Density functional theory approximations for charge separation processes
The continuous increase in global energy demand needs sustainable sources, and organic photovoltaics (OPVs) are promising solutions due to their cost-effectiveness and adaptability. A critical challenge lies in optimizing charge separation processes in OPVs, which depend on photo-excitation followed by electronic interactions within donor-acceptor (D-A) molecular systems in the active layer. Conventional density functional theory (DFT) approximations struggle to model low bandgap D-A systems due to delocalization errors and inadequate representation of solvation effects. This manuscript addresses these issues as follows: (1) the agreement between experimental data and computational predictions including solvation effects and (2) the development of novel density functional approximations (DFAs) to reliably predict electronic properties of small band gap molecules and D-A systems composed by them. For that, we employ range-separated hybrid (RSH) DFT and time-dependent DFT (TD-DFT) to implicitly incorporate solvent effects through a first-principles screening protocol based on the solvent’s dynamic dielectric constant, avoiding further parameterization and enabling an accurate calculation of vertical excitation. Additionally, we evaluate spatially-resolved approaches -- such as local range separation (LRSH) and range-separated local hybridization (RSLH) -- with distinct building principles for the description of target properties that are important for the design of D-A systems and polarizabilities in conjugated chains. Our findings reveal that the integration of solvation effects via dielectric screening resolves discrepancies between computed and experimental absorption spectra and ionization energies, achieving deviations below 0.20 eV in low-bandgap solvated systems, as S,N-Heteroacenes and oligothiophenes. We also established that spatially resolved RSH frameworks (LRSH and RSLH) compare to the accuracy of the optimally-tuned functionals in predicting ionization energies and improve the description of linear polarizabilities. These results shed light on the strengths and limitations of these functional approximations, paving the way for more accurate and efficient electronic structure calculations in the DFT framework.