techniques and state-of-the-art multilevel solution methods to yield the most efficient and reliable numerical solution. New parallelization techniques and software tools capable of working with adaptive grids and multiphysics computational methods will be developed for advanced computer platforms to simulate realistic materials. The project team will work closely with the Pacific Northwest National Laboratory on interfacial catalysis, metal organic framework materials for hydrogen storage, and fusion materials able to withstand intense neutron radiation. Each of these applications presents multiscale challenges that are likely to be overcome by our hybrid ab initio and quasicontinuum methods.
Project Details
Project type
Exploratory Research
Start Date
2006-05-08
End Date
2007-06-01
Status
Closed
Released Data Link
Team
Principal Investigator
Related Publications
Bylaska EJ, KL Tsemekhman, SB Baden, JH Weare, and H Jonsson. 2011. "Parallel Implementation of Gamma-Point Pseudopotential Plane-Wave DFT with Exact Exchange." Journal of Computational Chemistry 32(1):54-69.