2019/02/28 by Lars Goerigk, Nisha Mehta · 1 citation
Chemistry · Materials Science · Physics and Astronomy · #Advanced Chemical Physics Studies #Crystallography and molecular interactions #Machine Learning in Materials Science
paper · pdf · doi:10.1071/ch19023
openalex publication_date 2019/02/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/01
This account is written for general users of density functional theory (DFT) methods as well as experimental researchers who are new to the field and would like to conduct such calculations. Its main emphasis lies on how to find a way through the confusing ‘zoo’ of DFT by addressing common misconceptions and highlighting those modern methods that should ideally be used in calculations of energetic properties and geometries. A particular focus is on highly popular methods and the important fact that popularity does not imply accuracy. In this context, we present a new analysis of the openly available data published in Swart and co-workers’ famous annual ‘DFT poll’ (http://www.marcelswart.eu/dft-poll/) to demonstrate the existing communication gap between the DFT user and developer communities. We show that despite considerable methodological advances in the field, the perception of some parts of the user community regarding their favourite approaches has changed little. It is hoped that this account makes a contribution towards changing this status and that users are inspired to adjust their current computational protocols to accommodate strategies that are based on proven robustness, accuracy, and efficiency rather than popularity.