AI-driven material screening and device fabrication of organic solar cells
Abstract
The development of high-performance organic solar cells (OSCs) is a critical step towards next-generation renewable energy, but progress is often hindered by the vast chemical space of potential materials and complex, multi-variable device fabrication processes. Traditional trial-and-error methodologies are both time-consuming and resource-intensive. This work presents an integrated, AI-driven workflow that accelerates both the discovery of novel photoactive materials and the optimization of device manufacturing. We employ machine learning models, trained on a curated dataset of experimental and computational data, to perform high-throughput virtual screening of potential donor and acceptor molecules, rapidly identifying candidates with optimal electronic and photophysical properties. Concurrently, a Bayesian optimization framework is coupled with an automated fabrication platform to efficiently navigate the high-dimensional parameter space of device engineering, including layer thickness, annealing temperatures, and blend morphology. This synergistic approach not only led to the identification of several promising material candidates but also demonstrated a significant reduction in the time required to achieve high power conversion efficiencies. Our findings establish a new paradigm for organic electronics, where intelligent automation streamlines the entire discovery-to-device pipeline, paving the way for the rapid realization of commercially viable organic solar cell technologies.
