I shall discuss progress in the design of a next generation of density functional theories. In contrast to most approaches to functional design, we have adopted a combinatorial approach, in which we have trained a huge number of functionals (over 100,000 in the case of the generalized gradient approximation models, and over 1010 in the case of meta GGA functionals). The functional from each class that performs best on independent test data (survival of the most transferable) is self-consistently trained to yield new generation functionals that seems very promising for application purposes. A key component of these functionals is the non-local VV10 density-density correlation functional that enables highly accurate treatment of non-bonded interactions, if the entire functional is self-consistently trained. The resulting functionals involve significantly fewer parameters than many of the best functionals of recent years. I shall discuss the results of large-scale benchmarking exercises that assess the new functionals against other existing functionals, including simpler atom-atom dispersion corrections with optimized damping functions. I will also pay particular attention to the applicability of the new functionals to periodic condensed phase electronic structure theory, which requires replacement of the VV10 functional by the revised VV10 (rVV10) non-local correction.

Advances in density functional theory for molecules and materials

253rd National Meeting (2017)

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