Proceedings of International Conference on Hybrid and Organic Photovoltaics (HOPV24)
DOI: https://doi.org/10.29363/nanoge.hopv.2024.016
Publication date: 6th February 2024
The translation of statistical techniques from the artificial intelligence community to materials science and engineering is helping to bridge the divide between traditional modelling and measurements [1]. In the study of metal halide perovskites, data-driven machine learning (ML) workflows are being used for diverse tasks ranging from novel materials discovery to closed-loop accelerated device optimisation. I will provide an introduction to this topic, with a focus on how the limits of materials modelling are being extended by incorporating ML techniques to develop deeper insights into the behaviour of perovskites across time and length scales [2,3]. In particular, I will discuss our latest understanding of compositional and structural disorder at the nanoscale, linked to multi-modal experimental characterisation [4]. The overarching goal is to shed light on the origins of the exceptional performance of these systems, as well as to identify routes to develop the next generation of perovskite-inspired materials.