Long-range correction for tight-binding TD-DFT

TL;DR

This paper introduces a long-range correction to tight-binding TD-DFT, improving charge transfer excitation energies and oscillator strengths, enabling better modeling of large conjugated systems.

Contribution

The authors develop a long-range correction for TD-DFTB that enhances accuracy for charge transfer states and oscillator strengths, addressing limitations of the standard method.

Findings

Charge transfer excitation energies are improved.

Oscillator strengths for localized excitations are more realistic.

Method performs better on conjugated polymers.

Abstract

We present two improvements to the tight-binding approximation of time-dependent density functional theory (TD-DFTB): Firstly, we add an exact Hartree-Fock exchange term, which is switched on at large distances, to the ground state Hamiltonian and similarly to the coupling matrix that enters the linear response equations for the calculation excited electronic states. We show that the excitation energies of charge transfer states are improved relative to the standard approach without the long-range correction by testing the method on a set of molecules from the database in J. Chem. Phys. (2008),128,044118. that are known to exhibit problematic charge transfer states. The degree of spatial overlap between occupied and virtual orbitals indicates where TD-DFTB and lc-TD-DFTB can be expected to produce large errors. Secondly, we improve the calculation of oscillator strengths. The transition dipoles are obtained from Slater Koster files for the dipole matrix elements between valence orbitals. In particular excitations localized on a single atom, which appear dark when using Mulliken transition charges, in this way acquire a more realistic oscillator strength. These extensions pave the way for using long-range corrected TD-DFTB (lc-TD-DFTB) to describe the electronic structure of large chromophoric polymers, where uncorrected TD-DFTB fails to describe the high degree of conjugation and produces spurious low-lying charge transfer states.