Point Defects and Localized Excitons in 2D WSe2

Abstract

Identifying the point defects in 2D materials is important for many applications. Recent studies have proposed that W vacancies are the predominant point defect in 2D WSe2, in contrast to theoretical studies, which predict that chalcogen vacancies are the most likely intrinsic point defects in transition metal dichalcogenide semiconductors. We show using first-principles calculations, scanning tunneling microscopy (STM), and scanning transmission electron microscopy experiments that W vacancies are not present in our CVD-grown 2D WSe2. We predict that O-passivated Se vacancies (OSe) and O interstitials (Oins) are present in 2D WSe2, because of facile O2 dissociation at Se vacancies or due to the presence of WO3 precursors in CVD growth. These defects give STM images in good agreement with experiment. The optical properties of point defects in 2D WSe2 are important because single-photon emission (SPE) from 2D WSe2 has been observed experimentally. While strain gradients funnel the exciton in real space, point defects are necessary for the localization of the exciton at length scales that enable photons to be emitted one at a time. Using state-of-the-art GW-Bethe-Salpeter-equation calculations, we predict that only Oins defects give localized excitons within the energy range of SPE in previous experiments, making them a likely source of previously observed SPE. No other point defects (OSe, Se vacancies, W vacancies, and SeW antisites) give localized excitons in the same energy range. Our predictions suggest ways to realize SPE in related 2D materials and point experimentalists toward other energy ranges for SPE in 2D WSe2.