Modern quantum chemistry has achieved a remarkable level of description of atoms, molecules, and their interactions. Theoretical approaches are particularly helpful when experimental studies are hampered or slowed down. For instance, experimental studies on the advancement of photovoltaic materials heavily rely on a trial-and-error approach. This severe disadvantage entails large workloads, depleting a significant amount of time and consumables. A more efficient approach is to utilize quantum chemistry to guide the synthesis of new and promising materials, thereby accelerating the development of novel compounds efficiently. The computational models are, however, difficult primarily because conventional, highly accurate approaches are technically limited to small—and to some extent simplified—model compounds and require user control at an expert level. In large-scale modeling of heavy-element-containing organic electronic materials, density functional theory (DFT) is hence considered as the method of choice. Yet, DFT may predict unreliable electronic structures and molecular properties as these systems may feature a substantial amount of strong correlation. Specifically, the failures of DFT include the overestimation of conjugation, torsional barriers, and electronic coupling, as well as the underestimation of bond-length alternations and excited-state energies in low-band-gap polymers. Charge-transfer excitations pose one of the significant challenges for DFT in accurately modeling organic light-emitting diodes (OLEDs). Thus, innovative new quantum-mechanical models are desirable that do not suffer from the technical bottlenecks and shortcomings of present-day quantum chemistry approaches and that open up new frontiers for computational chemistry, particularly in wave function approaches.
To break the current paradigm of computational chemistry, novel and neat approximations are desirable. One such innovative approach models many-electron systems using electron pair states. Current electron-pair methods are, however, insucient to reach chemical or spectroscopic accuracy for large molecules of organic electronics and must be extended to (i) accurately describe interactions beyond the simple pairing eects, especially in cases where conventional corrections break, (ii) reliably predict molecular properties and elucidating the structure-properties relationship using an understandable language, and (iii) provide an intuitive and black-box platform for non-expert users. The synergy between an inexpensive but reliable quantitative description and the qualitative interpretation of molecular interactions will accelerate the discovery of new materials in organic electronics. Thus, this project will shift the current paradigm in computational chemistry, large-scale modeling, and theoretical materials design of organic electronics towards novel and systematically improvable approaches (beyond DFT) implemented in modern QC codes that use progressive programming models to deploy workloads across various units and accelerators (CPUs, GPUs) and that are freely and broadly available according to the open-source model and FAIR features. Specifically, we aim to break the current paradigm in computational organic electronics by developing innovative hybrid wave-function-based approaches that are designed to be computationally inexpensive, robust, and black-box-like, requiring minimal user–software interaction. Our methodologies are based on the pair coupled cluster doubles (pCCD) approach, a wave function model with electron pairs as fundamental building blocks. The proposed models are implemented in PyBEST, a Python-based open-source program designed to tackle large-scale modeling of organic electronics featuring heavy elements.
