Advanced Energy and Sustainability Research Environmental Science (miscellaneous)

Bridging the Scalability Gap: From the Fourth Paradigm to Interconnected Discovery Ecosystems in Perovskite Photovoltaics

Jamal Uddin Ahamed, Maitry Barua, Tahamina Najnin, Intefar Samim Sami
Wiley · 2026-11-01 · DOI 10.1002/aesr.70297

Abstract

Rapid efficiency gains in laboratory‐scale perovskite photovoltaics have not translated into reproducible, scalable manufacturing because of persistent scientific, experimental, and process‐level barriers. Existing reviews cover machine learning and AI in materials screening, property prediction, and self‐driving laboratories, but links between laboratory demonstrations and prospectively validated, manufacturing‐relevant systems remain unresolved. This review examines five bottlenecks: new‐material discovery, device efficiency, operational stability, reproducibility and knowledge transfer, and scalable manufacturing, using evidence beyond retrospective accuracy and champion‐device efficiency, including experimental realization, uncertainty, independent batches, spatial uniformity, process‐window robustness, module yield, and durability. The review proposes an operational, falsifiable Fifth‐Paradigm framework defined by five attributes: closed‐loop experimental agency, material‐to‐manufacturing lifecycle integration, bidirectional physical‐digital synchronization, provenance‐aware data infrastructure, and validated transferable or distributed learning across domains. Each attribute is necessary but individually insufficient; ecosystem‐level status requires coordinated, prospectively validated evidence across all five. Generative modeling, autonomous experimentation, digital twins, and distributed learning are enabling mechanisms rather than indicators of a distinct paradigm. Current platforms demonstrate laboratory automation and closed‐loop optimization, but material‐device autonomy remains in its early stages, and a cross‐site discovery ecosystem has not yet been demonstrated. Distinction from Fourth‐Paradigm workflows requires measurable gains in prospective validation, reproducibility, cross‐domain transfer, and manufacturing‐relevant decision‐making.

Read at publisher →