Efficient time-dependent density functional theory approximations for hybrid density functionals: analytical gradients and parallelization | JoVE Visualize

Area of Science:

Computational Chemistry
Quantum Chemistry
Theoretical Chemistry

Background:

Accurate calculation of excited-state properties is vital for understanding molecular behavior.
Traditional time-dependent density functional theory (TDDFT) methods can be computationally expensive, especially for large systems.
Hybrid density functionals require efficient approximations for both Coulomb and exchange terms.

Purpose of the Study:

To implement and validate efficient approximations for TDDFT within the Tamm-Dancoff approximation (TDA) using hybrid density functionals.
To assess the accuracy and computational speedup of the combined Resolution of Identity (RI-J) and Chain-of-Spheres Exchange (COSX) algorithms (RIJCOSX).
To extend the methodology for excited-state property calculations, including analytical gradients and frozen-core approximations.

Main Methods:

Combined RI-J algorithm for Coulomb terms and COSX algorithm for exchange terms in TDDFT/TDA calculations.
Assessment of adiabatic transition energies, excited-state structures, and vibrational frequencies using configuration interaction singles and hybrid TDDFT/TDA.
Implementation of parallel algorithms for energy and gradient calculations, including Z-vector equation solutions.
Extension of Lagrangian formalism for frozen-core approximation in TDDFT excited-state property calculations.

Main Results:

The RIJCOSX approximation achieves speedups of up to two orders of magnitude for extended basis sets compared to traditional methods.
Typical errors in transition energies are around 0.01 eV, with minimal deviations in excited-state geometries and vibrational frequencies.
The COSX algorithm demonstrates near-perfect parallelization, enabling rapid calculations for large molecules like Ag-TB2-helicate.
The implemented algorithms show insignificant errors compared to inherent approximations in TDDFT and basis set truncation.

Conclusions:

The RIJCOSX approximation provides a computationally efficient and accurate approach for TDDFT/TDA calculations.
The parallel implementation significantly reduces computation time for complex molecular systems.
These advancements are integrated into the ORCA electronic structure system, offering valuable tools for researchers in computational and theoretical chemistry.