The accuracy and efficiency of large-scale density functional theory (DFT) calculations is an issue that is central to the ability of computational chemistry to attack the central problems associated with the nanoporous materials genome problem.  At the heart of this issue is the design of more accurate and more transferable density functionals that yield the very high accuracy needed for both intermolecular and intramolecular interactions.  We argue that the tremendous flexibility of third rung functionals of the meta generalized gradient approximation (mGGA) type demands a combinatorial search approach to the mGGA functional design problem.  We then characterize the enormous dimension of the functional space within which combinatorial design can be performed, and describe our approach to a partial search of this space.  The resulting functionals are compared against leading existing functionals on a broad variety of test cases.  It is shown that substantially fewer parameters are required than have been commonly employed, and that very significant improvements in accuracy for intermolecular interactions can be obtained, as well as significant improvements in accuracy for bonded interactions.

New density functionals obtained by a genome-scale search approach to functional design

249th National Meeting (2015)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.