Advancing accurate and scalable electronic structure formalisms for light-harvesting materials

Light-harvesting applications (photovoltaics, photocatalysis, or photo-electrocatalysis) rely on physical processes that are directly amenable to first-principles simulations, facilitating detailed understanding and predictions of new materials at the very scale that matters in experiment. Major challenges for simulations include system size and complexity, accuracy of the underlying numerical descriptions, and accuracy of the physical description of ground- and excited state phenomena. This talk outlines our current reach based on a high-accuracy, scalable framework of numeric atom-centered basis functions (the FHI-aims code) and a general, open-source software infrastructure ELSI (http://elsi-interchange.org) that connects seamlessly to electronic structure solvers for different scales and problem classes. Specific application areas addressed in this talk include: (i) New multinary chalcogenide semiconductors Cu₂BaSn(S,Se)₄ and, more generally, I₂-II-IV-VI₄ (I=Cu,Ag; II=Ba,Sr; IV=Ge,Sn; VI=S,Se) for photovoltaics, designed to overcome limits of the Cu₂ZnSn(S,Se)₄ kesterites; (ii) Predicting the structural, electronic and optical properties of new crystalline layered organic-inorganic perovskites with large, electronically active organic functionalities such as oligothiophenes, allowing one to tune the detailed carrier properties by varying both the organic and the inorganic components. We particularly highlight the power of large-scale hybrid density-functional calculations (here covering crystalline materials with over 400 atoms per unit cell) for predictions of carrier properties with excellent qualitative accuracy, providing a reliable means to identify promising new materials for future experimental syntheses.