Quantum mechanics at the core of multi-scale simulations

Rodney J. Bartlett, Josh McClellan, J. C. Greer, Scott Monaghan

Research output: Journal PublicationArticlepeer-review

5 Citations (Scopus)

Abstract

Quantum mechancial forces at the core of multi-scale simulations, require a one-electron Hamiltonian approach whose solution provide electronic energies, forces, and properties for > 1,000 atoms fast enough that it can drive large scale molecular dynamics. Such a transfer-Hamiltonian is hoped to be as predictive as accurate, ab initio quantum chemistry for such systems. To design the Hamiltonian requires that, we investigate rigorous one-particle theories including density functional theory (DFT) and the recently proposed, correlated orbital potential (COP) approach that has been developed solely from wavefunction considerations. The latter insists upon exact, principal ionization potentials and electron affinities for a system, while DFT insists upon the exact density and the HOMO ionization. These two complementary approaches help identify the essential quantities that an exact one-particle theory of electronic structure requires. The intent, then, is to incorporate these into a simple approximation that can provide the accuracy required but at a speed four orders of magnitude faster than today's DFT. The theory is presented and its neglect of diatomic differential overlap (NDDO) realization is illustrated for select systems.

Original languageEnglish
Pages (from-to)89-109
Number of pages21
JournalJournal of Computer-Aided Materials Design
Volume13
Issue number1-3
DOIs
Publication statusPublished - Oct 2006
Externally publishedYes

Keywords

  • Born-Oppenheimer forces
  • Chemical realism
  • Multi-scale simulations
  • Predictive simulations
  • Quantum mechanical forces
  • Semi-empirical Hamiltonians

ASJC Scopus subject areas

  • General Materials Science
  • Computer Science Applications
  • Computational Theory and Mathematics

Fingerprint

Dive into the research topics of 'Quantum mechanics at the core of multi-scale simulations'. Together they form a unique fingerprint.

Cite this