From Data to Dimers: Engineering Acene Derivatives for Photovoltaic Singlet Fission

Singlet fission (SF) is a photophysical process of interest in photovoltaic research with the potential to increase solar cell efficiency by reducing thermal energy losses. Acenes are prototypical SF chromophores and acene dimers have been critical to unravel the mechanisms and rates of SF. However, their practical deployment in photovoltaic devices is hindered by poor kinetic stability and low triplet energies, which limit both device lifetimes and open-circuit voltages. Here, we target the design of acene dimers that retain efficient SF while exhibiting higher triplet energies and improved stability, making them suitable as triplet sensitizers for hybrid solar cell architectures. Using a combinatorial assembly strategy, we generate a library of over 15,000 acene-based derivatives. Transformer-based machine learning models, trained on computed energetic properties, are then employed for high-throughput virtual screening against criteria relevant to SF efficiency and photovoltaic integration. A protocol to approximate acene dimer energetics from monomer data is presented, enabling efficient screening of this large design space. We identify numerous promising candidates for experimental synthesis and evaluation and distill a key design principle: incorporation of (benzo)furan and (benzo)thiophene units to yield 5- and 6-ring heteroacenes systematically tunes the energetics toward the SF regime while maintaining high triplet energies.