Machine-Learning-Driven in-Device Optimization of All-Printed Perovskite Solar Cells
Emha Bayu Miftahullatif, Shreyas Dinesh Pethe, Andre K. Y. Low, Ayan A. Zhumekenov, Natalia Yantara, Priyanka Kajal, Qinjie Wu, Darrell Jun Jie Tay, Divyam Sharma, Saumya Sebastian, Jose Recatala‐Gomez, Abhishek Nambiar, Nripan Mathews, Kedar Hippalgaonkar
Abstract
Metal halide perovskites offer a vast but largely unexplored compositional and processing space. High-throughput experimentation (HTE) integrated with machine learning (ML) is ideal for efficient exploration, preferably at the device level. However, multilayer deposition challenges often limit HTE to stand-alone materials. We address this by employing a screen-printed triple-mesoscopic architecture, offering stability and low-cost fabrication, enabling rapid in-device screening of up to 81 unique devices per batch. Our platform accelerates experimental throughput over 100× and reduces data variance to 25% of manual methods. We present a ML-driven workflow to identify optimal additive compositions within MAPbI 3, MAPbI 3 /AVAI, and MAPbI 3 /MACl compositional space that simultaneously enhance device efficiency and stability. Prior additive studies were performed individually in conventional contexts, whereas our HT/ML-assisted approach on full devices is unprecedented. Our approach achieves a 5.75-fold improvement over pristine MAPbI 3, validated across two experimental batches. Further analysis shows AVAI and MACl act synergistically─AVAI aids infiltration and early crystallization, while MACl suppresses long-term PbI 2 formation─together enhancing carrier dynamics and stability.