Accelerating photovoltaic polymer discovery via machine learning and synthetic accessibility analysis

Talal M. Althagafi, Muhammad Ramzan Saeed Ashraf Janjua, Fatimah Mohammed A. Alzahrani, Rashid Iqbal, M.S. Al-Buriahi,
Accelerating photovoltaic polymer discovery via machine learning and synthetic accessibility analysis,
Solid State Communications,
Volume 406,
2025,
116182,
ISSN 0038-1098,
https://doi.org/10.1016/j.ssc.2025.116182.
(https://www.sciencedirect.com/science/article/pii/S0038109825003576)
Abstract: It is quite challenging to design photovoltaic polymers with the best light emission characteristics. A thorough framework for creating photovoltaic polymers and predicting light emission maxima is presented in this research. Four machine learning models are trained for this. The most successful model is the Random Forest model (R2 = 0.94). Additionally, a feature importance analysis is conducted. A database containing 10,000 polymers is assembled and shown. Thirty polymers with higher emission maximum values are selected following screening. The majority of the selected polymers are easily synthesised based on synthetic accessibility. The discovery process is sped up by combining machine learning with polymer design, making it possible to quickly identify potential polymers for solar applications.
Keywords: Machine learning; Polymers; Emission maxima; Photovoltaics