Model Complexity Reduction Method for Multiscale Device Modeling

Back to all technologies
Download as PDF
Methods of electronic and atomic structure calculations have greatly contributed to expanding the field of computational material science. Density functional theory (DFT) methods for the simulation of electronic devices have a major disadvantage in poor scaling of DFT with system size. Even order-N DFT methods scale poorly when having atom sizes greater than a few thousand atoms. A model is needed for the variation of semiempirical tight binding (SETB) parameters with changes in material environment due to alloying, strain, and confinement.

Researchers at Purdue University have developed a new mapping scheme algorithm, which has improved accuracy and transferability of tight-bonding (TB) parameters. This algorithm has desired features of mapping from DFT to TB using alternate information from DFT, e.g., waveforms and DOS, providing a physically justified model for variation of on-site elements within environment, and provides relationships between SETB parameters. Mapping ab initio to reduce TB allows for the application of new materials outside of the normal scope of NEMO5 software.

-Completely bypasses the fitting-to-targets process by mathematically mapping from DFT to TB using alternate information (wave functions, DOS, etc.)
-Provides a physically justified model for variation of on-site elements with environment (strain, alloying, and interfaces to improve transferability)
-Provides mathematical relationships between SETB parameters to reduce the number of independent SETB parameters required to be optimized

Potential Applications
-Electronic and atomic structure calculations
Aug 28, 2018
United States
Jun 4, 2019

Dec 31, 2017
United States
Aug 28, 2018

Oct 24, 2014
Utility Patent
United States
Jan 2, 2018

Oct 24, 2013
United States
Purdue Office of Technology Commercialization
The Convergence Center
101 Foundry Drive, Suite 2500
West Lafayette, IN 47906

Phone: (765) 588-3475
Fax: (765) 463-3486