Abstract

We present a benchmark of the performances of the density functional tight-binding model (DFTB) and its time-dependent counterpart (TD-DFTB) in describing both the ground state (GS) and excited state (ES) geometries of a panel of 30 organic molecules. Thanks to high-level wave function reference calculations, we are able to quantitatively assess the strengths and weaknesses of four DFTB models, using either second- or third-order self-consistent charge procedures, as well as different sets of parameters. The performances of the different DFTB models are found to be largely dependent on the type of bond considered, but the global mean absolute error remains acceptable for such "cheap" calculations, as it slightly exceeds the one obtained with DFT (PBE and B3LYP) or CC2 models for the GS of the same set of compounds. When considering the ES of the molecules, the TD-DFTB errors are surprisingly not systematically larger than their GS counterparts. However, the trends, either when going from the GS to the ES or within a homologous chemical series, are less consistently reproduced with TD-DFTB than with the considered ab initio models. This work therefore validates the use of TD-DFTB for describing ES geometries while highlighting that care has to be applied when looking at subtle variations.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call