ChatGPT-assisted rational design for iterative performance optimization of perovskite solar cells

Artificial intelligence (AI) in materials science has rapidly driven breakthroughs in material optimization, significantly enabling the development of high-efficiency perovskite solar cells (PSCs). However, the exploration of co-assembled self-assembled monolayers (co-SAMs) in PSCs remains limited due to the high-dimensional complexity of their design and the scarcity of comprehensive experimental data. Traditional trial-and-error approaches further impede the efficient discovery of optimal co-SAMs configurations. To address these challenges, we employed the ChatGPT-4o model to assist in the iterative performance optimization of using co-SAMs strategy. By applying structured prompt engineering within a Question-Answer-Refinement (Q-A-R) framework, we attained a power conversion efficiency (PCE) of 24.48% within 25 days, which was more than four times faster than traditional methods. Beyond guiding experimental design, ChatGPT-4o also proved effective in streamlining characterization techniques to understand the performance advantages of co-SAMs. Overall, this study demonstrates that ChatGPT-assisted rational experimental design offers significant potential for PSC devices iterative performance optimization while providing valuable insights into the design and characterization of advanced materials.