Machine Learning Enabling the Prediction of Activation Energies of SPAAC

ABSTRACT This study leverages machine learning to predict the activation energies of strain‐promoted azide‐alkyne cycloaddition (SPAAC) reactions. Using DFT calculations, 631 sets of bond angles and Mulliken charges from two acyclic alkynes were collected. Multiple machine learning models were trained on these data, achieving high accuracy ( R 2 > 0.95). Both bond angle and charge were crucial for reliable predictions. The models successfully predicted activation energies for SPAAC reactions with unseen cycloalkynes, within certain limitations.