Validation of local hybrid functionals for TDDFT calculations of electronic excitation energies.

J Chem Phys

Technische Universität Berlin, Institut für Chemie, Theoretische Chemie/Quantenchemie, Sekr. C7, Straße des 17. Juni 135, D-10623 Berlin, Germany.

Published: February 2016

The first systematic evaluation of local hybrid functionals for the calculation of electronic excitation energies within linear-response time-dependent density functional theory (TDDFT) is reported. Using our recent efficient semi-numerical TDDFT implementation [T. M. Maier et al., J. Chem. Theory Comput. 11, 4226 (2015)], four simple, thermochemically optimized one-parameter local hybrid functionals based on local spin-density exchange are evaluated against a database of singlet and triplet valence excitations of organic molecules, and against a mixed database including also Rydberg, intramolecular charge-transfer (CT) and core excitations. The four local hybrids exhibit comparable performance to standard global or range-separated hybrid functionals for common singlet valence excitations, but several local hybrids outperform all other functionals tested for the triplet excitations of the first test set, as well as for relative energies of excited states. Evaluation for the combined second test set shows that local hybrids can also provide excellent Rydberg and core excitations, in the latter case rivaling specialized functionals optimized specifically for such excitations. This good performance of local hybrids for different excitation types could be traced to relatively large exact-exchange (EXX) admixtures in a spatial region intermediate between valence and asymptotics, as well as close to the nucleus, and lower EXX admixtures in the valence region. In contrast, the tested local hybrids cannot compete with the best range-separated hybrids for intra- and intermolecular CT excitation energies. Possible directions for improvement in the latter category are discussed. As the used efficient TDDFT implementation requires essentially the same computational effort for global and local hybrids, applications of local hybrid functionals to excited-state problems appear promising in a wide range of fields. Influences of current-density dependence of local kinetic-energy dependent local hybrids, differences between spin-resolved and "common" local mixing functions in local hybrids, and the effects of the Tamm-Dancoff approximation on the excitation energies are also discussed.

Download full-text PDF

Source
http://dx.doi.org/10.1063/1.4941919DOI Listing

Publication Analysis

Top Keywords

local hybrids
32
hybrid functionals
20
local hybrid
16
excitation energies
16
local
14
hybrids
9
electronic excitation
8
tddft implementation
8
valence excitations
8
core excitations
8

Similar Publications

Data-Driven Improvement of Local Hybrid Functionals: Neural-Network-Based Local Mixing Functions and Power-Series Correlation Functionals.

J Chem Theory Comput

January 2025

Technische Universitát Berlin, Institut für Chemie, Theoretische Chemie/Quantenchemie, Sekr. C7, Straße des 17. Juni 135, Berlin D-10623, Germany.

Local hybrid functionals (LHs) use a real-space position-dependent admixture of exact exchange (EXX), governed by a local mixing function (LMF). The systematic construction of LMFs has been hampered over the years by a lack of exact physical constraints on their valence behavior. Here, we exploit a data-driven approach and train a new type of "n-LMF" as a relatively shallow neural network.

View Article and Find Full Text PDF

Gamete killers are genetic loci that distort segregation in the progeny of hybrids because the killer allele promotes the elimination of the gametes that carry the sensitive allele. They are widely distributed in eukaryotes and are important for understanding genome evolution and speciation. We had previously identified a pollen killer in hybrids between two distant natural accessions of Arabidopsis thaliana.

View Article and Find Full Text PDF

Polymorphic short insertions and deletions (INDELs ≤ 50 bp) are abundant, although less common than single nucleotide polymorphisms (SNPs). Evidence from model organisms shows INDELs to be more strongly influenced by purifying selection than SNPs. Partly for this reason, INDELs are rarely used as markers for demographic processes or to detect divergent selection.

View Article and Find Full Text PDF

Territorial aggression is widespread across the animal kingdom and is expressed in diverse ecological and social contexts. In addition, there are marked variations in the degree of male reproductive territoriality within and between species. These differences are often attributed to genetic components.

View Article and Find Full Text PDF

Rainfall prediction is a crucial aspect of climate science, particularly in monsoon-influenced regions where accurate forecasts are essential. This study evaluates rainfall prediction models in the Eastern Thailand by examining an optimal lag time associated with the Oceanic Niño Index (ONI). Five deep learning models-RNN with ReLU, LSTM, GRU (single-layer), LSTM+LSTM, and LSTM+GRU (multi-layer)-were compared using mean absolute error (MAE) and root mean square error (RMSE).

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!