Filipp Furche of the University of California, Irvine is supported by an award from the Chemical Theory, Models and Computational Methods Program (Division of Chemistry), the Condensed Matter and Materials Theory Program (Division of Materials Research) and the Computational and Data-Enabled Science and Engineering Program (CDS&E) to develop, implement, test, and apply computational methods that deliver predictive accuracy for an important class of molecular systems that are very difficult to characterize either by experiment or by current computational approaches. These systems, known as small band-gap systems, have a small energy gap between the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO). Molecules and ions containing transition metals such as iron or platinum or the even heavier lanthanides and actinides can fall into this category. The rational design of catalysts crucially relies on our ability to model such molecules. However, accurate simulations of small-gap systems with more than a few atoms have been very elusive. The methods and computer programs developed by Furche and his research group enable simulations of chemical structures, processes, and materials of fundamental and technological importance. The methods developed in this project are made available to the public through the Turbomole quantum chemistry software. This project also involves undergraduate curriculum development at UC Irvine and an outreach program for high school students in disadvantaged neighborhoods.
Previous work in the Furche group has established that random phase approximation (RPA)-Renormalized many-body perturbation theory is capable of systematically improving semi-local DFT results for small-gap systems. The proposed project builds on these results and aims to transform the way computational and experimental chemists and materials scientists approach small-gap molecules by developing an armamentarium of robust and widely applicable computational tools. A frequency-dependent RPA renormalized Bethe-Salpeter kernel is proposed to boost the accuracy of RPA-type methods. Electronically excited states and frequency-dependent response properties are accessed via time-dependent response theory. Algorithmic developments aim to further reduce the cost of RPA and beyond-RPA calculations and extend their scope to systems with hundreds of atoms.
