Density functional theory (DFT) has become a popular method to model materials. DFT calculations can be used to guide experiments since they are able to predict a wide range of physical properties and chemical trends. Since their inception, DFT methods have been known to underestimate the band gap, and this is known as the "band gap problem". With the advent of highly parallelized computer architectures, it has become possible to perform more DFT calculations in a shorter amount of time, and this has led to high-throughput DFT studies as a mode of materials discovery. A major roadblock to the large scale implementation of high-throughput DFT calculations is the lack of a flexible method to calculate accurate properties derived from the electronic band structure, such as the band gap, without a prohibitive computational cost. One method that provides a more accurate determination of electronic properties at little computational cost is the addition of an on-site Coulomb repulsion term called the Hubbard U. This term corrects for the overdelocalization of electron density that occurs in DFT methods and can be derived from first-principles using linear response. In this work we derive the Hubbard U term for the 3d, 4d, and 5d rows of transition metals for the GBRV set of ultrasoft pseudopotentials in a variety of well-known structure types in which we can systematically vary the oxidation state of the transition metal. We then discuss the results in terms of basic periodic trends and go on to present how the results of our investigation can be used in complex oxide materials that contain multi-valent and/or multiple transition metals that are redox active.

Systematic determination of Hubbard U for high-throughput DFT calculations

255th National Meeting (2018)

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