Algorithm-Guided Experimentation for Optimization of High-Performance Perovskite Solar Cells

Achieving high performance in perovskite solar cells (PSCs) is challenging due to the numerous complex variables involved across multiple fabrication steps. The strong interdependence among design variables, such as fabrication processing parameters and material composition, further complicates the search for optimal designs in a vast design space. In this work, we present a data-driven, algorithm-guided experimentation framework for the systematic optimization of PSC performance. We employed a model-based, derivative-free optimization algorithm to explore the design space of key processing parameters and identify promising regions for experimentation. By optimizing up to six processing parameters across the device structure and requiring fewer than 100 tested designs, we improved reverse-scanning power conversion efficiency from 20.3% to 23.1% without altering chemical ingredients or device configuration. These results demonstrate an effective combination of mathematical optimization and experimental research, which can benefit applications across disciplines that involve complex, interrelated variables requiring systematic optimization.