Dicarbon defect in hexagonal boron nitride monolayer—a theoretical study

Abstract

A comprehensive theoretical study of the lowest electronic vertical excitations of the CBCN defect in the monolayer of hexagonal boron nitride has been performed. Both the periodic boundary conditions approach and the finite-cluster simulation of the defect have been utilized at the density-functional theory (DFT) level. Clusters of increasing sizes have been used in order to estimate artefacts resulting from edge effects. The stability of the results with respect to several density functionals and various basis sets has been also examined. High-level ab initio calculations with methods like equation-of-motion coupled cluster method with single and double excitations (EOM-CCSD), algebraic-diagrammatic construction to the second order (ADC(2)), and time-dependent approximate coupled cluster theory to the second order (TD-CC2) were performed for the smallest clusters. It turns out that TD-DFT with the CAM-B3LYP functional gives similar lowest excitation energies as EOM-CCSD, ADC(2), and TD-CC2. The lowest excitation energies resulting from the periodic-boundary calculation utilizing the Bethe–Salpeter equation are in agreement with the results for finite clusters. The analysis of important configurations and transition densities shows that for all studied methods, the lowest excited state is localized on two carbon atoms and their closest neighbours and has a large dipole transition moment. The optimized geometries for the lowest two excited states indicate that in both cases, the carbon–carbon bond becomes a single bond, while for the second excited state, additionally one from boron–nitrogen bonds loses its partially double character. The calculation of the excitation energies at the respective optimal geometry reveals that these two energies become about 0.5 eV lower than vertical excitations from the ground-state geometry.