Dispersion-corrected DFT calculations on C60-porphyrin complexes

Abstract

The quality of the newly added, empirical dispersion correction in density functional theory (DFT) calculations is examined for several supramolecular complexes of fullerene (C(60)) with free-base and metal porphyrins (Por). The benzene dimer (C(6)H(6))(2), naphthalene dimer (C(10)H(8))(2), and anthracene dimer (C(14)H(10))(2) were also included in the study for comparison. Three density functionals, two damping functions, and two types of basis sets were employed in the computations. The estimated dispersion energies in the fullerene-porphyrin systems are rather large, ranging from 0.5 eV in C(60).ZnP to 1 eV in C(60).H(2)TPP. Any dispersion-corrected DFT (DFT + E(disp)) method is shown to perform well for C(60).H(2)TPP, C(60).ZnTPP, and C(60).ZnP, where the intermolecular distances are relatively large. But large basis sets, e.g. TZP (triple-zeta + one polarization function), are required in order to obtain reliable results with DFT + E(disp). In the case of C(60).FeP, where the intermolecular distance R is short, the DFT + E(disp) calculated R depends on the damping function as well as on the DFT method, and all the DFT + E(disp) calculations lead to significant changes in the relative energies of the various spin states. The quality of the DFT + E(disp) calculated results on C(60).FeP is hard to judge here without detailed experimental data on a C(60).FePor complex. Owing to error cancellation, the pure DFT calculations with a smaller DZP (double-zeta + one polarization function) basis set without any correction are shown to give quite accurate results.