A machine learning-gaussian process screening of carbazole based donors to design efficient organic polymers for photovoltaic applications

Rapid industrialization is creating a serious threat to natural resources due to the overuse of fossil fuels. This is not only destroying the environment, but their extinction is also expected soon. Such a situation has increased the interest of scientists in designing new photovoltaic (PV) materials with tailored applications. In this study, a machine learning (ML)-assisted approach was proposed for the identification of optimal carbazole-based donor materials aimed at increasing the efficiency of organic PVs (OPVs). An extensive dataset comprising 592 carbazole-derived organic compounds was curated from existing literature, and their open circuit voltage (V_oc), is calculated. Through a targeted analysis, the top-performing donors exhibiting the highest V_oc values are identified. These selected donors are subsequently employed to design new TIC-based polymers to contribute a notable enhancement of the V_oc in the resulting PV devices. Results demonstrate ML potential to accelerate the discovery and optimization of organic solar materials, paving the way for the development of more sustainable and effective PV technologies.