Excitons - bound electron-hole pairs - dominate the optical response of van der Waals layered materials assemblies, yet modelling them in these materials has remained computationally prohibitive due to their large unit cells containing hundreds to thousands of atoms.
We pioneered an efficient and accurate approach to solve the Bethe-Salpeter equation by exploiting the localization of atomic Wannier functions. This enables exciton calculations in systems with 5,000+ atoms - previously out of reach for first-principles methods.
Applying this method to WS₂/WSe₂ heterostructures, we resolved a longstanding puzzle: the origin of three absorption peaks near 1.7 eV observed in experiments. We discovered that these peaks arise from distinct spatial localizations of Wannier and charge-transfer excitons. The underlying mechanism is a competition between these two characters, driven by stacking-dependent variations in direct and indirect bandgaps from atomic relaxations, combined with environmentally tunable electron-hole interactions. Our results achieved meV-level agreement with experiment - the first quantitatively accurate atomistic description of moiré excitons.
Current & Recent Collaborators:
Theory: Johannes Lischner and Arash Mostofi (Imperial College), Ángel Rubio (Max Planck).
Experiment: Sufei Shi (Carnegie Mellon), Archana Raja (Berkeley), Fang Liu (Stanford), Jared Maxson (Cornell).