Publication date: 15th December 2025
In novel printed organic solar cells, the device performance depends heavily on the donor–acceptor nanomorphology. We focus on two ways nanomorphology affects organic solar cells. First, we study domain connectivity by gradually diluting the donor material, an aspect of functional morphology that is linked to the tortuosity of charge transport. We observe how this affects charge generation, transport, and recombination in a state-of-the-art model system. Second, we investigate a different approach of linking nanomorphology to performance. Instead of using simple 1-dimensional drift–diffusion simulations, we now combine complex 3-dimensional simulations of realistic bulk heterojunction morphologies with machine learning. This approach has two goals. First, to find realistic solar cell parameters based on 3D models rather than effective ones based on 1D fits. Second, since 3D fitting is very time-consuming, machine learning lets us quickly predict solar cell performance and morphology from limited experimental data, reducing computation time from days to seconds.
We thank the Deutsche Forschungsgemeinschaft (DFG) for funding this work (Research Unit FOR 5387 POPULAR, Project No. 461909888).
