Modeling the electronic properties of solvated chromophores is challenging due to the computational cost associated with performing <i>ab initio</i> calculations on large systems. A popular way to overcome this issue is to use hybrid quantum mechanics and molecular mechanics methods (QM/MM). In QM/MM only the chromophore is modeled with an <i>ab initio </i>method, while the solvent or surroundings are modeled with a force field, such as the effective fragment potential (EFP). EFP computes the interaction energy between solvent molecules or effective fragments as the sum of electrostatics, polarization, dispersion and exchange repulsion terms. However, the interaction between the effective fragments and the QM chromophore is typically described by electrostatics and polarization terms only, corresponding to the polarizable embedding. In this work, we augment the QM/EFP embedding scheme by introducing the exchange-repulsion term and adding a charge-penetration correction to the electrostatics term. We compare the performance of these schemes to full quantum calculations by computing excitation energies in 38 different systems. Our results show that while the charge-penetration correction of the electrostatics term increases the mean error of QM/EFP, adding exchange-repulsion significantly reduces both of the mean and variance of the QM/EFP error distribution. Furthermore, we find that exchange-repulsion is particularly important for describing transitions involving charge transfer to and from solvent, making QM/EFP reliable for a broad range of electronic transitions.

Exchange-repulsion for QM/EFP excitations: Aye or nay?

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.