Machine Learning‐guided Design of Conducting Polymers for UV–vis Photodetectors

For fabricating the high‐performance wide bandgap polymeric ultraviolet visible (UV–vis) photodetectors, the selection of the polymer to be utilized as an active photo‐detection material in such devices is quite difficult task. The efficiency of these photodetectors is directly linked with the property of bandgap ( E g ) associated with these polymers. Here, we utilized the machine learning (ML) approach as a robust technique to generate the novel library of polymers possessing desirable E g values for UV–vis photodetector application. The molecular descriptors, significantly impacting the parameter of E g , were generated and identified for these polymeric materials and ML analysis was performed to train the numerous ML models. The CatBoost model exhibited the best results. New database of polymers is created. Moreover, the synthetic feasibility of the newly generated polymers was also theoretically assessed via synthetic accessibility analysis and similarity analysis performed for the polymers.