Exchange-correlation functional with good accuracy for both structural and energetic properties while depending only on the density and its gradient

Roberto Peverati, Donald G. Truhlar

Research output: Contribution to journalArticlepeer-review

201 Scopus citations

Abstract

The generalized gradient approximation (GGA) has been a workhorse exchange-correlation functional for electronic structure studies of extended systems (liquid-phase reactions, solids, heterogeneous and enzymatic catalysis, biopolymers) because its dependence on only the spin-labeled electron densities and their reduced gradients makes it the most affordable choice that produces realistic results for thermochemistry. However, much recent research has focused on its poor performance for solid-state lattice constants; the results for lattice constants can be improved but only at the cost of making the energetic predictions worse. In the present article, we propose a new density functional, called N12, which may be thought of as a generalization of range-separated functionals. The N12 functional depends only on the spin-labeled electron densities and their reduced gradients, but with a new kind of nonseparable term that gives it much greater flexibility. The N12 functional is the first exchange-correlation functional depending only on the spin-labeled electron densities and their reduced gradients that simultaneously provides good accuracy for the four key energetic and structural properties of solids and molecules, namely, solid-state cohesive energies and lattice constants and molecular atomization energies and bond lengths.

Original languageEnglish (US)
Pages (from-to)2310-2319
Number of pages10
JournalJournal of Chemical Theory and Computation
Volume8
Issue number7
DOIs
StatePublished - Jul 10 2012

Fingerprint Dive into the research topics of 'Exchange-correlation functional with good accuracy for both structural and energetic properties while depending only on the density and its gradient'. Together they form a unique fingerprint.

Cite this